De Icing Guidelines

THE NATIONAL ACADEMIES PRESS
This PDF is available at http://nap.edu/23178 SHARE

   
Guidelines for the Selection of Snow and Ice Control

Materials to Mitigate Environmental Impacts
DETAILS
194 pages | 8.5 x 11 | PAPERBACK

ISBN 978-0-309-42233-8 | DOI 10.17226/23178
CONTRIBUTORS
GET THIS BOOK National Cooperative Highway Research Program; Transportation Research Board;
National Academies of Sciences, Engineering, and Medicine
FIND RELATED TITLES

Visit the National Academies Press at NAP.edu and login or register to get:
– Access to free PDF downloads of thousands of scientific reports

– 10% off the price of print titles
– Email or social media notifications of new titles related to your interests
– Special offers and discounts

Distribution, posting, or copying of this PDF is strictly prohibited without written permission of the National Academies Press.
(Request Permission) Unless otherwise indicated, all materials in this PDF are copyrighted by the National Academy of Sciences.
Copyright © National Academy of Sciences. All rights reserved.

Guidelines for the Selection of Snow and Ice Control Materials to Mitigate Environmental Impacts
NATIONAL COOPERATIVE HIGHWAY RESEARCH PROGRAM
NCHRP REPORT 577

Guidelines for the Selection of
Snow and Ice Control Materials to
Mitigate Environmental Impacts
LEVELTON CONSULTANTS LIMITED

Richmond, BC
Subject Areas
Energy and Environment • Maintenance
Research sponsored by the American Association of State Highway and Transportation Officials
in cooperation with the Federal Highway Administration
TRANSPORTATION RESEARCH BOARD

WASHINGTON, D.C.
2007
www.TRB.org
Copyright National Academy of Sciences. All rights reserved.

NATIONAL COOPERATIVE HIGHWAY NCHRP REPORT 577

RESEARCH PROGRAM
Systematic, well-designed research provides the most effective Project 6-16
approach to the solution of many problems facing highway ISSN 0077-5614
administrators and engineers. Often, highway problems are of local ISBN: 978-0-309-09880-9
interest and can best be studied by highway departments individually Library of Congress Control Number 2007925771
or in cooperation with their state universities and others. However, the © 2007 Transportation Research Board
accelerating growth of highway transportation develops increasingly
complex problems of wide interest to highway authorities. These
problems are best studied through a coordinated program of COPYRIGHT PERMISSION
cooperative research.
Authors herein are responsible for the authenticity of their materials and for obtaining
In recognition of these needs, the highway administrators of the written permissions from publishers or persons who own the copyright to any previously
American Association of State Highway and Transportation Officials published or copyrighted material used herein.
initiated in 1962 an objective national highway research program Cooperative Research Programs (CRP) grants permission to reproduce material in this
employing modern scientific techniques. This program is supported on publication for classroom and not-for-profit purposes. Permission is given with the
understanding that none of the material will be used to imply TRB, AASHTO, FAA, FHWA,
a continuing basis by funds from participating member states of the
FMCSA, FTA, or Transit Development Corporation endorsement of a particular product,
Association and it receives the full cooperation and support of the method, or practice. It is expected that those reproducing the material in this document for
Federal Highway Administration, United States Department of educational and not-for-profit uses will give appropriate acknowledgment of the source of
any reprinted or reproduced material. For other uses of the material, request permission
Transportation.
from CRP.
The Transportation Research Board of the National Academies was
requested by the Association to administer the research program
because of the Board’s recognized objectivity and understanding of
NOTICE
modern research practices. The Board is uniquely suited for this
purpose as it maintains an extensive committee structure from which The project that is the subject of this report was a part of the National Cooperative Highway
Research Program conducted by the Transportation Research Board with the approval of
authorities on any highway transportation subject may be drawn; it the Governing Board of the National Research Council. Such approval reflects the
possesses avenues of communications and cooperation with federal, Governing Board’s judgment that the program concerned is of national importance and
state and local governmental agencies, universities, and industry; its appropriate with respect to both the purposes and resources of the National Research
Council.
relationship to the National Research Council is an insurance of
The members of the technical committee selected to monitor this project and to review this
objectivity; it maintains a full-time research correlation staff of
report were chosen for recognized scholarly competence and with due consideration for the
specialists in highway transportation matters to bring the findings of balance of disciplines appropriate to the project. The opinions and conclusions expressed
research directly to those who are in a position to use them. or implied are those of the research agency that performed the research, and, while they have
been accepted as appropriate by the technical committee, they are not necessarily those of
The program is developed on the basis of research needs identified
the Transportation Research Board, the National Research Council, the American
by chief administrators of the highway and transportation departments Association of State Highway and Transportation Officials, or the Federal Highway
and by committees of AASHTO. Each year, specific areas of research Administration, U.S. Department of Transportation.
needs to be included in the program are proposed to the National Each report is reviewed and accepted for publication by the technical committee according
Research Council and the Board by the American Association of State to procedures established and monitored by the Transportation Research Board Executive
Committee and the Governing Board of the National Research Council.
Highway and Transportation Officials. Research projects to fulfill these
needs are defined by the Board, and qualified research agencies are The Transportation Research Board of the National Academies, the National Research
Council, the Federal Highway Administration, the American Association of State Highway
selected from those that have submitted proposals. Administration and and Transportation Officials, and the individual states participating in the National
surveillance of research contracts are the responsibilities of the National Cooperative Highway Research Program do not endorse products or manufacturers. Trade
Research Council and the Transportation Research Board. or manufacturers’ names appear herein solely because they are considered essential to the
object of this report.
The needs for highway research are many, and the National
Cooperative Highway Research Program can make significant
contributions to the solution of highway transportation problems of
mutual concern to many responsible groups. The program, however, is
intended to complement rather than to substitute for or duplicate other
highway research programs.
Published reports of the
NATIONAL COOPERATIVE HIGHWAY RESEARCH PROGRAM

are available from:
Transportation Research Board

Business Office
500 Fifth Street, NW
Washington, DC 20001
and can be ordered through the Internet at:

http://www.national-academies.org/trb/bookstore
Printed in the United States of America


COOPERATIVE RESEARCH PROGRAMS
CRP STAFF FOR NCHRP REPORT 577

Christopher W. Jenks, Director, Cooperative Research Programs
Crawford F. Jencks, Deputy Director, Cooperative Research Programs
Christopher J. Hedges, Senior Program Officer
Eileen P. Delaney, Director of Publications
Hilary Freer, Senior Editor
NCHRP PROJECT 6-16 PANEL

Field of Maintenance—Area of Snow and Ice Control
Daniel P. Johnston, South Dakota DOT, Pierre, SD (Chair)
Jeanne Hewitt, New York State DOT, Albany, NY
Robert Chenier, Environment Canada, Hull, QC
Edward Fink, Colorado DOT, Grand Junction, CO
G. Michael Fitch, Virginia DOT, Charlottesville, VA
Rod Frederick, U.S. Environmental Protection Agency (Retired)
Jeanne Hewitt, New York State DOT, Albany, NY
Deborah J. O’Bannon, University of Missouri–Kansas City, Kansas City, MO
Leland D. Smithson, Iowa DOT, Ames, IA
Paul Pisano, FHWA Liaison
Frank N. Lisle, TRB Liaison
AUTHOR ACKNOWLEDGMENTS
The research report herein was performed under NCHRP Project 06-16 by Levelton Consultants, Ltd.,
in cooperation with Ice and Snow Technologies; Dr. William M. Lewis, Jr.; and Dr. Brian Chollar. The
work undertaken by Ice and Snow Technologies, Dr. William M. Lewis, Jr., and Dr. Brian Chollar was
under subcontract with Levelton Consultants Ltd.
The Principal Investigator for this report was Brent T. Mussato, Levelton Consultants, Ltd. The other
authors for this report were Oliver K. Gepraegs, Phil T. Seabrook, John P. Davidson, and Robert S. Char-
lton, Levelton Consultants Ltd.; Richard D. Parker and Dale G. Keep, Ice and Snow Technologies;
Dr. William M. Lewis, Jr., University of Colorado, Boulder; and Dr. Brian Chollar. The work was con-
ducted under the general supervision of Wayne C. Edwards, P.Eng, Principal, Levelton Consultants Ltd.

FOREWORD
By Christopher J. Hedges
Staff Officer
Transportation Research Board
This report presents guidelines for the selection of snow and ice control materials through
an evaluation of their cost, performance, and impacts on the environment and infrastruc-
ture. The guidelines should be useful in helping maintenance managers develop a program
that will minimize the environmental impacts of snow and ice control without compromis-
ing effective maintenance strategies. The Guidelines will help highway agencies fill their dual
role of providing safe roadways for the driving public while serving as stewards to protect
and enhance the natural environment.
Every year, considerable quantities of snow and ice control products are applied to high-
ways, and environmental and regulatory agencies have questioned the environmental
impact of these products. Transportation agencies are asked to use “environmentally
friendly” or less toxic alternatives wherever possible, but there is no commonly accepted
guidance for determining which products are acceptable. The traditional use of roadsalt has
been prohibited in some locations, leaving highway agencies uncertain about how traffic
safety can be maintained in bad weather. For example, Environment Canada has concluded
that inorganic chloride roadsalts are harmful to the environment under the Canadian Envi-
ronmental Protection Act, thus requiring development and implementation of improved
management practices.
Studies of the most common chemical alternatives—sodium chloride (salt), magnesium
chloride, calcium chloride, calcium magnesium acetate, potassium acetate, and urea—have
focused on performance and cost under various weather conditions without evaluating
their relative impacts on the environment. Several new chemical preparations, including
some that are trademarked, have entered the market as snow and ice control chemicals for
use by transportation agencies, but there is limited information about their environmental
impacts. There is a need for rational decision-making guidelines to assist DOT maintenance
managers in selecting the most appropriate snow and ice control materials for the condi-
tions that exist in their jurisdictions.
Under NCHRP Project 06-16, a research team led by Levelton Consultants Ltd. con-
ducted an extensive review of relevant technical information, then conducted a program of
analytical laboratory testing on a wide range of products currently available in the North
American market. The results were used to develop three products for the end user: a deci-
sion tool for the selection of the most appropriate snow and ice control materials to suit the
specific needs of any given highway agency, a purchase specification that can be used by an
agency once agency staff have selected materials for use, and a quality assurance monitor-
ing program that includes procedures and standard test methods to characterize snow and
ice control products before their purchase or use.

There are two major components to this report. Chapters 1 through 11 (a) summarize
the review of technical information on the performance and environmental impacts of cur-
rently available snow and ice control materials, and (b) present the results of an analytical
laboratory testing program designed to address gaps in the current literature and to develop
a matrix of properties of 42 chemicals. The second component presents guidelines for the
user with three products: a decision tool for the selection of materials, a purchase specifica-
tion, and the quality assurance monitoring program. An electronic version of the decision
tool and several appendixes to the final report can be downloaded from the TRB website at
http://www.trb.org/TRBNet/ProjectDisplay.asp?ProjectID=883.

CONTENTS
1 Summary
8 Chapter 1 Introduction
8 1.1 Problem Statement and Research Approach
8 1.2 Scope of Study
10 Chapter 2 Operational and Performance Considerations
10 2.1 Products Used and Properties
10 2.1.1 Chloride Salts
11 2.1.2 Organic Products
13 2.1.3 Sand or Abrasives
13 2.2 Storage
13 2.2.1 Solid Snow and Ice Control Materials
13 2.2.2 Liquid Snow and Ice Control Materials
13 2.2.3 Sand or Abrasives
13 2.3 Handling and Application Methods
13 2.4 Agency Survey—Current and Future Status of Use in the
United States and Canada
14 2.4.1 Product Use and Methods of Application
15 2.4.2 Natural Environment Concerns
16 2.4.3 Corrosion Concerns
16 2.4.4 Purchasing Issues
18 2.5 Application Strategies
18 2.5.1 Anti-Icing
19 2.5.2 Deicing
19 2.5.3 Pre-wetting
20 2.5.4 Dry Abrasive/Chemical Mixes
20 2.5.5 Abrasive Use
20 2.5.6 Stockpile Treatment
21 2.5.7 Application Strategy Studies
22 2.6 Materials Loading
22 2.6.1 Product Selection
23 2.6.2 Application Amounts
23 2.6.3 Strategy Selection
25 CHAPTER 3 Impacts to the Natural Environment
25 3.1 Environmental Pathways
25 3.1.1 Soil Transport
27 3.1.2 Air Transport
27 3.1.3 Water Transport
27 3.2 Framework for Protection of Environmental Systems
27 3.2.1 U.S. Regulations
28 3.2.2 Canadian Regulations
29 3.2.3 Others

29 3.3 Components of Concern

30 3.3.1 Ferrocyanides
30 3.3.2 Radioactivity
30 3.4 Aquatic Environments
30 3.4.1 Protection of Aquatic Resources
32 3.4.2 Potential Types of Impairment of Aquatic Resources by Snow and
Ice Control Materials
42 3.4.3 Evaluation of Specific Snow and Ice Control Materials
46 3.4.4 Overview of Specific Snow and Ice Control Materials
46 3.4.5 Conclusions
47 3.5 Soils
48 3.5.1 Chloride
48 3.5.2 Sodium
49 3.5.3 Calcium
49 3.5.4 Magnesium
49 3.5.5 Potassium
50 3.5.6 Acetates
50 3.5.7 Organic Products
50 3.5.8 Urea
51 3.5.9 Abrasives
51 3.5.10 Ferrocyanide Products
51 3.5.11 Conclusion
51 3.6 Soil Biotic
51 3.6.1 Chloride-Based Salts
52 3.6.2 Urea and Potassium Acetate
52 3.6.3 Ferrocyanides
53 3.6.4 Conclusion
53 3.7 Vegetation
53 3.7.1 Aerial Deposition
54 3.7.2 Ion Uptake and Accumulation from Soil
54 3.7.3 Osmotic Stress
55 3.7.4 Species Sensitivities
55 3.7.5 Other Factors Affecting Impacts to Vegetation
56 3.7.6 Specific Effects Associated with Select Snow and Ice Control
Materials
59 3.7.7 Conclusion
59 3.8 Air Quality
60 3.9 Animals
60 3.9.1 Wildlife
61 3.9.2 Birds
61 3.9.3 Toxic Effects
61 3.9.4 Bioaccumulation of Heavy Metals
62 3.9.5 Conclusions
62 3.10 Human Health and Exposure
63 CHAPTER 4 Impacts to Infrastructure and Vehicles
63 4.1 Concrete and Concrete Infrastructure
64 4.1.1 Corrosion of Concrete-Embedded Reinforcing Steel
68 4.1.2 Effects of Snow and Ice Control Materials on Cement Paste
70 4.1.3 Concrete-Embedded Reinforcing Steel and Concrete Paste
Conclusions

71 4.1.4 Atmospheric Corrosion Costs

71 4.1.5 Forms of Corrosion
72 4.1.6 Atmospheric Corrosion of Motor Vehicles, Roadside Infrastructure,
and Metal Bridges
73 4.1.7 Motor Vehicle, Roadside Infrastructure, and Bridge Materials
74 4.1.8 Snow and Ice Control Chemicals
74 4.1.9 Corrosion Inhibitors
75 4.1.10 Discussion of Atmospheric Corrosion of Snow and Ice Control
Chemicals
76 4.2 Other Impacts to Infrastructure
76 4.2.1 Railway Traffic Control Signaling
77 4.2.2 Power Distribution Lines
79 CHAPTER 5 Analytical Methods Used to Measure Attributes
of Snow and Ice Control Materials
79 5.1 Specification Sources
79 5.1.1 Pacific Northwest Snowfighters (PNS)
79 5.1.2 Strategic Highway Research Program (SHRP)
80 5.1.3 American Society for Testing and Materials (ASTM)
80 5.1.4 American Association of State Highway and Transportation
Officials (AASHTO)
80 5.2 Physical Attributes
80 5.2.1 Specific Gravity
80 5.2.2 Freezing Point and Eutectic Temperature
80 5.2.3 Gradation
80 5.2.4 Particle Degradation
81 5.2.5 Settleability
81 5.3 Chemical and Environmental Attributes
81 5.3.1 Active Ingredient
82 5.3.2 Metals
83 5.3.3 Total Phosphorus
83 5.3.4 Nitrogen
83 5.3.5 Organic Matter (BOD and COD)
83 5.3.6 pH
84 5.3.7 Cyanide
84 5.3.8 Aquatic Toxicity
85 5.4 Evaluation of Concrete-Related Test Methods
85 5.4.1 SHRP H-205.8, Test Method for Rapid Evaluation of Effects of
Deicing Chemicals on Concrete
85 5.4.2 SHRP H-205.9, Test Method for Evaluation of Scaling Effects of
Deicing Chemicals on Concrete Surfaces
86 5.4.3 SHRP H-205.12, Test Method for Evaluation of Corrosive Effects of
Deicing Chemicals on Steel Reinforcement in Concrete
86 5.4.4 Summary of Concrete-Related Test Methods
87 5.5 Evaluation of Atmospheric Corrosion Testing
87 5.5.1 Immersion Testing
89 5.5.2 Poultice Corrosion Testing
89 5.5.3 Salt-Spray Testing
90 5.5.4 Electrochemical Testing
90 5.5.5 Test Track Testing
90 5.5.6 General Comments Regarding Laboratory Corrosion Tests

92 CHAPTER 6 General Assessment of Information

92 6.1 General Conclusion
92 6.1.1 Natural Receiving Environment
92 6.1.2 Infrastructure Corrosion
93 6.1.3 Other Infrastructure
94 6.2 Draft Material Selection Framework
94 6.2.1 Approach and Assumptions
95 6.2.2 Candidate Product Selection and Matrix of Significant Properties
96 6.2.3 Preliminary Qualification
96 6.2.4 Decision Tool
98 6.2.5 Product Selection
99 6.3 Information Gaps
99 6.3.1 Information Gaps—Natural Environment
101 6.3.2 Information Gaps—Infrastructure
102 6.3.3 Information Gaps—Materials
102 6.4 Revised Work Plan
102 6.4.1 Task 7 (Analytical Testing Program)
106 6.4.2 Task 8 (Matrix of Significant Properties)
106 6.4.3 Task 9 (Purchase Specification)
107 6.4.4 Task 10 (Decision Tool Guidelines)
107 6.4.5 Phase II Reporting
109 CHAPTER 7 Analytical Program Overview
111 CHAPTER 8 Chemical Analysis
111 8.1 Organic Matter (BOD and COD)
112 8.2 Nitrogen
112 8.3 Phosphorus
112 8.4 Cyanide
112 8.5 Heavy Metals
112 8.5.1 Arsenic
115 8.5.2 Cadmium
115 8.5.3 Chromium
116 8.5.4 Copper
116 8.5.5 Lead
117 8.5.6 Mercury
117 8.5.7 Nickel
118 8.5.8 Selenium
118 8.5.9 Silver
119 8.5.10 Zinc
119 8.6 Chemical Analysis Summary
122 CHAPTER 9 Aquatic Toxicity Testing
127 CHAPTER 10 Effects of Snow and Ice Control Materials on
Reinforcing Steel Corrosion
127 10.1 Introduction
127 10.2 Background
128 10.3 Laboratory Program
128 10.3.1 Concrete Properties
128 10.3.2 Snow and Ice Control Materials
129 10.3.3 ASTM G 109 Exposure

131 10.3.4 Chloride Diffusion—ASTM C 1556

131 10.4 Results and Discussion
133 10.5 Conclusions and Recommendations
135 CHAPTER 11 Atmospheric Corrosion Testing
135 11.1 Introduction
136 11.2 Technology Overview
136 11.2.1 Electrochemistry
136 11.2.2 LPR
136 11.2.3 Weight Loss
136 11.3 Test Method
137 11.4 Test Program Evaluation
139 11.5 Data Assessment
140 11.5.1 Structural Steel, A36
140 11.5.2 Wrought Aluminum, AA6061-T6
141 11.5.3 Cast Aluminum AA356.2
142 11.5.4 Free Machining Brass, CDA 360
143 11.5.5 Magnesium Alloy, AZ91C-T6
143 11.6 Conclusions
145 11.7 Proposed Testing Program
146 References

SUMMARY
Guidelines for the Selection of

Snow and Ice Control Materials to
Mitigate Environmental Impacts
Every year, considerable quantities of snow and ice control products are applied to high-
ways; environmental and regulatory agencies have questioned the environmental effects of
these products. Transportation agencies are asked to use “environmentally friendly” or less
toxic alternatives wherever possible, but there is no commonly accepted guidance for deter-
mining which products meet these criteria. The traditional use of road salt has been pro-
hibited in some locations, leaving highway agencies uncertain about how traffic safety can
be maintained in winter conditions.
Studies of the most common chemical alternatives—sodium chloride (salt, NaCl), mag-
nesium chloride (MgCl2), calcium chloride (CaCl2), calcium magnesium acetate (CMA), and
potassium acetate (KA)—have focused on performance and cost under various weather con-
ditions without evaluating their relative effects on the environment. Several new chemical
preparations, including some that are proprietary formulations, have entered the market as
snow and ice control chemicals for use by transportation agencies, but there is limited infor-
mation about the environmental effects of these preparations.
Rational decision-making guidelines are needed to assist winter maintenance managers
in selecting the most appropriate snow and ice control materials for the conditions that
exist in their jurisdictions. A transportation agency must consider a range of factors when
assessing snow and ice control materials for use, including performance, cost, potential for
the material to impair the natural receiving environment, and potential for the material to
impair infrastructure. Each agency will have unique objectives and conditions that influ-
ence the importance of each of these factors in the decision-making process. The overall
objective of this project is to develop guidelines for selecting snow and ice control materi-
als based on their properties and common site-specific conditions near roadways on which
these products would be used.
This study was completed in two phases. Phase I involved collection, compilation, and
quantification of relevant technical information related to snow and ice control materi-
als. Phase II supplemented the information compiled in Phase I through analytical test-
ing programs. The full body of knowledge was applied toward a snow and ice control
material selection process that included a material decision tool, a purchase specification,
and a quality assurance monitoring plan.
NCHRP Project 06-16 Phase I

Phase I of NCHRP Project 06-16 involved a comprehensive search to assemble infor-
mation related to snow and ice control materials. The review included published infor-
mation, current practices, research in progress, and a survey of highway maintenance
professionals. The following topics were covered:

• Available materials, including typical costs, handling concerns, and application issues;
• Impacts to the environment and infrastructure related to common use, application, stor-
age and disposal;
• Site-specific conditions that affect choices of materials;
• Product loading and fate/transport processes;
• Current and proposed specifications and regulations for material acceptance; and
• Parameters and test methods used as measures of performance and environmental
properties.
The gathered information was categorized and quantified. Relevant product categories
included NaCl, CaCl2, MgCl2, acetates (CMA and KA), organic matter from biomass (OMB),
and abrasives. The most significant factors to be considered when selecting snow and ice con-
trol materials were determined to be cost, performance, effects on the environment, and effects
on infrastructure.
How great an effect a snow and ice control material may have depends on such conditions
as material type, application amounts, exposure pathways, chemical-specific impacts, and site-
specific characteristics. A simple model was employed to determine that runoff from the road-
way would be diluted 500 times at the point a snow and ice control material leaves the roadway.
Areas of greatest environmental concern included water quality and aquatic life, air qual-
ity, and vegetation impacts. Contamination of drinking water supplies could occur under
some circumstances, especially for wells very close to areas of routine snow and ice control
materials applications. Direct effects on soils are generally not a concern, but snow and ice
control material can influence soil structure, and the presence of these materials in soil can
further affect water systems and vegetation.
Atmospheric corrosion processes on exposed metals are extremely complex. Most informa-
tion depends on specific conditions, and study conclusions can be wide ranging and, in some
cases, contradict one another. Corrosion related to a given snow and ice control material may
show various responses between product concentrations, metal types, and even between alloys
within the same metal group. Despite the weakness in specific information, some general
assessments exist:
• Chloride-based snow and ice control chemicals are the most corrosive to metals with the
hygroscopic chlorides of magnesium and calcium being the most aggressive because of
the longer time of wetness;
• Acetate-based and organic snow and ice control materials tend to cause less corrosion than
chloride-based chemicals, but can still accelerate corrosion by increasing the conductiv-
ity of the moisture;
• Additives used to inhibit corrosion that are tested in the laboratory can show reductions
in corrosion rate of the metals being tested, but they may show little or no inhibiting effect
on other metals (even those in the same metal group) and may accelerate corrosion of
other metals.
Snow and ice control chemicals can have detrimental effects on concrete infrastructure.
These can be separated into three general categories:
• The enabling and/or acceleration of corrosion of concrete reinforcing (rebar) resulting

from chloride ion penetration. This effect is primarily related to bridges, but there is a
possibility of a secondary risk in concrete pavements with doweled joints or continuous
reinforcement.
• Reactions of snow and ice control materials with the cement paste, resulting in deteriora-
tion of the cement matrix. This effect is primarily a risk for concrete pavements and the

amount of deterioration is directly related to the quality of concrete. Current evidence

suggests that magnesium chemically reacts with the cement paste in a process that reduces
concrete strength.
• All snow and ice control chemicals by definition of their action increase the number of
freezing and thawing cycles on the concrete in service. This can result in deterioration in
the form of freezing and thawing and/or scaling.
It was expected that the findings of the information assessment conducted through Phase I
of this study would identify gaps in information. In general
• Technologies and products employed for snow and ice control are continually evolving,
and current information is not always available.
• For many of the laboratory studies conducted, limited field correlations are available.
• Many materials have had a relatively short history of use (less than 10 to 15 years) and it
is unclear what the potential for long-term adverse effects are over decade time scales.
NCHRP Project 06-16 Phase II

Phase II of this study began with an analytical testing program. The information gathered
through this program was combined with the information developed through Phase I to
develop a practical snow and ice control material selection process that included a material
decision tool, an environmental purchase specification, and a quality assurance monitoring
program.
Analytical Program
The analytical program was designed to help address gaps in information and to help
establish a matrix of significant properties for common snow and ice control chemicals
in the areas of environmental chemistry, aquatic toxicity, concrete corrosion, and atmos-
pheric corrosion.
Forty-two products were obtained that represented the 2004 U.S. and Canadian market
in terms of product type (NaCl, MgCl2, CaCl2, CMA, and KA), manufacturers, regional
sources, and additives. Abrasives were not included in the test program, given that material
sources are too numerous and varied. All products received full chemical analysis. Fifteen
materials representing each material type and common additive combination were selected
for aquatic toxicity, concrete corrosion, and atmospheric corrosion assessments.
The benefit of conducting a large-scale analytical assessment on a wide range of snow and
ice control materials using the same test methods is that it provides data that can be directly
compared for trends so that significant material attributes can be identified.
Chemical Analysis
From the chemical analysis program, the following general conclusions can be made:
• High organic matter concentrations, which can lead to biochemical oxygen demand
(BOD) concerns in aquatic systems, are associated with acetate materials and with mate-
rials containing organic matter from biomass (OMB). For the latter, levels depended on
the proportion of OMB in the blend. Solids with OMB generally had much lower organic
matter levels than liquid counterparts.
• Nutrient (nitrogen and phosphorus) levels found in many of the materials tested can be
considered a potential environmental concern. Nutrient levels had a strong relationship

to the presence of OMB materials in the product blend. Higher nutrient levels were also
present in other material types.
• Cyanide could be a potential environmental concern for materials, such as solid NaCl,
that routinely contain iron-cyanide anti-caking agents. More research and information
is necessary to assess the bio-availability of iron-cyanide materials under field condi-
tions conclusively.
• Overall, metals levels were below levels considered to be a significant environmental con-
cern, although the literature indicates that higher metals levels have routinely been
observed in snow and ice control materials.
• The variability in the data between products, and the general lack of strong trends, out-
side of those already listed, suggests that it is important to continue to monitor chemical
attributes of snow and ice control materials through a quality assurance monitoring plan.
Aquatic Toxicity
Aquatic toxicity testing was conducted on fifteen snow and ice control products follow-
ing U.S. EPA standard test methods for chronic exposure to various levels of aquatic biota,
including vertebrates, invertebrates, and algae:
• Fathead Minnow 7-day Survival and Growth Test, which also included acute toxicity data
for Fathead Minnow 96-hour Survival Test;
• Ceriodaphnia dubia 3-brood Survival and Reproduction Test; and
• Selenastrum capricornutum 96-hour Growth Test
The tolerance thresholds determined from the testing were compared with the expected
dosage that organisms would receive in the field based on dilution at the margin of a road-
way of 500:1—this would be close to the amount of dilution for many applications.
Although it is likely that the most sensitive organisms would be affected in some way at
dilutions that could be expected at the margin of a roadway for many snow and ice control
materials, it is also true that dilutions greater than 500:1 can be expected within a relatively
short distance beyond the roadway. Therefore, the toxicity data imply that there could be an
impact zone relatively close to the roadway, bounded by dilutions that offset acute or chronic
toxicities close to the roadway.
Atmospheric Corrosion
Linear Polarization Resistance (LPR), a mature, rapid, and sensitive corrosion rate meas-
urement technique, was used to assess atmospheric corrosion rates. Of particular benefit was
the potential suitability of LPR for evaluating the corrosiveness of inhibited snow and ice
control materials and the effect of roadside dilution.
Fifteen snow and ice control materials with and without inhibitors were tested on five
commonly available alloys representative of materials used in highway infrastructure and
vehicles. These included structural steel alloy (A36), magnesium alloy (AZ91C-T6), wrought
aluminum alloy (AA6061-T6), cast aluminum alloy (AA356.2), and brass alloy (CDA 360).
To represent material concentrations expected in the field, LPR assessments were conducted
on liquid snow and ice control materials at 100 percent (v/v–percent volume per volume)
(as received), 10 percent (v/v) and 1 percent (v/v) concentrations for each alloy. The follow-
ing conclusions were obtained through this study:
• Overall, the structural steel alloy (A36) demonstrated the lowest corrosion rate values
while the magnesium alloy (AZ91C-T6) demonstrated the highest corrosion rate values.

Wrought aluminum alloy (AA6061-T6), cast aluminum alloy (AA356.2), and brass alloy
(CDA 360) demonstrated intermediate corrosion rate values.
• With the exception of structural steel (A36) and brass (CDA 360), corrosion rates
increased with increasing snow and ice control material concentration. Structural steel
corrosion rates were sufficiently low that identification of trends may have been obscured
by measurement variability at these levels. Brass corrosion rates were highest when snow
and ice control materials were at the 10-percent concentration level.
• Overall, chloride-based snow and ice control materials (NaCl, CaCl2, MgCl2, and blended
chlorides) displayed similar corrosion rates for each metal type tested at each concentra-
tion level. This suggests little difference in corrosion rates between chloride-based snow
and ice control materials. Field conditions along with other snow and ice control material
properties (such as hygroscopicity) may contribute more to the relative corrosiveness of
these materials in the service environment. Other corrosion mechanisms not emulated by
aqueous testing may be significant.
• Acetate materials (CMA and KA) demonstrated a range of LPR corrosion rates for the
alloys assessed:
– The lowest corrosion rates were for structural steel (A36). Observed values bordered
on the minimum practical detection ability of test instrumentation for all acetate types.
– Acetates were equally or slightly less corrosive than chlorides to magnesium (AZ91C-T6).
– Acetates were equally or slightly more corrosive than chlorides to brass (CDA 360).
– For aluminum alloys (AA356.2 and AA6061-T6), acetates containing potassium (KA)
demonstrated corrosion rates equal to chloride-based snow and ice control materials.
CMA demonstrated a much lower corrosion rate for these alloys.
• The effectiveness of corrosion inhibitors was varied and poorly quantified. In some
cases, these materials provided no detectable corrosion improvement; in other cases,
corrosion improvement appeared to depend on the concentration level of the test solu-
tion. This suggests that although some benefit may be realized from the use of these
materials, the overall effectiveness under field conditions (including dilution) and after
exposure to numerous potential alloy types is questionable.
Concrete Corrosion
The concrete corrosion analytical testing program was initiated in June, 2004, and was
based on the following standard tests:
1. ASTM G 109 with the objective of relatively assessing fifteen snow and ice control mate-
rials for their potential to initiate corrosion on embedded reinforcing;
2. ASTM C 1556 with the objective of providing a relative assessment of chloride ion diffu-
sion rates for MgCl2, CaCl2, and NaCl.
In designing this program, it was recognized that similar studies conducted in the past
used poor quality concrete, mortars, and pastes in order to reduce the test duration. These
materials do not represent concrete used in real-world structures and it was important that
the current project use a more realistic concrete mix design (higher quality, w/c < 0.40, air
content 5–8%).
Initial estimates, extrapolated from past studies, were that corrosion initiation would
occur 1 to 2 years after test commencement. Regrettably, after 2 years, test samples have not
yet produced detectable changes in corrosion potential. Acknowledging the time and budget
constraints of NCHRP Project 06-16, it was decided to complete the project without the con-
tribution from the concrete corrosion testing component.

Decision Tool
When selecting a snow and ice control material, the factors affecting the decision processes
can be numerous and complex. The framework provided here is relatively straightforward
and addresses the most common and significant items. Decision categories include cost,
performance, potential to impair the natural environment, and potential to impair infra-
structure. These balance economic value related to cost and performance with potential con-
sequences of use related to environmental and corrosion impacts. User input is a key element
to the tool to ensure that maintenance area conditions are represented. The end product is
a practical tool that generates a numerical evaluation that can be used to compare snow and
ice control materials.
Each agency will have unique objectives and priorities that will affect their selections for
the decision categories and these will be expressed through percentage weighting of impor-
tance. Decision subcategories that support the natural environment and infrastructure deci-
sion categories are relatively straightforward and represent the most significant components
contributing to the higher level decision category.
To quantify risks of impairment for environmental and infrastructure subcategories
resulting from each snow and ice control material type, numerical rankings were developed.
These were based on a material’s potential to elicit an effect on a receptor, as developed from
a comprehensive information review and analytical test results. These values were applied
where sensitive receptors existed in the maintenance area as identified through user-com-
pleted receptor inventory worksheets for each decision subcategory.
When assessing relative cost and performance of snow and ice control materials, many
factors are influential and considerable research continues to be done here. The Decision
Tool uses the following approaches:
• Cost is based on the purchase price of the dry active ingredient. Active ingredients include
the five primary snow and ice control chemicals: NaCl, MgCl2, CaCl2, CMA, and KA.
• Performance is based on melting potential derived from phase curve information and
expected winter temperature in the maintenance region. Melting potential represents the
chemical’s ability to melt ice at a specified temperature.
These approaches may not be ideal; however, they provide a standardized and practical
method that reasonably represents the dominant component for each decision category.
For each decision category, a weighted score is generated. These are summed for each
material to produce a numerical product score that is used to compare snow and ice control
material types objectively. Once a material type has been selected, various sources may exist.
Considerable variation in certain attributes can exist among similar material types from dif-
ferent sources or between batches from the same source. A Purchase Specification contain-
ing concentration-based guidelines for key attributes will allow transportation agencies to
screen out unsuitable products. The Purchase Specification is supported by a Quality Assur-
ance Monitoring Program to assess received materials and by standard test methods to meas-
ure these attributes.
Product selection occurs after a review of the Purchase Specification and supplemental
considerations that may include but are not limited to product availability (delivery times
and quantities), quality testing results, and equipment availability.
Purchase Specification
This Purchase Specification was developed to assist transportation agencies with mitigat-
ing effects to the natural environment by allowing the user to qualify and compare snow and
ice control materials based on key environmental attributes. The Purchase Specification is

to be applied after the agency has completed the Decision Tool and a snow and ice control
material type has been selected for use.
Aquatic systems are the primary environmental concern associated with snow and ice con-
trol materials. The Purchase Specification focuses on snow and ice control material chemi-
cal attributes that could impair aquatic systems under normal use but are not addressed
through the Decision Tool. Numerical criteria specific to snow and ice control materials have
been established based on
• U.S. EPA numerical water quality criteria for the protection of aquatic life and drinking
water (based on U.S. EPA National Recommended Water Quality Criteria: 2002, EPA-822-
R-02-047, November 2002).
• Site-specific factors, specifically receiving water type, size, and current nutrient levels.
• The default value for expected roadway runoff dilution is estimated to be 500 times at the
point the materials leave the roadway. This may be adjusted up or down to reflect any other
assumed or measured dilution factor.
The Guidelines for Maximum Concentrations (GMCs) of constituents in snow and ice
control materials are shown in Table 3-2. These concentrations were derived assuming
that the concentrations of the applied materials are diluted 500 times at the point these
materials leave the roadway. Three general categories exist according to type of snow and
ice control material:
1. The first category includes chemical products with a calcium or magnesium base,
which are the two components of hardness for water. Concentration limits (i.e.,
aquatic life criteria) for certain metals (e.g., Cd, Cu, Pb, Ni, Ag, Zn) depend on hard-
ness, which can reduce their relative toxicity. Therefore, the maximum allowable con-
centrations for these metals are highest when calcium and magnesium are major
components of ice and snow control materials.
2. A second category applies to snow and ice control chemicals that contain neither calcium
nor magnesium as major constituents (NaCl, KA, agriculturally derived organic materi-
als). Such products lack any hardness-inducing component in large quantities, which
requires more restrictive limits for metals whose toxicity is affected by hardness.
3. The third category of snow and ice control materials is abrasives. These materials are not
soluble, but may contain contaminants that are soluble. Abrasives obtained from indus-
trial processes (such as metallurgical slag, mining waste, and furnace slag) can contain
high toxicant levels. GMC values are based on the assumption that abrasives to be tested
will be leached in pH-adjusted (acidic) water according to U.S. EPA Method 1311, Toxi-
city Characteristic Leaching Procedure (Revision 0, July 1992), which incorporates a liq-
uid-to-solid ratio of 20:1(50 grams abrasive plus 1 L extraction fluid).
Quality Assurance Monitoring Program

The Quality Assurance (QA) Monitoring Program takes the form of a suite of qualifying
tests and submittals for the agency to characterize snow and ice control products before pur-
chase or use. This section includes procedures for collecting and testing received snow and
ice control chemicals. Standard test methods used to quantify material characteristics are
also referenced.
To ensure suitability of the products with agency purchase specifications, it is suggested
that QA procedures be included in contract documentation.

CHAPTER 1
Introduction
1.1 Problem Statement aquatic life; vegetation, animals, and aquatic environments;
and Research Approach air quality; vehicles; and physical infrastructure including
bridges, concrete roadways and structures, railway electronic
Every year, considerable quantities of snow and ice control traffic signaling systems, and power distribution lines.
products are applied to highways, and environmental and
regulatory agencies have questioned the environmental
1.2 Scope of Study
effects of these products. Transportation agencies are asked to
use “environmentally friendly” or less toxic alternatives wher- Phase I of NCHRP Project 06-16 involved collection, com-
ever possible, but there is no commonly accepted guidance for pilation, and quantification of relevant technical information
determining which products meet these criteria. The tradi- about snow and ice control materials. This information was to
tional use of roadsalt has been prohibited in some locations, be used in preparing a draft report to include a framework for
leaving highway agencies uncertain about how traffic safety a proposed material selection guideline, an assessment of
can be maintained in bad weather. For example, Environment applicable analytical test methods, identification of areas where
Canada has concluded that inorganic chloride roadsalts are insufficient information is currently available, and a revised
harmful to the environment under the Canadian Environ- work plan. Phase I covered Tasks 1 through 6 of the NCHRP
mental Protection Act (CEPA), thus requiring development Project 06-16 research plan. These tasks were as follows:
and implementation of improved management practices.
Studies of the most common chemical alternatives— 1. Task 1 (Information Search). This task involved a com-
sodium chloride (salt), magnesium chloride, calcium chlo- prehensive search to assemble information about snow
ride, calcium magnesium acetate, potassium acetate, and and ice control materials. The review included published
urea—have focused on performance and cost under various information, current practices, and research in progress
weather conditions without evaluating their relative effects on and covered the following topics:
the environment. Several new chemical preparations, includ- • Available materials, including typical costs, handling
ing some that are proprietary formulations, have entered the concerns, and application issues;
market as snow and ice control chemicals for use by trans- • Environmental impacts related to common use, appli-
portation agencies, but there is limited information about cation, storage, and disposal;
their environmental impacts. There is a need for rational • Site-specific conditions that affect the choices of materials;
decision-making guidelines to assist DOT maintenance man- • Product loading and fate/transport processes;
agers in selecting the most appropriate snow and ice control • Current and proposed specifications and regulations for
materials for the conditions that exist in their jurisdictions. material acceptance; and
The overall objective of this project was to develop guide- • Parameters and test methods used as measures of per-
lines for selection of snow and ice control chemicals and abra- formance and environmental properties.
sives, based on their composition, performance, potential One component of this task involved developing a
environmental effects, and cost of the product together with survey questionnaire that was distributed to state TRB
common site-specific conditions near roadways on which research advisory panel members, provincial highway
these products would be used. For this study, environmental agencies, other roadway maintenance agencies, and
impacts were defined to include effects on human health; vendors and distributors of snow and ice control products.

The survey sought information on the current status of use was used in Phase II of the study to create a decision tool
for snow and ice control materials, effects on the natural for selecting products for use in common situations.
receiving environment, effects on infrastructure, product 4. Task 4 (Review of Analytical Test Methods). This task
selection processes, and products commercially available identified and justified the test methods that were most
for use. The research team sought follow-up information effective for measuring the properties and constituents
from survey respondents via telephone as required. The that determine the environmental impacts of the current
findings are provided in Chapters 2 through 4. range of snow and ice control materials. The potential
2. Task 2 (Assessment of Information). This task summa- for environmental impacts resulting from snow and ice
rized the information gathered from Task 1 and catego- control products can be assessed through analytical meas-
rized and quantified the most significant factors that urement of product properties and constituents and com-
should be considered in the selection of snow and ice paring these measurements to known specifications and
control materials. A summary of findings is included at criteria. Currently numerous test procedures are being
the end of each area reviewed, and a general assessment used to measure performance and environmental impacts
is provided in Chapter 6. of snow and ice control materials, and it is not uncommon
3. Task 3 (Draft Framework for Product Selection). This for laboratories to provide technically valid, yet conflict-
task applied the most significant factors identified in ing test results. In order to assess the potential for risk of a
Task 2 in a draft framework for the selection of snow and product and to ensure comparability of test results among
ice control materials. The objective of this task was to products, test facilities, and jurisdictions, the parameters
provide the user with a practical process and the necessary of concern and the methods for measurement must be
information to make a decision given a user-defined set of well defined. Review of the analytical test methods is pro-
goals—it was not the intent to be completely prescriptive. vided in Chapter 5.
Product rankings were developed by assessing user- 5. Task 5 (Identification of Gaps in Information). This task
identified site-specific factors and product attributes identified data gaps and needs, from Tasks 1 through 4,
known to affect these factors. Product attributes were that exist because snow and ice control strategies are
further developed through Phase II of this study. For continually evolving, new products are being developed,
example, a material attribute known to adversely affect and research efforts are providing new information. Infor-
aquatic environments might receive a less favorable mation gaps are summarized in Section 6.3.
ranking if such areas were present. Product rankings 6. Task 6 (Interim Report). After completing Tasks 1 through
were further processed through user-weighted product 5, the research team prepared an interim (Phase I) report
selection categories. Final product scores were developed that summarized the information obtained and assessed,
and used to compare the merits of individual products. provided a draft framework for snow and ice control mate-
The draft framework for materials selection is provided rial selection, reviewed the analytical test method used to
in Section 6.2. The draft framework for product selection assess these materials, and identified gaps in information.

10
CHAPTER 2
Operational and Performance Considerations
2.1 Products Used and Properties product sales volumes is believed to result from there being
no trade association for the suppliers of the numerous pro-
Snow and ice control materials include various chemical prietary blenders and resellers.
product types as well as abrasives. These have been classified
as shown in Table 2-1.
2.1.1 Chloride Salts
Chloride salts include sodium chloride (NaCl), calcium
chloride (CaCl2), and magnesium chloride (MgCl2). Organic Chloride-based snow and ice control materials are gener-
products include acetates (e.g., calcium magnesium acetate ally produced from the mining of surface or underground
[CMA] and potassium acetate [KA]) and other organic mate- deposits, extracting and fractionating well brines, industrial
rials that mostly include agricultural by-products. Manufac- by-products, or through solarizing saltwater. Natural well
tured organic materials (e.g., glycols, methanol, and nitrogen brines are usually, but not always, associated with petroleum
products) are most commonly used for airport applications and natural gas production and are found in locations
and see limited roadway applications. throughout the United States and Canada, including Alberta
Some hybrid mixtures containing chloride salts and and Texas. Generally, these brines are composed of several
organic products are often available in both solid and liquid chemicals and usually are processed to extract various com-
form. These typically involve the addition of an agriculturally mercial products, including snow and ice control materials.
derived organic product to a chloride salt with the intent of Solarization is the ponding of natural salt brines from sea-
reducing corrosion. Additionally, snow and ice control mate- water or salt lakes (e.g., Salt Lake in Utah and the Dead Sea in
rials (i.e., NaCl) are often purchased by an agency in solid Israel) and allowing solar energy to evaporate the water. The
form and made into brine using onsite equipment. various products crystallize at different concentrations
Snow and ice control materials can be characterized by (depending on solubility), allowing the manufacturer to sep-
their primary components, the form in which they are used, arate the various chemicals from the original brine.
and the optimum eutectic temperature. Table 2-2 lists the
general properties for chloride-based snow and ice control
2.1.1.1 NaCl Solid (Roadsalt) and Liquid (Brine)
chemicals. The optimum eutectic temperature for a given
product is the lowest temperature at which a product will Roadsalt or sodium chloride is the most commonly used
freeze when at the optimum ratio of chemical to water (1). snow and ice control chemical. Most of the roadsalt used in
This can also be stated as the lowest freeze-point achievable the United States and Canada comes from underground min-
by a given chemical through an optimum ratio of chemical to ing, solution mines, and solarization.
water. The prices shown in Table 2-2 are from contracts pub- Solution mining is a process of injecting hot water into a
lished in 2003 available on the Internet. These prices should subsurface deposit of salt or other chlorides through a well.
be considered as representative values only and may vary, pri- The hot water forms a solution that is pumped to the surface
marily because of shipping costs. For example, Salt Lake City, and evaporated to recover the various chemicals. Salt is mined
Utah, salt prices were $13.80 in U.S. dollars (USD) per U.S. in Alberta, New Brunswick, Nova Scotia, New York, Kansas,
ton and in Yellow Knife, Northwest Territories, the prices were Louisiana, and other areas of the United States and Canada.
$85.37 (USD) per ton. Annual totals for quantities used are Solarization is used to produce salt in San Francisco, Salt Lake
not available, except for roadsalt. The lack of data on the City, Louisiana, Mexico, and Chile.

11
Table 2-1. Snow and ice control materials. salt works in San Francisco Bay. Some magnesium chloride is
produced as a by-product from metallurgical processes, but
Snow and Ice
Material Type
Control Material
Primary Components these sources may contain sufficient trace metal contami-
Sodium Chloride (NaCl) Na, Cl nants (e.g., arsenic) to be of concern and should be analyzed
Chloride Salts carefully before being accepted for use.
Calcium Chloride (CaCl2) Ca, Cl
Magnesium Chloride (MgCl2)
Mg, Cl
Calcium Magnesium Acetate 2.1.1.4 Blended Chlorides
Ca, Mg, C2H3O2
(CMA)
Potassium Acetate (KA) K, C2H3O2 There are two types of blended chlorides: those that
Organic Products
Agricultural By-Products Complex sugars
occur naturally and those that are custom-blended. The
Manufactured Organic Varies with product (i.e.
natural-blended chlorides come from complex well brines
Materials glycol, methanol) and natural deposits of mixed chlorides such as the salt flats
Nitrogen Products Urea Urea, Ammonia in Utah. The natural sources are usually not highly
Varies with the source processed and thus their chemical constituents can be vari-
Abrasive Abrasives
of the material
able. Custom-blended chlorides usually consist of two to
three chemicals. NaCl is typically the major component,
2.1.1.2 Calcium Chloride (CaCl2) with other chlorides added to improve low-temperature
performance and/or the hygroscopic properties of the
Calcium chloride (CaCl2) is primarily produced from
blend. Many of these custom-blended products also include
natural well brines and as a by-product of the Solvay process.
a corrosion inhibitor.
The Solvay process involves combining sodium chloride and
calcium carbonate (CaCO3) to produce sodium carbonate
(Na2CO3) and CaCl2. 2.1.2 Organic Products
Organic snow and ice control products are, for the most
part, manufactured products. Some are wholly synthesized
2.1.1.3 Magnesium Chloride (MgCl2)
(e.g., CMA and KA); others are refined from agricultural
Magnesium chloride (MgCl2) is most commonly obtained sources (e.g., by-products from grain processing, brewing,
through solarizing natural salt brines. Magnesium chloride winemaking, and similar sources). Table 2-3 lists the general
brine, along with magnesium sulfate (MgSO4) is the most properties for organic snow and ice control products. The
soluble of the components in the natural brines. The most optimum eutectic temperature for a given product is the
common sources of magnesium chloride in the United States lowest temperature at which a product will freeze when at
and Canada are the Great Salt Lake in Utah and the extensive the optimum ratio of chemical to water (1). This can also be
Table 2-2. Chloride salts, general properties.

Median Cost
Optimum Eutectic Approximate
(USD) per
Chemical Temperature Annual usage
Material Forms Used Ton (survey
Formula °C (°F) @ % Common Sources Tonnes (Tons)
of Internet
Concentration 1 North America
contracts)
2
Sodium Chloride NaCl Primarily solid, -21 (-5.8) @ 23.3% Mined from natural deposits, 21,080,000
but increasing use solarization of natural brines (22,291,000) $ 36
of liquid (Salt Institute)
Calcium Chloride CaCl2 Mostly liquid -51 (-60) @ 29.8% Natural well brines, by-product
brine, some solid of the Solvay process
Not Available $120
flake
Magnesium MgCl2 Mostly liquid -33 (-28) @ 21.6% Solarization of natural brines,
Chloride brine, some solid natural well brines, by-product
Not Available $ 95
flake of metallurgical process
Blended Chlorides Varies with Solid and liquid Varies with product Natural well brines, solarization
product of natural brines, mined from
Not Available $142
natural deposits
1
Source: (2)
2
as of October 2003

12
Table 2-3. Organic products, general properties.

Median Cost
Optimum Eutectic Approximate
(USD) per
Chemical Temperature Annual usage
Material Forms used Ton (survey
Formula °C (°F) @ % Common Source(s) Tonnes (Tons)
of Internet
Concentration 1 North America
contracts)2
Reaction of Highly
Calcium Concentrated Acetic Not Available $1280
Mostly liquid Acid with Dolomite
Magnesium CaMgAc -27.5 (-17.5) @ 32.5%
with some solid Limestone
Acetate
Reaction of Highly
Concentrated Acetic Not Available Not
Acid with caustic Available
Potassium
KAc Liquid only -60 (-76) @ 49% potash (KOH). This
Acetate
reaction produces
potassium acetate and
water.
Refined from Blends
Agricultural Usually blended with Agricultural base Not Available $ 108
NA Liquid only materials
By-Products chloride-based products
Other Organic Glycols Not Available Not

Liquid only Varies with product Varies Available
Materials Methanol
1
Source: (2)
2
As of October 2003
stated as the lowest freeze-point achievable by a given chem- 2.1.2.2 Potassium Acetate (KA)
ical through an optimum ratio of chemical to water. The
Potassium acetate is a commercially available chemical
prices shown in Table 2-2 are from contracts published in
used in numerous chemical and pharmaceutical processes.
2003 available on the Internet. These prices should be con-
It is a non-chloride, high-performance product originally
sidered as representative values only and may vary, primarily
designed for use as a runway deicer. Potassium acetate is com-
because of shipping costs.
mercially manufactured using the same high-grade acetic
acid as CMA; therefore, cost is an issue when considering
2.1.2.1 Calcium Magnesium Acetate (CMA) roadway snow and ice control.
CMA was the result of an FHWA effort to find a low-
corrosion biodegradable substitute for roadsalt. Although 2.1.2.3 Agricultural By-products (Organic Biomass)
CMA meets the project objectives, it is relatively costly to
Agricultural additives for snow and ice control are refined
produce. CMA is produced by reacting high-grade acetic
from various agricultural feedstock, including corn, wheat,
acid with dolomite limestone. Acetic acid (the acid in vine-
and rice. Currently, these are all proprietary products and,
gar) is the product of fermentation of agricultural (usually thus, little is known of the actual manufacturing/refining
corn) products. Under normal conditions, the fermenta- process. However, in most cases, these products are reduced
tion process can produce a 4- to 5-percent solution of acetic organic biomass. These products are usually used in con-
acid. The production of CMA requires a 25-percent solu- junction with chloride-based products, although stand-alone
tion of acetic acid to be viable. Subsequently, much of the products have been marketed. These products have demon-
cost of CMA is the result of producing high-concentration strated significant corrosion-inhibiting characteristics under
acetic acid. certain conditions, as well as claims of increased overall prod-
Acetic acid can be made in various ways. For example, fer- uct performance.
mentation of corn or other organic products yields a solution
of approximately 4 percent. Other methods, such as air oxi-
2.1.2.4 Other Organic Materials
dation of butane, produce a much higher yield of acetic acid.
The acid is then distilled from the other ingredients to obtain Other organic compounds are used for snow and ice control.
glacial acetic acid. CMA is now made from the direct reaction One group of compounds is alcohols (e.g., methanol or
of acetic acid and dolomitic lime. ethanol). These are usually distilled from organic feedstock,

13
although some can be synthesized from petroleum sources. ing a runoff control system for proper discharge or collection
Alcohols are volatile and flammable and present some storage of materials for reuse. Permanent storage structures should be
and handling concerns. Glycols are another common group of constructed to withstand the pressure from the material and
organic chemicals used for snow and ice control. The most the stress of loaders pushing materials against the inside walls.
common glycol-related compounds encountered are ethylene Alternatively, temporary barriers can be used to protect the
and propylene glycol. These are found in commercial automo- storage structure walls when the original walls alone may not
bile antifreeze products. These compounds are used as aircraft have adequate strength (4).
deicers and have seen limited roadway application.
2.2.2 Liquid Snow and Ice Control Materials
2.1.3 Sand or Abrasives
Liquid storage details include adequate tank capacity,
Sanding materials or abrasives are used in most areas of proper-sized pumps and hoses for quick loading, and recir-
Canada and the United States, commonly as the primary culation capability to maintain product consistency should
method to enhance traction on ice-covered roadways. Abra- settling occur. Liquid chemical storage should include con-
sives provide no ice-melting capabilities. The chemistry of tainment barriers sufficient to contain and recapture spills or
sanding materials depends on the original source. Sanding the volume released from a tank rupture.
materials or abrasives are produced by crushing natural
deposits of stone, screened or crushed from natural deposits
2.2.3 Sand or Abrasives
or from scoreaceous volcanic deposits (sometimes called
cinders in the western areas of the United States and Canada). The storage of sanding materials entails having appropri-
Other sources of sanding materials are metallurgical slag ate loading equipment and adequate access to the stockpile.
(often referred to as cinders in the eastern areas of the In many cases, abrasive piles are treated with liquid snow and
United States and Canada), “clinker” ash from coal-fired ice control chemicals to prevent freezing. In such cases,
boilers, and natural river sands. The approximate annual use proper runoff protection is required.
is unknown because of the large number of commercial and
non-commercial sources.
2.3 Handling and Application
Methods
2.2 Storage
In general, the exposure effects of snow and ice control chem-
Material storage is an important operational and cost con- icals are relatively mild and key concerns are stated on Material
sideration and may also be of concern because of releases of Safety Data Sheets (MSDS). Most products can produce dust in
stored materials into the natural receiving environment. their dry form and may irritate the respiratory system. Eye and
Unwanted releases of snow and ice control materials from skin irritation is a common concern when handling snow and
storage facilities have contributed to roadsalts being placed on ice control chemicals in liquid form. Eye, skin, and respiration
the Priority Substances List under the Canadian Environ- protection is recommended under certain conditions.
mental Protection Act (CEPA) (3). Appropriate equipment is needed for handling and apply-
ing snow and ice control chemicals. Application equipment
can vary greatly from new equipment to equipment built to
2.2.1 Solid Snow and Ice Control Materials
customized versions of existing equipment. Calibration and
Proper storage of solid snow and ice control material regular re-calibration of equipment is of the utmost impor-
entails having appropriate receiving and loading equipment, tance to achieve control of the rate of application.
adequate access to the stockpile, and proper protection
against escape of chemicals or leachate (4). Solid snow and ice
2.4 Agency Survey—Current
control chemicals should always be stored inside to prevent
and Future Status of Use in the
runoff of salts dissolved by precipitation. Storage structures
United States and Canada
should be constructed on an impermeable pad of asphalt,
concrete, or other suitable material graded away from the A survey was distributed to assist in defining snow and ice
center of the storage area for drainage. The pad should extend control material use in North America. The questionnaire
around the exterior of the structure and be graded away from was distributed to U.S. DOTs (through the members of the
the building, so runoff does not enter the structure. Storage TRB Research Advisory Council); Canadian provincial trans-
structures should also allow for gathering any stormwater or portation agencies; and various other agencies involved in
runoff that may occur. This can be accomplished by provid- using and distributing snow and ice control materials. The

14
survey (included as Appendix A, which is available for down- among materials. Table 2-4 summarizes the results. Some
load from the TRB website) was designed to gather the fol- respondents listed more than one material as their primary or
lowing information: secondary preference. As an example, three respondents listed
both NaCl solid and NaCl brine as their primary choice. This
• Product use and methods of application, could indicate that the users might be developing storm-
• Environmental concerns, specific strategies for each of the chemicals.
• Corrosion concerns, and
• Product selection process.
2.4.1.1 Chloride-Based Snow and Ice Control
Materials
Twenty-eight organizations (i.e., 22 states, 3 provinces, and
3 cities) responded. Detailed summaries of responses are As expected, solid NaCl was the most commonly used snow
provided in Appendix B, which is available for download and ice control chemical and was considered a first preference
from the TRB website. With a few exceptions, the survey for 57 percent of the respondents. The total quantity of salt used
respondents included the primary U.S. and Canadian agen- (4.7 million tons) when compared with other products further
cies involved in snow and ice control and constituted a good supports this finding. NaCl brine, in most cases produced on
geographical cross section from east to west. site by the agency, was a first or second choice for 43 percent of
the respondents. Several respondents had both salt and salt
brine as their first preference.
2.4.1 Product Use and Methods
Magnesium chloride, used as a liquid, was the next pre-
of Application
ferred chemical with 14 of the respondents rating it as either
Respondents were asked what snow and ice control mate- their first, second, or third preference (50 percent). Calcium
rials were used and what their preferences (rankings) were chloride, primarily used as a liquid, was the next most popu-
Table 2-4. Snow and ice control material preference.
1st 2nd 3rd 4th 5th 6th

Material
Choice Choice Choice Choice Choice Choice
NaCl solid 57% (16) 18% (5) 4% (1) 0 0 0
NaCl brine 11% (3) 32% (9) 7% (2) 0 4% (1) 0
Salt-based solid products

4% (1) 4% (1) 0 0 0 0
plus other ingredients
Chloride-based brines plus

0 4% (1) 0 4% (1) 7% (2) 0
organic additive
CaCl2 7% (2) 18% (5) 18% (5) 14% (4) 0 0
MgCl2 14% (4) 7% (2) 29% (8) 0 14% (4) 0
CMA 0 4% (1) 0 7% (2) 0 0
KA 4% (1) 7% (2) 0 0 0 4% (1)
Abrasives 21% (6) 18% (5) 7% (2) 11% (3) 7% (2) 4% (1)
Abrasive/NaCl mixture 4% (1) 0 0 0 0 0
Sand mixed with salt solids

0 4% (1) 0 0 0 0
plus inhibitor
Source: Agency survey (based on current practices)

15
lar, with 12 of the respondents rating it as either their first, storage practices, and applicable environmental regulations
second, or third preference (43 percent). for use of snow and ice control materials.
For the most part, respondents reported MgCl2 and CaCl2
to be used as blended products containing organic materials
2.4.2.1 Areas of Environmental Concern
used as corrosion inhibitors. A natural product containing a
combination of sodium, magnesium, and potassium chlo- The survey asked respondents to prioritize their environ-
rides had some high preferences from western states. This mental concerns by allocating percentage weighting to each
product was used as a stand-alone product and blended with concern. They were also asked if they had modified their
an organic-based corrosion inhibitor. operations to address these problems. Weighted averages
were used because not all of the respondents completed this
section. Survey results are provided in Table 2-5. Water qual-
2.4.1.2 Organic Deicers
ity appears to be the greatest environmental concern, fol-
The purely organic snow and control chemicals include lowed by air quality, vegetation, and endangered species. Air
calcium magnesium acetate (CMA), potassium acetate (KA), quality issues are more important in western regions,
and some proprietary organic products. These materials are whereas water quality is important in eastern regions. A few
not as popular as chloride-based products, but are used as agencies separated subsurface (wells) from surface contam-
either stand-alone liquids, blended with inorganic liquids, or ination as a concern, but for this study, they were combined
as stockpile treatments. In general, they are not the most pre- as aquatic impacts. Survey results indicate that some agen-
ferred products; they tend to be used for special cases (e.g., for cies modified operations to address environmental con-
low-corrosion applications such as bridge decks). cerns; however, no specific information was given as to how
this was accomplished.
2.4.1.3 Hybrid Deicers
2.4.2.2 Materials Storage
Use of hybrid organic/inorganic deicers is limited. This
class of chemicals, which uses a blend of significant amounts Respondents were asked to describe current storage pro-
of both organic and inorganic deicing chemicals, is usually cedures for abrasives, solid chemicals, and liquid chemicals
designed for low-corrosion applications. (see Table 2-6). The data reflect the total number of respon-
dents. Some agencies listed multiple methods of storage for
each category, whereas some did not respond. Responses
2.4.2 Natural Environment Concerns
indicate that, although most abrasives are stored outdoors
The survey covered some issues pertaining to the environ- and uncovered, significant amounts are stored indoors.
ment, including areas of environmental concerns, material Most solid chemicals are stored indoors on impermeable
Table 2-5. Environmental concerns.
Environmental Concern Priority Ranking Modified Operations

Aquatic Impacts (Surface and Subsurface) 40 % 46% (13)
Air Quality 27 % 36% (10)
Vegetation 20 % 25% (7)
Endangered Species 13 % 18% (5)
Other – Public Input 1% 4% (1)
Table 2-6. Materials storage.
Outside Outside Indoor Impermeable Runoff

Materials
Uncovered Covered Storage Pad Control
Abrasives 68% (19) 32% (9) 29% (8) 25% (7) 29% (7)
Solid
11% (3) 36% (10) 79% (22) 54% (15) 50% (14)
Chemicals
Liquid
NA NA NA 50% (14) 22% (11)
Chemicals

16
pads with some form of runoff control. However, some conducting or had conducted related research. Weighted
respondents reported storing solid chemicals outdoors averages were used because not all of the respondents com-
with and without cover. Some store solid chemicals indoors pleted this section.
without impermeable pads and/or runoff control. To a lim- Corrosion to concrete reinforcing steel was found to be the
ited extent, liquid chemicals are stored in areas with runoff primary concern, followed by corrosion of vehicles, general
control. concrete damage, structural steel, and roadside structures.
The survey results also indicate that some agencies modified
2.4.2.3 Environmental Regulations operations to address corrosion concerns; however, such
modifications were not explained.
The survey asked the respondents if environmental regu-
lations governed their operations (see Table 2-7). Many
respondents (38 percent) indicated that environmental reg- 2.4.3.2 Railway Traffic Control Signaling
ulations were not a concern, despite national policies in Four respondents (i.e., Colorado, Missouri, Montana, and
both the United States and Canada. It is unclear whether the Ontario) indicated incidences where snow and ice control
survey question was well understood or whether a general materials have affected the operation of railway traffic control
lack of understanding exists regarding specific environmen- signaling. Follow-up efforts were unable to substantiate a
tal regulations. relationship beyond anecdotal accounts.
2.4.3 Corrosion Concerns 2.4.3.3 Power Distribution Lines

The survey addressed issues pertaining to corrosion Only one agency (South Dakota) reported incidences
impacts, including prioritizing corrosion concerns, along where snow and ice control materials have affected power
with the less common side effects on railway traffic control lines. Based on media reports, such problems have been
signaling and electrical power lines. observed in several areas.
2.4.3.1 Areas of Corrosion Concerns 2.4.4 Purchasing Issues

The respondents were asked to rank their corrosion con- The respondents were asked to provide information on
cerns by allocating percentage weightings to each concern selection processes for snow and ice control materials, includ-
(see Table 2-8). They were also asked if they had modified ing information on applicable specifications and details of
their operations to address these problems and if they are quality assurance protocols.
Table 2-7. Environmental 2.4.4.1 Purchase Selection Criteria

regulations governing snow
and ice control practices. The respondents were asked to rank present purchasing
criteria for snow and ice control materials by assigning per-
Regulating Agency Responses centages to each criterion (see Table 2-9). They were also
Federal 10 (36%) asked to do the same for future purchases. Weighted averages
State/Provincial 7 (25%) were used because not all of the respondents completed this
Other 3 (11%) section.
Table 2-8. Corrosion concerns.
Concerns Priority Ranking Modified Operations

Concrete Reinforcing 30 % 22% (6)
Vehicles 22 % 26% (7)
Concrete Damage 20 % 22% (6)
Structural Steel 16 % 26% (7)
Roadside Structures, Utilities, and Equipment 12 % 11% (3)
Other (no response) (no response)

17
Table 2-9. Product selection process.
Present Priority Future Priority

Criterion
Average Average
Environmental 7.3 % 9.6 %
Corrosion 8.5 % 9.8 %
Human Exposure 3.5 % 3.8 %
Purchase Price 38.7 % 35.2 %
Cost of Use (i.e., Capital and Operational) 6.7 % 6.8 %
Storage and Handling 7.7 % 7.8 %
General Performance and Ease of Use 14.6 % 14.0 %
Climatic Requirements 10.5 % 11.1 %
Tradition 2.1 % 1.5 %
Others (e.g., friction, odor, wildlife attraction) 0.4 % 0.4 %
Purchase price was the most important criterion for the attributes such as fracture and durability are also common.
respondents and probably will remain so. Seven of the Purchases of solid chemicals were primarily based on state
respondents cited “tradition” as a current purchase criterion specifications usually derived from AASHTO or ASTM
and four respondents indicated they would continue to use specifications. These specifications tend to involve physical
tradition as a purchase criterion in the near future. attributes (e.g., size distribution and moisture content).
Environmental issues were most important in western Purchases of liquid chemicals were primarily based on PNS
regions, while corrosion issues were more important in the or state specifications. In many cases, non-PNS members
Midwest and eastern regions. Some movement toward have used the PNS specifications as a model for agency
increasing future attention paid to corrosion and the envi- specifications. Further descriptions of test protocols asso-
ronment concerns was evident; however, the changes were ciated with AASHTO, ASTM, and PNS specifications are
small. provided in Chapter 5.
Survey responses detailing agency quality assurance test-
ing are shown in Table 2-11. Most respondents (93 percent)
2.4.4.2 Purchasing Specifications and
indicated that some form of QA testing is being conducted.
Quality Assurance
Fifty-seven percent of the respondents reported the level of
Respondents were asked to describe applicable purchase effort (i.e., number of loads tested). QA test frequency
specifications for snow and ice control materials. Purchase ranged from 1 to 100 percent of loads, with the median being
specifications are used by agencies to identify minimum qual- 10 percent. The wide variation in the level of quality control
ification before considering a particular product for purchase. and quality assurance may result from one or more of the
Table 2-10 provides details. following factors:
Ninety-three percent of respondents indicated that some
levels of product specifications were in place. The specifi- • The prevalent and long-term use of solid NaCl, which
cations for purchasing abrasives were mostly agency speci- tends to be a relatively uniform product with fairly simple
fications and tend to involve measurement of simple specifications as compared with liquid chemical products
physical attributes (e.g., particle size distribution), but with more complex specifications;
Table 2-10. Purchasing specifications for snow and ice control

materials used by agencies.
AASHTO ASTM PNS Agency Other

Abrasives (Not Available) 3 (11 %) (Not Available) 16 (57 % ) 4 (15 %)
Solid Chemicals 4 (15 %) 6 (22 %) 6 (22 %) 12 (43 % ) 3 (11 %)
Liquid Chemicals (Not Available) 2 ( 7%) 10 (37 %) 11 (39 %) 2 ( 7%)
AASHTO – American Association of State Highways and Transportation Officials
ASTM – American Society for Testing and Materials
PNS – Pacific Northwest Snowfighters

18
Table 2-11. Quality testing.
Agencies Conducting Level of PNS SHRP Agency

Others
Quality Testing Effort Protocols Protocols Protocols
26 (93 %) 10 % 7 (25 %) 2 (7 %) 12 (43 %) 3 (11 %)
• The cost of testing, particularly when extensive laboratory necessary. For this study, snow and ice control application
testing is needed (This is particularly true when using the strategies are classified as anti-icing, deicing, and abrasive use
PNS specifications that are far more extensive in scope); (see Table 2-12).
• The use of AASHTO and ASTM specifications, which
have been a standard for solid NaCl for a number of years
2.5.1 Anti-Icing
(these specifications primarily address gradation and
moisture content, which are not normally tested in the Anti-icing is a proactive strategy in which snow and ice
field); and control materials are applied before a snow, ice, or frost
• The use of agency specifications that tend to involve test- event. This strategy prevents precipitation from bonding
ing during the production of the materials and enforce- (i.e., freezing) with the pavement surface, or, at the very least,
ment on initial delivery. (Little or no follow-up testing is weakens bonds that may be formed, for easier removal than
common.) with deicing. Anti-icing applications are not limited to liquid
chemicals. Properly timed applications of either dry or pre-
wetted chemicals can prevent bond formation if these mate-
rials can be held in place and not removed by traffic action.
2.5 Application Strategies
As such, anti-icing with solids can be very effective on side-
Various strategies are available for applying snow and ice walks and other areas not subject to vehicle action.
control materials to roadways. Depending on weather condi- Application rates in most of the research literature are
tions, site-specific factors, and level-of-service goals, an expressed in either dry weight per lane mile or per lane kilo-
agency’s options include materials selection as well as the meter. However, in actual practice, the rates are usually
timing, rate, and frequency of application. Just as no one tool expressed in terms of liters or kilograms per lane kilometer or
is suitable for every job, no one application strategy is suitable gallons or pounds per lane mile, depending on whether liq-
for all weather conditions. Application strategies should, uids or solids are being applied. Table 2-13 shows typical
therefore, be viewed as tools that can be selected and used as application rates used for anti-icing, based on the actual mass
Table 2-12. Application rates for various snow and ice control strategies.
Strategy/ Materials Pavement Application Rates2
Method Temperature
Ranges1
Anti-Icing Liquid Chemicals, 0° C to -12° C 18-110 Kg /Lane /Km
Solid Chemicals, (32° F to 10° F) (65 – 400 Lbs / Lane/ Mile)
Pre-wet Solid Chemicals
Deicing Pre-wet Solid Chemicals, 0° C to -18° C 113 – 400 Kg /Lane /Km
Dry Solid Chemicals (32° F to 0° F) (200-700 Lbs / Lane/ Mile)
Abrasives Pre-wet Abrasives, No limits 225 – 2,700 Kg /Lane /Km

Dry Abrasives (500-6,000 Lbs / Lane/ Mile)
Abrasive/Salt Mixes 0° C to -18° C 225 – 2,700 Kg /Lane /Km

(32° F to 0° F) (500-6,000 Lbs / Lane/ Mile)
1
Source: (6)
2
Source: (6, 7)

19
Table 2-13. Typical chemical application rates for anti-icing activities at

–10°C to –0°C (15°F to 32°F).
of solids applied for 0°C to –10°C. Within this temperature viable and desirable alternative to dry application of solid
range, most snow and ice control products have similar levels materials (1).
of performance. If the solid material is a chemical, such as NaCl, the wet-
A properly managed anti-icing program could reduce these ting agent initiates the formation of brine faster than would
application rates through analysis of actual experience. When occur without pre-wetting. Once brine is formed, it is more
anti-icing programs began in Washington State and Oregon, likely to stay on the roadway surface than be displaced by
application rates were 65 L (18.2-kg dry chemical using a 28- traffic action. A Danish study (9) assessed the retention of
percent solution) per lane-km (30 gal per lane-mile) for morn- NaCl applied as a 20-percent brine versus pre-wet NaCl. Sta-
ing frost. As crews gained experience and evaluated operations, tistical analysis of 1,800 test results indicated that about 90
rates as low as 18 L (6.5-kg dry chemical using a 28-percent percent of the salt from brine remained on the roadway
solution) per lane-km (8.5 gal per lane-mile) could be used compared with 60 to 65 percent for pre-wetted NaCl. Much
while maintaining an equivalent level of performance (1). lower retention is expected after application of dry chemi-
These low application rates were then used in cases where rain cals. Roadway retention for both application methods was
was predicted to immediately follow the frost event. This strat- also found to decrease at similar rates as traffic levels
egy reduced the amount of chemical applied and consequently increased. Solid chemicals, most commonly salt, can be pre-
released to the environment with storm runoff (1). wet with any of the available liquid chemicals, including salt
brine. In present practice, dry salt is pre-wet with 35 to 45 L
of liquid per metric tonne (8 to 10 gal per U.S. ton). These
2.5.2 Deicing
rates will and should vary with experience and temperature.
Deicing is a reactive snow and ice control strategy that Some application rates for the total pre-wet product from
involves the application of a chemical on top of a layer of snow, Ontario (10) are shown in Table 2-14.
ice, or frost that is already bonded to the surface of the pave- When pre-wetting abrasives, the wetting agent adds
ment (1). Typical application rates for dry chemicals range from weight as well as cushions the impact as the particle hits the
25 kg per lane kilometer (100 lbs. per lane mile) to as much as roadway. There is also evidence that the liquid may help the
130 kg per lane kilometer (500 lbs. per lane mile); however, material “stick” to the roadway. If the roadway surface has
sometimes more are used, depending on conditions. a layer of compacted snow or ice, the pre-wet liquid will
Deicing typically involves the application of solid NaCl partially melt the ice and allow the abrasive particle to
and, as with any dry application, traffic action can result in a become embedded in the ice layer. The net effect is that the
significant (up to 80 percent or more) loss of material to the abrasive particle becomes far more resistant to displace-
roadside environment. A Montana study (5) found that deic- ment from traffic action, further reducing material loss
ing typically requires 5 times the amount of chemical prod- from the roadway. A Montana study (5) reported that pre-
uct that anti-icing requires. wetting of abrasives before application can reduce the
amounts applied by 50 percent in cold temperatures. Abra-
sives can be pre-wet with any of the available liquid chem-
2.5.3 Pre-wetting
icals. As with solid chemicals, dry abrasive is pre-wet with
Pre-wetting is injecting or spraying a liquid chemical on 35 to 45 L of liquid per metric tonne (8 to 10 gal per U.S.
solid chemicals or abrasives to enhance their effectiveness ton). These rates will and should vary with experience,
and reduce material loss and other forms of waste. It is a temperature, and storm conditions.

20
Table 2-14. Variable salt application rates with on-board pre-wetting.
Road Surface Temperature Range °C (°F)

Precipitation
Type Warmer than -5° (23°) -5° (23°) to -10° (14°) Less than -10° (14°) to -18° (-1°)
Kg/lane-km (Lbs/lane-mile) Kg/lane-km (Lbs/lane-mile) Kg/lane-km (Lbs/lane-mile)
Frost 25 (97) 35 (136) 35 (136)
Light snow 35 (136) 50 (195) 65 (253)
Heavy snow 65 (253) 65 (253) 85 (331)
Freezing rain 65 (253) 85 (331) 85 (331)
Source: (10)
2.5.4 Dry Abrasive/Chemical Mixes Abrasive gradation (particle size) can also affect roadway
retention. In one study, researchers (8) measured the reten-
Application of abrasive/salt mixes, a popular strategy, typ- tion of sand particles resulting from traffic and wind under
ically involves approximately 1:1 mixtures of abrasive and various snow and ice conditions. They found retention
salt. This strategy is not as cost-efficient as using chemicals ranged from 10 percent for natural fine sand to 50 percent for
only and is less effective at reducing accidents. Kuemmel and manufactured coarse sand. Typical gradations for abrasive
Bari (11) concluded that accident reductions for two-lane material are shown in Table 2-16.
highways using the salt/abrasive mixtures were less than with In addition to performance-related issues, other costs of
salt only. Accident frequency on freeways was higher when abrasive use can include increased insurance claims and envi-
salt/abrasive mixtures were used than when only salt was ronmental impact. A Colorado study (13) concluded that “the
used. Cost/benefit calculations showed the application of major benefits of increasing use of liquid chemical deicers
salt/abrasive mixtures did not recover winter maintenance include the reduction of the insurance cost because of
costs on two-lane highways during the 11-hour analysis reduced chance of windshield damage and cleaner air.” Since
period and freeways operations recovered costs in 6 hours, the implementation of the findings from the research on
substantially slower than with salt only. “Environmentally Sensitive Sanding and Deicing Practices,”
in 1994, the Denver metropolitan region has not violated the
2.5.5 Abrasive Use U.S. EPA PM-10 Standard.
Sanding has long been a strategy of choice by many
agencies, because it is a very visible low-cost approach to 2.5.6 Stockpile Treatment
managing pavement friction. However, it has been shown
Stockpile treatment involves treating stockpiles of abrasive
that when abrasives are placed “dry” on the road surface
materials or solid chemicals with an appropriate chemical to
(without significant pre-wetting) they provide at best, a
ensure that under cold-weather conditions these materials do
very short term increase in road surface friction (12). As
not freeze into clumps that make the materials difficult to
roadway traffic levels increase, and vehicle speeds are
handle and apply (1). Typically chloride mixtures (liquid or
greater than 30 mph (48 kph), any benefit from abrasive use
solid) are added at 5 percent (by weight) chemical to a stock-
diminishes. Table 2-15 presents the practices for use of dry
pile for “frost proofing.”
abrasives recommended by the Iowa Highway Research
A stockpile of abrasive materials can be frost-proofed by
Board (12).
mixing a small amount of chemical into it as the stockpile is
being made. This is the first step in reducing the incidence of
frozen lumps. When frost-proofing sanding material, it is
Table 2-15. Abrasive use. advisable to use only enough chemical to keep the material
Road Type Use of Dry Abrasives workable. Both liquids and solids chemicals can be used to
Freeways Inappropriate treat stockpiles of abrasive materials.
Rural roads, paved Inappropriate
Rural roads, gravel Only on low speed sections (perhaps hills and curves)
Solid chemicals can and often do attract moisture, causing
Rural intersections Only on low speed approach length of gravel roads the stockpile to freeze into lumps that present a workability
High speed urban roads Inappropriate
Low speed urban roads Only in certain locations and when snowpack will persist
problem. Stockpiles of solid chemicals can be “frost-proofed”
Urban intersections Only when snowpack will persist by adding a small amount of liquid chemical with a lower
Source: (12) ultimate eutectic temperature than the stored solid.

21
Table 2-16. Typical abrasive gradations
Oregon
British Columbia
Department of
Ministry of Transportation and Highways
Transportation
Sieve Size
Designated 6.75 mm
Designated 12.5 mm Designated 16 mm Designated 19 mm (.25 in)
% passing % passing % passing % passing
19.000 mm (3/4”) – – 100 –
16.000 mm (5/8”) – 100 –
12.500 mm (1/2”) 100 – –
9.500 mm (3/8”) – 80–100 80–100 100
6.750 mm (1/4”) – – – 80–100
4.750 mm (#4) 50–95 50–95 50–95 –
2.360 mm (#8) 30–80 30–80 30–80 –
1.500 mm (#10) – – – 8–22
0.600 mm (#25) 10–50 10–50 10–50 –
0.375 mm (#40) – – – 0–10
0.300 mm (#50) 0–25 0–25 0–25 –
0.075 mm (#200) 0–6 0–6 0–6 0–2
Source: (1)
“–” indicates data not available
2.5.7 Application Strategy Studies the magnesium chloride used in the after period. The chloride
quantities were calculated using the molecular weights shown
Numerous accounts illustrate the value of using more
in Table 2-19, adjusting for moisture content in the solid salt
proactive snow and ice control strategies (i.e., anti-icing and
and other non-chloride ingredients in the magnesium chlo-
pre-wetting) and using liquid snow and ice control materi-
ride solution.
als. Three different studies outlined below illustrate the
The Cypress Bowl study, which benefited from highly
value of using all of the available application strategies
detailed weather data, was able to use a weather index devel-
appropriately. In each case, multiple strategies were used to
oped by the Washington DOT to determine weather differ-
achieve the desired results in a cost-effective and environ-
ences. The weather index used in this study was derived from
mentally sensitive manner. These three studies are typical of
the original SHRP weather index. This index produced a
results obtained elsewhere in the United States and Canada
result in the range of 50 (severe) to 50 (mild).
when traditional methods are replaced with more cost-
effective strategies. The result is more cost-effective and effi- At Cypress Bowl, the months of March 2000 and March
cient use of chemicals and sand. 2001 yielded weather index numbers of 12.30 and 11.36,
respectively. Because the winter index values are not linear,
2.5.7.1 Cypress Bowl one cannot infer anything except that the 2 months were rel-
atively equal in weather severity. Given the small difference in
A case study (14) conducted for the Cypress Bowl ski area indexes and the range of possible index results, the compar-
near Vancouver, BC, illustrates how an agency can vary appli- isons made were much more accurate than other studies.
cation strategies to address changing storm conditions and
meet desired levels of service goals. Traditional strategies
2.5.7.2 Idaho DOT
involving dry abrasives, dry salt, and abrasive/salt mixes were
replaced by techniques that involved using various combina- An Idaho DOT study (15) found significant benefits when
tions of chemicals and application methods (i.e., anti-icing traditional methods (i.e., deicing using NaCl and abrasive
and pre-wetting of salt and/or abrasives) to address each spe- use) were replaced with liquid chemical anti-icing. The area
cific storm event. This change yielded a 39-percent reduction studied was on US 12 near Orofino in the winding Clearwater
in material costs and a 28-percent reduction in operation River canyon. The results of 3 year’s operations (1997 to 2000)
costs. Volumes of applied materials were also reduced, further are shown in Table 2-17 and demonstrate significant reduc-
reducing the potential for effect on the natural receiving envi- tions in labor, abrasive use, and accidents. The Idaho study
ronment. Abrasive use decreased by 64 percent and chlorides used a 5-year average for setting the baseline of comparison.
use decreased by 73 percent. Although this is a workable approach and yields a reasonable
The reduction in chlorides was calculated by inventorying comparison, it is still not as accurate as truly one-to-one
the roadsalt used in the before and after periods and adding weather comparisons.

22
Table 2-17. Idaho results. materials used in winter maintenance is related to the type of
product used and the amount applied. Any measure to reduce
Average Annual Average Annual Average Annual
Labor Hours Abrasives Used Accidents volumes of materials applied can reduce the magnitude of
Before 1997 650 1475 cu. M 16.2 environmental effects.
1997-2000 Period 248 247 cu. M 2.7
Percent Reduction 62% 83% 83%
2.6.1 Product Selection
Source: (15)
The product used determines the “constituent speciation”
Table 2-18. Colorado DOT materials use from (what is in the product) and “constituent concentration”
1992 to 2000. (how much of each constituent is present). As an example,
Table 2-19 shows the amounts of component elements and
Sand Salt (NaCl) Liquid Deicer
Year
(tons/mile) (tons/mile) (gallons/mile) ions in the most commonly used chloride-based snow and ice
1992 0.25 1.1 <0.5 control chemicals. The selection of a particular product can
1993 0.20 0.9 <0.5 affect the amount of a constituent chemical released to the
1994 0.17 1.3 <0.5
1995 0.29 0 <0.5 environment. Regardless of loading, a certain level of service
1996 0.19 0 1 must be maintained. Depending on the weather conditions,
1997 0.23 0 3 certain materials may be more effective than others. It is pos-
1998 0.17 0 5
1999 0.09 0 8 sible to gauge performance by assessing eutectic or phase
2000 0.04 0 11 curve information and calculating differences in dilution fac-
Source: (13) tors between products at a given temperature. Dilution of a
chemical takes place as it melts water and subsequently
reduces the concentration of the solution until it will freeze.
The performance index (PI) is a comparison tool that
2.5.7.3 Colorado Department of Transportation determines how much melting takes place before the solution
(CDOT) will freeze at a given temperature. A higher PI indicates more
A CDOT study (13) investigating the effects of abrasive use melting can occur before freezing and thus better perfor-
on air quality cited a general improvement in air quality as evi- mance. The formula for a PI is
denced by the fact that Colorado has not exceeded a U.S. EPA
PI (BC/EC) 1
PM-10 standard for seven winters. This improvement corre-
lated with a shift in winter maintenance activities, away from Where
abrasive and salt (NaCl) and toward snow and liquid snow ice BC the concentration of chemical as applied to
control chemicals. This study concluded that the “optimal the roadway surface.
maintenance strategy involves an optimum mix use of liquid EC the concentration (read from the phase curve)
deicers and sand without the total elimination of the latter.” at which the brine will freeze after dilution at
Because of the study’s statewide scope, it did not attempt to the stated surface temperature.
compensate for different weather conditions among regions;
however, the author noted where weather could have been a fac- Example: NaCl (23.3 percent) and KA (50 percent) compared
tor in differences in results. Table 2-18 provides further detail. at 6° C (21°F).
For NaCl (23.3%), the beginning concentra-
2.6 Materials Loading tion (BC) is 23.3% and the ending concentra-
tion (EC) is 9%. This provides a PI of 1.59.
In winter maintenance operations, the amount of material
applied for roadway affects the product loading to the natural For KA (50%) the BC is 50 percent and the EC
receiving environment significantly. Environmental impact of is 13 percent. This provides a PI of 2.85.
Table 2-19. Molecular weight calculations for chloride-based chemicals.

23
Using the above comparison, one could reasonably con- the environmental effects of materials, the effect of climate
clude that a smaller application of KA would yield the same could be somewhat normalized because materials were
results in terms of ice removal at this temperature. PI is compared with one another under the same climate and
used to estimate relative performance of products, but can weather conditions (1).
be affected by changes in temperatures, so care should be
taken in its use. 2.6.3 Strategy Selection
Selection of a strategy, or sequence of strategies, can affect
2.6.2 Application Amounts the total loading of a snow and ice control material. There is
The amount of snow and ice control materials entering the significant anecdotal evidence that the use of anti-icing and
environment is directly related to the application rate compre-wetting strategies singly, or in concert, can reduce the total
bined with the number of applications being made. A key fac- environmental loading of materials significantly (14–16). The
tor in application rate and frequency of application is weather Cypress Bowl Case Study (14) illustrates the reduction in envi-
condition, which will dictate either the application rate or the ronmental loading that can be achieved by proper strategy
number of applications, or both. Thus, weather conditions selection. The study’s validity results from the fact that the
will, in most cases, dictate the total amount of a particular weather conditions for both study periods were nearly identi-
snow and ice control material applied. The effect of weather cal, allowing a one-to-one comparison of the before and after
is illustrated by the weather- and temperature-dependent strategies. No other published studies were found with identi-
application rates shown in Tables 2-13 and 2-14. Hygroscopic cal weather conditions between compared periods. Anti-icing
properties, pavement type, pavement texture, and traffic vol- and pre-wetting of solid materials can reduce the total
umes can also affect application rates and frequencies, but are environmental loading significantly. Table 2-20 shows that
relatively minor in effect. reduction in the application rates cited for the various strate-
Application rates are determined to suit specific site and gies can affect the total loading for the materials of choice.
weather conditions and, as a result, are highly variable. These estimates are based on data presented in Tables 2-13 and
Chemical application is not a “do it by the formula” process 2-14. Liquid application data have been converted to equiva-
and, if correctly done, should be variable with the goal of lent solid mass of chemical. Figure 2-1 illustrates the data in
adjusting to conditions. The most important of these Table 2-20 and shows how selection of snow and ice control
conditions is weather, which is also the least predictable. strategies can affect the overall environmental loading of ice
However, in the context of this study, which was to compare and snow control materials.
Figure 2-1. Comparative application rates for select snow and ice control
strategies.

24
Table 2-20. Reduction in material application loading from changes in strategies.
Original Loading 1 Pre-wetting Solids Loading 2 Anti-Icing Loading 3
Low High Low High Low High

Original
Strategy Kg (Lb) Kg (Lb) Kg (Lb) Kg (Lb) Kg (Lb) Kg (Lb)
Dry
225 (500) 2700 (6000) 113 5 (250) 1350 5 (3000) 34 4 (75) 405 4 (2100)
Abrasives
Dry Solid 5 5
113 (200) 400 (700) 57 (100) 200 (350) 18 1 (65) 110 1 (400)
Chemicals
1
From Tables 2-12 and 2-13.
2
Assumes pre-wetting of abrasives or solid chemicals as indicated by original strategy.
3
Assumes liquid or solid chemical is used. Abrasives may have been used under certain circumstances.
4
Source: Average from Idaho (15) and Cypress Bowl (14) studies
5
Source: Montana (5)

25
CHAPTER 3
Impacts to the Natural Environment
3.1 Environmental Pathways cause solid deposits to migrate to soils, surface waters, and
groundwater as described above.
All snow and ice control materials can affect the natural
receiving environment. The magnitude of the effect depends
3.1.1 Soil Transport
on conditions such as the types of materials, quantities being
applied, transport pathways, exposure pathways, exposure Snow and ice materials are deposited on roadside soils
duration, chemical-specific impacts, and other site-specific from splashing, spraying, gravity drainage, or plowing. The
environmental characteristics. The distribution of snow and magnitude and the spatial extent of effect depend on local
ice control materials into the environment involves complex features and conditions, such as the slope of the roadside,
and interrelated processes that can be simplified into a direction of drainage, type of highway drainage system, soil
conceptual model showing primary transport mechanism type, vegetative cover, presence of snow and ice, and precipi-
and pathways and areas of effect (see Figure 3-1). tation (18). In addition to accumulating in the snowpack,
Initial discharge of snow and ice control materials into the some solids percolate vertically into the soil or are trans-
receiving environment can occur during transport, storage, ported horizontally through runoff. Sites with steep slopes in
or application on roadways. Roadway application is the pri- addition to compacted soils (i.e., clays) are characterized by
mary source of discharge based on the volumes of materials high runoff and low infiltration rates and promote the hori-
applied; however, spills and discharges from storage or trans- zontal movement of runoff (and soluble ions) in preference
port can cause significant localized effects. Transport to the to vertical movement (19). On the other hand, sites with low
environment begins when the snow and ice control material slopes and drainage, low runoff volumes, and permeable soils
is applied to the roadway. Materials will either leave the road- all contribute to the downward movement of solubilized ions
way as (1) a liquid, through splashing, spraying, or gravity into the soil. Jones et al. (19) determined that compacted soils
drainage; or (2) as a dried residue mobilized or re-suspended could result in 75-percent runoff and 25-percent infiltration,
by traffic action and wind. Product leaving the roadway as while permeable soils could result in 35-percent runoff and
runoff may directly discharge to receiving surface water bod- 65-percent infiltration.
ies or percolate into soils. Product released into soils will Movements of salts through frozen ground have shown
either accumulate (in soil solutions or through soil adsorp- near-normal infiltration up to 10.2 cm in heavy-textured soils
tion) or permeate the soil and combine with groundwater. (19). Light-textured soils (i.e., typical farming and pasture)
Where soil solution or groundwater comes in contact with have shown almost no infiltration. These are extremely com-
the root zone of vegetation, uptake through the roots can plex processes. Methods of predicting salt infiltration in
occur. Chemical carried as runoff, that has percolated frozen soil are not available, and the movement of moisture
through the soil, or that has reached the groundwater will through soil is not yet fully understood.
often be transported to surface water bodies. Products forced Novotny et al. (20) determined that most solids (i.e., total
into the air by vehicle motion, wind, or plowing will be solids and dissolved solids) accumulation occurs within 1
deposited on vehicles, the road surface, the roadside (e.g., meter from the edge of the roadway in snow piles formed by
roadside cover and ditches), adjacent soils, vegetation, the plowing operations and traffic. The concentration of solids
snow layer, surface waters in the surrounding vicinity, or during winter increases linearly until the snow pile melts in
combinations thereof. Secondary transport processes will the spring. Enrichment and “first flush” effects describe the

26
Figure 3-1. Environmental pathway model modified from TRB (17).
physical and chemical process occurring in the roadside severely limited by low precipitation rates or restricted
snowpack. Ions are rejected from the crystalline lattice of access to percolating water. The greatest effects are, there-
snow during repeated freezing and melting and become avail- fore, associated with areas with the highest loads, with the
able for washoff during the first stage of snowmelt. This can lowest soil percolation rates. Effects are calculated to be
result in high runoff concentrations during early spring greatest on soils with the highest levels of sodium satura-
washoff. tion when subjected to low EC water (21).
A Canadian study (21) mapped salt concentrations Studies of sodium and chloride levels in soil generally indi-
throughout the Canadian landscape and found concentra- cate that the greatest concentrations are found within 2 to 3
tions to be highest in areas near salt depots and along road- meters of the pavement edge (22). However, depending on
sides. On a regional level, salt accumulations tended to vary local conditions, impact areas can be more extensive. For
with climate. The highest salt accumulations in soil instance, one study (23) reported higher-than-background
occurred in the Prairie provinces, whereas in areas with concentrations of sodium and chloride as far as 10 meters
considerable rainfall, such as British Columbia or Atlantic from the roadway. High concentrations of sodium and chlo-
Canada, effects of roadsalt applications were either negligi- ride are not usually found at depths below 1 meter (22, 24).
ble or are not cumulative over the course of the snow-free McBean and Al-Nassri (25, 26) also investigated the dis-
season. Effects are cumulative in areas where dilution tance that salt applied to roads can travel from the road and
resulting from surface runoff and soil percolation is contaminate the surrounding soil or receiving water. They

27
determined that a 10-m setback is sufficient to protect most exponentially with height from the roadway, with most
vegetation from the effect of snow and ice control material deposition occurring up to 3 meters in height.
applications; beyond 10 meters, concentrations of snow and Topography can also contribute to the extent of aerial
ice control chemicals dropped to background levels. If the soil transport. A TRB study (32) found that wooded areas down-
is well drained, the extent of impact may be as little as hill within 40 to 60 feet of heavily traveled salt-treated high-
5 meters from the road for deep gravelly, stony soil (27). ways are primary candidates for salt-related damage as
Earlier studies concluded that, although salt levels in compared with similar uphill areas.
roadside soils may be high in spring, salt is readily leached
from the top meter of soil and disappears in the summer,
3.1.3 Water Transport
whereas other studies document a steady buildup of salts in
the soil (20). Whereas the extent of deposition of snow and ice control
Cain et al. (28) summarized 10 studies, including studies materials through soil or through the air pathways is some-
referenced above, that provided measurements of substrate what predictable, transportation of these materials into
(i.e., soil, soil water, or solutions) of chloride, sodium, or NaCl groundwater and surface waters is more involved and
following application at specific distances from the roadway. depends on site-specific conditions. Materials may reach
The assessment found, for the most part, that elevated levels surface water bodies through direct runoff, groundwater
of sodium (over 500 ppm) tended to occur within 10 meters percolation, and, to a lesser degree, airborne deposition.
of the roadway. Elevated chloride levels (over 500 ppm) were Studies have been done to monitor highway runoff. Results
also found to occur within the same distance. indicate that in highway drainage systems, the lack of base
Based on an assessment of these studies of snow and ice flow, uncertainty in dates and times of expected precipita-
control materials over roadside soil (from splash, spray, and tion, the short time scale of runoff, and the first-flush
runoff), the effects on soil are concluded to typically drop phenomenon may prevent meaningful data from being
quickly beyond the first 10- to 20-m-wide strip of land bor- obtained using a manual sampling program (33). Many
dering the roadway or along paths of overland flow. studies have attempted to model the concentrations of
direct runoff (34–36); however, findings have been highly
variable and dependent on site-specific factors such as pre-
3.1.2 Air Transport
cipitation amounts, first-flush effects, soil infiltration,
Snow and ice control materials can be transported through hydraulic conductivity, and dilution effects. In general it can
the air and deposited as a splash or spray. Aerosols and re- be conservatively estimated that snow and ice control mate-
suspended dried materials can be transported by vehicle tur- rials will be diluted 500-fold as they leave the roadway and
bulence or winds and can be carried further by winds blowing can be diluted substantially more as the distance from the
across the highway. Airborne materials can affect soil and roadway increases, but site-specific studies are critical for
water concentrations; however, deposition of materials on support of state or regional programs. Further discussion of
vegetation tends to be a primary concern with this transport runoff and dilution effects is provided in Section 3.4.3.1.
mechanism. Traffic and weather patterns play a key role in
determining the distance that materials are transported from
the roadway. Kelsey and Hootman (29) found that dispersion 3.2 Framework for Protection of
of aerosols from a highway is related to traffic volume and Environmental Systems
speed. The findings suggest that NaCl aerosol dispersion is 3.2.1 U.S. Regulations
extensive and could be at least 15-m high within 67 meters of
a highway. Kelsey and Hootman also found that the higher the The United States has general environmental protection
plume formed by traffic, the farther it is likely to be carried by legislation that prohibits contamination of air, land, and
winds. A study conducted by the Connecticut DOT found water; however, few of these address roadsalt specifically.
airborne salt to travel as far as 300 feet laterally from the road-
way under heavy traffic conditions. Chung (30) found that
3.2.1.1 National Pollutant Discharge Elimination
salt spray could also be transported for distances greater than
System (NPDES)
500 feet under high wind conditions. A European report
determined that more than 90-percent of spray or splash next Since 1987, the U.S. Environmental Protection Agency
to a primary motorway was deposited along a corridor about (U.S. EPA) has been required by Congress to control
15-m wide and that the amount deposited decreased stormwater runoff. NPDES requires permitting of storm-
exponentially with increasing distance (31). This study also water discharges from industrial and construction activities
found vertical distribution to be distributed more or less and separate municipal storm sewer systems (MS4s). In

28
November 1990, the U.S. EPA adopted Phase I regulations applied would require that the applicators (i.e., states or local
pertaining to stormwater discharges from MS4s serving governments in most cases) specifically evaluate the potential
more than 100,000 people, construction activities disturb- threat to the species. The listing agency might require consul-
ing greater than 5 acres, and industrial runoff into surface tation on ESA issues, or an evaluation could be undertaken
waters. The program was expanded when the Phase II final voluntarily (e.g., Lewis, 1999 (52)) as a means of ensuring that
rules were passed in December 1999, through permitting the activity would not be later found in conflict with the ESA
requirements for construction activities disturbing 1 to or similar state legislation.
5 acres and small MS4s in urbanized areas. Some states have
classified road maintenance stockpiles and operational sites
3.2.1.3 Clean Air Act
as industrial sites requiring separate permitting and pollu-
tion control plans. Road maintenance, including snow and Protecting public health, the environment, and quality of
ice control activities, is controlled as part of stormwater per- life from the detrimental effects of air pollution has been a
mitting. Regulation outside stormwater regulation zones is strong national priority since 1970, when Congress passed
much weaker, but may involve best management practices the Clean Air Act. Under this authority, the U.S. EPA has
to minimize polluted runoff. established air quality standards—known as National
Ambient Air Quality Standards (NAAQS)—for six pollu-
tants. Recently, the U.S. EPA completed its review of the
3.2.1.2 Endangered Species Act
current air quality standards for ground-level ozone (com-
The U.S. Congress passed the Endangered Species Act monly known as smog) and particulate matter (or PM). Of
(ESA) in 1973 to establish federal legal authority for prevent- particular interest to users of sanding materials or abrasives
ing extinction of species within the United States and for con- are the PM-10 regulations. Annual and 24-hour NAAQS for
tributing to the prevention of extinction globally by PM were first set in 1971. Total suspended particulate (TSP)
prohibiting personal or commercial traffic in endangered was the first indicator used to represent suspended particles
species or their by-products from other parts of the world. in the ambient air. Since July 1, 1987, however, the U.S. EPA
The ESA is administered by the U.S. Fish and Wildlife Service has used the indicator PM-10, which includes only those
for non-marine species and by the National Marine Fisheries particles with an aerodynamic diameter smaller than 10
Service for species that have marine affinities. These two micrometers. These smaller particles probably are responsi-
agencies decide, based on legal and regulatory criteria, which ble for most of the adverse health effects of PM because of
species should be listed. Listing is initiated through informa- the ability of the particles to reach the thoracic or lower
tion gathered by the agencies or information provided by regions of the respiratory tract.
other parties, which may be governmental or private. Species
are classified as “endangered” if they are in imminent peril of
3.2.1.4 Other Programs
extinction and as “threatened” if they are progressing notably
toward the status of endangerment. Many states have laws The U.S. EPA has, to date, not developed methods to
that reinforce or enhance the scope of the ESA. address toxicity to soils (37). The U.S. EPA is considering
The ESA and its state counterparts restrict human interac- adding sodium to the Contaminant Candidate List (CCL);
tions with listed species. The act prohibits “take,” which for however, the primary stressor for this initiative is the protec-
the purposes of the ESA is defined as harvesting or causing tion of drinking water and aquatic environments (38).
unintended mortality or stress as a by-product of any activity
within the range of the species. The act also requires that “crit-
3.2.2 Canadian Regulations
ical habitat” (defined by the listing agencies) not be destroyed
and requires the agencies to prepare “recovery plans” that set Environment Canada investigated Canadian programs,
guidelines for an extension of abundance or range of the initiatives, and guidelines related to roadsalts (39) and
listed species. Thus, both private and governmental activities found relatively few statutes or regulations directed at
may be restricted within the range of a listed species. reducing the adverse environmental impacts of snow and
Although federal water-quality regulations within the ice control materials. Although most jurisdictions have
United States are not applied to road application materials, general environmental protection legislation that prohibits
with the exception of urban areas where stormwater permits contamination of air, land, and water, few of these address
are required, the ESA is directly applicable to such materials, roadsalt specifically. Regulations for use and release of snow
just as it would be to any other major activity that could have and ice control materials into terrestrial environments tend
environmental effects. Therefore, the presence of endangered to be less of a concern from a regulatory standpoint when
species near roads where snow and ice control materials are compared with effects on aquatic environments. Some

29
jurisdictions have implemented specific groundwater The Canadian Council of Ministers of Environment
protection strategies that could affect the location of salt (CCME) provides Environmental Quality Guidelines based
storage facilities. Even fewer regulations are in place directly on toxicological data for potential receptors (e.g., plants,
addressing effects on soils and terrestrial biota from snow invertebrates, wildlife and microorganisms) and land use cat-
and ice control materials. egories (37). Soil criteria for sodium, magnesium, calcium,
Roadsalts were put on the Priority Substances List under and chloride have not been developed (42).
the Canadian Environmental Protection Act (CEPA) in 1995
because of concerns about large quantities of chlorides
3.2.3 Others
affecting the environment. After a thorough scientific
assessment (3), it was determined that roadsalts pose an The Pacific Northwest Snowfighters (PNS) is a consortium
environmental risk and that management plans should be of states and provinces (i.e., Idaho, Montana, Oregon, Wash-
developed to reduce the effects of roadsalts on the environ- ington, and British Columbia) that has developed a list of
ment. Environment Canada’s “Risk Management Strategies physical, environmental (primarily based on aquatic stan-
for Road Salts” (40) suggests that the environmental objec- dards), and corrosion-related specifications for qualification
tive for protection of soil is to protect soil integrity. This of snow and ice control materials (43). Products found to meet
ensures that application of roadsalts does not result in dele- minimum specifications are included on a qualified products
terious effects on physical and chemical parameters (e.g., list (QPL). Participating highway agencies, when issuing ten-
soil structural stability, soil dispersion, soil permeability, soil ders, may request that vendors be included on the QPL before
swelling and crusting, soil electrical conductivity, and soil consideration. Many jurisdictions and agencies not directly
osmotic potential). It is further assumed that an environ- involved in the PNS consortium have, to some degree, also
mental objective developed for vegetation will be protective adopted the PNS specifications for product selection.
enough for soils.
In Canada, most published guidelines for salt contamina-
3.3 Components of Concern
tion in soil are based on measures of electrical conductivity
and the sodium adsorption ratio (41). Notable exceptions This report evaluates a range of snow and ice control prod-
include a recent guideline put forward by the Alberta Energy ucts that can be classified by their primary components (see
and Utilities Board limiting chloride concentrations in sand Table 3-1). In general, the primary components dictate the
used for roadbed construction to 3,000 mg Cl/kg of sand. type and magnitude of effect that may result when introduced
British Columbia has draft criteria for sodium and chloride to the natural receiving environment. Secondary attributes
in soil based on site-specific factors (i.e., hydraulic conduc- are material properties typically associated with the product
tivity and species-specific impacts) and land use (i.e., agri- as an additive or contaminant and that influence the type and
cultural, urban park, residential, commercial, and industrial). magnitude of effect (44). Table 3-1 lists the most commonly
These were developed through analysis of toxicity data for soil encountered secondary attributes associated with each mate-
invertebrates and plants. rial, but it is not an exhaustive list.
Table 3-1. Primary components and secondary attributes of selected snow

and ice control materials.
Snow and Ice Control
Material Type Primary Components Secondary Attributes
Material
Sodium Chloride (NaCl) Na, Cl Heavy metals, CN, P
Chloride-Based
Salts Calcium Chloride (CaCl2) Ca, Cl Heavy metals, P
Magnesium Chloride (MgCl2) Mg, Cl Heavy metals, P

Calcium Magnesium Acetate
Acetates Ca, Mg, C2H3O2 BOD
(CMA)
Potassium Acetate (KA) K, C2H3O2 BOD
Agricultural By-products Organic Matter (complex BOD, Heavy metals,
Organic Products (Organic Biomass) sugars) Phosphorus, Nitrogen
Manufactured Organic Organic Matter – varies with
BOD
Materials product (i.e., glycol, methanol)
Urea, Ammonia (i.e.,
Nitrogen Products Urea Not Available
Nitrogen)
Air Quality - PM10, PM2.5
Abrasive Abrasives Heavy metals, P
Water Quality - Sedimentation

30
Chloride-based salts are the most common chemical prod- Ferrocyanide ions can decompose to cyanide ions when
ucts used for snow and ice control, followed by acetates (see exposed to ultraviolet light in a process termed “photode-
Section 2.4.1). Many commercially available chloride-based composition.” The degree to which this occurs is unclear, as is
products are blended with corrosion inhibitors. Most corro- the extent of volatilization of HCN ions after photodecom-
sion inhibitors are proprietary, although they typically con- position. It is unlikely that photodecomposition will occur
sist of reduced sugar by-products or co-products from once ferrocyanide ions are in soils or groundwater. Labora-
agriculture and food industries. Urea products and manufac- tory experiments have shown that decomposition of ferro-
tured organic materials (e.g., glycols) are typically associated cyanide complexes in soils (in the dark) is slow. The half life
with airport deicing. These products are more costly than for ferrocyanide in soil ranged from 1 year at a low pH of soil
chloride products and are not commonly used for roadway water to 1,000 years at levels greater than pH 7 (20).
application. Abrasives used for snow and ice control are typ-
ically derived from regional sources. Concerns include poorer
3.3.2 Radioactivity
air quality (PM-10, PM-2.5), stream sedimentation, and
potential to contribute heavy metals. Radioactivity is expressed as a concern with some brines
Primary components and secondary attributes listed in Table obtained through groundwater wells—specifically for those
3-1 can be combined into common groupings. The six primary associated with oil and gas drilling operations. The U.S. EPA
components of environmental concern are chloride (Cl), estimates that 30 percent of oil and gas operations in the
sodium (Na), calcium (Ca), magnesium (Mg), potassium (K), United States have naturally occurring radioactive matter
and acetate (C2H3O2). Secondary attributes associated with (NORM) in brine and brine solids to be of sufficient concern
these products include heavy metals, ferrocyanide yielding the (45). Radiation levels for petroleum (i.e., oil and gas) pro-
cyanide anion (CN), phosphorus (P), nitrogen (N), and organic duced water range between 0.1 and 9,000 pCi/g (46). Regions
matter quantified as biochemical oxygen demand (BOD). of concern include, but are not limited to, the U.S. Gulf States,
the U.S. Midwest (e.g., Illinois, Michigan, Kansas, and Okla-
3.3.1 Ferrocyanides homa) and the Appalachian states. In most cases, NORM-
containing brines produced through oil and gas drilling
If roadsalt becomes exposed to air with fluctuating humid- operation are re-introduced into the subsurface. In some cases,
ity while in storage, it can cake, making it difficult to handle however, these materials are reused for snow and ice control.
and apply to roadways. When atmospheric humidity is high A New York State study (47) found NORM in wastes pro-
(i.e., exceeding 70 percent) a brine solution forms on the sur- duced from oil and gas drilling operations. Of these wastes,
face of the salt crystals. When relative humidity drops, water the highest concentrations of radium were found in brines,
evaporates, and the brine solution recrystallizes between the which are traditionally “disposed of ” for reuse as roadway
salt crystals, causing aggregation of the salt crystals. Ferro- snow and ice control materials. The study concluded, how-
cyanide can be added to decrease the solubility of sodium ever, that, even if it is assumed that all brines contain the high-
chloride in the absorbed moisture, which reduces the amount est detected level of radioactivity, the resulting dose would be
of recrystallization. These compounds include less than 30 percent of the 10 millirem per year standard pre-
sented in the New York State Department of Environmental
• Sodium ferrocyanide, Na4Fe(CN)6⋅10H2O, commonly Concern (NYSDEC) Technical Administrative Guidance
known as yellow prussiate of soda; Memorandum (TAGM) 4003. Brines can have significant
• Sodium ferrocyanide, Na3Fe(CN)6⋅H2O; levels of radioactivity and their reuse on roadways could be
• Potassium ferrocyanide, K4Fe(CN)6⋅10H2O, commonly hazardous to human health and the environment. Limited
known as yellow prussiate of potash; and information is available detailing more specific regions of
• Ferric ferrocyanide, Fe4[Fe(CN)6]3, commonly known as concern for NORM materials or what percentages of pro-
Prussian Blue. duced brines used for snow and ice control contain NORM.
Individual states have jurisdiction over the production and
Effective concentrations of these products are typically less subsequent use of NORM materials.
than 0.01 percent by weight.
The fate of ferrocyanide ions in the environment is unclear.
Negatively charged ions tend to be rejected by soils and move 3.4 Aquatic Environments
readily with percolation and groundwater; however, a study
3.4.1 Protection of Aquatic Resources
conducted by Olson and Ohno (98) speculated that, because
the relative proportion of cyanide with respect to salt was low, In Canada and in Europe, the general pattern for protect-
the ferrocyanide may be adsorbed onto soil particles. ing aquatic resources is legislation and regulation directed at

31
specific activities or specific classes of pollutants. A good stream segment may invoke strict standards for heavy metals,
example under the CEPA of 1999 is the development of man- which are harmful in very small amounts to coldwater fishes.
agement plans that deal specifically with roadsalts (3). In the Thus, the classification system leads to a matrix of require-
United States, the approach has been somewhat different in ments, most of which are of a chemical nature, for support of
that aquatic resources are protected by broad federal legisla- the use, which in turn sets the stage for regulation of sources
tion, which is then applied incrementally to specific activities of impairment. The Clean Water Act regulates “point sources”
until the overall goals are achieved. Both systems are effective of pollution—those that originate from a specific discharge
if rigorously implemented. point (i.e., typically a pipe and typically excluding such
The most relevant federal legislation in the United States for unengineered sources as snow piles). Point sources of dis-
the protection of aquatic resources is the Federal Water Pollu- charge are required to hold a valid NPDES permit (see pre-
tion Control Act (1948) and its amendments. The act origi- ceding section), which prescribes the maximum amounts of
nally was focused primarily on public health through regulated substances in or attributes of a discharge. These
protection of domestic water supplies, but was drastically limits are established by application of the standards (i.e., by
broadened in 1972 and again in 1977, at which time the Act calculating the amount of a particular substance that could be
and its amendments became known as the Clean Water Act assimilated into a receiving water without causing a violation
(48). An amendment of the Clean Water Act in 1987 added of the standard). States or local governments can enhance
permitting requirements for certain stormwater drainages. federal regulations in some cases. The same is true in Canada
The goal of the Clean Water Act, as given in 33 USC Section for provinces. For example, the province of Quebec has
101, is to “maintain and restore the chemical, physical, and bio- detailed regulations dealing explicitly with snow dumps and
logical integrity of the nation’s waters”and, ultimately, to elim- snow removal.
inate pollution from the waters of the United States. Progress The U.S. Clean Water Act also allows for some segments
toward the achievement of this goal will be incremental. not to be classified according to use. In such cases, the states
The U.S. Clean Water Act places a set of requirements on must follow antidegradation regulations in issuing permits
states for protection of water quality, aquatic life, and aquatic for discharge. Antidegradation limits, which substitute for
environments, as summarized below (see http://www.epa. standards derived from specific uses, vary by state, but may
gov/region5/water/cwa.htm). To a large extent, implementa- require that the amount of any regulated substance cannot
tion and enforcement of the regulations occur at the state be changed by more than a specific percent (e.g., for Col-
level. State laws and state regulatory bodies typically imple- orado, it is 15 percent of the available assimilative capacity
ment the requirements of the Clean Water Act, but there is available at the time the regulation was adopted) from the
some flexibility, allowed through oversight procedures exer- ambient value. Some categories of nonpoint source pollution
cised by the U.S. EPA, for differences in the application of are not yet regulated through the Clean Water Act. Agricul-
Clean Water Act requirements from one state to another, ture, for example, is an acknowledged diffuse source of
according to regional circumstances. In some states, the U.S. pollution, but it is not explicitly regulated by the Clean Water
EPA maintains direct application of the Clean Water Act, Act, although the federal government has begun to invest
either because the state has opted not to seek authorization to money in incentives to reduce agriculturally derived non-
implement the act or because the U.S. EPA considers state point source pollution.
efforts to do so inadequate. Use of road application materials for snow and ice control
States are required to divide their waters into segments is not explicitly regulated by the U.S. Clean Water Act, but
(e.g., portions of drainage basins or clusters of water bod- could be regulated indirectly in urban areas through
ies that have similar characteristics) and to classify the stormwater permitting. Even so, individual states could use
waters thus subdivided according to use. Uses vary from the Clean Water Act as a template for state legislation that
state to state and may be elaborately subdivided in some would control application of materials to roads. Even in the
cases. Uses recognized by most states include domestic water absence of regulations, the framework of the Clean Water Act
supply, agricultural water supply, support of aquatic life, is useful as a basis for assessing the potential for impairment
and recreation. and the thresholds for impairment at specific locations. For
The Clean Water Act requires that each state have basic example, application of snow and ice control materials by a
standards for the protection of all uses in all waters of the state agency could be assessed on the basis of the rate of appli-
state. These standards, which are reviewable by the U.S. EPA cation, the rate of transport from the roadway to waters of the
and must be reviewed periodically by the state itself along state, and the resulting concentrations of regulated substances
with the classification system, set technically based thresholds in relation to the established standards for those substances as
for impairment of use. For example, a coldwater fish fauna used by the state in its point source regulations. This
that is part of the aquatic life classification for a particular approach has been taken by the state of Colorado and other

32
states and is used here as the main basis for evaluating the chloride, and dissolved organic matter, as well as other less
effects of road application materials to aquatic resources (see common possibilities.
Figure 3-2). A second category of constituents can be designated as
additives (reviewed by Cappelow et al. 1993 (49)). These sub-
stances, which often are proprietary and thus not described
3.4.2 Potential Types of Impairment of
specifically by vendors except under unusual circumstances,
Aquatic Resources by Snow and Ice
typically are intended to reduce corrosion or, in the case of
Control Materials
solids, caking of the main ingredients. They are additives in
From the view of potential impairment of aquatic that they are introduced intentionally at concentrations of 1
resources, it is useful to think of snow and ice control percent to a few percent of the major ingredients. Examples
materials as consisting of three groups of constituents (see vary widely, but anti-corrosion agents can include heavy met-
Table 3-2). The major constituents, aside from water in the als such as zinc, inorganic ions such as phosphate, and organic
case of solutions, are those that account for the bulk of the substances such as citrate; anti-caking agents include salt
dissolved or solid materials. Examples in the case of chemi- (NaCl) to prevent caking of sand and ferrocyanide to prevent
cal solutions could include magnesium, calcium, sodium, caking of bulk sodium chloride.
Figure 3-2. Flowchart showing concepts derived from the U.S.

clean water act for evaluation of snow and ice control materials.

33
Table 3-2. Three groups of constituents for consideration in

assessing impairment of aquatic resources.
Constituent Group Examples

Major Constituents Sand, chloride, acetate, magnesium, organic matter
Additives Phosphate corrosion inhibitor, ferrocyanide
Contaminants Heavy metals, phosphorus, nitrogen
A third category of constituents can be called contami- situation are best determined by an approach similar to that
nants. Substances belonging to this category are not intro- outlined in Figure 3-2.
duced intentionally; they are a by-product of the source Toxicity is not the only mechanism of environmental impair-
material and its processing. For example, heavy metals are ment. Impairment can be physical, as in the case of excessive
found in natural sources of salts used in many control mate- sand added to flowing waters, or it can be biogeochemical, as in
rials. Similarly, sand and other abrasives carry soluble or fine the case of eutrophication through addition of nutrients or
particulate components that will appear on the roadway through stimulation of microbial respiration by the addition of
when these materials are applied. organic matter. Impairment may even consist of complex multi-
The major ingredients of snow and ice control chemicals step processes such as those resulting from the displacement of
typically are chosen in such a way as to ensure that they are one ionic substance by another within a soil, thus causing a
benign in small to moderate amounts. For example, sand as change in the chemistry of surface waters. Some mechanisms of
an abrasive or calcium and magnesium as components of impairment are well documented; others are not.
chemical mixtures would not commonly be suspected of Along with mechanisms, gradients of concentration must
causing environmental damage through toxicity or other be considered as an aspect of environmental impairment. For
direct mechanisms. When characteristically benign sub- example, a heavy metal included as an additive or contami-
stances are used in very large amounts, however, potential nant in a deicing compound could reach toxicity thresholds
environmental damage should be considered. Additives, in undiluted deicer but still not cause receiving waters in the
although used in relatively small amounts compared with the environment to rise above the toxicity threshold because of
main ingredients of deicers, can be harmful in smaller dilution processes occurring on or near the roadway. Dilution
amounts. This is the case for heavy metals, which are toxic to is further discussed in Section 3.4.3.1.
aquatic life above specific thresholds that often are quite low,
and phosphate, which is a potential source of nutrient enrich-
ment that can cause eutrophication (see below). Contami- 3.4.2.1 Major Constituents
nants present perhaps the greatest problem for evaluation Major constituents that are most common in snow and ice
because they often are not quantified fully by vendors and control materials include particulate abrasives, alkali or alkali-
encompass a wide range of compounds, some of which (e.g., earth cations (i.e., calcium, magnesium, sodium, potassium),
mercury and cadmium) can be a matter for concern at very the anion chloride, and organic substances that may either be
low concentrations. synthesized (e.g., acetate) or derived from biomass (e.g., corn
The following text discusses potential environmental by-products).
impairment associated with each of the three main groups of
constituents shown in Table 3-2. Toxicity is perhaps the first Particulate Abrasives. Particulate abrasives include sand,
type of impairment to come to mind and needs to be consid- crushed rock, cinder, and potentially any other particulate
ered for all constituents. Furthermore, judgments on toxicity substance available in quantity at a specific location. Such
need to reflect the current sophistication of knowledge about materials present two general types of potential impairment:
toxicity in relation to concentration. It is not appropriate to chemical and physical. Evaluations of potential for chemical
designate specific substances as toxic and others as nontoxic, impairment could involve toxicity tests (see Section 3.4.3.2)
given that toxicity is a matter of concentration. For example, and tests for chemical composition (see Table 3-3). The first
the metal zinc is an essential nutrient for aquatic plants and priority among these is toxicity testing. A general toxicity test
animals, but is also toxic to aquatic life at relatively low con- should involve not only the raw application materials with
centrations (50). Thus, it does not make sense to classify zinc appropriate amounts of dilution, but also pulverized material
as a toxic substance, but rather to think in terms of the to represent the actual condition of the abrasive on the road-
amounts and forms of zinc that can be present in a given envi- way under traffic. Any indication of general toxicity would
ronment without impairment of aquatic life or other uses of lead to more specific testing to determine specific con-
water. The maximum amounts of any substance that can be stituents responsible for the general toxicity. Consideration of
present without environmental impairment for a specific organic and nutrient content is separate from consideration

34
Table 3-3. Chemical composition of three sources of sand (with salt

added) used by CDOT during winter, 1997–1998.
µg/g Dry Mass

Salt with Sand
Source 1 Source 2 Source 3
Major Ions
Ca 197 146 101
Mg 138 <5 <5
Na 52,000 18,000 9,000
Cl 79,000 27,000 14,000
Nutrients
P Soluble Reactive 0.27 1.01 0.34
P Total Soluble 0.31 1.20 0.38
P Particulate 1.60 2.03 2.09
P Total 1.91 3.23 2.47
Other Inorganics
Cu soluble 0.15 0.15 0.25
total 0.30 2.5 4.3
As soluble <0.5 <0.5 <0.5
total <0.5 <0.5 <0.5
Cd soluble <0.00 <0.00 0.01
total 0.06 0.06 <0.02
Cr soluble <1 <1 <1
total <1 <1 <1
Hg soluble <0.01 <0.01 <0.01
total <0.01 <0.01 <0.01
Ni soluble <1 <1 <1
total 1 <1 <1
Pb soluble <0.5 <0.5 <0.5
total 0.6 <1.7 <4.2
Se soluble 0.15 <0.1 0.1
total <0.1 <0.1 <0.1
Zn soluble <1 <1 <1
total 1.0 5.5 8.6
Source: (52)
Large variation in Na and Cl corresponds to varied additions of NaCl, which prevents caking
of toxicity. Organic content for abrasives typically would be percentage of fines through reduction in average particle
low and thus usually not important. Nutrient content is eas- size on the roadway under traffic. Fines typically reach sur-
ily determined and, although unlikely to raise impairment face waters in suspended form and remain in suspension
issues in most cases, deserves consideration. until they reach an area with low water velocity, particularly
Physical impairment caused by abrasives can be attrib- lakes and wetlands, where they are removed by gravity. Fines
uted either to particulate bed load or to suspended solids. may contribute to blockage of interstitial spaces in stream
Most particles that fall within the size category of sand beds, particularly if the spaces are already partially blocked
(0.0625 to 2.0000 mm) typically move through a stream as by sand (51).
bed load, except under the highest flow conditions. Bed load Suspended solids (often measured as turbidity) are regu-
consisting of sand can block interstitial spaces between par- lated for the protection of aquatic life and sometimes also for
ticles of larger size, such as gravel or cobbles. Under these aesthetic reasons (Federal Register, 10 December 1998, Vol. 63,
conditions, substantial impairment of invertebrate and fish Number 237). Thus, it is possible to use dilution testing in
communities may occur because the interstitial spaces are combination with site-specific water-quality standards for
important either as habitat or, in the case of fish, for spawn- suspended solids in evaluating the suitability of a particular
ing (51). Interception of sand before it enters the stream abrasive.
is possible through retention of storm flow because parti-
cles of this size settle relatively rapidly, but interception of Chloride. Chloride is the third most common anion in
sand typically is expensive and requires maintenance involv- inland waters (53). Its natural concentrations range from
ing the removal of accumulated particles from retention approximately 1 mg/L at large distances from marine sources
basins. in areas that lack sedimentary rocks up to concentrations that
Abrasives typically contain fines (here used to mean equal or exceed those of seawater under arid conditions
particles smaller than sand), but also show an increase in where concentration by evaporation occurs. Most commonly,

35
background chloride concentrations fall within the range of flow) is low. Concentrations are suppressed by spring runoff.
1 to 10 mg/L (54). Because chloride is consistently present in Even during the summer after runoff, chloride concentra-
inland waters, the tolerance of aquatic life for chloride, which tions are notably above background. Despite the large change
is a component of protoplasmic fluids, is quite high (31, 49). in chloride concentrations in a relative sense, however, chlo-
The freshwater aquatic life criteria for chloride as deter- ride concentrations in this case did not exceed the aquatic life
mined by the U.S. EPA (Federal Register, 10 December 1998, chronic standard, even though the stream is relatively small
Vol. 63, Number 237) are 860 mg/L for acute exposure (one and the rate of application for the road, which is at the
hour) and 230 mg/L for chronic exposure (96 hours or more); Continental Divide, was very high.
standard application of the criteria allows one exceedance Numerous cases have been recorded in which surface waters
every 3 years. Chloride is highly mobile in soils (55). Thus, a receiving road runoff have concentrations of chloride far
large portion of the chloride leaving a road (40 to 70 percent) exceeding the aquatic life standards (summarized by Environ-
can be expected to be flushed from the adjacent watershed in ment Canada (3, 99)). These examples involve situations for
the same year (19, 52). which relative dilution is especially low. As explained in Sec-
The U.S. EPA’s aquatic life criteria or associated state stan- tion 3.4.3.1, relative dilution is under the combined influence
dards can be used in evaluating whether or not a snow and ice of rate of application, road surface per unit of drainage area,
control product leaving the roadway under the influence of climate, and related factors.
precipitation will impair aquatic life and, if so, the distance Pollutants that accumulate in terrestrial or aquatic envi-
beyond the roadway that would be encompassed by any zone ronments can be especially troublesome if pollution contin-
of exceedance. Because some freshwater organisms (e.g., adult ues over long intervals. Among dissolved solids found readily
rainbow trout) are much more tolerant than others of chlo- in surface waters, chloride is among the least likely to accu-
ride, concentrations of chloride exceeding the criteria are mulate. As stated by Drever, chloride is not removed or sup-
more likely to change the composition of aquatic communi- plied significantly by interaction with rocks and is not
ties than to impair all species (19, 56); even so, such changes precipitated as a salt until very high salinities are reached.
could be of major ecological concern. At even higher concen- Thus, the annual addition of chloride will raise the mean and
trations, chloride could be harmful to the potability of water, maximum concentrations of chloride in the waters of a par-
as can be the case of drinking water wells near a road where ticular watershed by adding a concentrated source, but will
chloride is added (21). not lead to significant accumulation of chloride in soils or
Where chloride is applied as a main ingredient of snow and waters on an annual basis, except where there is not enough
ice control materials, substantial augmentation of chloride water to produce runoff, or where the deep waters of lakes or
concentrations in receiving waters is likely. Figure 3-3 shows ponds become stabilized and insulated by excess density
an example for two Colorado streams. Background concen- derived from salt (see below). Seasonal variations in chloride
trations of chloride in these streams are very low (ca. 1 mg/L), may occur, of course, in response to seasonality of applica-
and the addition of snow and ice control materials raises the tion of chloride or the immobilization of runoff by very cold
concentrations by a factor of as much as 100. Peak concen- weather.
trations occur during the application season, when the sup- Chloride may liberate mercury bound to wet sediments
ply of chloride is high and the amount of dilution (stream (57). This phenomenon has been little studied, however, and
Figure 3-3. Comparison of chloride concentrations (mg/L) in (1) Straight

Creek, Colorado, which receives runoff from I-70 affected by sand and
salt and magnesium chloride solutions, and (2) Laskey Gulch, a tributary
of Straight Creek that does not receive road drainage (Source: (52)).

36
unusual mercury concentrations in water have not yet been waters with sodium at thresholds that might meaningfully
reported for streams receiving road drainage containing affect daily intake of sodium by humans (61). To be regulated
chloride. through the U.S. Safe Drinking Water Act, a substance must be
known to have adverse effects on human health, be likely to
Alkali and Alkaline-Earth Cations. The cation forms of occur in public water systems at levels of public health con-
calcium, magnesium, sodium, and potassium are common cern, and be reducible by regulation by an amount correspon-
ingredients, either singly or in mixtures, of liquid snow and ding to a meaningful reduction in health risk for persons
ice control solutions. These are also the four major cations by served by public water systems. The U.S. Safe Drinking Water
concentration in unpolluted inland waters (53). Conse- Act amendments of 1996 required the U.S. EPA to publish a
quently, tolerances of aquatic life and of human beings for list of unregulated contaminants designated “Contaminant
these ions are quite high, and, because all four of the ions are Candidate List,” or CCL. The contaminants on this list are dif-
constituents of protoplasm (e.g., Cowan (58)), they can be ferent from contaminants included in actual or proposed
listed as essential nutrients for all animals and plants. national primary drinking water regulations. Thus, the list is
Calcium and magnesium are held moderately well by soils, essentially a means of keeping track of contaminants not reg-
where they may be bound firmly to phosphate or carbonate ulated but could be regulated in the future. The 1998 CCL,
or to organic complexes or minerals (59). Soils also release which followed from the 1996 regulation, listed sodium
calcium and magnesium to some degree. This release sustains among 50 chemical constituents.
concentrations of calcium and magnesium in inland waters No health advisory has ever been issued by the U.S. EPA for
that range from a few mg/L in waters derived from granitic sodium. A drinking water guidance is 20 mg/L, which is non-
lithology to much higher concentrations where rocks are sed- enforceable and based on a recommendation of the American
imentary or where concentration occurs by evaporation (55). Heart Association; the U.S. EPA believes that the guidance is
Calcium and magnesium are among the first ions to precipi- excessively restrictive, but has not made official changes in it
tate from water as salts are concentrated (60) and thus tend to as yet. The U.S. EPA also has used the National Research
be depleted in highly saline waters. Council’s dietary guidance of 120 mg/L, which corresponds
There are no aquatic life standards or other general use to 10 percent of average total sodium intake at a drinking
protective upper limits for calcium and magnesium in water intake of 2 liters per day, for evaluating sodium con-
water, except as related to hardness associated with very centrations in drinking water. In a national survey of 59,000
high concentrations of Ca and Mg in relation to domestic groundwater-based systems, approximately 13 percent
water supply for maintenance of cleaning performance. showed concentrations in excess of 120 grams per milliliter.
Thus, it is unlikely that addition of calcium or magnesium Causes for exceedances were not diagnosed for this study,
to water would be seen as an environmental threat or a however.
threat to the domestic use of water. As in the case of chlo- In its most recent drinking water advisory, the U.S. EPA
ride, addition of calcium and magnesium to soils could notes that about 3 percent of the U.S. population is on
change either the physical features of the soil or cause ion sodium-restrictive diets, which range in stringency between
exchange reactions to occur that would result in the libera- 250 and 2000 mgd. Drinking water concentrations of sodium
tion of other substances that would not otherwise be would be mandatory for inclusion in calculations of the daily
released (59). However, no published literature indicates sodium intake in these cases. Typical dietary intake of sodium
that the calcium and magnesium derived from snow and ice in the United States is 4000–6000 mgd. At concentrations of
control materials would be sufficient to induce large about 500 mg/L, the palatability of water is notably reduced,
changes involving these mechanisms. which would likely cause a search for an alternate drinking
Sodium and potassium are weakly held by soils under most water supply. Thus, except for highly restrictive diets, drink-
circumstances (59). They are present in inland waters at con- ing water intake is likely to be a minor proportion of the total.
centrations that range from about 1 mg/L up to very high The U.S. EPA concludes that regulation may not present a
concentrations in saline waters (55). Sodium and potassium meaningful opportunity for reduction of health risk, but con-
are less likely to precipitate chemically than calcium and mag- tinues to believe that the drinking water advisory is useful
nesium and so may be concentrated by evaporation to a high where water is contaminated with sodium for roadsalts; spe-
degree. cific sodium contamination of surface water supplies that
There are no general aquatic life standards for sodium or would impair raw water for treatment and distribution is
potassium. Because high daily intake of sodium is contraindi- much less likely than contamination of individual wells near
cated for individuals who have high blood pressure, there are roadways. If contamination of a central supply did occur, evi-
reasons to consider potential for contamination of ground- dence of it would be known from mandatory testing of cen-

37
tralized water supplies as conducted in the United States. breaks down. Under extreme conditions (long winters with
Exceedance of guidelines is most likely for individual shallow high use of sodium chloride), the isolation of the saline bot-
wells very near roadways where sodium-containing deicers tom layer may be essentially permanent, and the accompa-
are used in quantity. In such instances, it is advisable that well nying bottom conditions would not support aquatic life
water be tested annually during winter or spring for sodium except for microbes. A lake’s permanent failure to mix fully
contamination. is designated meromixis (53), which can be maintained by
The effect of applying any of the four major cations as a recurrent runoff of saline water. Although there is no easy
major ingredient of snow and ice control materials is to modeling or other well-established approach to predict the
increase the supply of these substances to receiving waters. development of salinity-induced layering caused by snow
Where the background concentrations of the major cations and ice control materials, the condition is relatively easily
are low to moderate, the result typically will be a notable diagnosed in the field with commonly available instruments
increase in concentration. For example, Lewis (52) docu- (e.g., dissolved oxygen meter, conductance meter).
mented changes in annual maximum magnesium concentra-
tion ranging from 50 percent to almost tenfold in Colorado Organic Matter. Organic matter may be synthesized indus-
streams receiving magnesium as a component of a magne- trially as small molecules to make up a bulk anionic compo-
sium chloride solution and increases in sodium concentra- nent of a chemical snow and ice control material (e.g., acetate
tions by a factor of as much as 25 when sand containing salt for CMA). Alternatively, unrefined organic matter from agri-
was applied to roads in the same watersheds. Although these culturally derived biomass (particularly corn) may be used as
changes in concentration represent a substantial chemical a nonionic additive to a saline mixture such as a magnesium
perturbation of the receiving waters, there is no evidence at chloride solution. In either case, the primary type of impair-
present that they cause any biotic responses, given the benign ment to be explored is an increase in biochemical oxygen
nature of the four major cations within the lower ranges of demand (BOD) in the receiving waters, which could produce
concentration. hypoxic (low oxygen) or anoxic conditions in these waters.
Biochemical oxygen demand is caused by the use of organic
Saline Mixtures. Liquid snow and ice control mixtures substrates by microbes, which oxidize the organic matter
typically are saline; their ionic content usually includes one or using oxygen obtained from the water, thus creating a strong
more of the four major cations mentioned above as well as oxygen demand. If the microbial demand is not offset by
chloride or a small organic anion. Total salinity of water is reaeration processes or by photosynthesis, hypoxia or com-
subject to regulation for human potability (500 mg/L, sec- plete loss of oxygen can be the result. These kinds of adverse
ondary drinking water standard). Salinity may affect the wel- effects involving BOD are most likely in waters that move
fare of aquatic life, and in part is controlled for this purpose slowly because such waters have minimal physical exchange
by chloride standards (see the discussion of chloride above). of oxygen with the atmosphere (i.e., low reaeration rates).
An indirect effect of salinity that occurs through its influ- Temperature also affects the rate of oxygen consumption, but
ence on water density has been documented and should be relative influence of water management is likely to be pre-
considered a potential type of environmental impairment dominant, given that temperature varies little for liquid water
associated with runoff containing salinities well above those during winter. Environments most likely to be affected would
of the ambient waters. At a given temperature, saline waters include pools, wetlands, and small lakes.
are denser than waters containing only small amounts of The amount of organic matter in a liquid mixture can be
salts (53). Waters containing moderate to high amounts of determined by carbon analysis (typically through combus-
salt, therefore, may form a distinct bottom layer in a small tion followed by infrared spectrophotometry: APHA 1998
water body such as a pond or wetland pool or in some cases Method 5310 B (67)). The potential oxygen demand of a spe-
even a lake (63–65). The formation of such layers is likely cific amount of carbon can be obtained from the stoichiom-
only where current velocities are low (i.e., it is unlikely in etry of respiration (slightly over one mole of O2 per mole of
streams or other waters flowing at velocities exceeding a few C). Thus, the total oxygen demand can be calculated for a
centimeters per second). If such a layer does form, however, given mass or volume of material applied to a roadway.
it can be residual in that mixing caused by wind may be Organic matter also can be measured indirectly with a stan-
insufficient to disrupt the layering. If so, the bottom layer, dard BOD test (67). The difficulty in using information from
being deprived of contact with the surface, could become either test is that adverse effects of organic matter on oxygen
anoxic (without oxygen). The development of anoxic condi- are determined in large part by the speed with which the
tions over sediment not normally anoxic can be harmful to organic material is decomposed by microbes in the natural
aquatic life, at least on a temporary basis until the layering environment and the likelihood that standing waters or

38
slowly flowing waters will intercept the organic matter as it Table 3-4. Comparison of two brands of MgCl2
passes through the drainage network. This issue is discussed deicer, one of which probably contains zinc and
further in Section 3.4.3.3. phosphate as additives.
Natural concentrations of total organic carbon in surface
MgCl2 MgCl2 Environmental
water range generally between 2 and 15 mg/L, although occa- Constituent Type 1 Type 2 Standard
sionally they may be higher (66). The corresponding BOD
Zn, mg/L 1 39 0.07
would be approximately 3 to 20 mg/L (67). Snow and ice con-
Total P, mg/L 12 82 0.015*
trol materials not containing organic matter as a major ingre-
dient may have organic carbon content as high as a few Source: (52)
*no state standard; typical stream concentration is shown
thousand mg/L (e.g., Lewis 2000 for magnesium chloride
(68)), which corresponds to BODs that could be as high as
5000 to 10,000 mg/L. In the context of expected dilution on Other Inorganic Substances. Phosphorus, in the form of
or near the roadway, however, the total augmentation of phosphate, may be used as an additive for corrosion control.
oxygen demand from this source would be noticeable only Snow and ice control materials and additives derived from
over a very small area or perhaps not noticeable at all, agricultural matter are also associated with high phosphorus
depending on the speed of water movement. concentrations. A complete discussion of environmental
The presence of organic matter as a major ingredient is problems associated with phosphorus is given below in con-
associated with total organic carbon concentrations within nection with contaminants. Phosphorus is present in only
the range of 20,000 to 70,000 mg/L (e.g., Lewis 2000 (68)), small amounts in unpolluted waters (typically 10 to 100
which corresponds to BOD exceeding 100,000 mg/L (see also µg/L). Therefore, an additive that contains phosphate, even if
data tabulated by Fischel 2001 (69)). In this case, organic mat- constituting less than 1 percent of total ingredient mass (e.g.,
ter subsequent to dilution on or near the roadway could Table 3-3), can increase phosphorus concentrations in the
exceed the amount of organic matter in water by a factor of receiving waters to an amount that might be undesirable from
10 or more and thus could affect a substantial volume of the viewpoint of eutrophication. As in the case of heavy
receiving water, especially if the water happened to be stand- metals, the best approach is to use stream standards or other
ing or moving slowly. Thus, inclusion of organic matter as a guidelines (e.g., antidegradation principles) to set a cap on
major ingredient in deicers raises the issue of oxygen deple- the amount of phosphorus that can be added and advising
tion as a potential impairment. Even though this potential vendors of this maximum amount.
obviously exists, it is not always manifest. For example, if Ferrocyanide salts (e.g., Na4FeCN6) are added as anti-
runoff immediately joins a large, rapidly flowing water body caking agents to bulk salts (primarily NaCl) in solid form.
that offers a very high amount of dilution, the effects of even The free cyanide ion (mostly joined to hydrogen as HCN) is
these amounts of organic matter may not be detectable. highly toxic, but it has a high affinity for metals and, there-
fore, often is present in bound form, in which case it is much
less toxic (50). Ferrocyanide itself has relatively low toxicity,
3.4.2.2 Additives but in the presence of solar irradiance or low pH, it may dis-
Although additives may include any substances that ven- sociate and thus liberate the cyanide ion (3). Therefore, fer-
dors find capable of enhancing performance, the most likely rocyanide is an additive of concern and is listed by the U.S.
prospects are selected heavy metals, inorganic anions, and EPA as a toxicity concern. There is insufficient evidence to
small organic molecules. date as to whether liberation of free cyanide from ferro-
cyanide in connection with roadsalts is a serious problem
Heavy Metals. Heavy metals may be used with liquid (3, 20, 70).
snow and ice control formulations as a means of reducing
corrosion. Zinc is an example of a metal used in this way Organic Matter. Small organic molecules such as citrate or
(Table 3-4). Potential risks associated with zinc or other triethanolamine (TEA) may be used to inhibit corrosion.
heavy metals are discussed below in connection with con- When added at 1 percent to a few percent, small organic sub-
taminants. Use of heavy metals at unacceptable concentra- stances such as this, although easily biodegraded and thus
tions in the context of designated uses, especially aquatic life potentially a source of BOD in surface waters (35), may be
uses, which are particularly restrictive for heavy metals, can acceptable because dilution reduces their concentrations rap-
be achieved by determining thresholds for heavy metals idly to such a degree that negative effects on oxygen concen-
through dilution calculations in conjunction with numeric trations are unlikely to cause impairment. Impairment is least
standards for specific metals on specific stream segments likely where receiving waters have high energy and where the
(Figure 3-2). added organic matter, after dilution on or near the roadway,

39
is near the natural inventory of organic matter in the water gives the combination of the original SRP plus the dissolved
(usually 2 to 15 mg/L (53)), and most problematic where organic phosphorus that was unreactive prior to digestion,
runoff enters small bodies of still water and where the added thus producing an estimate of total dissolved phosphorus
concentrations greatly exceed the natural inventory of dis- (TDP). Subtraction of SRP from total dissolved phospho-
solved organic matter. rus produces an estimate of DOP. Both SRP and DOP often
are quite scarce in unpolluted natural waters (e.g., < 10 µg/L
for each of these fractions (72)). Constant uptake of phos-
3.4.2.3 Contaminants phorus in these forms, which are available for uptake by
Contaminants can be divided for convenience into key plants, suppresses the ambient concentrations, even though
nutrients, inorganic substances other than key nutrients, and soluble phosphorus is also being generated by the decom-
organic substances. Although any of the 20 or more elements position process.
required by organisms for the synthesis of biomass can be Particulate phosphorus may be structurally bound to
classified as nutrients, the nutrients of greatest importance in organic material or it may be bound (mostly externally) to
aquatic systems are nitrogen and phosphorus. These nutri- particles; clays are particularly good carriers of phosphorus
ents often are called “key nutrients” or “limiting nutrients” in adsorbed form (59). Total phosphorus typically is
because they are in lowest supply relative to biotic demand analyzed by a digestion procedure that liberates the phos-
(71). Addition of key nutrients to surface waters often causes phorus as soluble reactive P, followed by application of the
a biotic response in the form of increased plant growth. SRP test (67).
Where this increased plant growth markedly increases the The dissolved fractions of phosphorus are more readily
amount of plant biomass, the result is generally considered available for uptake by plants than the particulate fractions.
undesirable and is termed “eutrophication.” Vendors whose products contain a substantial amount of
phosphorus may explain that a substantial portion of the
Phosphorus. Phosphorus (P) is often considered the ele- phosphorus is in the particulate fraction or a stable compo-
ment most likely to limit plant growth in inland waters. Phos- nent to the organic fraction and therefore should not be
phorus has a strong multiplier effect on plant biomass counted as a potential agent of impairment for the environ-
because it constitutes only about 1 percent of the dry mass of ment. This is incorrect in that these forms of phosphorus
plant matter (72). Therefore, a plant community (consisting often can be mobilized through biogeochemical processes
of rooted aquatic plants or of algae) that has exhausted its (53). It is customary in estimating the critical phosphorus
phosphorus supplies, which is often the case under natural loading of lakes, for example, to make evaluations of impair-
circumstances, will respond to the addition of a pound of ment based on total phosphorus load (75). Therefore, the
phosphorus by the creation of about 100 pounds of new dry environmental characteristics of snow and ice control mate-
biomass, or approximately 500 pounds of wet biomass. rials should be evaluated primarily in terms of total phos-
Because of this multiplier effect, relatively small amounts of phorus, rather than specific phosphorus fractions.
phosphorus can change aquatic ecosystems drastically. Unde- Phosphorus can be expected as a contaminant in all mate-
sirable side effects include excessive organic matter leading to rials derived from natural sources, including salt deposits
deoxygenation of poorly circulated waters, aesthetic nuisance, and biomass. Abrasive mixtures (e.g., sand) and salt deposits
harmful algal blooms, and change in the community compo- used in preparing liquid snow and ice control mixtures typ-
sition of organisms, including fishes (73). ically contain relatively small amounts of P contamination
It is traditional for environmental purposes to divide total (Table 3-5), although each new material for proposed use
phosphorus in surface waters into three categories: dissolved should be evaluated for phosphorus content. In contrast,
inorganic, dissolved organic, and particulate (67). The sum of when biomass is a bulk component (main ingredient)
these categories is total phosphorus. Dissolved inorganic of control materials, the materials often show very high
phosphorus is present in water primarily in the form of ion- phosphorus content (Table 3-5).
ized phosphoric acid (60). It is often referred to as “soluble The vulnerability of natural aquatic environments to per-
reactive phosphorus” (SRP) because the test by which it is turbation by phosphorus loading varies according to the
determined leads to the detection not only of dissolved inor- natural phosphorus supply to the environment. Waters that
ganic phosphorus, but also an indeterminate fraction of small receive small amounts of phosphorus from natural sources
phosphorus-bearing organic molecules (74). are termed “oligotrophic” and can be sensitive to small addi-
Dissolved organic phosphorus (DOP) typically is deter- tions of phosphorus (72). Water bodies that lie within
mined through digestion of a filtered sample. The digestion watersheds with naturally rich phosphorus supplies can tol-
procedure reduces all organic dissolved phosphorus to erate more anthropogenic phosphorus without showing sig-
inorganic form. Subsequent application of the SRP test nificant change. Therefore, the thresholds for phosphorus

40
Table 3-5. Concentrations of nutrients in three classes of snow and ice

control materials.
Sand with Salt, µg/g MgCl2, mg/L MgCl2 + Organic, mg/L
Brand 1 Brand 2 Brand 1 Brand 2 Brand 1 Brand 2
Soluble Reactive P 0.27 1.01 0.4 4.3 28 40
Total Dissolved P 0.31 1.2 0.7 11.7 67 116
Particulate P 1.6 2.0 0.3 2.7 3 18
Total P 1.9 3.2 1.0 14.4 70 134
Ammonia - - 4 - 31 43
Source: (52, 68)
“ - ” indicates data not available
addition can be expected to vary significantly from one loca- ularly fishes and invertebrates) introduces an additional
tion to another. potential for impairment (78).
Because phosphorus is not a toxicant, many states in the The thresholds for impairment of aquatic life by ammonia
U.S. have no standards for phosphorus except perhaps on a have been thoroughly studied, but they must be evaluated on
site-specific basis for especially sensitive areas. There is a a site-specific basis. Cold water fauna, for example, are much
general conclusion among aquatic scientists, however, that more vulnerable than warm water fauna. Furthermore, the
controls are needed on nutrients in the U.S., and particularly toxicity of ammonia in water varies with pH and tempera-
phosphorus (71). The U.S. EPA has been directed to develop ture. In 1999, the U.S. EPA prepared new criteria for ammo-
and circulate nutrient criteria that it proposes as the basis for nia (80) to replace the older criteria (1985 (78)) that had been
control of nutrients (phosphorus and nitrogen) in surface used throughout the United States for almost twenty years.
waters (76). The U.S. EPA likely will insist that states include The new criteria are being adopted by states through require-
limits on nutrients in their basic standards. When this occurs, ments of the U.S. EPA. Based on either the new criteria or the
there will be numeric standards for phosphorus and nitrogen older criteria that still govern many state standards, the
in all states. Until then, advisory information on nutrient appropriate maximum ammonia concentration for a given
concentrations can be obtained from the U.S. EPA criteria or site can be evaluated numerically.
from the application of an antidegradation concept. Nitrate also is readily assimilated by plants in water and can
serve as a primary source of nitrogen to support protein syn-
Nitrogen. Like phosphorus, nitrogen can be divided for thesis when ammonia is scarce, as is commonly the case.
environmental purposes into dissolved inorganic, dissolved Nitrate typically is more abundant in inland waters than
organic, and particulate forms. The composite of these is total ammonia, although unpolluted waters typically show con-
nitrogen. Dissolved inorganic nitrogen in inland waters concentrations below 200 µg/L (79). Nitrate, in contrast to
sists almost entirely of ammonia and nitrate. Dissolved ammonia, is not toxic to aquatic life at any concentration
organic forms are partially labile (easily used by microbes and likely to occur even in wastewaters, although at very high con-
readily recycled in the environment) and partly refractory centrations (10 mg/L) there is a human-health threshold for
(77). Particulate nitrogen typically is a component of living or nitrate content in drinking water supplies (http://www.epa.
nonliving biomass. gov/safewater/mcl.html).
Ammonia is easily assimilated by plants, for which it is an Dissolved organic nitrogen in natural systems consists of a
ideal nitrogen source. Ammonia is also a toxicant, however, mixture of tissue degradation byproducts and excreted
and can cause impairment or death of aquatic life above spe- organic compounds that come from organisms. Examples of
cific threshold concentrations (78). Thus, concentrations of substances that would be included under this heading are
ammonia are regulated for protection of aquatic life. Natural amino acids and nucleotides. In addition, more complex
concentrations of ammonia in toxic waters are very low organic materials containing nitrogen are derived from soils
(<50 µg/L; Lewis (79)) because of the high plant demand for and mostly fall under the heading of humic and fulvic acids.
ammonia and the tendency of microbes to convert ammonia, These compounds are resistant to degradation (53). Typically
in the presence of oxygen, to nitrate by the process of nitrifi- organic nitrogen is similar in concentration to inorganic
cation. Enrichment of the low concentrations of ammonia in nitrogen in inland waters, although dissolved organic nitro-
natural waters may stimulate plant growth where nitrogen is gen may become the primary dissolved form if the demand
a limiting nutrient, and in this way ammonia could be con- on the inorganic pool is strong. Natural concentrations typi-
sidered a potential source of impairment related to eutroph- cally are below 400 µg/L (79).
ication (71). At even higher concentrations (usually >1 mg/L, Particulate nitrogen usually is the smallest component of
but variable), the toxicity of ammonia to aquatic life (partic- total nitrogen in natural waters (79). Because it comes from

41
biomass, it is typically associated with various natural organic As a result, the list consists of heavy elements recognized as
compounds that compose organisms. most likely to cause impairment. This list of elements com-
As in the case of phosphorus, total nitrogen should be the monly is referred to as “heavy metals.” Strictly speaking,
main index by which nitrogen content is evaluated. A separate however, some of the elements on the list are not metals
evaluation of ammonia and nitrate content is warranted (Table 3-6).
because ammonia is potentially toxic to aquatic life at high con- The substances listed in Table 3-6 can impair waters either
centrations and nitrate is limited for drinking water supplies. for support of aquatic life or for human consumption. For
Ammonia is unlikely to be a significant component of snow most substances, thresholds of impairment for aquatic life
and ice control mixtures that are primarily inorganic. For exam- are lower than the thresholds for impairment of domestic
ple, salt and sand mixtures or liquid control materials derived water supply (e.g., see U.S. EPA criteria cited above). The
from natural salts typically do not contain large amounts of substances differ substantially in thresholds of concentration
ammonia (Table 3-5). Biomass, when present as a bulk con- considered to be toxic. For example, thresholds for mercury
stituent of deicers, may contain large amounts of ammonia, typically are in the range of 10 parts per trillion because mer-
however, which requires evaluation from the viewpoint of tox- cury is multiplied in concentration as it is transferred from
icity and potential stimulation of plant growth (Table 3-5). one level in the food web to another. Cadmium and silver
Nitrate is an unlikely major contaminant of snow and ice have somewhat higher but still very low thresholds, whereas
control materials, although analysis for nitrate is warranted most of the other substances have thresholds considerably
whenever a new type of material is considered for use. Simi- higher, but still low on an absolute scale. The actual toxicity
larly, dissolved organic nitrogen and particulate nitrogen are of the substances shown in Table 3-6 typically is conditional
likely to be present only in small amounts in inorganic snow on the environmental matrix. For example, hardness (i.e.,
and ice control materials, but many be more problematic for calcium and magnesium content) affects the toxicity of cop-
materials containing bulk organic matter from biomass; little per and zinc (see U.S. EPA criteria cited above). Also, inor-
is known of these components. ganic dissolved forms generally are far more toxic than
bound forms to aquatic life (60).
Inorganics Other Than Key Nutrients. Although the list Those elements from the list in Table 3-6 most likely to
of inorganic substances that can be found in the environ- create problems by way of contamination in snow and ice
ment runs to many thousands of compounds, environ- control materials include zinc, copper, lead, and cadmium (3,
mental impairment typically can be evaluated from the 52, 69). All four of these elements tend to become concen-
perspective of a relatively short list of elements that have trated, probably by evaporation mechanisms, in salt deposits
the highest possibility to cause impairment through toxic- used as sources of magnesium chloride for liquid snow and
ity either in elemental form or as a constituent of an inor- ice control materials. Other elements on the list may create
ganic compound. The elements of low atomic mass do not problems under specific circumstances. Variation in the con-
appear on the list because they are either rare or are known taminant concentrations for elements in Table 3-7 from one
not likely to cause impairment through toxicity. Many vendor to another or even from one batch to another should
heavy elements also do not appear because of their scarcity. be anticipated.
Table 3-6. Heavy metals and other inorganic substances most likely to
contaminate snow and ice control mixtures.

42
Table 3-7. Chemical composition of two potential to cause eutrophication or oxygen depletion, and
sources of liquid MgCl2 used by CDOT during (4) field monitoring.
winter, 1997–1998.
Milligrams per Liter 3.4.3.1 Dilution Analysis

MgCl2 Solution
Brand 1 Brand 2 Because the harm that can be caused by a given substance
Cu sol 0.2 0.6 that might enter an aquatic environment from a road
tot 0.7+ 0.6+
As sol 2.4 6.4
depends strongly on concentration of the substance in the
tot 2.2* (Not Applicable) environment, the amount of dilution that will occur between
Cd sol <0.01 0.01 the road and the receiving waters (e.g., streams, lakes, and
tot 0.11* <0.05*
Cr sol <2 <2 wetlands) is estimated as a first step in evaluating potential
tot <2 <2 impairment (52, 68). The simplest approach is to use histor-
Hg sol <0.02 <0.02
tot <0.02 <0.02 ical information for application rates in relation to the total
Ni sol <2 <2 amount of winter precipitation. Such an examination of
tot 6 <2
Pb sol <1 <1
historical data leads to calculation of a characteristic ratio of
tot <1* <1* annual application rate to annual winter precipitation. For
Se sol 0.2 0.02 simplicity, it can be assumed that most or all winter precipi-
tot <0.2 <0.02
Zn sol <2 <2 tation falling on the road becomes runoff and that it carries
tot <2 4* away all of the application materials. These generalizations
Source: (52) are not entirely correct but, when used in conjunction with
sol = soluble; tot = total recoverable a margin of safety to offset estimation errors, they provide a
*Lower than the same brand in 1996-1997
+
Higher than the same brand in 1996-1997 basis for screening. More detailed types of quantification can
be conducted if screening shows that the likelihood of
impairment is not clear from simple approximations. In
3.4.3 Evaluation of Specific Snow and Ice
most cases, however, either there is clear potential for impair-
Control Materials
ment or there is no reason to suspect impairment from a
Evaluation of specific deicers involves the use of screening simple dilution analysis. Figure 3-4 illustrates the steps taken
protocols with sufficiently broad coverage to consider all for determination of expected dilution for snow and ice con-
potential types of impairment and to take site-specific control materials. The substance for which the calculations are
siderations into account. There are four steps in the evalua- being made in this case is chloride. The area available for
tion of a specific type of snow and ice control material: (1) dilution prior to exit of chloride to the environment is here
dilution analysis, (2) toxicity screening, (3) screening for assumed to equal the area of the road as well as shoulders and
Figure 3-4. Steps for determining predicted concentration of a

component of MgCl2 solution as it leaves the roadway (based on: (52)).

43
embankments. Characteristic winter precipitation is used to Table 3-8 summarizes the main factors that influence the
compute runoff per mile for an entire winter. The total concentrations of snow and ice control materials in receiving
amount of MgCl2 solution characteristically applied in a waters beyond the roadway. The amount of road surface per
winter of characteristic precipitation is obtained from unit of watershed area is a critical factor that is most evident
application records, which then allows a dilution ratio to be in the contrast between urban and rural conditions. Concen-
computed. In this case the ratio was 500:1. trations of snow and ice control materials reach their maxi-
Given the initial concentration of chloride in the solution, mum under urban conditions or at other locations where the
plus the dilution ratio, the predicted mean concentration can amount of road structure is very high in relation to the
be estimated at the exit point (i.e., margin of the roadway). In amount of watershed area (e.g., bridges and interchanges).
this case, the estimate indicated 400 mg/L, which is above the Under such conditions, concentrations of chloride and other
chronic standard (i.e., 290 mg/L, long-term exposure) but soluble constituents, or the augmentation of bedload in the
below the acute standard for chloride for protection of case of abrasives, can far exceed any limits that might be set
aquatic life (i.e., 840 mg/L, short-term exposure). Use of the for the protection of aquatic life (examples given by Environ-
chronic standard for judging possible impairment allows a ment Canada (3)).
considerable margin of safety over use of the acute standard, A more subtle influence on concentrations of snow and ice
which might be more applicable to most actual occurrences control materials is the amount of runoff from the roadway
of runoff. Although the runoff is anticipated to exceed the tar- in relation to the rate of application. Because some materials
get concentration (290 mg/L in this case), additional dilution may be applied to prevent the formation of ice (e.g., MgCl2),
of considerable magnitude is anticipated over short distances they may, under some climatic conditions, be applied when
beyond the exit point. As shown by Figure 3-5 for this partic- the amount of runoff that ultimately results from an antici-
ular application, the dilution factor rises rapidly to 1:10,000 pated icing event is minimal. In such cases, the dilution of the
within less than 1meter from the exit point. Dilution might snow and ice control material will be less than would be the
not rise so rapidly under some conditions, such as those case for conditions that typically produce large amounts of
involving the presence of a hydrologically isolated wetland, precipitation.
pond, or small lake near the roadway. Thus, evaluation of the The characteristics of receiving waters also affect the final
significance of a computed concentration at the exit point concentrations of snow and ice control materials that appear
involves knowledge of the site, particularly as regards the in surface waters. The most adverse conditions involve waters
expected dilution immediately beyond the roadway. that flush slowly, are of low volume, and are close to the road-
Figure 3-5. Dilution of deicer as related to distance along four tran-

sects (numbered 1–4) from side of a paved surface (I-70, near the
Continental Divide, Colorado). Dilutions above 10,000 are likely to
be underestimated. Source: (52).

44
Table 3-8. Factors that influence concentrations of widely available for use in testing such as Ceriodaphnia, a
snow and ice control materials in receiving waters water flea; the fathead minnow, a small fish; rainbow trout;
(streams, lakes, wetlands). and Selenastrum capricornutum, a common type of sus-
Effect on Snow and Ice
pended algae (see Lewis 1999 (52) for examples of these tests).
Factor Control Materials in Species of special concern ((i.e., threatened, endangered)
Receiving Waters should be tested as well (e.g., Lewis 1999 (52)). Further dis-
Higher road surface per unit watershed Increases concentrations
cussion of aquatic toxicity protocols is provided in Section
Higher roadway runoff Decreases concentrations 5.3.8. Testing of Ceriodaphnia is perhaps most common
Higher rate of application Increases concentrations because it is less complicated than the testing of fish and gives
Greater distance to receiving water Decreases concentrations a good representation of the response of some of the most
Greater volume of receiving water Decreases concentrations sensitive kinds of organisms in aquatic environments.
Greater flushing of receiving water Decreases concentrations Although U.S. EPA documents can be referenced for the
detailed protocols, the basic concept is application of the
material being tested on the test organisms over a sequence of
concentrations to the test organisms under controlled condi-
way. At the other end of the spectrum are water bodies of tions for specific time intervals followed by tabulation of
high volume at considerable distance from the roadway that mortality or sub-lethal effects. The exact conditions of such
show a high rate of water renewal. For example, a large tests are tightly proscribed so that the results are easy to eval-
stream with a high winter flow located several thousand feet uate on a standardized basis. In the case of a snow and ice con-
from a road would be much less responsive to the application trol material, the range of appropriate dilutions would begin
of snow and ice control materials than a shallow wetland or at concentrations higher than those expected under actual
pond adjacent to a roadway and not connected to a perma-
circumstances and extend to concentrations lower than
nent stream.
expected under actual circumstances, thus bracketing the
Although the factors in Table 3-8 could be quantified, it
expected concentrations.
would be difficult to produce an algorithm that would be
If bioassay testing indicates toxicity of the mixture at dilu-
sufficiently flexible to allow reliable predictions in most situ-
tions near those expected beyond the roadway, the materials
ations. More feasible is a qualitative inventory of the kinds of
can simply be rejected for use or, if reasonable alternative for-
aquatic environments near a specific segment of road in con-
mulations are not available, further analysis can be conducted
junction with field monitoring that involves indicators of
to isolate the cause of toxicity, thus providing guidance to a
dilution (see below).
vendor about required changes in the formulation. Analysis
for suspected toxicants would first include the most likely
3.4.3.2 Toxicity Screening causes of toxicity. If these tests did not indicate that any of the
In principle, screening for toxicity could be conducted on analytes were likely causes of toxicity, more unlikely causes of
the basis of a comprehensive chemical analysis, which would toxicity would be analyzed.
be evaluated in the context of the expected dilution and the The analytical protocols for individual constituents should
known toxicity thresholds, often given as numeric standards be taken from analytical methods approved by the U.S. EPA,
for particular substances. In practice, this approach has sev- by Standard Methods for the Examination of Water and Waste-
eral drawbacks. First, the toxicity of mixtures is not always the water (67), or by other widely used, government-approved
same as might be expected from the list of concentrations for authorities. Because these methods change from time to time,
individual components, because of interaction of compo- the list of approved methods should be reconsulted periodi-
nents. Second, a strictly analytical approach may omit analy- cally. Further discussion of environmental analytical test
sis of an unusual toxicant, thus failing to detect toxicity. In protocols is provided in Section 5.3.
view of similar problems for effluents, which are mixtures of There is one especially important limitation to the use of
uncertain composition, the U.S. EPA introduced whole efflu- approved analytical methods. The approved method in ques-
ent testing (WET) as a screening tool. Although WET testing tion must be capable of providing a quantitative estimate of
does not entirely displace analysis of known toxicants, it adds the concentration of a substance of interest within the range
a degree of certainty to the conclusion that a particular mix- of environmental concern. In other words, it is wrong to
ture is not likely to cause impairment beyond a specific apply a method approved but actually intended for applica-
amount of dilution. tion to concentrations far above the thresholds of interest.
Protocols for testing of mixtures are well established. They In such cases, the results may indicate no detection when
are based on bioassay tests applied to specific organisms now the substance is present above concentrations that are

45
environmentally significant. Detection limits should always the standard BOD tests, however, the test is based on incu-
be given with the results of individual tests. bation at a characteristic temperature (e.g., relatively low
If a bioassay test shows no indication of toxicity, it is still temperature for a coldwater environment), and the incuba-
advisable to conduct analyses for substances most likely to be tion medium is water drawn from the receiving water (e.g.,
present at concentrations that could cause toxicity. This is a stream, lake, or wetland). The snow and ice control material
check on the bioassay test and also indicates substances that is added within the range of expected dilution or somewhat
may be approaching concentrations of concern. Such infor- above that range, and the progressive loss of oxygen is mon-
mation can be passed to a vendor for pre-emptive action. itored over a 24-hour or longer interval to produce a time
course of oxygen depletion. The results of such tests typi-
cally show a lag in the response of bacteria to the presence
3.4.3.3 Testing for Nutrients and Oxygen of bulk organic matter, probably because the microbes most
Depletion Potential able to use specific kinds of organic matter are present ini-
Because impairment is not limited to toxicity, testing must tially in small quantities and build up exponentially as they
extend beyond toxicity. The concentrations of total phospho- exploit the organic matter. Thus, oxygen demand may be
rus and phosphorus fractions, as well as total nitrogen and relatively low for as long as several days and then become
nitrogen fractions, should be obtained for snow and ice con- extremely high. Knowledge of this time lag may be impor-
trol materials. Through dilution analysis, the amounts of tant for evaluating the location where effects may be noticed
in the receiving waters. For example, if dissolved oxygen
these plant nutrients passed from the roadway can be calcu-
depletion is observed only after a period of 2 or 3 days,
lated. Where numeric standards are available, the likelihood
downstream environments that hold waters temporarily
of causing the receiving waters to exceed a standard also can
separated from the sources by 2 or 3 days would be expected
be estimated. Where no standards are available, the U.S. EPA
to show the maximum response.
nutrient criteria can be used, or an approach based on anti-
degradation can be used as an index of potential impairment.
Organic matter, when present as a bulk constituent, poses 3.4.3.4 Field Monitoring
the possibility of oxygen depletion in receiving waters. The
Some types of environmental impairment are difficult to
total oxygen demand associated with water leaving the road
forecast from chemical analyses or toxicity testing. Therefore,
should be estimated. The estimate can be conducted through
field monitoring is an important adjunct to laboratory stud-
analysis of total carbon (67), through a standard 5-day BOD ies. Field monitoring can be limited to relatively simple meth-
test (67), or through a specialized BOD test (68). ods, but must include broad spatial coverage of the areas
As mentioned in Section 3.4.2.1, the analysis of total car- receiving substantial amounts of runoff from the roads and
bon can be used to estimate maximum potential oxygen ideally would include continuous monitoring of selected
demand based on the stoichiometry of respiration. Stan- sites. Most important is sampling of slowly flowing waters by
dardized BOD tests, which are routine for wastewater efflu- use of meters that can detect specific conductance, which
ents, have a similar intent and, in some senses, are more indicates the presence of salts and dissolved oxygen. Indica-
realistic in that they account for the fact that some of the tors of potential effects of snow and ice control materials
organic matter in the mixture may be resistant to use by bac- would include substantial depression of dissolved oxygen
teria and therefore not a significant short-term contribution below saturation or selective accumulation of saline water,
to oxygen demand. The deficiency of the standard 5-day especially near the bottom of standing waters or slowly flow-
BOD test is that it is conducted under conditions that devi- ing waters. Low oxygen would suggest the possibility of exces-
ate strongly from ambient conditions for unpolluted streams sive organic matter emanating from the road, and saline layers
because it involves the addition of nutrients and mainte- would indicate excessive salt reaching slowly flowing waters.
nance of a high, constant temperature. Therefore, its appli- Where wells are present, well sampling done in a compar-
cation to a relatively unpolluted water for purposes of ative manner involving other wells not affected by road runoff
estimating oxygen drawdown is problematic. In addition, the or more distant from the runoff source can be used in testing
standard BOD test does not provide information on the time the groundwater intrusion of runoff waters influenced by
course of oxygen depletion. snow and ice control materials. The simplest index would be
An alternative to the standard BOD test has been intro- specific conductance, which would increase in response to
duced by Lewis (68) specifically for evaluating snow and ice any significant intrusion of roadway runoff influenced by
control compounds. As in the case of the standard BOD snow and ice control materials. Initial evidence of intrusion
tests, water samples are held in closed containers that do not could be followed by more detailed analysis of constituents of
admit the passage of oxygen from the atmosphere. Unlike concern from the viewpoint of drinking water.

46
3.4.4 Overview of Specific Snow and Ice • Nutrient-poor streams typically contain organisms not tol-
Control Materials erant of oxygen depletion. Nutrient-rich streams are more
likely to experience oxygen depletion and, therefore, typi-
Likelihood of specific kinds of environmental impairment cally contain organisms more tolerant of oxygen depletion.
varies greatly among snow and ice control materials. Antici- Therefore, the tolerance of nutrient-poor streams for oxy-
pation of the most likely problems may save time and effort gen demanding substances, such as acetate and OMB, is
in evaluating specific materials. Table 3-9 summarizes the rated lower for nutrient-poor streams than for nutrient-
foregoing text. rich streams, particularly if the streams are small.
• The probability that the chloride standard for aquatic life
will be exceeded is highest for standing waters of small vol-
3.4.5 Conclusions ume and lowest for flowing waters of large volume.
The suitability of a particular snow and ice control mate-
rial for a specific location depends on its general type, its for- When considering drinking water obtained from groundwa-
mulation, its application rate, and the characteristics of the ter, the following conclusions exist:
receiving environment. These four considerations can be
• Magnesium, calcium, and chloride are not associated with
brought together according to the approaches outlined above.
impairment of human health.
The following general conclusions can be made:
• Sodium in groundwater is a potential threat to persons
with cardiovascular impairment, although the threshold
• Small water bodies offer less dilution than large water
concentration for health effects is relatively high.
bodies. Consequently, acceptability of any given material • Addition of organic carbon compounds (e.g., acetate or
with a potentially adverse environmental effect tends to mixed organic matter from biomass) may cause deoxy-
increase as the size of the receiving water increases. genation of groundwater, which in turn could cause release
• Abrasives block interstitial spaces within coarse substrates, of potentially harmful substances such as heavy metals into
thus damaging habitat. This particular category of damage the groundwater and could be a cause of taste and odor
is not an issue for aquatic bodies with fine substrates. problems.
• Organic matter and nutrients cause oxygen depletion or
eutrophication where potential dilution is low, but are In some instances, a clear likelihood of impairment will be
much less likely to be harmful where dilution is high. identified. Whether such impairment falls outside regulatory
Higher nutrient levels tend to be associated with OMB limits or not depends on law (i.e., state law within the United
materials. Nutrient-poor water bodies have a lower toler- States). For governmental agencies, however, it is advisable to
ance than nutrient-rich water bodies to additional nutrient consider water-quality standards as a strong guideline for the
loadings. application of snow and ice control materials even when the
Table 3-9. Generalized potential for impairment and its most likely
causes for broad categories of snow and ice control materials.
Potential for
Name Most Likely Cause of Impairment
Impairment
Excessive chloride loading; additives containing metals
CaCl2 Moderate or phosphorus; contamination with copper, zinc, lead,
or cadmium.
MgCl2 Moderate or phosphorus; contamination with copper, zinc, lead,
or cadmium.
Excessive chloride loading, metal contaminants;
NaCl Moderate
ferrocyanide additives.
or phosphorus; contamination with copper, zinc, lead,
Chloride-based salt
High or cadmium. Organic matter leading to oxygen
plus bulk organics
demand; nutrient enrichment by phosphorus and
nitrogen.
CMA Moderate Organic content leading to oxygen demand.
KA Moderate Organic content leading to oxygen demand.

Sand or sand with
High Turbidity; increased bedload.
salt

47
regulations have not been directly applied to such materials. a soil can displace calcium ions, even though calcium has a
In most instances, effective control materials that minimize higher adsorption affinity than sodium. The process can be
risk of impairment can be found. reversed through introduction of additional calcium, which
has a greater adsorption potential to CEC sites and will easily
displace the sodium.
3.5 Soils
Anionic species such as chlorides have a negative charge
The primary reason to consider the effects of snow and ice and do not bind well to like-charged soil particles. Anions are
control material on soil is that the presence and movement of typically transported through soil with soil water percolation
these materials through the soil can further affect vegetation, or with groundwater movement.
groundwater, and surface water. Soil conditions such as per- CEC has been shown to be a function of soil type. A com-
meability, erosion, and soil stability can also be influenced. pilation of CEC data for different soil types from across the
The effects of snow and ice control materials on soil struc- United States has been prepared by RCI (81). In general, sand
ture and chemistry have been well documented. Although and sandy loam soils tended to have very low CEC (i.e., 2.0 to
many studies have directly assessed the effects of snow and ice 17.1 meq/100g dry soil), silt loams had intermediate values
control materials on roadside soils, a great deal of information (i.e., 9.4 to 26.3 meq/100g dry soils), and clay and clay loam
is also available by adapting information from studies that had soils had the highest values (i.e., 4 to 57.5 meq/100g dry soil).
very different purposes—most commonly agriculture-related It was further demonstrated that clay soils could adsorb up to
studies investigating the effects of salinity from intensive irri- 5 times more sodium chloride than sand.
gation, saltwater intrusion, or inorganic fertilizer applications. Application of snow and ice control chemicals, specifically
Two general mechanisms for impact to soil exist: sodium, to soils can influence the physical properties of the
soil. Sodium exchange is a measure of a soil’s capacity to bind
• Modification to soil chemistry, including accumulation of sodium ions. High levels of sodium exchange are particularly
ionic species, cation exchange processes, and mobilization detrimental to soil structure in that soil particle sizes are
of metals from soil; and reduced, which can lead to increased soil compactness.
• Modification of physical soil characteristics. Increased compactness further reduces soil permeability and
can result in increased surface runoff conditions and erosion
After roadway application, some snow and ice control of soils.
materials will be deposited on roadside soils. Depending on In certain cases, cationic species (Na, Ca, Mg, and K) may
the specific products applied, increased levels of primary displace heavy metals already bound to soil particles in a
components will occur (refer to Table 3-1). For example, process commonly referred to as metals mobilization.
application of NaCl will elevate sodium and chloride concen- Increased mobility implies an increased dissolution of metals
trations, whereas CMA will increase calcium, magnesium, to soil pore water, where they may either resorb onto other
and acetate concentrations. Associated increases will also be soil sites, interact with soil organic material, form metal
realized for secondary attributes such as BOD, heavy metals, acetate pairs, bioaccumulate, or move with the hydraulic gra-
cyanide, and phosphorus, where applicable. The degree to dient in groundwater with eventual discharge to surface water
which a certain material accumulates in the soil depends on (82). Mobilized metals are more biologically available than
product properties, soil permeability, adsorption, water per- soil-bound metals and pose a greater risk to vegetation,
colation, vegetative uptake, and biodegradation. terrestrial organisms, and aquatic life.
Soils are known to contain negatively charged binding sites The potential for metals mobilization is highest where the
on the surfaces of soil particles. Positively charged cations soil already contains high levels of heavy metals and where
(e.g., sodium, calcium, magnesium, potassium, heavy metals) concentrations of applied cations (from snow and ice control
introduced into soil water will adsorb onto the negative bind- chemicals) are greatest. These conditions are characterized by
ing sites. Cation Exchange Capacity (CEC) is the ability of a the roadway environment, where soils tend to already have
cation to replace on a soil particle another cation of weaker higher-than-background levels of cationic heavy metals from
affinity or smaller concentration. The relative adsorption vehicle and pavement attrition (82, 83). Metals commonly
affinity of select cations is as follows (81): found in roadside environments are listed in Table 3-10. The
risk of heavy metal mobilization in roadside soils is therefore
High Adsorption Affinity H+> Ca2+> Mg2+> K+> Na+ Low
likely if sufficient quantities of cations from snow and ice con-
Adsorption Affinity
trol materials are available.
Hydrogen ions have a greater affinity for soil binding than cal- Novotny et al. (20) found the partitioning coefficient for
cium and so forth, although the process is also affected by metals is inversely proportional to the salt concentration of
concentration. For example, excessive loading of sodium onto the snowmelt. This means that during winter, salt-laden

48
Table 3-10. Heavy metals released by vehicles. is limited. A field study in southern Ontario found that con-
centrations of heavy metals in groundwater beneath roadside
Heavy Metal Traffic Source
soils were similar to those in non-highway locations (86), sug-
Tire wear, lubricating oils and grease, bearing
Lead gesting that the effect is not significant. Another field study
wear
Zinc Tire wear, motor oil, grease conducted in British Columbia failed to detect any trends in
Iron Vehicle rust, highway structures, engine parts soil metals concentrations after winter applications of MgCl2
Copper Metal plating, brake lining, bearing wear and NaCl (34).
Cadmium Diesel oil, tire wear
Chromium Metal plating, brake lining wear
Source: (84)
3.5.2 Sodium
Effects of sodium on soil structure and chemistry have
snowmelt or soil pore water may have a higher proportion of been well documented. Elevated concentrations of sodium in
dissolved metals. Dissolved trace metals are more toxic to soil have long been an agricultural concern, and many stud-
biota, whereas the adsorbed or particulate fraction is consid- ies have investigated salinity effects related to intensive irriga-
ered less biologically available. tion, saltwater intrusion, and inorganic fertilizer applications.
Temporal effects also influence the degree of metal mobi- NaCl use for snow and ice control has further increased
lization resulting from high cation concentrations. Field stud- interest in sodium-related impacts to soil.
ies have observed “extreme differences [higher levels]” in Of the primary cations associated with snow and ice con-
cadmium and zinc soil concentrations between winter/spring trol materials ((e.g., Na, Ca, Mg, K), sodium is most com-
and summer/fall (20). This difference was attributed to leach- monly used and also has the greatest potential to negatively
ing of metals from the soil caused by high levels of NaCl from affect soil properties. Accumulation of sodium in soil can
deicing salts in the winter/spring period. bring about undesirable properties such as increased osmotic
pressure, reduced permeability, and the potential for cationic
exchange processes (including nutrient imbalances and met-
3.5.1 Chloride
als mobility).
Chloride is the primary anionic component of NaCl, The most significant effect sodium has on soil is its ability
CaCl2, and MgCl2. Chloride does not easily precipitate, is not to adversely affect soil structure. Excessive amounts of
biodegradable, is not readily involved in biological process, sodium will exchange with calcium and magnesium already
and does not adsorb significantly to mineral/soil surfaces. bound to soil particles. Calcium and magnesium ions tend to
Chloride is therefore extremely mobile and tends to move aggregate (bind) organic and inorganic particles within the
with soil water percolation and groundwater flow. Chloride is soil, whereas sodium causes them to break apart. When soil is
often used as a tracer to delineate the extent of salt movement broken up, the particles wash through the soil; pores become
through the subsurface because of its ability to move without sealed and permeability is reduced. Clay soil structure is also
significant impedance. affected because sodium adsorption causes clay particles to
The literature does not directly associate chloride with any swell—this also reduces soil permeability.
negative impacts to soil properties; however, chloride, in con- The effect of sodium on a particular soil type can be clas-
junction with an associated cation (i.e., Na, Ca, Mg), can sified by the soil’s ability to bind sodium on its negatively
increase soil conductivity, which can further lead to increases charged soil sites. The term Exchangeable Sodium Percentage
in soil osmotic pressure. (ESP) is the amount of cation exchange capacity taken up by
There is some evidence from laboratory experiments that sodium (Naexch) on a particular soil type and can be demon-
chloride passing through soil contributes to mobilization (to strated by the following equation (87):
groundwater) of some heavy metals, such as cadmium, zinc,
ESP Naexch/CEC 100
and lead (83). The premise is that chloride can form com-
plexes with metals in soils, potentially releasing them into the Soils with high ESP will disperse and also plug up easily,
environment (83, 84). Doner (85) found that higher chloride whereas low ESP soils do not disperse readily and will main-
concentrations in soils increased the rate of mobility of tain soil permeability (21). In general ESP values of 13 or
nickel, copper, and cadmium through soil through the for- higher are considered potentially hazardous to soil quality
mation of chloro-complexes. It is however extremely difficult (29). Morin et al. (21) further classified soils based on per-
to separate the influence of chloride ions from the cationic ceived potential impacts from sodium (Table 3-11).
species initiating metals mobilization. At best, the chloro- Over time as sodium is applied to roads, sodium will
complexes reduce the affinity of the mobilized metal to resorb accumulate in roadside soils, cumulatively. Higher-than-
on to other soil CEC sites. However, supporting field evidence normal salt concentrations can be further exacerbated by

49
Table 3-11. Soil types classified based on potential (e.g., permeability, plasticity, and strength). Brandt (90)
impact from sodium. further suggested that CaCl2 and NaCl mixtures used for
snow and ice control could prevent the need for adding
Level of Impact Soil Type materials like gypsum to periodically decrease the sodium
Low
Highly permeable, coarse textured. Sands, loamy sands, contamination of roadside soils.
sandy loams, silts
Low permeable. Compacted clays, clay loams, loams, sandy Several studies have investigated the potential for calcium
Medium to Low
clay loams, silt loams to mobilize metals in soil. Many of these studies involve
Highly permeable. Organically enriched clay, clay loams,
Medium
loams, sandy clay loams, silt loams assessments of CMA, which includes calcium as well as mag-
High
Medium permeability, medium texture. Non-compacted nesium. A California laboratory study (91) found that a CMA
loams, sandy clay loams, silt loams
solution that leached through soils tended to exchange cal-
Highest Soils enriched with swelling clay cium and magnesium with other metals already present in the
Adapted from: (21) soils. Results indicated that CMA could extract iron, alu-
minum, sodium, and potassium from roadside soils. Labora-
tory and field tests conducted by Horner (36) also indicated
low precipitation conditions. As reductions in permeability potential for CMA to mobilize and release trace metals from
related to sodium take effect, there will be lower water soils. Laboratory experiments conducted by Elliot and Linn
movement through the soil, further increasing the potential (82) using CMA also found preferential adsorption of Ca and
for sodium accumulation and further still decreasing per- Mg at soil exchange sites and displaced heavy metal ions in
meability. The net result of decreased soil permeability soil solution. Of particular interest were increased desorbed
means that more water can flow over land and cause erosion H+ ions, which will reduce pH and further promote the dis-
(22, 88). This can be of particular concern when consider- solution of metals adsorbed to soil particles. More recent
ing structural stability for roads. studies by Amrhein et al. (92) have compared the mobiliza-
Excessive levels of sodium can also affect soil pH. When tion of trace metals in soils treated with solutions of NaCl and
the exchange capacity of a soil becomes greatly saturated by CMA. The results of these investigations indicated that CMA
sodium and depleted of calcium and magnesium, the soil is probably has less effect than NaCl on trace metal movement
considered to be an alkali soil, and the pH of the soil may in soil. Tests also showed that calcium and magnesium ions
increase (19). Other studies have demonstrated reversed were advantageous for the permeability and structure of soil.
results, indicating that pH change depends on other factors Sodium, on the other hand, tended to destroy soil structure
as well. by dispersion of soil aggregates.
Sodium can also cause metals mobilization; however the
extent depends on specific soil conditions. Areas with the
3.5.4 Magnesium
highest impact will be where there is a high percentage of
clay content with a high exchangeable sodium ratio and low There is limited information directly discussing the effects
precipitation (21), as well as areas already containing high of magnesium on soils. Most information is available though
levels of available metals such as roadside environments are investigations conducted for CMA, which involve the com-
of concern. bined effects of calcium and magnesium ions. As discussed
for calcium (Section 3.5.3), magnesium is also thought to be
beneficial to soil permeability and soil structure and can
3.5.3 Calcium
reduce dispersion of organic matter and clay (20). Magne-
The effect of calcium on soil has been adequately docu- sium also can exchange on soil particles in a process that may
mented in the literature, including studies of CaCl2 and mobilize heavy metals. Magnesium is however not as strongly
CMA use for snow and ice control, CaCl2 used as a dust sup- adsorbed as calcium to soil particles (81, 89).
pressant, and CaCl2 used as a soil stabilizer for agricultural
applications.
3.5.5 Potassium
Calcium is generally thought to have beneficial effects on
soil physical properties when compared with sodium. A Few published reports detail the effects of potassium appli-
high concentration of calcium causes fine clay particles to cations on soil. Potassium is an important plant nutrient
clump together and aggregate. The presence of soil aggre- commonly applied in fertilizer blends. It can be tolerated at
gates creates macropores in the soil that permit better water high concentrations by most plants; however, at high appli-
drainage and improved aeration (89). A University of cation rates it can increase osmotic pressure of the soil (93).
Washington study (36) found that Ca associated with CMA Like the other cations typically associated with snow and ice
did not significantly affect the physical properties of soil control materials (i.e., Na, Ca, and Mg), excessive levels of

50
potassium should have the potential to influence cation produced. This will increase soil pH and assist in immobiliz-
exchange processes and increase the mobility of heavy metals. ing solubilized metals and results in a net decrease of leach-
ing from soil (82, 83). Acetate itself has also been found to
provide similar buffering capacity (82).
3.5.6 Acetates
Elliot and Linn (82) also found with CMA that preferential
Information regarding the effects of acetates on soil adsorption of Ca and Mg at exchange sites displaced bound
properties has been drawn from studies conducted on CMA metal ions into solution and desorbed H+, which promoted
and, to a lesser extent, KA. The acetate ion (CH3COO) dissolution of adsorbed metals. This was counteracted by the
is organic and can be used by many species of micro- pH neutralizing effect of acetate ions, so that within 10-pore
organisms as a nutrient. Generally, all acetates will degrade volume displacements, the presence of CMA resulted in a net
by microbial action to carbon dioxide and water. A Univer- suppression of leachate metal ions. Soils with higher acidity
sity of Washington study (36) revealed that, at temperatures may resist this neutralization.
above 10°C, a high level of CMA-acetate decomposition Many of the study conclusions cited above are based on
was achieved within 2 weeks, while at 2°C, 4 weeks were laboratory testing, and limited information is available that
required for full decomposition. Therefore, the direct replicates these findings when subjected to the highly variable
effects from acetates are relatively short lived and confined conditions experienced in the field.
to an area around the roadway—the extent of transport of
acetates through the soil subsurface does not typically
3.5.7 Organic Products
exceed 10 meters (36).
CMA has been shown to influence the mobilization of Organic products used for snow and ice control include a
heavy metals from soil. This process is predominantly related range of manufactured materials such as glycol and agricul-
to the cationic exchange processes related to calcium and tural by-products—typically agriculturally derived organic
magnesium, but acetate is thought to have some influence on biomass products blended with chloride-based salts as a cor-
the process. Ammonium acetate is commonly used to leach rosion inhibitor. There is limited specific information detail-
metals from soils for analytical testing purposes. Ammonium ing the effects of organic products on soil properties, but,
initiates the ion exchanges with the bound metal after which based on their organic constitution, it is expected that they
the acetate complexes (binds) the metal, helping to maintain would be subject to degradation by soil micro-organisms.
it in solution. There have therefore been concerns that Degradation could lead to anaerobic conditions in the soil,
acetates introduced as snow and ice control materials can which can decrease soil pH and increase the solubility of
increase the solubility of metals in soil water, preventing adsorbed metals.
potential readsorption to soil. Laboratory studies conducted Klecka et al. (95) investigated biodegradation of five differ-
by Amrhein and Strong (83) found that, after metals are ent glycol-based aircraft deicing fluids in soil. Glycol
mobilized from cation exchange processes, acetate forms biodegradation was observed in soil at concentrations rang-
stronger metal pairings than chloride, potentially increasing ing from 392 to 5278 mg/kg, suggesting that high levels of
the trace metal concentration in solution. However, this the deicing fluids are unlikely to be inhibitory to soil micro-
process is deemed unlikely to persist under field conditions organisms. Regardless of the type of glycol present (e.g.,
because of biochemical processes involved with acetate ethylene glycol, propylene glycol, diethylene glycol), bio-
degradation (83). degradation rates were very similar. Average rates were 66.3 to
Laboratory tests comparing the effect of CMA and NaCl on 93.3 mg/kg soil per day at 25°C, 19.7 to 27.0 mg/kg soil per
trace metals mobility found mobility associated with CMA day at 8°C, and 2.3 to 4.5 mg/kg soil per day at 2°C.
did occur but was less than for NaCl (83, 94). The addition of Although biodegradation rates decrease as temperature
CMA to soil initially may increase the mobility of metals and decreases, it was expected that biodegradation would play a
their leaching from contaminated soil—both calcium and major role in removing residual levels of glycols from soils.
magnesium have higher soil adsorption affinities than Use of glycols for roadway snow and ice control is not practi-
sodium. There is also concern that rapid, microbial decom- cal because of cost, and they are rarely used for this purpose.
position of CMA could lead to low redox (oxygen deficient)
conditions in roadside soils and to the leaching of acetate to
3.5.8 Urea
groundwater. Low redox conditions resulting from CMA
decomposition could have a strong influence [increase] on Few studies have been conducted detailing the effects of
trace metal solubility as well. urea-based snow and ice control materials on soil properties.
Conversely, an important effect of acetate degradation is Urea is a common component of commercial fertilizers and,
that it creates neutralization conditions when bicarbonate is when placed in the soil, is quickly hydrolyzed (usually within

51
2 to 10 days depending on soil temperature) to ammonium. cern when considering soils as the sole impact receptors.
When soil micro-organisms convert urea to ammonium, Increases of concentrations of snow and ice control materials
there can be a short-term increase in pH in the area immedi- in soil tend to have a far greater effect on terrestrial vegetation
ately around the fertilizer granule. This effect is more pro- and to a lesser extent aquatic quality.
nounced at low soil moisture levels, at high temperatures, and Physical effects are more of a concern, specifically when
in light sandy soils. At pH values greater than 7.5, ammonia considering sodium accumulation, which is known to break
gas also can be generated. This ammonia gas will usually dis- down soil structure. Breakdown of the soil structure can
solve in the available soil water or adsorb to soil particles and reduce water and air permeability and can lead to erosion
not pose any problems. If conditions are dry, however, free conditions. However, magnitude of impact depends on the
ammonia gas can persist in the soil pore space, and this can types of soil present (Table 3-11). Documented cases of NaCl
damage plant roots when fertilizer has been banded near the applications adversely affecting the stability of roadside soils
root system (96). were not found.
Snow and ice control materials containing calcium or mag-
nesium tend to improve soil structure and permeability by
3.5.9 Abrasives increasing aggregation of fine soil particles. Products con-
Abrasives have few direct effects on soil because they do not taining organic components are partially or fully biodegrad-
chemically react and tend to remain on the surface of the soil. able in the soil and have little known effect. Mobilization of
However, concerns have been raised about the trace con- heavy metals is a concern when excessive levels of cationic
stituents associated with the abrasive materials. Some mate- components are introduced to roadside soils; however, lim-
rials have shown high levels of available trace materials (e.g., ited field evidence has been obtained to support this.
metals that may leach into the environment). Oberts (97)
noted that sand obtained for the Minneapolis metropolitan 3.6 Soil Biotic
area contained high levels of phosphate related to the calcium
phosphate mineral, apatite, which was present in the source Much of the information relating the effects of snow and
material. ice control materials to soil microbial species has been sum-
marized in a literature review conducted by Butler for the
Environment Canada CEPA Priority Substances List Roadsalt
3.5.10 Ferrocyanide Products Assessment (99). This study specifically sought information
Ferrocyanide complexes are commonly associated with for NaCl, KCl, and CaCl2; however, limited literature was
NaCl, and have limited direct effect on soil properties. Ionic found directly assessing exposures and effects of these snow
effects are negligible when compared with NaCl, because and ice control materials to soil microbial communities.
Study findings were thus developed by adapting information
approximate concentrations are typically less than 0.01 per-
from studies that had different purposes, most commonly the
cent by weight of NaCl. The fate of ferrocyanide ions in soil is
effects of high salinity from intensive irrigation or inorganic
unclear. Negatively charged ions tend to be rejected by soils
fertilizer applications. Therefore, the lack of published infor-
and move readily with percolation and groundwater; how-
mation on the effects of snow and ice control materials on soil
ever, a study conducted by Olson and Ohno (98) speculated
invertebrates was clearly identified as a data gap in the Envi-
that, because the relative proportion of cyanide with respect
ronment Canada CEPA Priority Substances List Roadsalt
to salt is low, the ferrocyanide may be adsorbed onto the soil
Assessment.
particle. The photodecomposition process requires ultravio-
let light to occur, and it is unlikely this process will occur once
ferrocyanides are in soils or groundwater. Laboratory experi- 3.6.1 Chloride-Based Salts
ments have shown that decomposition of ferrocyanide com-
plexes in soils (in the dark) is slow. The half life for High concentrations of salts in the soil can affect soil
ferrocyanide in soil ranged from 1 year at a low pH of soil microbial activity via osmotic stress as well as apparently ion-
water to 1,000 years at levels greater than pH 7 (20). specific stresses, though the magnitude of effect is very diffi-
cult to gauge. Organisms tend to be highly adaptable and soil
characteristics (e.g., pH, ion exchange capacity, ammonium
3.5.11 Conclusion
level and fertilization state, and matric water potential) can
Snow and ice control materials can affect soils both chem- influence the magnitude of effect. The organisms most likely
ically (i.e., by affecting concentrations of available ionic com- to be at risk are those in intimate contact with the dissolved
ponents in the soil) and physically (i.e., by affecting soil solids in the soil water (e.g., protozoa, nematodes, earth-
structure). Chemical effects to soils are not a significant con- worms, and arthropods) (99).

52
Affected soil strips along roadsides are typically narrow soil concentrations based on the 25th percentile of ranked
and, even if total sterilization occurs as a result of snow and study concentration data for various species, the authors
ice material application, rehabilitation probably will occur developed toxicity thresholds that found lethal effects for
after conditions return to normal (99). A European study NaCl to occur at 4,700 mg/kg whereas non-lethal effects
cited by Butler and Addison (99) found that some changes to would occur at 1,200 mg/kg. The authors also conducted lab-
the soil faunal community should be expected as a result of oratory tests that compared the toxicity of commercial road-
roadsalts; however, even in the absence of salt impingement, salt (99 percent NaCl, containing approximately 15 ppm
the strips of soil next to the roadway must already be con- ferrocyanide) with analytical grade NaCl for two soil inverte-
sidered disturbed environments. Organisms inhabiting the brates (F. candida and O. folsomi). They found no discernible
roadside community probably are drawn from a pool of difference between responses for both organisms, suggesting
species adapted to disturbed environments and are not nec- that reagent grade NaCl was a valid surrogate for commercial
essarily indicative of the “natural” population through which salt products and, perhaps more important, suggesting that
the road passes. ferrocyanide as used in commercial NaCl preparations pro-
Few studies have assessed the potential process for effect to duces no measurable effect on the species tested. This study
soil microbial species. Gunter and Wilke (100) determined that also found no discernible differences in toxicity between KCl
enzyme activity of soil bacteria decreased during the period of and NaCl for selected soil invertebrates.
sodium chloride runoff. The decease resulted from a general
decrease in microbial activity (i.e., metabolism) and not from
3.6.2 Urea and Potassium Acetate
enzyme shutdown or death. The effects were temporary. Activ-
ities returned to normal levels after the ions had leached from Transport Canada (101) assessed the effects of urea and
the roadside soils. It was not determined whether the effect was potassium acetate used as runway deicers on soil microbial
caused by the sodium or chloride ion, or both, but it does rep- communities at Halifax International Airport. They found no
resent an effect on the balance of soil microbial chemistry. discernible differences between the sites tested with respect to
Bright and Addison (41) developed British Columbia
matrix numerical soil standards for NaCl by constructing • The in situ concentrations of total heterotrophic bacteria,
species sensitivity distributions based on mortality (LC20) and whether these bacteria were either anaerobic or aerobic or
EC50 (Median Effects Concentration) for a range of soils and psychrophilic (low-temperature) or mesophilic (moderate
moisture contents in British Columbia (Table 3-12). Using temperature) bacteria;
• The in situ concentrations of fungi, whether these fungi
were either psychrophilic (low-temperature) or mesophilic
Table 3-12. Relationship between percent of soil (moderate temperature) mold; and
invertebrate species potentially affected and soil • Soil respiration characteristics of the rate of carbon diox-
NaCl concentrations. ide evolved, the proportion of organic carbon metabolized,
or the temperature coefficient.
Predicted Soil Concentration of NaCl
(mg/kg, dry weight basis)
Percent of Species One of the less-studied aspects of soils with respect to snow
Corresponding to
Sensitivity Distribution
Effect Type and Level and ice control materials is the role of soil bacteria and fungi
LC20 EC50
5 3200 660
in nutrient cycling and transfer (101). Bacteria in soils are
10 3500 770 responsible for the decomposition of organic matter to forms
15 3900 890 that can be used by plants. Fungi, specifically microrhizae, are
20 4300 1000
25 4700 1200 associated with the roots of many species of plants and enable
30 5200 1400 the transfer of essential nutrients (e.g., zinc, manganese, and
35 5700 1600 copper) from soils to the roots. The interference with the
40 6300 1900
45 6900 2200 metabolism of bacteria or microrhizae may adversely affect
50 7600 2500 vegetation and warrants further investigation.
55 8300 2900
60 9200 3400
65 10000 4000 3.6.3 Ferrocyanides
70 11000 4600
75 12000 5300
80 13000 6100
Ferrocyanide compounds have limited bioavailability and
85 15000 7200 are not considered toxic (20). However, they can decompose
90 16000 8300 to a more toxic free-cyanide form. Fast breakdown of com-
Source: (41) plex ferrocyanide is caused by photodecomposition by ultra-

53
violet light and results in the formation of HCN gas, which is vegetation has been investigated in the Adirondack region in New
volatile. If the rate of formation of the gas is less than the rate York, the Lake Tahoe basin of California and Nevada (107, 108)
and Europe (109, 110).
of volatilization, an increase in the aqueous concentration for
available cyanide will occur. This may be transported to soils
next to the roadways. Studies have found significant effects on plants because of
intake of salt ions from contaminated soil and/or water
Photodecomposition requires ultraviolet light in order to
through the roots and damage from aerial deposition on
occur, and it is unlikely this process will occur once ferro-
leaves and stems. Often, damage may result from both soil
cyanides are in soils or groundwater. Laboratory experiments
uptake and aerial deposition and, in many cases, may not be
have shown that decomposition of ferrocyanide complexes in
noticeable until a buildup of salt ions in the plant tissue
soils (in the dark) is slow. The half life for ferrocyanide in soil
reaches toxic levels over the course of several years. Effects can
ranged from 1 year at a low pH of soil water to 1,000 years at
occur through various modes (e.g., biochemical, physiologi-
levels greater than pH7 (20).
cal, physical, and chemical processes). The literature clearly
indicates that effects can be significant, but depend on a wide
3.6.4 Conclusion range of factors unique to each site. Most research on the
effects of snow and ice control materials on roadside vegeta-
Cain et al. (28) summarized various field studies that meas-
tion generally finds that damage is limited to areas both (1)
ured NaCl concentration in roadside soils. At distances less
exposed to high levels of chemical splash/spray or runoff and
than 10 meters from the roadway, cases were cited where NaCl
(2) containing salt-sensitive species.
concentrations were greater than the toxicity thresholds The effects snow and ice control compounds have on ter-
developed by Bright and Addison (41). However, it was found restrial vegetation, as identified in the literature, can be clas-
that NaCl concentrations at distances greater than 10 meters sified into three groups:
from roadways were generally greatly reduced. The roadside
soil environment is affected by a wide range of stressors not 1. Aerial deposition;
related to snow and ice control. It is also likely that the organ- 2. Ion uptake and accumulation from soil; and
isms inhabiting the roadside community are drawn from a 3. Osmotic stress.
pool of species adapted to disturbed environments and are
not necessarily indicative of the “natural” population through
which the road passes. It is suggested that routine snow and 3.7.1 Aerial Deposition
ice control practices do not pose a significant risk to roadside When salt is deposited on foliage, it may either stay on the
microbial populations outside what would already be consid- exterior plant parts (e.g., needles, leaves, and stem) or be
ered “disturbed environment.” transported to the interior of the plants either through young
shoots, buds, or over-wintering needles (i.e., conifer leaves).
3.7 Vegetation Direct injury from deposition on mature bark has not been
observed, even in salt-sensitive species (28). Moran et al. (62)
The adverse effects of snow and ice control materials on (29) found salt spray to be responsible for a condition called
roadside vegetation have been known for some time. Unlike “leaf burn,” a process where leaf tissue is damaged through
effects on soil and surface water, damage to roadside vegeta- contact with highly ionized salts. Salt spray was also found to
tion is highly visible and well publicized, especially in forests cause desiccation of plant tissues, a process where cell water
and parklands next to highways. Numerous reports and diffuses from the cell to the area of greater ion concentration.
reviews have been written during the past three decades doc- In addition to drought or desiccation, Cain et al. (28) also
umenting the effect of these materials on the environment. found injury related to penetration of phytotoxic ions of
The following is a historical summary based on TRB Special NaCl through the stem, bud, and leaf tissues. General symp-
Report 235 (17): toms of injury related to aerial deposition of salts on vegeta-
tion include (28)
Incidents of vegetation injury were first reported in Minnesota dur-
ing the 1950’s, when trees along city boulevards started to show • Death of stems and buds on young shoots;
signs of salt-related decline (102). New Hampshire also reported • Delayed resumption of spring growth;
the death and removal of nearly 14,000 trees along 3700 miles of
• Reduced new growth compared with uninjured plants;
salt-treated highways (103). Investigations of salt’s impact on veg-
• Reduced vigor and growth of trees (because of repeated
etation were conducted during the 1960’s and 1970’s in New
Hampshire, Minnesota, Michigan, and several other snowbelt injury over time); and
states, usually in response to concerns about damage to roadside • Premature leaf drop for conifers, further reducing photo-
trees (103–106). More recently, damage to trees and other roadside synthetic capacity.

54
Hanes et al. (111) determined that susceptibility of plants however, sodium may contribute to plant injury by increas-
to absorption of roadsalt spray and splash depends on the ing osmotic stress (84).
growth stage. Salt deposited on foliage is more readily Any effects that may occur depend on plant sensitivities to
transported into an actively growing plant than a dormant the specific ion. Calcium, magnesium, and potassium are all
one. Snow and ice control chemicals are applied in winter essential elements for plant growth. Because of this, most
when plants are dormant, and therefore effects to dormant plants can tolerate relatively high levels. Westing (88) did,
deciduous trees and grasses should be low because of this however, demonstrate that excessive calcium in soils may
mechanism. Salt ions that accumulate on plant surfaces can decrease both sodium and potassium uptake.
be washed off by precipitation, reducing the effect from this Sodium and chloride, on the other hand, have been shown
mechanism; however, the author further stated that salt to induce toxic effects at levels above 0.3 percent dry weight
ions will be deposited on surrounding soils and can still in leaves for sodium (88) and above 0.5 percent dry weight for
continue to be taken up through their roots during the chloride (22, 88). Extreme sodium levels in soils can interfere
winter. with potassium uptake by plants and cause calcium deficiency
Susceptibility of plants to aerial deposition can also be (22, 88).
related to plant vigor (i.e., young or aging trees may be
affected more than healthy mature trees). Also, the age and
3.7.3 Osmotic Stress
size of a plant may influence the magnitude of effect. For
example, young trees may be more sensitive to salt spray In general, an increase in ion concentration in soils will
than mature trees because, on mature plants, the crown, increase soil salinity and cause osmotic stress to plants. This
the foliage, and the new shoots may be above the zone of is a situation where a high concentration of salts causes an
spray (28). osmotic potential between the soil water and the plant.
Osmotic stress can inhibit the plant’s ability to absorb water
and can cause nutritional imbalances by disrupting the
3.7.2 Ion Uptake and Accumulation
uptake of nutrients (22, 28, 107). Not only will osmotic stress
from Soil
reduce the root’s ability to uptake water from the soil, it can
Commonly used snow and ice control chemicals (e.g., lead to dehydration and collapse of plant root tissues.
sodium chloride, magnesium chloride, calcium chloride, Osmotic stress can also cause cellular damage, which can
potassium chloride, CMA, and KA) introduce the potential lead to reduced shoot growth and drought-like symptoms
for high levels of constituent ions (e.g., sodium, magnesium, (22, 28, 107).
calcium, potassium, and chloride) in the soil environment. Water availability can affect osmotic stress conditions.
These ions will accumulate or migrate though the soil based Species that have a marginal tolerance to salt stress may
on precipitation, soil permeability, cation exchange capacity, become salt intolerant under drought conditions. In addition,
and hydraulic conductivity. When these ions reach the root drought stress has been shown to aggravate vegetation with
zone, they are taken up by the root system at a rate equal to pre-existing salt damage and may lead to increased severity of
the transpiration rate from leaves where they will accumulate symptoms (112) (108).
in plant tissues (22, 88, 111). General symptoms in response Excessive ionic concentrations in soil will also disrupt
to elevated soil levels of Na, Ca, and Cl include general plant nutrient balances. Crop yield has been shown to decrease
decline, reduction in leaf size and plant growth, leaf chloro- with increasing levels of soil conductivity, regardless of the
sis, leaf burn, and tissue death (28). Seed germination can also level of sensitivities (21). The ionic dynamics in soils are
be reduced or delayed. complex, and specific mechanisms remain poorly under-
Studies attribute specific damages to vegetation to exces- stood; however, in general, any excessive concentration of
sive chloride levels, not sodium levels. Hanes et al. (111) ions can disrupt natural balances.
examined the effect of salt on silver maples and found that, Experiments have shown that plasmatic resistance to salts
although the amount of sodium absorbed by the plant was is lowest during the growth season and highest in the winter
proportionally higher than chloride, it appeared that less when the plants are dormant. Lab experiments have also
sodium was accumulated over time. Although the accumula- shown that repeated prolonged above-ground contact causes
tion from high levels of sodium can be detrimental to the far more damage than repeated short-lived contacts (31).
plant, the study results indicated that there is more danger to Where excessive soil concentrations of sodium exist, not
the plants from chloride accumulation. Another study con- only is osmotic stress a concern, the soil structure can deteri-
cluded more chloride than sodium is absorbed by plants orate—soil pores become clogged, permeability decreases,
because sodium is more readily adsorbed onto soil particles; and soils can crust (refer to Section 3.5.2). Clogged soil pores

55
reduce the space available for penetration and permeation of cussed in Section 3.1. Secondary factors include distance from
air and water through the soil. Reduced water supply can roadway, season, duration of exposure, age and health of
induce drought-like conditions, further exacerbating the plant species, and soil conditions. The effects of snow and ice
effects associated with an osmotic stress condition. Reduced control chemicals to vegetation can also be affected by a com-
soil air supply will affect root growth. Soil crusting reduces plex set of site-specific factors and conditions. The Trans-
shoot emergence of subsurface seeds and root penetration of portation Research Board (32) developed the following list
both surface and subsurface seeds, resulting in reduced plant identifying factors that affect the degree of vegetation damage
establishment (28). resulting from salts:
3.7.4 Species Sensitivities • Temperature: The effects of salts are compounded by

higher temperatures that cause increased dehydration
The effect of snow and ice control materials on roadside through foliage, faster movement of salt to the plant, and
vegetation varies by species, and several investigators have increased salt absorption through the roots.
compiled lists of relative tolerances of native and horticul- • Light: Exposure to direct sunlight increases the rate of
tural species. NaCl “sensitivity lists” have been prepared by dehydration.
Hanes et al. (111) and Sucoff (103). Agencies have used this • Humidity: High humidity lowers the rate of dehydration
information to assist in selecting tolerant species for areas and, therefore, helps alleviate water stress.
with high-expected levels of roadsalt application. A literature
• Wind: Exposure to wind may increase the rate of
review conducted by Cain et al. (28) found that vegetation
dehydration.
most sensitive to the effects of chlorides include native
• Soil water: As salinity increases, the soil water available to
grasses; wetland species, such as sphagnum moss and sedges;
plants decreases.
and maple, pine, Douglas Fir, dogwood, peach, and plum
• Soil texture and drainage: The ability of soil to retain salt is
trees. In Canada, 53 percent of principal forest trees species
partially determined by soil texture and drainage charac-
have been rated as sensitive to roadsalt and 73 percent of prin-
teristics. For example coarse-textured soils are quickly
cipal genera of forest trees contain species that have been
leached of salt, and steep slopes may not absorb salt
rated as sensitive to roadsalt (28).
because of rapid runoff.
In general, grassy vegetation is more salt tolerant than are
• Precipitation: Rainfall and other precipitation can flush salt
trees and woody plants (88). Hanes et al. (111) found the fol-
lowing general salt sensitivities: deposits from foliage and dilute salt solutions in soil water.
On the other hand, precipitation can transport salts via
• Most grasses are fairly adaptable to high concentrations of
surface runoff to roadside soils.
Na and Cl;
• Woody trees are fairly tolerant; Effects on vegetation can be subject to temporal effects.
• Conifers are least tolerant; and Laboratory experiments have shown that plasmatic resist-
• Younger plants are more susceptible than aged plants. ance to salts is lowest during the growth season and highest
in the winter when plants are dormant (31). Most plants have
Species that have a marginal tolerance to salt stress may completed their active growing cycles before the arrival of
become salt intolerant under drought conditions. In addition, winter. Therefore, the primary exposure is uptake of salt via
drought stress has been shown to aggravate vegetation with the soil once growth resumes in the spring, which generally
pre-existing salt damage and may lead to increased severity of occurs within the first 10 meters of the roadway, or where salt
symptoms (108, 112). levels are elevated because of runoff, drainage patterns, or
soil water movement (28). Elevated levels of snow and ice
control materials in soil have generally been found to
3.7.5 Other Factors Affecting Impacts
decrease over the growing season as a result of rainfall and
to Vegetation
runoff (23, 24).
The magnitude of effect that snow and ice control materi- Backman and Folkeson (109) found vegetation damage “to
als may elicit on roadside vegetation depends on a wide range a remarkable and previously unknown extent” that occurred
of conditions. First, effect is directly related to the exposure in spring 1994 alongside Swedish roadways. Effects were
concentration, which depends on application amounts and attributed to three interacting factors. First, roadsalt con-
specific transport mechanisms. General details on the mech- sumption was higher than usual during the preceding two
anisms and extent of soil, water, and aerial transport were dis- winter seasons. Second, applications occurred relatively late

56
in the season—the last application was in early April. Third, • Injury occurs in a linear pattern along roads or in areas
subsequent months reported near drought conditions with where runoff from roads collects;
unusually high temperatures and low precipitation. Salts were • Injury is more severe on the side of the plant facing the
not sufficiently washed away from foliage or the soil before road;
spring growth (bud unfolding). This combined with unfa- • Injury is more severe on the downhill side of the road;
vorable weather conditions caused a high degree of impact. • Injury decreases with the distance from the road;
Though the study concluded that salt contributed to the dam- • Injury is worse on the downwind side of the road;
age, it was also indicated that roadside vegetation is exposed • Parts of woody plants covered by snow or sheltered from
to severe conditions as a result of several interacting stress fac- spray lack injury symptoms;
tors. In many cases, this may lead to generally weakened con- • Parts of trees above the salt spray zone are not injured or
ditions, predisposing the vegetation to damage from other are injured less;
factors that include exhaust emissions, high wind velocity, • Salt spray injury only extends a short distance into dense
higher soil and air temperatures, and longer periods of sun- plants;
light. • Injury in coniferous plants becomes evident in late winter
In a study assessing roadside tree injury and decline at 17 and continues into the growing season; and
sites in interior British Columbia (113), common stressors • Injury in deciduous trees becomes evident in spring when
potentially involved in roadside decline were investigated, growth resumes and continues into the growing season.
including (1) highway winter and summer maintenance;
(2) climate; (3) herbicide use; (4) vehicle metal emissions;
(5) point source pollution; (6) insects and diseases; and 3.7.6 Specific Effects Associated with
(7) animal discharges. Roadsalt was found to be a major Select Snow and Ice Control Materials
inciting stressor; however, other factors such as climate
3.7.6.1 Sodium Chloride
(i.e., summer drought) and vehicle trace metal emissions
were significant independently or in combination with Sodium chloride toxicity to plants has been widely studied,
other factors. and evidence reveals that the damage to plants is more from
Blomqvist (114) found symptoms of salt damage in conif- high chloride concentrations than from high cation concen-
erous trees to not only impact the biota itself (e.g. browning trations. Hanes et al. (111) conducted a study examining
needles and needle loss), but also participate in other causal NaCl damage on silver maples and found that salt-damaged
relationships such as lower photosynthetic capacity, increased plant tissues contained sodium concentrations between 0.03
amount of salt in the soil water, osmotic stress, inhibition of percent and 0.08 percent, whereas chloride concentrations
water uptake, stress avoidance of the plant, and energy expen- were between 0.67 percent and 1.18 percent. Whereas 0.08
diture. Most of these effects will, in the end, diminish the percent sodium is not considered a toxic level, chloride levels
growth and can predispose the tree to damage from fungi or greater than 0.20 percent will result in damage to leaves and
insects. stems and 0.50 percent chloride will ultimately result in plant
Eppard et al. (115) assessed overstory vegetation in the death. Study conclusions indicated that, although accumula-
environmentally sensitive Lake Tahoe basin and found tion of sodium is detrimental to plants, there is more danger
approximately 45 percent of the study trees were affected by to plant life from accumulation of chloride ions.
one or more factors such as insect infestation, fungus, ozone, Although sodium is not an essential element, it is com-
or salt. The two main types of damage were attributed to monly found in plants and has improved the quality of cer-
insects (17.46 percent) and salt (15.05 percent). tain vegetable crops (28). Chloride is a micronutrient
Separating the different stress factors can be difficult given required in small amounts for plant growth (adequate plant
that one factor may have predisposed the tree to damage, tissue concentration is around 100 ppm). Excessive levels of
another factor may have triggered the damage, and a third may chloride in plant tissues may lead to necrosis, burning of leaf
have contributed to the actual killing of the tree. Effects also tips and margins, and eventual death. Some plants, however,
may not be realized for many years until accumulation of ions can screen out chloride through their root membranes.
is sufficient to elicit toxic effects to the vegetation. Because of Most annual crops and short-lived perennials are moder-
the wide range of circumstances and conditions, it is difficult ately to highly tolerant of chlorides. Trees, vines, and woody
to develop a general model to predict salt damage to vegeta- ornamental plants are often sensitive to chlorides (116).
tion. However, the literature indicates that certain conditions Similar effects and concerns can therefore be anticipated
are most likely to result in problems. The following injury pat- from other chloride-containing snow and ice control mate-
terns are based on those developed by Cain et al. (28): rials, namely CaCl2, MgCl2, and KCl.

57
Cain et al. (28) developed threshold soil concentration the soil. Elevated concentrations of CaCl2 in soil cause
values from estimates of effects concentrations providing osmotic stresses. Direct contact with the vegetation by salt
25-percent effects level, No Observed Effects Level (NOEL) spray and runoff can also cause injuries, including a general
and Low Observed Effects Level (LOEL) values and critical inhibition of growth, specific branch and vegetation die-off,
toxicity values estimated from experimental evaluations or or, in the extreme case, the death of the tree or plant. An Envi-
field sampling applications of NaCl. Threshold values for ronment Canada study (118) found that most of the observed
root exposure to Na ranged from 215 to 300 ppm, and for Cl calcium chloride-induced vegetation damage has been attrib-
was 300 ppm. Threshold values for tissue exposure to Na uted to direct contact—the absorption of salt splashed on the
ranged from 200 to 650 ppm, and for Cl, values ranged from foliage.
800 to 1650 ppm. A Belgian study (119) investigated the effect of CaCl2 on
Bright and Addison (41), during the development of four common roadside tree species. The study monitored
British Columbia matrix numerical soil standards for NaCl, timing of bud-break, development and spread of foliar necro-
constructed species sensitivity distributions based on lethal sis, and changes in leaf and soil chloride content for each
impacts (LC20) and non-lethal impacts (25th percentile of species. Effects varied among species, but were typically
species sensitivities distribution based on EC50) for a range of found to be a result of chloride ion concentrations.
plant species. Lethal effects for NaCl in soil were developed to In a study investigating CaCl2 used as a dust suppressant in
be 840 mg/kg (275 mg/kg Na, 425 mg/kg Cl) whereas non- the Yukon, the authors reported that there had been no
lethal effects were realized at 700 mg/kg (220 mg/kg Na, observed or reported instances of roadside vegetation injury
440 mg/kg Cl). or die-off related to areas where calcium chloride had been
applied (120). CaCl2 applied as a dust suppressant is not
transported as splash and spray to the same degree as when it
3.7.6.2 Calcium Chloride
is applied for snow and ice control and any effect observed
Calcium chloride has been used as a snow and ice control would be related to soil transport.
chemical and as a road dust suppressant. A literature review A European study (121), comparing the uptake of sodium,
revealed that there have been far fewer studies of the envi- calcium, and magnesium chloride by spruce seedlings, con-
ronmental impact of calcium chloride than sodium chloride. cluded that the presence of calcium ions had beneficial effects
Many of the studies have evaluated the effects of calcium on ion regulation. A reduction of accumulation of sodium in
chloride as a dust suppressant. Although the rates and tim- needles and young shoots occurred when calcium chloride
ing of application and transport mechanisms differ when was added to the sodium chloride. There was also a trend for
comparing snow and ice control and dust suppressant use, reduced uptake of chloride in the presence of calcium ions
the resulting ion-specific environmental impacts should be compared with sodium, magnesium, or potassium.
comparable.
Because of its widespread occurrence in rocks and soils and
3.7.6.3 Magnesium Chloride
its ready solubility, calcium is present in nearly all waters and
is more abundant than any of the other alkaline-earth metals Magnesium chloride has also been used in recent years
(e.g., sodium, potassium, and magnesium). The addition of both for snow and ice control and for roadway dust suppres-
calcium from snow and ice control activities would typically sion. Few published reports directly relating to either the veg-
be low when compared with the amount already present in etation impacts of magnesium chloride or its environmental
the environment. In contrast, although chloride is present in impact exist. Many inferences can, nevertheless, be made
all natural waters, the concentration is quite low in most based on the constituent components, namely Mg and Cl.
areas. The primary cause of damage to roadside plants is the The environmental impact of magnesium chloride should be
accumulation of chloride in plant tissues to toxic levels and, similar to that of CaCl2, given that both are chloride salts, with
therefore, CaCl2 poses a negative risk to vegetation (62). a cation with little known specific ion adverse effects.
Calcium is considered one of the major plant nutrients, Magnesium is an important plant nutrient because it is an
given that it is a structural component of cell walls (117). essential component of chlorophyll and is not considered
Most plants have between 1 percent and 2 percent by weight toxic even at high concentrations (35, 61, 89). The content of
of dry matter of calcium in their leaves. Calcium is a magnesium in plant leaves typically ranges from 0.1 percent
macronutrient required in larger concentrations for plant to 0.4 percent of dry matter (117).
growth, and adequate plant tissue concentration is 5000 ppm A discussion of the potential toxicity to plants resulting
(0.5 percent) (28). For calcium chloride to adversely affect from the application of liquid MgCl2 as a dust suppressant is
vegetation, the calcium chloride must migrate into the soil given in a report by McLaren Environmental Engineering
water or groundwater or significantly increase the salinity of (122). The author states that the potential toxicity to plants

58
from the use of magnesium chloride is based on the type of Concentrations of 5,000 and 10,000 mg/L reduced yield or
plant and on the amount of exposure to the salt solution. killed the plants tested. Horner further concluded that CMA
Exposure to plants could occur through direct contact of at concentrations used for snow and ice control does not
magnesium chloride to plant leaves and stems or by indirect harm terrestrial vegetation, except when concentrations in
contact via increasing the concentration of soluble salts— the root zone are high enough to cause osmotic stress in the
osmotic stress impacts—within the plant root zone away root zone.
from the area of application. Transport Canada (123) conducted field tests to compare
Direct contact of plant leaves with MgCl2 would adversely the environmental impact of KA with other deicers at Halifax
affect plants by creating an osmotic imbalance and subse- airport from 1990 through 1994. A limited number of tests
quent dehydration, resulting ultimately in defoliation of the were done to evaluate the effects of KA on vegetation. Green-
plant. The primary effect of increased concentrations of sol- house growth studies indicated little or no inhibition of sur-
uble salt in the plant root zone, which could occur from mag- face vegetative growth when KA was added at concentrations
nesium chloride application, is also osmotic, because high likely to be encountered during normal operating conditions
salt concentrations make it difficult for the plant to obtain (i.e., below 500 mg solids/L). However, at concentrations
water for growth. Excessive magnesium levels can also cause approximately double this (1000 mg/L), there was almost a
nutritional imbalances in plants. For example, excessive 50-percent inhibition of growth.
levels of magnesium in soil may produce calcium deficiencies
in plants (89).
3.7.6.5 Organic Products
In summary, the few published articles dealing with MgCl2
applications suggest that there is a low potential for damage Few assessments have looked at the effects of organic-based
to vegetation. However, there have been no laboratory or field snow and ice control materials on vegetation. This is perhaps
studies reported in the literature to confirm this hypothesis. a result of the number and variability of available proprietary
Given the similarities of MgCl2’s components and physical products. Many of these products are biodegradable, and it is
properties to those of CaCl2, it is suggested that the impacts generally assumed that they pose minimal concern for vege-
associated with MgCl2 use will be very similar to CaCl2 use. tation when compared with other types of snow and ice con-
trol materials. Many of the organic corrosion inhibitors
contain appreciable concentrations of phosphorus, a plant
3.7.6.4 Acetate Products (CMA and KA)
nutrient that can stimulate growth in small quantities. These
Acetate-based snow and ice control materials include CMA products have also been shown to contain high levels of
and KA. CMA was developed in the early 1980s as an alterna- ammonia, which also can serve as a plant nutrient to stimu-
tive to chloride-based snow and ice control chemicals, late plant growth.
whereas KA has seen use for airport deicing. Both these prod-
ucts partially consist of an acetate anion component and asso-
3.7.6.6 Other Materials (Urea, Potassium
ciated cations: Mg, Ca, and K. Any effects directly associated
Carbonate)
with these cations as components of CaCl2 or MgCl2 can be
related to CMA use. Potential effects include osmotic stress Monitoring of urea used at airports has shown no negative
and some ion-specific impacts. effect to vegetation because chemically, urea breaks down as
Extensive studies have been conducted to evaluate the urea molecules, as opposed to chlorides (29). The drawback
effects of CMA on terrestrial vegetation (36, 91) and prelim- of urea is the excessive fertilization it causes at the roadside
inary laboratory tests and literature reviews indicated that (31). Excessive nitrogen levels can also result in increased
CMA is harmless to plants. Winters et al. (91) conducted thatch build-up and the possible incidence of more insects
(1) tests using foliar spray applications to potted plants and and disease. As well, at pH values greater than 7.5, ammonia
(2) soil-leaching experiments. The results demonstrated that gas can be generated. This ammonia gas will usually dissolve
CMA was less harmful to plants than NaCl. Of the 18 tree in the available soil water or adsorb to soil particles and not
species tested, only 1, the Russian olive, was damaged more by pose any problems. If conditions are dry, however, free
CMA than NaCl. Horner (36) conducted a laboratory study ammonia gas can persist in the soil pore space, and this can
consisting of various applications of CMA to vegetation by damage plant roots when fertilizer has been banded near the
spraying and flooding and found CMA did not affect the root system.
yield, cover, vigor, or rooting of various herbaceous and An Austrian study found the application of up to
woody plants. All species tested withstood root zone applica- 100 gm2yr1 K2CO3 exerted a fertilizing effect on a test plot,
tions of up to 2500 mg/L, which is a much higher concentra- with slightly higher biomass yields than the untreated control
tion than expected from routine deicing treatments. (93). When application was increased to 200 gm2yr1 K2CO3,

59
the fertilizing effect disappeared. When annual doses of 200 servative approach ensures the greatest level of environmen-
to 400 gm2yr1 K2CO3 were applied, biomass dropped by 23 tal protection.
percent. In comparison, annual doses of 200 to 400 gm2yr1 Chloride materials have the greatest potential to impair
NaCl reduced biomass yield by up to 41 percent. This study vegetation. Sodium can degrade soil structure, which can
also found that, over time, accumulation of K2CO3 led to impede water uptake and root growth. Evidence suggests that
osmotic stress in the plants that led to cellular adjustments acetate materials can also affect vegetation through osmotic
made at the expense of growth. stress damage, although this does not occur to the same
Snow and ice control materials may render certain metals degree as for chlorides. It is generally agreed that OMB mate-
more available for uptake by plant roots, possibly resulting in rials pose minimal concern to vegetation when compared
plant toxicity, depending on the type and amount of metals with inorganic salts. Abrasives are not completely innocu-
present in the soil (28). Limited information is available on ous—they can cause smothering concerns in some instances.
the effects of heavy metals associated with snow and ice con- It is suggested that a wide range of stressors not related to
trol materials on roadside vegetation. snow and ice control already affects the roadside environment
and, in certain cases, the species present are not necessarily
indicative of the “natural” population through which the road
3.7.6.7 Abrasives
passes. Therefore, some damage may be manageable. That
Use of abrasives for snow and ice control can stress and being said, it is better that the snow and ice control practices
smother roadside vegetation. Small particles can adhere to the used reduce the amounts of detrimental products being
stomates and affect the regulation of oxygen, carbon dioxide, transported to the vegetation.
and water vapor between the plant and the atmosphere. Fis-
chel (69) cited information from Audesirk and Audesirk
3.8 Air Quality
(124) indicating that blocked stomates can also reduce or
inhibit photosynthesis and can prevent the plant from With conventional snow and ice control products such as
adsorbing nutrients from the soil. abrasive/salt mixtures, airborne fine particulate from the
entrainment of dried abrasives and salt left on roadways can
contribute to local air quality deterioration and may even
3.7.7 Conclusion
affect the health of some members of the population. Fine
An assessment of the gathered information suggests that particulate can penetrate deep inside the lungs, resulting in
most impacts associated with roadside vegetation impair- respiratory damage. Airborne particulate may also cause vis-
ment can be linked to (1) extremely high levels of ions con- ibility deterioration in an area. The U.S. EPA standard for
tacting foliage and/or the root zone and (2) ion-specific PM-10 is 50 µg/m3 annually, and 150 µg/m2 (in a 24-hour
damage related to chloride ions. As a result, most highway period) with three allowable exceedances over a 3-year period
agencies are faced with either accepting some degree of salt- (69). Mountain regions typically contain localized air sheds
related vegetation damage or modifying operations to reduce that have limited recharge and are of particular concern for
the amounts of detrimental materials reaching vegetation. air quality issues. The survey conducted for this report (see
Cain et al. (28) summarized the zone of impact as linear Section 2.4.2.1) returned similar findings: transportation
along highways with the effect from aerial dispersion of NaCl agencies in the mountainous western United States listed air
extending up to about 40 to 100 meters from the road edge, quality as a primary environmental concern related to snow
although this depends on local conditions such as salt load, and ice control materials.
traffic volume and velocity, local topography, and prevailing The City of Boulder, Colorado, found that vehicles con-
winds. Effects on woody communities have been noted up to tributed to 87 percent of particulate matter less than 10 (m in
100 meters, and severe effects have been noted up to 40 diameter (PM-10) and that most of this (90 percent) consists
meters. Elevated soil levels of roadsalts normally occur up to of road dust and sand thrown in the air by passing vehicles
the first 10 to 20 meters of the highway or along paths of over- (125). The percentage of total PM-10 emissions from road
land drainage and this is where most effect occurs. dust has been shown to increase between 50 percent to 90 per-
Effects associated with roadside vegetation impairment cent during the 24-hour period following road sanding (69).
from snow and ice control materials are linked to extremely Denver has poor air dispersion during the winter months
high levels of ions contacting foliage and roots and damage and was designated to be in non-attainment for the PM-10
related to elevated chloride levels. Vegetation damage is of standard with the passage of the Clean Air Act Amendments
greatest concern within the primary material deposition zone of 1990 (126). Testing of filter samples in the Denver area
10 meters from the roadway. Although certain types of vege- showed that a large fraction of the particulate problem
tation are more tolerant to these stresses than others, a con- resulted from re-entrained road dust. Denver historically

60
relied on abrasives for winter maintenance, but improved air 3.9 Animals
quality by using a harder sanding material (less prone to
crushing from vehicle action) and incorporating liquid snow Impairments to animal populations from roadsalts have
and ice control materials in some areas. included traffic hazards caused by salt-hungry animals as they
An Idaho DOT study (15) found abrasive use to be seek roadside salt accumulation (19). It is also speculated
reduced by 86 percent after traditional methods (deicing through repeated observations of apparent behavioral toxic-
using NaCl and abrasives) were replaced with liquid chemi- ity as well as new information on the toxicology of oral salt
cal anti-icing. Similarly, a Colorado study (13) investigating ingestion in birds that salt toxicity per se is contributing to the
the effects of abrasive use on air quality cited a general vulnerability of birds to road traffic (127).
improvement in air quality as shown by the fact that
Colorado has not exceeded a U.S. EPA PM-10 standard for 3.9.1 Wildlife
seven winters. This improvement correlated with a shift in
winter maintenance activities, away from abrasive use, Sodium is an indispensable component of physiological
toward snow and ice control chemicals. processes in all vertebrates, but is toxic when taken in excess.
There is limited information assessing the effect of both Animal tolerance to salt is usually quite high, although
dry and liquid snow and ice control chemicals on air quality. increased concentrations of salt can cause concerns in smaller
It is generally thought that use of these materials has a net animals. Some wildlife species actively seek out sources of salt
benefit on air quality. Newer application methods (e.g., anti- to ingest, most notably large ungulates such as moose, deer,
icing and pre-wetting) can also lead to significant reductions elk, and bighorn sheep. Roadways with applied salt are an
in abrasive use. It is possible that liquid products can dry and obvious source for these animals.
be re-suspended or be released as an aerosol, potentially caus- The predominant hazard that roadsalt poses to animals
ing localized air quality concerns. This is likely more of a con- has been expressed to be vehicle injury when these animals
cern for NaCl than for MgCl2 and CaCl2, both of which are are attracted to salt residuals on the roadside. Roadsalt
hygroscopic and will tend to retain water, potentially pre- attraction has been identified as a contributor to vehicular
venting re-entrainment. kills of bighorn sheep and elk in Jasper National Park (127)
Accounts of lung, skin, and eye irritation have been and to moose-vehicle accidents in northern Ontario and
linked to aerosols and spray produced by roadsalts. Data Quebec (19).
shown in Table 3-13 indicates that snow and ice control The peak volume for moose sightings near roadways and
materials can cause some level of inhalation irritation. The for moose-vehicle accidents occurs in the spring, which is
extent or magnitude of impact of these pathways has not when roadway salt is abundant in roadside puddles—
been well developed. researchers believe that moose are attracted to these accu-
In general, use of abrasives can result in air quality degra- mulations because of the increased dietary need for salt
dation because of increases in airborne fine particulate. This owing to horn growth in males and lactation in females
is particularly true in areas with limited recharge or existing (128). Fraser (129) found that 70 percent of moose-vehicle
air quality issues. Therefore, reduced use of abrasives will accidents in northern Ontario occurred between May and
result in reduced airborne fine particulate and improved air June. It has been expressed that when moose are drinking
quality. Application strategies such as anti-icing and pre- salty water, they tend to lose their fear of humans and vehi-
wetting, and use of liquid snow and ice control chemicals may cles (19). Deer have also been reported to exhibit similar
significantly reduce abrasive application quantities. behavior.
The Wisconsin DOT has monitored motor vehicle-deer
accidents beginning in 1979 (130) and its findings contradict
Table 3-13. Rat toxicity data (LD50) those of Fraser (129). Findings indicate that deer crashes peak
for selected liquid deicers. during October–November (38.1 percent of annual crashes)
with a secondary peak in May–June (16.2 percent of annual
Chemical LD50 [mg/kg] (rats)
crashes). The high incidence of crashes during the October–
Magnesium Chloride 8100
November period does not coincide with normal roadway
CMA 3150
snow and ice control activities and is suggested to be related to
Calcium Chloride 4000
behavioral changes associated with mating.
Sodium Chloride 3750
It is not known what the effect on other small mammals
Potassium Acetate 3250
may be. The only documented study found was in a review by
Adapted from (81, 131) Jones et al. (19) that observed salt toxicosis in cottontail

61
rabbits in Wisconsin. Laboratory experiments testing for salt Finches tend to be the most susceptible to salt ingestion.
tolerance showed behavioral symptoms characterized by Brownlee et al. (127) determined that as few as 3 cubic or 5
depression and thirst, followed by excitement, tremors, torti- spherical 2.4-mm salt particles could provide a lethal dose to
cillis, opisthotonus, retropulsion, complete incoordination, a Cardueline Finch. Birds can be sub-lethally impaired fol-
coma, and eventually death. Examination of the specimens lowing ingestion of a single salt particle and could be killed by
also found similar physiological impairments. as few as two particles. Birds are at a higher risk of salt toxi-
Jones et al. (19) also cited studies that assessed the effects of cosis when water availability is reduced during severe winters
salts in domestic animals. Sheep were able to tolerate salt- (19). Melted snow on the roadway may be the most obtain-
water concentrations up to 13,000 mg/L of NaCl, much higher able source of water, which depending on the concentration
than would be expected to result from roadsalt application. of salt may compound the problem.
With turkeys and chickens, very high levels of NaCl were Conclusions from the review conducted by Brownlee et al.
required to cause death—levels up to 4,500 mg/L and 5,000 (127) suggest that the contribution of roadsalt to wildlife kills
mg/L in food respectively. Swine, cattle, and horses have also has long been underestimated by wildlife managers and
been shown to be quite tolerant to high levels of salt. The transportation officials.
reports of salt toxicosis in free-ranging vertebrate wildlife
have been the result of exposure to roadsalt, drought, ice,
hypersaline lakes, and discharge ponds for various industries. 3.9.3 Toxic Effects
In most cases, it was found that drinking more water could
compensate for salt overdoses. The acute oral toxicity level of a chemical is an indicator of
its toxicity and can be expressed as a LD50 value. This value is
the dosage which is fatal to half of the test animal population.
3.9.2 Birds
The higher the LD50 value the less toxic the chemical. The
An extensive literature review assessing the effects of road- acute oral toxicity in rats for several chemicals present in
salts on birds was conducted for Environment Canada by snow and ice control formulations is listed in Table 3-14.
Brownlee et al. (127). In the literature, 12 published reports Using rats as test subjects Hiatt (131) compared acute tox-
were found associating bird kills with salted roads. Two inci- icities for CMA and NaCl in Table 3-13. CMA exhibited sim-
dents were formally diagnosed as salt poisoning and observa- ilar toxicity to NaCl for acute oral toxicity, as well as skin and
tions of potentially aberrant behavior suggestive of toxicosis eye irritation potential. No inhalation or dermal toxicity was
were made in several other cases. Brownlee et al. (127) found available for NaCl at the time of study; CMA exhibited low
several authors made observations that birds exhibited toxicity by both of these routes of exposure. For the sub-
behavioral deficits—most commonly unusual fearlessness chronic oral toxicity test, CMA produces no toxicity at a
and could be approached easily—related to salt ingestion. “limit” testing dose, thus indicating low toxicity upon
These severe behavioral impairments were observed as well as repeated exposure.
the finding of elevated brain sodium levels. The most com- LD50 values (rats) in the range of 500 to 5,000 mg/kg are
mon observation of unusual fearlessness documented in both considered slightly toxic, whereas values within the 5,000 to
birds and mammals could be an early symptom of intoxica- 15,000 mg/kg range are practically non-toxic (131). Although
tion or it could be the manifestation of animals under the table shows that chloride deicers are slightly toxic to
extreme salt hunger modifying their usual behavior toward rats, tolerance in animals is related to body size. Animals
potential danger (127). have a relatively high tolerance for elevated chloride levels
It is often assumed that salt is ingested in order to fulfill a and, if adequate drinking water is available, ingestion of
physiological need associated with a largely vegetarian diet. excessive chlorides should not adversely affect wildlife and
However, the taking of salt crystals as grit (in order to aid in birds (120).
grinding food and to provide supplementary minerals in
birds with low-calcium diets) cannot be ruled out (127).
Birds select grit on the basis of size, color, and shape, and
3.9.4 Bioaccumulation of Heavy Metals
crushed salt particles are irregular in shape and translucent
to white in appearance, characteristics that have been As discussed in Section 3.5, heavy metal ions can be dis-
observed to be features of choice in tests carried out with placed by specific ions such as calcium and magnesium. These
House sparrows. Particle sizes of normally applied NaCl metal ions may bioaccumulate in the food chain (81); how-
roadsalt (between 1 to 10 mm) overlap broadly with the pre- ever, information is lacking in public literature relating this
ferred grit size for the birds. potential effect to snow and ice control materials.

62
Table 3-14. Comparative acute toxicity of select snow and ice control materials.
Toxic Effect NaCl CMA CaCl2 MgCl2 Description

Acute Oral 3,150 mg/kg 3,750 mg/kg 1000- 4000- Test relevant to
Toxicity LD50 Slightly toxic Slightly toxic 2000 mg/kg 8,100 mg/kg occasional human
Slightly toxic over-exposure.
Acute N/A > 5000 mg/m3 Corrosive Non-irritating Toxicity test with
Inhalation exposure from
Toxicity LD50 breathing.
Eye Irritation Mild to Moderate Corrosive Slight irritant Measures irritation or

moderate irritant damage from direct
irritant eye contact.
Skin Irritation Non-irritating Mildly irritating N/A N/A Tests ability to
irritate or damage
skin on contact.
Acute Dermal N/A > 5000 mg/kg Corrosive Minor irritation Tests whether
Toxicity LD50 absorption of enough
compound across the
skin can lead to
systemic toxicity.
Dermal None found in No reaction N/A N/A Predicts ability to
Sensitization literature elicit allergic skin
reactions.
Adapted from (131), (81)
3.9.5 Conclusions sheets. Although most liquid snow and ice control formula-
tions are generally considered non-toxic and are only mildly
Snow and ice control materials have been linked to reports irritating, recommended handling practices by suppliers
of salt toxicosis from ingestion and to reports of increases in should be followed to minimize workers’ exposure. Indoor air
vehicle-animal collisions. Some information suggests salt tox- quality data is generally lacking in published literature
icosis from ingestion is a concern for small animals such as because effects tend to be site-specific.
birds, although most terrestrial animals have high salt toler- Toxicity is the measure of the capability of a chemical to
ances, especially when adequate drinking water is available. It cause injury. Toxicity parameters including acute oral toxic-
is unclear whether snow and ice control materials increase ity, acute inhalation toxicity, acute dermal toxicity, eye irrita-
vehicle-animal collision rates. Some research supports this, tion, skin irritation, and sub-chronic oral toxicity have been
but most current information suggests other conditions con- reported for several chemicals used in deicer formulations
tribute much more significantly to vehicle-animal collisions. and are provided in Table 3-13. For oral toxicity, the LD50
value given in Table 3-13 represents the chemical dosage pre-
dicted to be fatal to one-half of the animals tested. Chemicals
3.10 Human Health and Exposure
with lower LD50 values are more toxic.
Snow and ice control materials can affect human health Caution should be exercised in extrapolating or comparing
through three main routes: ingestion, inhalation, and dermal toxicological data presented in Table 3-13 and 3-14 given that
contact. The degree of potential harmful impact is further some of the data was based on tests on commercial deicing
influenced by factors such as the chemical concentration formulations while other results were generated using the
(dose), duration of exposure, frequency of exposure, and solid/pure form of the deicing chemical. In addition, the tox-
individual variability (16). icities of other additives or impurities present in deicing for-
Indoor air quality associated with snow and ice control mulations have not been investigated. Hence, a more in-depth
material storage, and processing also affects workers’ health examination of the published toxicity data is warranted.
and exposure. If the product is shipped to the user in powder/ Human health concerns associated with handling of snow
solid form, dust may be generated during the preparation of and ice control materials are associated with very high doses
the liquid formulation. Manufacturers and suppliers typically and any real concerns can easily be mitigated by using proper
recommend ventilated storage and processing areas on MSDS safety measures (e.g., protective clothing and ventilation).

63
CHAPTER 4
Impacts to Infrastructure and Vehicles
4.1 Concrete and Concrete cement paste is porous and, under normal ambient condi-
Infrastructure tions, the pores are filled with a highly concentrated solution
of ions that dissolve from the cement. The two major com-
Snow and ice control chemicals are known to affect con- ponents of the hydrated cement paste are calcium-silicate-
crete structures—either through deterioration of the con- hydrate (C-S-H) and calcium hydroxide [Ca(OH)2]. The
crete paste or corrosion of the reinforcing steel. Corrosion C-S-H phase provides the bond strength of concrete. The
of reinforcing steel has typically been the primary deterio- Ca(OH)2 does not contribute significantly to the strength, but
ration mechanism and has been linked to use of chloride- buffers the pH of the pore solution to about 12.5. At this pH
based snow and ice control chemicals such as NaCl, CaCl2, level, a protective oxide film forms at the steel surface (passive
and MgCl2. NaCl has been the most widely used salt, layer), reducing corrosion rates to negligible levels in the
although CaCl2 and MgCl2 use has increased in recent years. absence of chlorides.
There have been claims regarding the relative corrosion of Winter conditions (i.e., freezing and thawing cycles and
these products; however, there is some disagreement in the application of chloride-based snow and ice control chemi-
literature. cals) can affect the durability of reinforced concrete in several
Regardless of the chloride-based salt used, corrosion of ways:
reinforcing steel is inevitable, although the time to corro-
sion initiation or rate may vary by product. To address these • Physical deterioration of the concrete surface through
effects, alternative non-chloride snow and ice control scaling,
chemicals have been introduced. Various research has been • Chemical reactions between the salt and the cement paste
conducted on these chemicals—significant efforts have or aggregates causing degradation of the cement paste, and
been conducted on calcium-magnesium-acetate (CMA). • Diffusion of chloride ions resulting in corrosion of the
Urea, potassium acetate, and glycols have been widely used reinforcing steel.
for airports; however, their associated costs have limited
their use on roadways. The level of research investigating Chloride-containing snow and ice control materials are
the effects on concrete is limited, but considering that these the predominant materials used for snow and ice control,
chemicals are used because they are non-corrosive to air- and chloride ion concentration in concrete is a widely used
craft parts, their effect on reinforced concrete is probably measure of assessing corrosion-related problems. Chloride-
not detrimental. containing materials will be discussed with respect to
NaCl has been used for more than 50 years and associated chloride diffusion properties, corrosion problems, and
long-term effects to reinforced concrete are fairly well under- effects on the concrete matrix. Effects of non-chloride snow
stood. Although there has been extensive research into the use and ice control materials will be discussed with respect to
of alternative materials (e.g., MgCl2, CaCl2, and CMA), there reinforcing steel corrosion and concrete degradation.
is limited data about their long-term effects. This literature Limited information is available for many of these materials
review emphasizes the effects of corrosion on reinforcing and their ability to corrode reinforcing steel—probably
steel. The effects on concrete are addressed to a lesser extent. because of their limited use when compared with chloride
Concrete is a composite material consisting of stone and products and because they contain no measurable common
sand held together in a matrix of hydrated cement paste. The corrosion indicator, such as chloride.

64
4.1.1 Corrosion of Concrete-Embedded reinforcing steel, chloride ions can be introduced from two
Reinforcing Steel general sources:
Reinforcing steel embedded in concrete is protected against 1. Contaminants in the concrete mix, unintentionally from
corrosion by the high alkalinity (pH > 12.5) of the concrete mix water or aggregates, or intentionally such as for an
pore solution, consisting of saturated Ca(OH)2, sodium ions, accelerator (e.g., CaCl2); and
and potassium ions. The high pH induces a “passive” layer 2. Exposure of hardened concrete to environmental condi-
(i.e., chemical coating) on the surface of the reinforcing steel, tions (e.g., sea water or snow and ice control salts), allow-
reducing the rate of corrosion to negligible levels—the con- ing ingress of chlorides into the pore structure of the
crete acts as a chemical as well as a physical barrier. The break- concrete.
down of the passive layer initiates active corrosion on the
reinforcing steel and can be caused by the following (132): Chloride-induced corrosion is extremely complex and not
fully understood. It is assumed that chloride ions destroy the
1. Chloride-induced corrosion; passive layer that forms on reinforced steel (depassivation) by
2. Neutralization (pH reduction) of the concrete paste fol- replacing some of the oxides in the passive film, which further
lowing exchange between Ca2+ and Mg2+ ions; and causes an increase in ion solubility and conductivity in the con-
3. Carbonation-induced corrosion, which is a reaction of the crete cover (132, 134). Without the passive layer, corrosion can
cement paste with atmospheric carbon dioxide. occur if sufficient water and oxygen are available. Chloride-
induced corrosion mainly causes pitting corrosion with local
When discussing snow and ice control materials, only anodic and cathodic areas along the rebar (see Figure 4-1).
chloride-induced corrosion and neutralization are of con- The localized small anodic areas will dissolve at a low
cern. Carbonation, although possible in addition to the other potential, whereas the larger passive area of the steel acts as a
mechanisms, is not normally a problem with the quality of cathode where oxygen reduction takes place at a compara-
concrete used in bridge decks and will not be considered tively higher potential, creating a galvanic cell on a very small
further here. scale. This localization is promoted by the typically uneven
distribution of chlorides in the concrete. The rate at which the
dissolution of iron will progress depends on the cathode-to-
4.1.1.1 Chloride-Induced Corrosion
anode ratio, the electrical resistivity of the concrete, chloride
Chloride-induced corrosion of reinforcing steel is the concentration, and availability of oxygen. This localized
most widely occurring cause of concrete deterioration and attack is particularly dangerous because typical signs of
is well documented in the literature. Chloride-induced corrosion are not readily visible—general expansion resulting
corrosion poses the highest risk to structural integrity in cracking of the concrete cover may not occur.
because it causes a very localized attack of the steel, referred If occurring over a longer section of bar, cracking of the
to as “pitting corrosion” (132). Chloride-induced corrosion concrete cover occurs as a result of the volume of the corro-
is also considered the major cause of corrosion of reinforc- sion products (i.e., various forms of iron oxide) that is sev-
ing steel in concrete resulting in high cost for maintenance eral times greater than that of iron. The tensile forces exerted
and repairs (133). When considering corrosion to concrete on the concrete are large enough to cause cracking of the
Figure 4-1. Localized chloride-induced pitting corrosion of steel in

concrete.

65
concrete cover, delaminations, and, ultimately, spalling of the considered the end of the service life of a structure, regardless
concrete. of the rate at which the steel is corroding.
The typical threshold value (concentration) for chloride Chloride is a relatively small, negatively charged anion that
in concrete to initiate rebar corrosion of black steel is 0.4 can diffuse through concrete as a result of concentration
percent by weight of cement; however, values as low as 0.15 gradients—chloride ions will tend to move away from areas of
percent may be considered dangerous. The threshold value high concentration. The pore structure and micro-cracks that
by weight of cement may be hard to determine for real are always present in concrete can influence chloride penetra-
structures, because the cement content of the concrete may tion significantly. However, chloride diffusion through con-
be unknown. Typically, threshold values for chloride- crete has been shown to not only depend on the quality of the
induced corrosion range from 0.025 to 0.040 percent by concrete but also on the type of salt used. Chloride penetration
mass of concrete. thorough larger (macro) cracks is typically much greater
The chloride threshold value also depends somewhat on because such cracks allow passage for water and salts. These
the pH of the concrete pore solution at the reinforcing steel cracks can be a result of shrinkage, loading, and corrosion.
level. Both MgCl2 and CaCl2 cause a drop in concrete paste Two experimental laboratory studies have determined
pH, whereas NaCl raises the pH (135). This suggests that the diffusion coefficients to be greatest for MgCl2, less so for CaCl2,
chloride threshold for NaCl may be somewhat higher than for and lowest for NaCl (136, 137). For ordinary Portland cement
MgCl2 and CaCl2. However, this argument requires that the (OPC) mortar samples, the effective diffusion coefficient of
cation diffuses through the concrete at a rate similar to that of MgCl2 was about 3 times greater than that of NaCl, and CaCl2
chloride. If the diffusion rate of the cation is lower than for was about 2.5 times greater that that of NaCl (Table 4-1).
chloride, the pH effects will not necessarily influence the When using slag-blended cement, the effective diffusion coef-
corrosion initiation threshold value. ficient is significantly smaller for all three salts used, although
Considering that chloride-induced corrosion is independ- the same ranking of the salts exists (Table 4-1). The lower dif-
ent of the associated cation, there is no reason to consider any fusion rates when using slag-blended cement can be attributed
relative differences in corrosivity between various chloride to the much denser pore structure that occurs—an average
salts—the steel does not differentiate between the source of pore radius of 16 nm for slag-blended cement versus 31 nm,
chloride. The amount of chlorides that will penetrate the con- at the same overall concrete porosity (136). Other supple-
crete and their rate of penetration do, however, affect corro- mentary materials (e.g., fly ash and silica fume) will have
sion. This process depends on the amounts of materials similar effects on the permeability of concrete.
applied as well as the type of application (i.e., liquid versus The ranking of diffusion coefficients between the two
solid). When applying chlorides in solid form, a large amount studies also appears to be independent of the salt concentra-
may be blown off the roadway by wind or passing cars, tions used. While Deja and Loj (136) used saturated solution
whereas spray-applied brine will adhere to the roadway for concentrations (36 percent NaCl, 34 percent CaCl2, and
extended periods. Surface concentration of materials can be 29 percent MgCl2), Kondo et al. (137) used 0.5 M solutions
further affected by washoff from precipitation. It is, therefore, (2.6 percent NaCl, 5.5 percent CaCl2, and 4.8 percent
difficult to estimate the chloride loading accurately when MgCl2)—almost 10 times more dilute. In a laboratory study
using different types of applications. conducted by Callahan (140), it was also determined that
CaCl2 penetrated concrete faster than NaCl when using the
ASTM G 109 ponding test.
4.1.1.2 Chloride Diffusion Properties
The water-to-cement (w/c) ratio of the mortar mixes used
The initiation of reinforcing steel corrosion depends on the in Deja and Loj’s (136) research was 0.5, whereas the w/c ratio
ingress of detrimental chemicals, such as chlorides, through of the pastes used by Kondo et al. was not stated. As men-
the concrete cover. Thus, the rate at which these chemicals tioned, when using slag-blended cement, the variation
penetrate the concrete determines the time at which the steel between the different salts is much smaller. This suggests that
will start to corrode. In some cases, the initiation time is the quality and type of concrete mix should be considered
Table 4-1. Chloride diffusion coefficients for different salts.
Diffusion Coefficient (m2/s) x 10-12

Salt Type OPC Mortar Slag Blended Cement
OPC Paste (139)
(138) Mortar (138)
NaCl 9.1 1.4 6.6
CaCl2 22.9 1.5 9.9
MgCl2 29.0 1.8 20.8

66
when selecting snow and ice control chemicals. It is likely that Larger amounts of snow and associated flushing by meltwater
a high-quality concrete mix with a w/c ratio of less than 0.40 or more rainfall during the summer were presented as plausi-
will result in lower diffusion coefficients. ble explanations. Considering the results of the study in the
Diffusion of chloride in porous materials (i.e., concrete) context of laboratory studies, the findings are not unrealistic.
has also been linked to moisture content. Experimental diffu- Higher chloride diffusion rate test results attributed to MgCl2
sion coefficients can vary between 2.7 1012 and 12.9 are supported by the fact that MgCl2 opens up the concrete
1012 m2/s, where increased saturation is associated with pore structure (144). In addition, possible deterioration of the
higher diffusion rates (141). This suggests that chloride ions microstructure, including micro-cracking as a result of min-
can penetrate faster when the concrete remains moist. eral growth, can increase concrete permeability and cause a
A field study by the Montana DOT Engineering Division higher rate of chloride diffusion (145).
showed chloride diffusion rates in concrete to be more than Similar observations were made by Arya et al. (146) when
twice as high in areas where both MgCl2 and NaCl were used submerging concrete samples in salt solutions. Although no
than in areas where only NaCl was applied (142). Diffusion diffusion coefficients were determined, chloride contents in
coefficients were calculated based on chloride sampling the samples varied considerably with the different salt types
at various depths of bridge decks for twenty bridges along a (Table 4-2). Furthermore, the percentage of bound chlorides
100-km-long stretch of Interstate 90 in western Montana. Fifty was higher for MgCl2 than for CaCl2 and for NaCl—probably
kilometers (ten bridges) of the highway were exposed to NaCl because of the more readily occurring chemical reactions of
only and fifty kilometers (ten bridges) were exposed to MgCl2 the Mg2+ with the cement paste phases. The age of the samples
and NaCl since about 1989. Diffusion coefficients for MgCl2 at the time of exposure was only 2 days and 6-mm-thick cement
were calculated to be 120 percent greater than for NaCl paste discs of 49 mm diameter were used. These very imma-
(1.81 1012 compared with 0.82 1012 m2/s). There was ture and thin specimens were used to achieve rapid saturation
some variation in the results of this study, depending on the as well as to minimize damage to the pore structure. Other
sampling location—higher diffusion coefficient values were experiments in the same study showed that curing time and
determined for wheel paths than for other locations of the w/c ratios had little influence on chloride binding.
roadway. These results are comparable in magnitude to the The Montana DOT field study (142) supports the labora-
laboratory results of the blended cements, suggesting that real- tory test results, suggesting that real conditions are very sim-
istic mix design properties are required in laboratory studies ilar to observations made in the laboratory; however, there are
to allow extrapolation to “real” structures. There is also a good several problems in making reliable predictions for possible
correlation of the relative ratios between salt types and the service life (e.g., high w/c ratios, sample size, sample age, and
field and the laboratory studies, further indicating that MgCl2 the use of mortar or cement paste instead of concrete).
has a diffusion coefficient between 2 and 3 times greater than Although it is not possible to determine an exact time to cor-
that for NaCl. rosion initiation reliably, a relative comparison is possible
Several problems have been suggested regarding the valid- based on the findings of these studies. Rather than using the
ity of the conclusions of the Montana study (143). The section diffusion coefficients derived in the laboratory and in the field
of the highway exposed to MgCl2 is subject to more severe study as absolute values, they should be considered in relative
weather conditions—about twice as much precipitation—and terms—the ratio of DMg/DNa.
is at higher elevations than the section exposed to NaCl only. The following limitations should be considered when
It is difficult to comment on the rebuttal by Williams without translating the laboratory results to practical applications:
being able to verify either application rates or possible sam-
pling variations. However, application rates will affect surface • Cement paste and mortar was used instead of concrete;
chloride concentration, unless they were at saturation levels. • Sample sizes were typically small;
Table 4-2. Effect of cation binding in OPC paste.
Chloride content (% by weight of cement)

NaCl CaCl2 MgCl2
Free chlorides 0.831 0.765 1.480
Bound chlorides 0.804 1.408 2.347
Total chlorides 1.635 2.173 3.827
Bound/total (%) 50 65 61
Source: (146)
w/c = 0.5, samples immersed in 20g Cl-/L solution

67
• Experimental w/c ratios were greater than for typical mix and –700 mVCSE (151). These potential shifts were measured
designs; after several months of exposure. Once active corrosion of
• Chloride concentrations were not necessarily representa- steel embedded in concrete is initiated, the corrosion rates
tive of application concentrations; and may not differ between concrete exposed to NaCl and CMA.
• Exposure conditions (e.g., temperature, precipitation) were These effects of macrocell corrosion have been contra-
not representative of natural conditions. dicted by Chollar and Virmani (154), who did not measure
any significant potential shifts or corrosion of reinforcing
The above-referenced laboratory studies and field study steel embedded in concrete slabs and ponded with CMA solu-
suggest that chlorides associated with MgCl2 diffuse through tion. Comparable slabs exposed to NaCl exhibited potential
concrete at the greatest rate, followed by chloride associated shifts of up to 500 mV over the 4-year test period. The pond-
with CaCl2, then NaCl. Such differences in relative chloride ing solutions used were 3-percent NaCl, 3-percent CMA, and
diffusion rates can result in a reduction in the time to corro- 5-percent CMA. The slabs contained only one mat of rebar
sion of reinforcing steel. Despite the stated limitations, the and thus macrocell effects between different layers of rebar
findings in these studies remain valid for practical applica- could not be evaluated, considering the relatively small
tions, at least on a relative basis, although adjustments, change in potential for CMA.
particularly with respect to time, are required. No studies were found in the literature measuring the dif-
fusion rates of CMA into concrete.
Callahan (150) investigated the corrosion-inhibiting prop-
4.1.1.3 Calcium Magnesium Acetate (CMA)
erties of CMA on NaCl for concrete-embedded reinforcing
There has been a significant amount of research applied to steel. When using the ASTM G 109 ponding test, Callahan
the use of CMA as a snow and ice control material. The found CMA to be ineffective as a corrosion inhibitor when
researchers generally agree that CMA is less corrosive than using two parts of NaCl and one part of CMA (150).
chloride-based materials and may even act as a corrosion Special Report 235 by the Transportation Research Board
inhibitor (147–152). Many of the corrosion studies involving summarized many of the findings from previous U.S. studies
CMA were carried out on reinforcing steel bars or steel and it was concluded that CMA was significantly less corro-
coupons directly exposed to CMA solutions, rather than sive than chloride-based materials (155). Concrete deteriora-
embedded in concrete. These conditions are more represen- tion was limited to salt scaling for both materials, although
tative of atmospheric corrosion and thus more applicable to deterioration was somewhat more severe when using rock
structural steel, light standards, or automotive components salt. Neither materials reportedly had any effects on the
and not necessarily representative of steel embedded in con- strength or hardness of the concrete (155, 156).
crete. In these studies, CMA was found to be less corrosive
than chloride salts. This led to studies where CMA was
4.1.1.4 Urea
blended with NaCl to evaluate its corrosion-inhibiting qual-
ities. Ihs and Gustafson (149) found that a 20/80 NaCl/CMA Urea is typically used as a runway deicer by airports. Its
mixture (w/w) reduced corrosion rates by 45 percent when chemical composition, CO(NH2)2, suggests urea is unlikely to
compared with pure NaCl when in solution. Conversely, have any corrosive effects on steel reinforcement. There is
other studies found limited corrosion inhibition in limited literature available discussing the effect of urea on
NaCl/CMA mixtures. Ushirode et al. (153) found CMA not to reinforced concrete, although its use as a corrosion inhibitor
be effective as an inhibitor for NaCl concentration over 0.1 M, has been suggested. The electrochemical behavior of urea
even at relatively high concentrations of CMA of 1.0 M. indicates that it provides protection of the steel from depas-
Although steel exposed to CMA solutions corrodes at sig- sivation (153). However, that study showed that urea only
nificantly lower rates than steel exposed to chlorides, it has acted as a corrosion inhibitor to NaCl concentrations of less
been suggested that steel embedded in concrete may react dif- than 0.1M. Thus the general use of chlorides mixed with urea
ferently. A study by Locke et al. (151) concluded that active is considered not to be effective at inhibiting corrosion.
corrosion rates for steel exposed to CMA solutions are about
2 to 4 times less than for steel exposed to NaCl solutions.
4.1.1.5 Proprietary Products
However, possible changes in the pore solution chemistry of
the cement paste may result in macrocell corrosion because Various proprietary salts for snow and ice control have
of potential differences between different layers of steel. been introduced to the market. These typically include
Potential shifts of more than 600 mV were observed for steel blends of various chloride salts and may or may not include
in mortar samples exposed to 10-percent CMA solutions with corrosion inhibitors. However, considering that the con-
all potentials being in the active range, typically between –600 stituents of these salts are not published, research data is of

68
limited scientific value. In many cases, the proprietary salts main frost-related concrete deterioration problem in
are closely related to the chloride salts mentioned above. several countries (157).
Often corrosion inhibitors are added to improve corrosivity Research by Lee et al. (159) has shown that different
tests on metal coupon samples. However, these test proce- chloride-based snow and ice control chemicals (i.e., NaCl,
dures are not applicable to steel in concrete. MgCl2, and CaCl2 ) can cause varying degrees of damage to
concrete. This is however, mainly a result of specific chemical
reactions between the associated cations (i.e., Mg2+, Na+,
4.1.1.6 Corrosion Inhibitors
Ca2+) with various phases of the cement paste. Additionally, if
No literature was found that evaluated the effectiveness of reactive aggregates (e.g., dolomitic limestone) are present,
corrosion inhibitors when used with snow and ice control deterioration from chemical reactions between the cation and
materials at reducing corrosion of steel embedded in con- the aggregate has also been observed (159).
crete. Inhibitor tests are typically carried out on steel coupon
samples and represent atmospheric corrosion conditions.
4.1.2.1 Magnesium Chloride
Many of the corrosion inhibitors commonly used for snow
and ice control are derived from agricultural by-products— Several researchers agree that MgCl2 causes more severe
these materials tend to be highly biodegradable. As the migra- deterioration to concrete than do NaCl or CaCl2, because of
tion properties of these and other inhibitor products are not the reaction of Mg2+ with components of the cement paste
known, their usefulness for inhibiting corrosion of steel (138, 145, 159–163). The MgCl2 reacts with the cementitious
embedded in concrete is questionable. Migrating corrosion C-S-H in the cement paste to produce non-cementitious
inhibitors have been developed for application to concrete; magnesium-silicate hydrate (M-S-H) and CaCl2:
however, these are designed to penetrate concrete through
application of relatively high dosages and allowance of suffi- MgCl2 C-S-H ➔ M-S-H CaCl2
cient time to impregnate the concrete.
Additionally, MgCl2 reacts with Ca(OH)2 in the cement
paste to produce magnesium hydroxide (Mg(OH)2), also
4.1.2 Effects of Snow and Ice Control
known as brucite, and CaCl2:
Materials on Cement Paste
In addition to chloride-induced corrosion of steel rein- MgCl2 Ca(OH)2 ➔ Mg(OH)2 CaCl2
forcing, snow and ice control chemicals can affect the dura-
bility of the concrete paste physically (i.e., by increasing the Both of these reactions are favored because M-S-H and
freezing and thawing cycles within the concrete) and chemi- Mg(OH)2 are thermodynamically more stable than C-S-H and
cally (i.e., by reaction of the salt with the concrete matrix) Ca(OH)2. The formation of M-S-H is particularly detrimental
(157, 158). to the concrete because its lack of binding capacity weakens the
Without salt application, scaling of concrete can occur as a cement paste, resulting in loss of strength as it replaces the
result of the expansive forces of freezing pore water near the C-S-H. This, combined with the expansive forces generated
concrete surface. As the number of freeze-thaw cycles through Mg(OH)2 formation, can accelerate concrete deterio-
increases, so does the potential for scaling. When salts are ration. Such processes eventually will lead to physical crumbling
applied to the concrete surface, they increase the frequency of of the concrete (145, 159, 160). In many of the exposure tests,
freezing and thawing cycles over what would be experienced complete loss of concrete strength was observed (138).
under ambient conditions. This effect occurs when the The replacement of Ca(OH)2 with Mg(OH)2 will also
applied salt reduces the freezing point at the concrete surface, reduce the pH of the pore solution [pH =12.6 for saturated
then the salt is diluted as a result of precipitation or melt- Ca(OH)2 and pH = 9.0 for saturated Mg(OH)2] which, if
water, after which the surface water is allowed to re-freeze occurring at the rebar level, will result in the loss of passiva-
because of its lower salt content. tion of the steel and allow the onset of active corrosion, even
Physical effects on concrete durability appear to be inde- in the absence of chloride ions (164). This effect, combined
pendent of the type of salt used and are more a function of with the presence of chloride ions, will further accelerate the
concrete quality (e.g., adequate air entrainment) and corrosion of the rebar. It is not known at what rate Mg(OH)2
weather conditions (e.g., the number of freezing and thaw- will replace Ca(OH)2, specifically at the rebar level.
ing cycles). Although physical processes contribute to Although most research studying the effects of salts on con-
scaling, the salt-related deterioration that occurs in concrete has used laboratory-prepared concrete specimens, the
junction with this process is of primary concern. This samples used by Cody et al. (145) were cores removed from
process, termed “salt scaling” has been recognized as the “actual” structures in Iowa that were between 8 and 40 years

69
old. These cores were exposed in the laboratory to various salts (w/c ratios of 0.50 and 0.70) even after 10 years of exposure
and subjected to repeated freezing and thawing cycles. Deteri- to 2,500 mg Mg2+/L MgCl2 solution.
oration from MgCl2 appeared to be independent of the con- Although there is some disagreement about deterioration
crete quality, which was described as “durable” and of the cement paste, there is consensus that magnesium read-
“non-durable” service life concrete. “Durable” concrete has ily reacts with the various cement phases. Some research has
had a service record of more than 40 years, whereas “non- found deterioration of cement paste within the concrete
durable” concrete has had a service record of less than 15 years. matrix; others observed deterioration to be limited to the sur-
The “non-durable” concrete contained relatively porous, fine- face. There also appears to be consensus that Mg(OH)2 and
grained reactive dolomite coarse aggregate, which contains M-S-H are the most predominant reaction products.
significant amounts of magnesium. All concrete types con- To date, there have not been any observed effects to “real”
tained dolomitic limestone aggregates, because it is common structures that can attribute concrete deterioration to MgCl2
throughout the State of Iowa. The Iowa study concluded that use. However, the relatively short experience with any MgCl2
MgCl2 and CaCl2 caused deterioration of the concrete by (1) use as a snow and ice control material (approximately less than
promoting expansion of the concrete through Mg(OH)2 for- 15 years) does not allow long-term predictions to be made.
mation and other mineral growth and (2) chemical reactions
with the cement paste, whereas NaCl proved to be less
4.1.2.2 Sodium Chloride
detrimental (145).
Experiments assessing the effects of magnesium products Concrete exposed to NaCl typically showed very little sur-
on concrete durability have shown that MgCl2 acts differ- face distress from salt scaling (138) (145). Surface distress of
ently than magnesium sulfate (MgSO4) in that MgCl2 causes concrete subject to freezing and thawing more likely may be
the cement pore structure to open whereas MgSO4 causes a caused by physical rather than chemical mechanisms if NaCl
densification (144). The densification caused by MgSO4 is used. This conclusion by the authors reflects the observa-
results from sulfate expansion, which results in the pores tion that 0.75 M NaCl solution caused more extensive deteri-
being filled, before eventually causing disintegration of the oration to the concrete surface than 3.0 M NaCl did.
cement paste. The opening of the pore structure is attrib- Generally, salt scaling appears to be limited to the concrete
uted to the leaching of Ca(OH)2 during the exchange from surface when NaCl is used.
Ca2+ to Mg2+ to form Mg(OH)2, as discussed above. This Long-term use of NaCl has not shown any deterioration or
was confirmed by examination using a scanning electron loss of strength of the cement matrix in concrete structures.
microscope. Similar observations were made by Wakeley et Samples immersed in 4-percent NaCl solution actually had
al. (163), although the greater porosity was only observed higher strength than samples exposed to freshwater after
in the first millimeter near the surface. Magnesium hydroxy- 12 months of exposure (165). Therefore, the primary detri-
chloride formed near the surface and the loss in strength mental effect of NaCl on the concrete appears to be through
was attributed to the formation of M-S-H. The researchers corrosion of the reinforcing steel.
could not identify Mg(OH)2 near the surface of the mortar
specimens using energy dispersive X-ray analysis. The
4.1.2.3 Calcium Chloride
absence of Mg(OH)2 was attributed to the Mg(OH)2 phase
being more transient than expected and that M-S-H formed CaCl2 has been widely used as a snow and ice control mate-
more readily. rial for many years. Detrimental effects on concrete structures
Helmy et al. (165) also determined MgCl2 to be more detri- as a result of CaCl2 use do not appear to be more significant
mental to concrete than NaCl, even when using blended than those of NaCl, although laboratory studies suggest that
cement. The formation of Mg(OH)2 and loss of strength were the extent of deterioration caused by CaCl2 falls between the
observed in samples exposed to MgCl2 after 12 months of levels produced by NaCl and MgCl2. Concrete samples
exposure. The samples were immersed in 4-percent solutions exposed to CaCl2 deteriorated in similar ways to samples
that were renewed every month. Specimens exposed to NaCl exposed to MgCl2, although the deterioration was slower and
had higher compressive strength values than those exposed to less severe for CaCl2 (138) (145). Calcium chloride primarily
MgCl2, even higher than those exposed to freshwater after affected concretes containing reactive dolomite aggregates
12 months of exposure. because CaCl2 enhanced the dedolomitization reactions,
Conversely, Rechenberg and Sylla (166) found that forma- releasing magnesium to form Mg(OH)2 and M-S-H (159).
tion of Mg(OH)2 at the concrete surface acts as a protective Considering that dolomitic limestone is not used in all juris-
layer, preventing the ingress of aggressive species. They dictions, there is less concern over this type of deterioration.
observed the formation of Mg(OH)2 exclusively near the Cody et al. (145) concluded that MgCl2 and CaCl2 most
surface and found no deterioration of concrete specimens likely will cause long-term deterioration in the concrete when

70
compared with NaCl; however, a relationship between “real” This effect is primarily related to bridges; however, there is
applications and laboratory data could not be established. a possibility of a secondary risk in concrete pavements with
There is no evidence of more pronounced deterioration on doweled joints or continuous reinforcement.
concrete structures in areas where CaCl2 has been applied for • Reactions of snow and ice control materials with the cement
several years. paste, resulting in deterioration of the cement matrix. This
effect is primarily a risk for pavements and the extent of
deterioration is directly related to the quality of concrete.
4.1.2.4 Calcium Magnesium Acetate (CMA)
Evidence suggests that magnesium readily reacts with the
Although most of the earlier studies involving CMA have cement paste in a process that reduces the concrete strength.
focused on its effects on steel corrosion, there is limited infor- • All snow and ice control chemicals, by definition of their
mation about its effect on the concrete itself. There is dis- action, increase the number of freezing and thawing cycles
agreement about its effects on the durability of the concrete, on the concrete in service. This can result in deterioration
especially when considering potentially detrimental effects of in the form of freezing and thawing and/or scaling.
magnesium on the cement paste matrix (see Section 4.1.2.1)
(149, 167, 168). Considerable research is being done on the effects of snow
Ihs and Gustafson (149) found that, compared with NaCl, and ice control materials on concrete paste. To date, there
surface scaling of concrete was reduced when using CMA. A have been conflicting findings as to how the various materi-
similar observation was made for CMA by Peterson (167), als affect the durability of the concrete. Traditional snow and
who found that a layer of Mg(OH)2 formed at the concrete ice control materials (e.g., NaCl and CaCl2) are known to have
surface exposed to CMA. Similar to all magnesium salts, it some detrimental effects on concrete, causing primarily scal-
was found that CMA causes a dissolution of hardened cement ing of the concrete surface, although no problems with the
paste, although the dissolution depended on temperature— bulk properties of the concrete are known. With good-quality
the dissolution was very slow at 5°C and rapid at 20°C. concrete, scaling problems can be avoided. Some laboratory
A laboratory study by the Ministry of Transportation of studies have shown deterioration of the cement paste and
Ontario (169) concluded that good-quality concrete resisted bulk properties of the concrete by CaCl2; however, the
scaling by both NaCl and CMA. Less scaling was observed when material is widely used, and these findings do not appear to
using CMA (compared with NaCl) on lower quality concrete be substantiated in field applications.
mixes that tended to scale. A further field study by the Ministry Laboratory studies with magnesium-based materials
of Transportation of Ontario did not identify any detrimental (e.g., MgCl2 and CMA) have found chemical degradation of
effects of CMA on concrete surfaces or corrosion, although the the concrete matrix as a result of magnesium reacting with
study period was limited to the winter of 1986–87 (170). the calcium-silica-hydrate (C-S-H) by replacing the calcium
In contrast with the previous studies, Santagata and with magnesium and, in time, resulting in strength loss of
Collepardi (168) found the cement matrix of concrete can be the concrete. There have been some reports of MgCl2 caus-
severely attacked by CMA and exposures of air-entrained con- ing deterioration of the concrete in the field; however, there
crete with a w/c ratio of 0.45 percent to a 25-percent CMA solu- does not appear to be a widespread problem. However,
tion resulted in significant degradation of the concrete. This led MgCl2 is used to a somewhat lesser extent. Considering the
to compressive strength losses of up to 50 percent. When using reactivity of magnesium, long-term effects are possible and,
slag-blended cement (i.e., 40-percent slag) these effects were given the findings of past studies, future work will be
somewhat mitigated, suggesting that the aggression is probably required. Close monitoring of structures where MgCl2 is
associated with leaching of Ca(OH)2 from the limestone Port- used is also recommended.
land cement. This agrees with other studies and can be attrib- Chlorides enable and/or accelerate corrosion of concrete
uted to the exchange of Ca(OH)2 with Mg(OH)2, which does reinforcing. Current evidence suggests that chloride associ-
not possess any strength characteristics (see Section 4.1.2.1). ated with magnesium may have higher chloride diffusion
coefficients than NaCl or CaCl2. Considerable research is
ongoing and, for practical purposes, all chloride materials
4.1.3 Concrete-Embedded Reinforcing Steel
have been ranked equally high in causing corrosion of the
and Concrete Paste Conclusions
reinforcing steel.
Snow and ice control chemicals can be detrimental to con- This study includes an analytical research program inves-
crete infrastructure in the following three ways: tigating the relative chloride diffusion coefficients of chlo-
ride-based snow and ice control materials. After 1.5 years of
• The enabling and/or acceleration of corrosion of concrete exposure testing, corrosion has not been initiated for any of
reinforcing (rebar) resulting from chloride ion penetration. the materials. The rankings will remain based on the

71
comprehensive literature assessment, because this time of infrastructures, and steel bridges (172–181). Forms of corro-
exposure can be considered short when viewed in terms of sion include uniform or general corrosion, crevice corrosion,
concrete durability where current service life requirements poultice corrosion, pitting, galvanic corrosion of dissimilar
in the order of 30 to 100 years are commonly specified. metals, and filiform corrosion of aluminum and magnesium
Acetate-based materials, OMBs, and abrasives are not con- alloys. All the above forms of corrosion require water and an
sidered corrosive to concrete reinforcing. ionic species such as chloride or another electrolyte for accel-
erated corrosion.
Crevice corrosion, poultice corrosion, and pitting are all
4.1.4 Atmospheric Corrosion Costs
forms of oxygen concentration cell corrosion. These forms of
Atmospheric corrosion is one of the most important corrosion all occur because of a difference in the concentra-
degradation processes that limits the life of motor vehicles tion of dissolved oxygen at one point as compared with
and roadside infrastructure where snow and ice control mate- another point on the metal surface. The direction of the cor-
rials are used for winter road maintenance. Corrosion dam- rosion reaction is always toward equilibrium, which occurs by
age results in extensive loss of capital. A study in 1976 (171) reducing the concentration of oxygen where it is highest. The
concluded that the damage caused by snow and ice control surfaces in contact with the solution containing the higher
chemicals on water supplies, health, vegetation, highway concentration of dissolved oxygen will become cathodic. The
structures, vehicles, and underground utilities exceeded surfaces in contact with the solution containing the lower
$3 billion per year or approximately 15 times the amount of concentration of dissolved oxygen will become anodic. The
money spent on purchasing and applying snow and ice con- anodic surfaces will suffer accelerated corrosion.
trol chemicals. Later studies have estimated the cost of corro- The various forms of corrosion that typically occur on
sion from these chemicals for motor vehicles and highway motor vehicles, steel bridge components, and roadside infra-
infrastructures to have increased. structure are summarized in the following sections.
Menzies (172) estimated the annual cost of corrosion from
snow and ice control chemicals for cars, light trucks, vans,
4.1.5.1 Uniform or General Corrosion
recreation vehicles, heavy trucks, and buses to be approxi-
mately $900 million for corrosion prevention on 1.4 million Uniform or general corrosion is defined as (1) a type of
new vehicles, and approximately $1.4 billion for the depreci- corrosion attack (deterioration) uniformly distributed over a
ation cost of the 86 million vehicles in operation. These costs metal surface or (2) corrosion that proceeds at approximately
total approximately $2.3 billion annually. the same rate over a metal surface (182). Uniform or general
A report commissioned by the FHWA in 1999 on the cor- corrosion is usually not a significant problem except on brake
rosion costs and prevention strategies in the United States components, structural members (e.g., frames and under-
(173) summarized the annual corrosion-related costs to bodies of motor vehicles) where thinning can result in a loss
motor vehicles to be approximately $23.4 billion per year. of strength, or electrical components where corrosion can
This cost was divided into three components: (1) increased cause failures. Most other components are relatively massive
manufacturing cost because of corrosion engineering and the or are protected by coatings.
use of corrosion-resistant materials ($2.56 billion per year);
(2) repairs and maintenance necessitated by corrosion ($6.45
4.1.5.2 Crevice Corrosion
billion per year); and (3) corrosion-related depreciation of
vehicles ($14.46 billion per year). The same study estimated Crevice corrosion is defined as localized corrosion of a
the annual dollar impact of corrosion on steel highway bridge metal surface at or immediately next to an area shielded
components to be $2 billion for maintenance and the cost of from full exposure to the environment because of proxim-
capital for bridge substructures and superstructures, and ity between the metal and the surface of another material
approximately $500 million for the maintenance painting (182). Crevice corrosion occurs in cracks or crevices formed
cost for steel bridges. Although all the costs associated with between mating surfaces of assemblies. The surfaces may be
vehicle and infrastructure impairments in the 1999 study the same or dissimilar metals or one surface may be a non-
(173) are not due to snow and ice control materials, they are metal (e.g., scale, surface deposits, washers, or gaskets).
thought to be the primary contributing factor. Crevice corrosion is a form of localized attack from oxygen
concentration cells where rapid perforation to failure can
occur on film-protected metals (e.g., aluminum, magne-
4.1.5 Forms of Corrosion
sium, and stainless steel). Crevice corrosion can also cause
Numerous authors and studies have discussed the various corrosion product wedging failures of fasteners because of
forms of corrosion that can occur on motor vehicles, roadside axial tensile overload.

72
4.1.5.3 Poultice Corrosion (i.e., voids and thin spots in the coating) and mechanical
damage on the organic coating surface.
Poultice corrosion is also called deposit corrosion or attack
and is a term used in the automotive industry to describe the
corrosion of vehicle body parts as a result of the collection of 4.1.6 Atmospheric Corrosion of Motor
roadsalts and debris on ledges and in pockets kept moist by Vehicles, Roadside Infrastructure,
weather and washing (182). Poultice corrosion is another and Metal Bridges
form of localized attack from oxygen concentration cells
Corrosion is an electrochemical reaction in which the
(183). Deposits that have accumulated can soak up corrosive
metal is oxidized in the presence of moisture and atmospheric
(ionic) substances (e.g., snow and ice control chemicals) and
oxygen. The addition of an electrolyte (e.g., salt or other snow
act as a sponge to trap moisture.
and ice control chemicals) facilitates electron transfer and
Little or no corrosion occurs while the entire assembly is
greatly accelerates the corrosion.
dry or wet. During the drying period, adjacent wet and dry
Atmospheric corrosion of metals occurs when the surfaces
areas develop and differential aeration cells develop at the
become wet. The rate of atmospheric corrosion on the wet
edge of the wet zone. This can lead to rapid corrosion.
surface is accelerated by the presence of ionic substances such
as chlorides. The total time of wetness is important because
4.1.5.4 Pitting Corrosion this affects the total corrosion, especially if chlorides are pres-
ent. The availability of oxygen is also important because oxy-
Pitting is a form of localized corrosion that results from gen concentration cells develop under deposits and in
local cell action that produces cavities starting at the surface. crevices and pits. The processes involved in atmospheric cor-
The cavities may or may not become filled with corrosion rosion are complex and involve numerous interacting and
products, and corrosion products may form nodules over the constantly varying factors (e.g., weather conditions, air pol-
pit cavities (184). Pitting occurs when a film-protected metal lutants, material conditions, and temperature).
(e.g., aluminum, magnesium, or stainless steel) is almost, but
not completely, resistant to corrosion. Only a small amount
of metal is corroded, but perforations can lead to costly 4.1.6.1 Atmospheric Corrosion of Motor Vehicles
repair of expensive equipment such as motor vehicles. Pitting Motor vehicle corrosion can be classified into two major
is a form of localized attack that primarily occurs on alu- categories: structural corrosion and cosmetic corrosion (185).
minum, magnesium, and stainless steels, especially if chlo- Structural corrosion is corrosion degradation that causes a
rides are present. Once pitting has initiated, the mechanism loss of integrity such as perforation of body panels, corrosion
is similar to crevice or poultice corrosion with the pit acting of brake components, and deterioration of frame, underbody
as the crevice. or bumper support systems. Cosmetic corrosion only affects
the aesthetic appearance of the motor vehicle such as rust
staining and blistering of painted body panels and discol-
4.1.5.5 Galvanic Corrosion
oration and pitting of trim. A general differentiation between
Galvanic corrosion is defined as accelerated corrosion of a perforation corrosion and cosmetic corrosion is that perfora-
metal because of an electrical contact with a more noble tion corrosion occurs from the inside of the body panels,
metal or non-metallic conductor in a corrosive electrolyte while cosmetic corrosion occurs on the outside of the body
(182). The more active metal becomes the anode and may suf- panels, trim, and wheels (175).
fer rapid corrosion in the vicinity of the contact. Snow and ice Localized corrosion from oxygen concentration cells (e.g.
control materials significantly increase galvanic corrosion by crevice, poultice, pitting, and filiform corrosion) and galvanic
increasing the conductivity of the electrolyte. The most corrosion are probably the most significant causes of corro-
prevalent problem in motor vehicles is contact of aluminum sion failures in motor vehicle service. All forms of localized
components with steel. corrosion can be defined as selective removal of metal by cor-
rosion at small special areas or zones on a metal surface in
contact with a liquid or moist environment. It usually occurs
4.1.5.6 Filiform Corrosion
under conditions where most of the metal surface is either
Filiform corrosion is another form of crevice corrosion not attacked or is attacked to a much lesser degree than the
that can occur on aluminum and magnesium alloy surfaces localized sites.
under an organic coating. It appears as a random distribution The motor vehicle environment is a dynamic condition
of threadlike filaments on the metal surface under organic because of road splash that carries snow and ice control
coatings (182). Corrosion typically initiates at holidays chemicals and other ionic materials to the metal surface, as

73
opposed to roadside components such as signage, guard rails, and noise barriers; underground utility lines, pipelines, and
bridges, and buried structures that are basically static. Motor storage tanks; and objects such as bronze statues and monu-
vehicle design has resulted in complex shapes, joints, and con- ments and metallic building siding and roofing (185).
figurations that result in areas where deposits can build up. Probably, there is damage to railroad-owned property at
These deposits retain moisture and ionic species, producing and near railroad crossings because of snow and ice control
oxygen concentration cells where poultice and crevice corro- chemicals. Such damage would include corrosion of rails and
sion can occur. switching sections, wiring, and so forth. The railroad systems
Crevice corrosion and poultice corrosion typically occur apparently do not consider this type of corrosion to be a
where dirt and moisture are trapped—between adjacent major expense and do not track this data (173). Further infor-
pieces of metal, under gaskets and at fasteners, or on the sur- mation detailing issues related to railway traffic control sig-
face of motor vehicle components. This is compounded by naling is provided in Section 4.2.1.
ingress of snow and ice control chemicals and other ionic
materials (e.g., acid rain) that increase the conductivity of the
4.1.7 Motor Vehicle, Roadside
trapped moisture. Aluminum alloys are more prone to crevice
Infrastructure, and Bridge Materials
corrosion and galvanic corrosion when coupled to steel.
The motor vehicle exhaust system has become a critical Motor vehicles, roadside infrastructure, and metal bridges
component because of government legislation requiring used a very wide range of metals, non-metals, and combina-
zero-leak performance with suggested design lives approach- tions of materials in intimate contact for their fabrication and
ing 240,000 km (150,000 miles) in the future (186). Stainless construction (175, 186, 187). Materials are described in the
steels have performed well, but with higher operating tem- following sections.
peratures from efforts to increase gasoline mileage and reduce
emission levels, much more corrosion-resistant materials will
4.1.7.1 Steel and Cast Irons
be required. Snow and ice control chemicals splashed on hot
exhaust system components can cause severe corrosion from Steel components include a very wide range of carbon,
hot (molten) salt corrosion and, as such, these chemicals will alloy, high-strength, low-alloy, and weathering steels. They
become a much more important aspect as a source of failure can be hot-rolled or cold-rolled sheet, plate, bar, pipe and tub-
in the future. ing; hot-rolled sections; forgings; or castings. The steel
components can be bare, metallic-coated (e.g., zinc- or
chromium-plated) or organic-coated (e.g., painted).
4.1.6.2 Atmospheric Corrosion of Roadside
Motor vehicles typically use hot-rolled steel sections for
Infrastructure and Metal Bridges
frames, bumpers, and other structural parts, which may or
General corrosion and crevice corrosion are the principal may not be coated; steel tubing for brake lines; cold-rolled
forms of corrosion observed on roadside infrastructures and coated steel sheet for body panels and fuel tanks; and steel
metal bridges. Roadside infrastructure and steel bridges suf- castings or forgings for fittings, steering components, and so
fer uniform corrosion, especially at coating failures, severe forth. Stainless steel sheet is used for motor vehicle trim, fuel
crevice corrosion at riveted and bolted joints, and localized tanks and exhaust systems. Stainless steel is also used for
corrosion under deposits (i.e., poultice corrosion). Severe tubing applications.
corrosion at riveted and bolted joints can cause tensile failures Roadside infrastructure and steel bridges use many grades
of fasteners because of corrosion product wedging (i.e., rust of structural steel that can be galvanized and/or painted.
jacking). Cast irons and ductile irons are used for motor vehicle
In the United States, approximately 15 percent of all engines and drive train components, brake drums and disks,
bridges are structurally deficient because of corrosion (173). and miscellaneous fittings. Roadside infrastructure and
The primary cause of bridge steel structural corrosion is bridges use cast iron and ductile iron cable eyes, fittings,
chloride-induced corrosion with the chlorides resulting from column bases, and so forth.
snow and ice control chemicals and marine exposure. General
corrosion damage also occurs on roadside metal guardrails,
4.1.7.2 Aluminum Alloys
signage, and so forth.
Corrosion can also be accelerated on numerous metallic Numerous aluminum alloys and forms of aluminum are
components alongside roads and highways because of snow used and can be bare, metallic-coated (e.g., chromium-
and ice control chemicals. Such components include drainage plated) or have organic coatings (e.g., painted). Motor vehi-
systems; fixtures such as sign posts, light standards, traffic sig- cles use various coated and uncoated sheet aluminum alloys
nal wiring, guardrails, wire fencing and metal retaining walls, for body panels, fuel tanks, trim, radiators, and so forth. Alu-

74
minum alloy castings and/or forgings are used for wheels, 4.1.8.1 Chloride-Based Snow and Ice Control
engine and drive train components, and miscellaneous fit- Chemicals
tings. Aluminum tubing is also used. Aluminum alloy tubing,
Chloride-based salts are all highly corrosive to motor vehicle
piping, extrusions, castings, and forgings are used for
components and roadside infrastructure but vary in degree,
guardrails, handrails, light standards, and so forth, while sheet
with the most corrosive being the hydroscopic chlorides (e.g.,
and plate aluminum alloys are used for building cladding, sig-
magnesium chloride and calcium chloride). The hydroscopic
nage, and so forth.
chlorides are more corrosive because deposits remain moist and
allow corrosion to occur for a much longer period. It has been
4.1.7.3 Magnesium Alloys demonstrated that when acid rain is present, there is a synergis-
tic effect (increase in corrosion rate) with chloride salts (200).
Magnesium alloy die or permanent mold castings and
forgings are used for motor vehicle wheels, transmission
housings, and other miscellaneous brackets and supports. 4.1.8.2 Acetate and Organic Snow and Ice
Control Chemicals
4.1.7.4 Copper and Copper Alloys The acetate-based and miscellaneous organic deicer
chemicals tend to result in significantly lower corrosion
The principal use of copper is for electrical wiring in rates than the chloride-based deicer chemicals (185, 200).
motor vehicles, roadside infrastructure, and bridges. Some Generally, even the most innocuous snow and ice control
copper tubing is also used for various applications. Copper chemical can still cause significant corrosion of metals if the
alloys (e.g., brass) are also used for electrical contacts, motor chemical increases the conductivity of the moisture in
vehicle radiators, and brake line fittings. Brasses and bronzes contact with the metal surface and/or increases the time of
are used for roadside heritage statues, markers, and so forth. surface wetness.
4.1.8 Snow and Ice Control Chemicals 4.1.9 Corrosion Inhibitors

There are extensive references in the literature discussing Any substance added in small amounts to a corrosive
atmospheric corrosion of various types of snow and ice con- environment that effectively reduces the corrosion rate of a
trol chemicals on motor vehicles and roadside infrastructure metal or alloy can be called a corrosion inhibitor. Corrosion
as well as comparisons of corrosion rates from field and in inhibitors are chemical substances or combination of sub-
accelerated laboratory tests (185, 188-199). stances that, when present, prevent or reduce corrosion without
Snow and ice control chemicals are typically ionic or non- significant reaction with the components of the environment.
ionic chemicals or mixtures of chemicals with or without cor- Many commercially available snow and ice control materials
rosion inhibitors. They typically have high solubility in water contain inhibitors. Agricultural by-products, proprietary in
at low temperature; relatively low molecular weight; and nature, are very popular additives to chloride-based chemicals.
eutectic temperatures lower than the expected ambient tem- Other materials such as triethanolamine (TEA) and phos-
perature (189). Numerous chemical substances could be used phates, carbonates, and silicates are not as popular, but have also
for snow and ice control, but the number actually used is rel- been used to reduce corrosion effects.
atively small because of cost, availability, environmental con- There are basically three types of inhibitors (174, 184,
straints, and so forth. 201–203):
In discussing atmospheric corrosion processes, snow and
ice control chemicals can generally be grouped as follows: • Anodic, which limit the anodic half-cell of the corrosion
chloride-based salts (e.g., sodium chloride, potassium chlo- reaction;
ride, magnesium chloride, calcium chloride, and mixtures • Cathodic, which impede the cathodic half-cell of the cor-
of chlorides or salt brines); acetate-based products (e.g., rosion reaction; and
calcium magnesium acetate and potassium acetate); and • Absorption or Passivating, which form a physical barrier
miscellaneous organic compounds (e.g., glycols, alcohols, on the metal surfaces.
and proprietary agricultural by-products—primarily
sugar-based organic biomass). Many commercially avail- Liquid phase or moist condition inhibitors are classified as
able products also consist of proprietary blends of snow anodic, cathodic, or mixed inhibitors, depending on whether
and ice control chemicals with and without corrosion they inhibit the anodic, cathodic, or both electrochemical
inhibitors. reactions.

75
Corrosion inhibitors tend to be metal and environment the cathode surfaces under localized conditions of high pH,
specific. No inhibitors retard corrosion significantly for all thus effectively isolating the cathode from the solution. Typ-
metals and alloys and all snow and ice control chemicals. ical precipitating inhibitors or cathodic inhibitors are calcium
bicarbonate, zinc ions, polyphosphates, and phosphonates.
They are all suitable for inhibiting snow and ice control
4.1.9.1 Anodic Inhibitors
chemicals.
Anodic inhibitors typically are used in near-neutral solutions
where insoluble products (e.g., oxides, hydroxides, or other
4.1.9.3 Mixed Inhibitors
salts) are formed. Anodic inhibitors form or cause the forma-
tion of passivating films that inhibit the anodic metal dissolu- Most commercial inhibitors are organic compounds that
tion reaction. When the concentration of an anodic inhibitor is cannot be designated as either anodic or cathodic and are
insufficient, corrosion can be enhanced rather than inhibited. known as mixed inhibitors. The effectiveness of organic
Typical anodic inhibitors are often called passivating inhibitors is related to the extent to which they reabsorb and
inhibitors and include chromates, nitrites, molybdates, phos- cover the metal surface. Absorption of the organic inhibitor
phates, carbonates, and silicates. In practice, only phosphates, depends on the structure of the inhibitor, on the surface
carbonates, and silicates tend to be suitable for snow and ice charge of the metal, and on the type of electrolyte. Mixed
control chemicals and they are used primarily to prevent cor- inhibitors protect the metal in three ways—physical absorp-
rosion of iron-based alloys. tion, chemisorption, and film formation (203).
They are the most efficient inhibitors because they can Organic compounds constitute a broad class of corrosion
almost completely prevent corrosion, but they are also dan- inhibitors that cannot be designated specifically as cathodic
gerous inhibitors because they can accelerate corrosion (i.e., or anodic. Cathodic inhibitors (e.g., amines) or anodic
cause severe localized attack) at low concentrations. Extensive inhibitors (e.g., sulphamates) will be absorbed preferentially
studies of inhibitors have shown the accelerated corrosion depending on whether the metal is charged negatively or pos-
effect of low anodic inhibitor concentrations, but apparently itively respective to the solution.
no studies have been carried out to evaluate the corrosive Numerous organic-based corrosion inhibitors are being
effect of snow and ice control chemicals when they are diluted widely used in conjunction with chloride-based snow and ice
and the anodic inhibitor concentration is less than optimal. control materials. These are typically by-products from a
The passive oxide film is cathodic to steel, and, when the wide range of processes, most typically processing of agricul-
passive film is penetrated by mechanical damage or by dis- tural biomass. As such, these materials may contain a wide
solution or there is insufficient inhibitor for the entire metal range of constituents, including amines, phosphates, hetro-
surface, the exposed metal becomes a small anodic area in the cyclic nitrogen compounds, sulfur compounds and numer-
middle of a large cathodic area of passive film. Rapid local- ous natural compounds such as proteins, plant extracts,
ized corrosion can then occur in the anodic area. Steel can be phytic acid (inositol hexaphosphoric acid), rice bran, soybean
easily passivated in any aqueous solution except those that cake, and grapeseed oil (195).
contain easily oxidized substances in solution or high con-
centrations of chloride ions.
4.1.9.4 Phosphorous-Based Inhibitors
Snow and ice control chemicals with anodic inhibitors can
show low corrosion rates in laboratory corrosion tests where Numerous inhibited proprietary deicing chemicals contain
the anodic inhibitor is at the proper concentration, but deic- relatively high levels of phosphorous. In most cases, it is
ing chemicals applied to roadways become diluted, which can unknown whether the phosphorous is present as a phosphate,
result in severe localized corrosion of motor vehicles and polyphosphate, or phosphonate. The form of phosphorous is
roadside infrastructure. extremely important because phosphates are anodic or dan-
gerous inhibitors while polyphosphates and phosphonates
are cathodic or safe inhibitors (174, 190, 201).
4.1.9.2 Cathodic Inhibitors
Generally, cathodic inhibitors are less effective than anodic
4.1.10 Discussion of Atmospheric Corrosion
inhibitors (174). Cathodic inhibitors reduce corrosion pri-
of Snow and Ice Control Chemicals
marily by interfering with or preventing the oxygen reduction
reaction. Cathodic inhibitors are considered to be safe The results of snow and ice control chemical corrosion
inhibitors because they decrease general corrosion but do not tests depend heavily on the corrosion test methods used and
stimulate pitting or localized attack. Precipitating inhibitors the metals tested. In addition, no corrosion inhibitors are
are cathodic inhibitors. They cause insoluble films to form on suitable for all metals and all snow and ice control chemicals.

76
These facts have led to numerous conflicting ranking results detrimental forms of corrosion are localized (i.e., poultice
for snow and ice control chemicals. For example and crevice) corrosion. Poultice corrosion occurs under
deposits and can be significant, but the most serious and
1. There is general agreement that calcium magnesium detrimental is crevice corrosion of structural steel riveted
acetate (CMA) is less corrosive to steel components than and bolted joints. Crevice corrosion in riveted and bolted
sodium chloride (185, 194, 196, 199, 204–208) but there is joints causes corrosion product wedging or rust jacking
some disagreement that it is less corrosive to aluminum which can cause the rivets and bolts to “pop” (fail cata-
components (185, 199). strophically because of axial tensile overload). Once crevice
2. Some work has shown magnesium chloride to be less cor- corrosion has initiated in a structural steel joint, it is almost
rosive than sodium chloride (191, 205, 206), while other impossible to prevent the corrosion from continuing, even if
work has shown it to be more corrosive (187). The varia- the structure is recoated.
tion in ranking probably is strictly because of test proce-
dures given that roadside infrastructure and vehicles will
stay wet longer when magnesium chloride is used. 4.1.10.4 Snow and Ice Control Chemicals
3. Calcium chloride and inhibited calcium chloride have The chloride-based snow and ice control chemicals are the
been shown to be less corrosive than magnesium chloride most corrosive with the hydroscopic chlorides of magnesium
and sodium chloride (197, 205, 210), have similar corro- and calcium being the most aggressive because of the longer
sion (195, 206, 211), or much more severe corrosion (198). time of wetness. The acetate-based (e.g., calcium magnesium
4. Inhibited calcium chloride is more corrosive than non- acetate) and miscellaneous organic chemicals all tend to cause
inhibited calcium chloride (210). significantly less corrosion than the chloride-based snow and
ice control chemicals, but they can still accelerate corrosion
Overall, chloride-based snow and ice control materials are because they increase the conductivity of the moisture.
corrosive to metal components. Numerous studies have
attempted to rank these materials relatively for corrosion but
complex contributing factors limited the development of 4.1.10.5 Inhibitors
definitive conclusions. From the standpoint of developing Snow and ice control chemical corrosion inhibitors that are
practical rankings, these materials have been ranked equally tested in the laboratory can show significant reductions in
high for their potential to cause corrosion. Information sug- corrosion rates of the metals being tested, but they may show
gests that acetate-based materials are less corrosive to steel little or no inhibiting effect on other metals (even those in the
than chlorides, but may lead to corrosion concerns for other same metal group) and may in effect accelerate corrosion of
alloys. While OMB materials are accepted as being non- some metals. To develop any meaningful general corrosion
corrosive to metals, limited evidence suggests that these mate- rate data for deicing chemical corrosion inhibitors, tests must
rials, used as corrosion inhibition additives, significantly be run over a wide range of concentrations for both the cor-
reduce corrosion rates under field conditions. rosion inhibitor and the deicing chemicals, and also under a
number of ambient conditions to simulate field conditions.
4.1.10.1 Motor Vehicles
The most detrimental forms of corrosion for motor vehicles 4.2 Other Impacts to Infrastructure
are localized corrosion (i.e., e.g., crevice, poultice, and pitting)
4.2.1 Railway Traffic Control Signaling
and galvanic corrosion—not uniform or general corrosion.
Concerns have been raised about the effect of snow and ice
control materials on the operation of railway traffic control
4.1.10.2 Roadside Infrastructure
signaling. Anecdotal information is available, but formal
Cosmetic corrosion is the most prevalent form of roadside studies addressing the frequency, mechanism, and magnitude
infrastructure corrosion. General or uniform corrosion also of occurrence are not. Various groups involved with railroad
occurs very frequently but usually is not a problem because of issues were contacted, including the Federal Highway
the massive metal sections involved. Administration (FHWA), the Federal Railroad Administra-
tion (FRA), and the Association of American Railroads
(AAR). Limited information beyond anecdotal accounts was
4.1.10.3 Metal Bridge Components
obtained, and inquiries were further referred to the American
Metal bridges and their components and similar struc- Railway Engineering and Maintenance-of-Way Association
tures can suffer general or uniform corrosion, but the most (AREMA). AREMA Technical Committee 36, Highway Rail

77
Grade Crossing Warning Systems, has the purpose of devel- identified that every effort should be made to minimize false
oping concepts, providing information, and recommending activations (213).
practices for the design, installation, operation, and mainte- Observed cases of snow and ice control materials affecting
nance of highway railroad grade crossing warning equipment railway traffic control signaling typically involve the use of
and systems. This group is aware of occurrences of snow and extremely high levels of chloride-based products. Other site-
ice control chemicals affecting railway traffic control devices, specific conditions can also influence the process. For illus-
but incidences are typically addressed on a case-by-case basis. tration, one example involves railroad tracks crossing a
Modern railway traffic control devices involve electric roadway at the foot of a hill. Heavy applications of roadsalts
circuit-based train detection to initiate traffic warning signals applied to maintain roadway mobility produced high vol-
and gates. Three general train detection systems are used (212): umes of salt brine that ran downhill and pooled across the
tracks. The presence of the electrically conductive solution
1. DC, AC-DC, or AFO Grade Crossing Island and between the tracks shunted the track circuit and resulted in
Approach Circuits. These train detection systems use a false signal activation.
battery or transmitter at one end of track section and a The Idaho and Missouri DOTs have conducted field trials
relay, receiver, or diode at the other end. A train on the sec- to trigger railway signaling with MgCl2 and NaCl applica-
tion of affected track will shunt the circuit and de-energize tions, but were unable to measure any changes in signal levels
the relay, further triggering operation of the signal mech- after application. It is unknown whether various snow and ice
anisms. This type of system will continue until the train control materials have higher or lower incidences or magni-
leaves the circuit. tudes of impact. However, given that false activations have
2. Motion-Sensitive Devices (MS). This train detection sys- been reported and the presumed mechanism of impact is
tem detects the presence of train movement within a given highly conductive ionic solutions affecting track circuitry,
electronic track circuit. This type of system will continue occurrences are likely linked to chloride-based products. It
until the train stops or leaves the circuit. can also be inferred that any measures taken to reduce appli-
3. Constant Warning Time (CWT) Systems. This train cations near rail crossings and pooling across the tracks can
detection system detects the speed and distance of a train minimize effects. However, measures to limit the use of ionic
entering the circuit and activates traffic control devices to (electrically conductive) snow and ice control materials at
provide a desired warning time. railway crossings are outweighed by safety concerns requiring
adequate traffic mobility across the tracks.
Each of these train detection systems relies on the transmit-
tance of electrical current through the rails for proper opera-
4.2.2 Power Distribution Lines
tion. These circuits are designed to be “fail-safe,” and any
break in continuity will trigger the most conservative event— As with railway traffic control signaling, concerns have
a traffic-signaling event. been raised about the effect of snow and ice control materials
High ionic salt solutions have low electrical resistivity (i.e., on the operation of power distribution lines. Anecdotal infor-
they can transmit electrical current through solution) and can mation is available; however, formal studies addressing the
affect track circuit electrical continuity, which can trigger sig- frequency, mechanism, and magnitude of occurrence are not.
nal operation without the presence of a train in the circuit. Salt Electrical power is most commonly distributed by wires
solutions can also cause early initiation of traffic signaling for suspended on poles or towers. Insulators, made of a fast-
CWT train detection systems. To counter these conditions, drying non-conducting material such as glass, porcelain, or
some signal maintenance engineers may increase the output of composite materials, are used to hold the wires and to mini-
the train detection systems. However, when conditions return mize current loss and grounding. Insulator failure because of
to normal, this can lead to over-energized circuits that, in the- “natural/ambient” conditions has been a concern and has
ory, can fail to detect oncoming train traffic at the proper time. resulted in loss of current, line shorting, and pole fires. A basis
Reasonable and consistent traffic warning times re-enforce description of the mechanism is as follows (214):
the railway traffic control system credibility. A major factor
affecting the credibility of these systems is an unusually high 1. Hot and dry weather conditions allow accumulation of
number of false activations at active crossings (213). Unrea- pollutants on the insulators and can dry wooden poles.
sonable or inconsistent warning times have been linked to 2. Damp conditions bring moisture, allowing the pollutants
undesirable driver behavior. Research has shown that when on the insulators to become slightly conductive.
warning times exceed 40 to 50 seconds, drivers will accept 3. Current leaks across the insulator and can cause line cur-
shorter clearance times at flashing lights, and a significant rent losses, shorting, or heating and potential fire of
number will attempt to drive around the gates. It has been wooden poles.

78
High ionic salt solutions can exacerbate the process. These are hygroscopic—these materials can combine with water
solutions have low electrical resistivity (i.e., they can readily and maintain a moist conductive coating on the insulator,
conduct electrical current through solution) and, when they even during dry conditions.
coat insulators, can allow transfer of current around the insu- The degree of aerial deposition plays a defining role in
lator to a greater degree than without salts. Incidents related whether or not effects occur. Major multi-lane and high-speed
to snow and ice control materials have included loss of cur- highway traffic can aerially disperse salt aerosols quite exten-
rent, shorting of transmission lines, and wooden pole fires sively. Kelsey and Hootman (29) estimated NaCl aerosols at
resulting from the leak of current across the insulator. least 15 meters high within 67 meters of a roadway. Vertical
All accounts of effects related to snow and ice control distribution was also found to be distributed more or less
materials involve chloride-based materials. Salts that have exponentially with height from the roadway, with most depo-
been deposited on an insulator and have dried pose little sition occurring up to 3 meters (31). Power lines subjected to
immediate concern. However, these materials can become high levels of aerial dispersion and locations where power lines
conductive when moisture is present, such as during damp are exceptionally close to the roadway are of particular con-
or foggy conditions. For this reason, it is suggested that CaCl2 cern. Extreme cases include power lines directly next to ele-
and MgCl2 pose the greatest potential concern because they vated roadways such as alongside an overpass or bridge.

79
CHAPTER 5
Analytical Methods Used to Measure

Attributes of Snow and Ice Control Materials
A snow and ice control material’s performance and 5.1 Specification Sources
potential for adverse effect can typically be defined by some
measurable attributes of the material. Much effort has been Many jurisdictions have developed internal specifications
put into developing and conducting tests to measure these and test procedures; however, these often have involved
attributes. In many cases, however, test results have not adapted existing standards. The following organizations are
been reproducible under field conditions. In other recognized as providing standard test methods developed
instances, two technically relevant test methods provide directly for testing snow and ice control materials:
conflicting conclusions. This causes confusion in the indus-
try by preventing direct product comparisons, which fur- • Pacific Northwest Snowfighters,
ther affects product selection processes and quality • Strategic Highway Research Program,
assurance efforts. • ASTM, and
Section 2.4.4 indicates many jurisdictions have purchase • AASHTO.
specifications detailing required product attributes and that
some level of quality testing is being conducted to ensure con- 5.1.1 Pacific Northwest Snowfighters (PNS)
formance with these specifications. The level of effort varies
from simple to detailed by jurisdiction. For example, many The PNS is a consortium of states and provinces (i.e.,
agencies qualify product shipments through simple hydrom- Idaho, Montana, Oregon, Washington, and British Colum-
eter measurements, whereas other groups conduct full chem- bia) that have developed a list of physical, environmental,
ical assessments for environmental attributes of concern. and corrosion-related specifications for snow and ice control
Whatever the level of effort, it is desirable that these tests be materials (43). Test methods have been defined to measure
done quickly, with minimal cost, and using standard available physical, chemical, toxicological, and corrosion attributes
test equipment. and, in many cases, use or adapt existing standard test meth-
The focus of this study is the assessment of environmental, ods. Many other jurisdictions and agencies have adopted the
not performance-related, attributes. However, there is some PNS specifications in the development of their own product
overlap of the two categories, and some benefit from the specifications and quality testing processes.
standpoint of product selection processes to discuss some
performance-related characteristics. The general areas of
5.1.2 Strategic Highway Research
snow and ice control material testing are categorized to
Program (SHRP)
include the following attributes:
SHRP, a unit of the National Research Council, was active
• Physical—includes specific gravity, freeze point, eutectic between 1988 and 1993 in evaluating and implementing
temperature, gradation, and settleability; new technologies for the nation’s transportation agencies. A
• Chemical and Environmental—includes heavy metals, “Handbook of Test Methods for Evaluating Chemical
chemical species, and toxicity; and Deicers” was prepared that contained a selection of specific
• Corrosion—atmospheric, concrete, and concrete rein- test methods for the complete characterization of chemical
forcing. deicers (215). Test methods include assessment of physio-

80
chemical properties, deicing performance, effect on metals products contain secondary constituents or consist of more
and concrete, and ecological impacts. than one base product (i.e., liquid salt plus a corrosion
inhibitor). In such cases, it is desirable that the vendor
provides a product specific gravity versus concentration curve
5.1.3 American Society for Testing and
for that product.
Materials (ASTM)
Although specific gravity is a quick and simple method for
ASTM International is a standards development organiza- providing a general idea of product concentration, direct
tion that provides a forum for the development and publication chemical analysis methods are more defensible.
of voluntary consensus standards for materials, products,
systems, and services. ASTM provides specifications and stan-
5.2.2 Freezing Point and Eutectic
dard test methods for snow and ice control materials; however,
Temperature
these tend to be limited in scope, focusing on measurement of
physical characteristics and corrosion-related test methods. ASTM D1177 describes a method (to determine the freez-
ing point of engine coolants) that can be applied to liquid
snow and ice control materials. Solutions are cooled with stir-
5.1.4 American Association of State ring while time and temperature are recorded. The freezing
Highway and Transportation point is taken as the temperature at the point of intersection
Officials (AASHTO) of temperature versus time and after the onset of freezing
AASHTO specifications for snow and ice control materials (215). Eutectic temperature is the lowest freeze point temper-
typically reference or modify ASTM procedures. ature achievable for a given product in solution (1). Eutectic
temperature, by definition, applies to pure products and in
practice (e.g., for commercially available blends) a phase
5.2 Physical Attributes curve can be developed for a range of active ingredient
concentrations. This involves measurement of the freezing
5.2.1 Specific Gravity
characteristics of a solution with a previously identified
Specific gravity is a material’s density (mass/volume) com- concentration.
pared with water. Many agencies use specific gravity to mon-
itor the percentage of the active ingredient in purchased
5.2.3 Gradation
liquid snow and ice control products. ASTM D 1429 details
various methods for specific gravity measurement for waters Methods for determining particle size distribution of solid
and brines: products typically are based on methods similar to ASTM
C 136, Sieve Analysis of Fine and Coarse Aggregates. The fol-
• Test Method A—Pycnometer: involves weighing a pyc- lowing test methods apply to snow and ice control products:
nometer (a fixed volume vessel) containing the product at
20°C ± 1°C. • ASTM D 632, Standard Specification for Sodium Chloride
• Test Method B—Balance: involves the use of a plummet to and
determine liquid weight by displacement. • ASTM D 98 and ASTM D 345, Standard Specifications for
• Test Method D—Hydrometer: involves a float-type device Calcium Chloride.
consisting of a weighted barrel and calibrated specific grav-
ity gradations. A hydrometer will float higher in a denser liq-
5.2.4 Particle Degradation
uid (i.e., one containing high solids) than a less dense liquid.
In the case of abrasives, hardness indicates resistance to
SHRP references ASTM D-1122-90a, which is a method mechanical degradation (i.e., e.g., abrasion). Harder materi-
developed for engine coolants, but is very similar to ASTM als are less prone to crushing from vehicular action, which can
D 1429, Test Method D-Hydrometer (215). reduce potential air quality concerns. Particle degradation
Agencies traditionally have compared specific gravity meas- testing is commonly referred to as “Los Angeles Abrasion”
urements to standard reference tables to determine the per- testing and standard test methods include
centage of active ingredient (i.e., CaCl2, MgCl2) in a purchased
liquid product. For example, standard reference tables state • ASTM C 131, Standard Test Method for Resistance to
that a 30-percent CaCl2 solution has a specific gravity of 1.284 Degradation Aggregate by Abrasion and Impact in the
(216). Standard reference tables should be used with caution Los Angeles Machine and
because they apply to pure products, and many available liquid • AASHTO T 96-02.

81
5.2.5 Settleability performance-based methodologies (PBMs). A PBM speci-

fies the data quality objectives (DQOs) that are to be
The PNS has developed a method to determine percent achieved but leaves the laboratory and its customer to deter-
Total Settleable Solids and percent Solids Passing a No. 10 mine the chemical and instrumental procedures by which
Sieve after a liquid chemical is stored at a specified cold tem- the DQOs are to be met. The goal of a PBM is to provide rea-
perature (17.8 ± 1°C for MgCl2 and 29 ± 1°C for CaCl2) sonable flexibility in the choice of analytical method and
without agitation for 168 hours (1 week) (43). The test instrument used. The intent of the PBM is to allow for tech-
procedure is designed to determine the potential for solids nical advances in measurement science to be implemented
formation under extreme field storage temperatures. Solids in the future and permits the analyst to modify methods to
formation in liquid products severely affects the operation deal with the realities of the sample such as matrix effects,
and performance of application equipment. interferences, and limitations in sample size. It also allows
for continuous improvement in the quality of laboratory
operations. That being said, certain benefits are realized by
5.3 Chemical and Environmental
specifying standard test methods, especially from the stand-
Attributes
point of maintaining a standardized program that involves
Many test methods applied to snow and ice control mate- product qualification and quality testing.
rials to determine chemical and environmental attributes
have been adapted from tests used for other purposes such as 5.3.1 Active Ingredient
for water and wastewater. PNS has taken leadership in
addressing environmental concerns attributed to snow and Active ingredient refers to the primary component of the
ice control materials, and many other jurisdictions have material. For example, when purchasing commercial liquid
incorporated their programs. calcium chloride, the chemical constituent, CaCl2, would be
Certain considerations are necessary when adapting water considered the active ingredient. An accurate assessment of
and wastewater methodologies to snow and ice control prod- the active ingredient is important to ensure “purchased” and
ucts. These include high specific gravity, extreme analyte con- “received” product concentrations are equivalent for contract
centrations, and the potential for chemical interferences. purposes. Many agencies also use active ingredient concen-
Standard test methods used for water and wastewater samples tration to select application rates for a desired snow and ice
are conducted on volumetrically measured sample aliquots. control performance.
Liquid products used for snow and ice control commonly Active ingredient concentrations may be estimated using
have densities approximately 15 to 35 percent greater than specific gravity charts or be directly determined through var-
water, and procedures must be modified to ensure represen- ious analytical techniques, including titrations and instru-
tative gravimetric aliquots are analyzed. Alternatively, volu- mental analysis. Measurable active ingredients include CaCl2,
metrically derived sample results (i.e., mg/L) must be MgCl2, NaCl, CMA, and KA, and can be estimated by direct
converted to a gravimetric basis after analysis. measurement of primary components, most typically the
Units of measurement commonly included in specifica- cationic species. For example, Ca would be measured for
tions and reported with test results have resulted in some CaCl2, and Ca and Mg are measured for CMA.
confusion. Ambiguous units such as parts per million Specific gravity measurements can be used to estimate
(ppm) or percent (%) should be clarified to reflect the basis product concentrations using standard reference tables when
of measurement—milligrams per liter (mg/L), milligrams near-pure products are being used. Alternatively, product
per kilogram (mg/kg), percent volume per volume (% v/v), suppliers may provide specific gravity versus concentration
percent mass per volume (% w/v) or percent mass per mass tables specific to a product.
(% w/w). Titration procedures are available for calcium and magne-
Chemical analysis of solid products typically requires dis- sium determination. ASTM E 449 involves an EDTA titration
solution of the product before analysis. The approach for CaCl2, and ASTM C 245 provides a titration procedure for
described by PNS requires a 25-percent (w/w) preparation MgCl2. These simple wet chemistry test procedures can be
with distilled or deionized water (43). Care is required to quite effective at measuring percent level concentrations as is
ensure material solubility is not exceeded when preparing common for snow and ice control products.
these solutions. Instrumental analysis procedures are more complex, but
The following assessments of standard chemical analysis provide a higher level of certainty in the measurement.
methods are not intended to be proscriptive. In many cases, Ca, Mg, Na, and K are metals that can be detected by vari-
more than one approach could provide acceptable test data. ous instrumental procedures including atomic absorption
Trends in analytical chemistry are moving toward spectroscopy, atomic emission spectroscopy, and mass

82
spectroscopy (refer to Section 5.3.2). PNS suggests atomic flame—thereby providing a low result. This is particularly
absorption (AA) spectroscopy (using lanthanum chloride true for Ca or Mg determinations when phosphates are
ionization suppression) or inductively coupled plasma (ICP) present. This is overcome by adding an ionization suppres-
spectroscopy when determining active ingredient concentra- sant such as lanthanum chloride (LaCl3) to samples, which
tions, but further states that product compliance or price binds the phosphate, releasing the Ca or Mg for detection.
adjustments will use the AA techniques (43). Detailed knowl- The PNS method for the determination of Ca and Mg by FAA
edge of interferences should be known, and special consider- references the use of lanthanum chloride when determining
ations should be taken to address extremely high solids Ca and Mg by atomic absorption (43).
concentration (i.e., sample dilutions).
5.3.2.2 ICP-AES
5.3.2 Metals
AES is the process in which the light emitted by excited
Many snow and ice control products are obtained from atoms or ions is measured. The emission occurs when suffi-
natural sources such as deposits or wells or are agriculturally cient thermal or electrical energy is available to exit a free
derived. In addition to primary constituents, other impuri- atom or ion to an unstable energy state. The energy is created
ties including heavy metals are often present. The term heavy through an ICP (applying an RF field to argon gas) that can
metals typically refers to any metallic chemical element that reach temperatures up to 10,000 °K. Light is emitted when the
has a relatively high density and is toxic, highly toxic, or poi- atom or ion returns to the more stable configuration in the
sonous at low concentrations. Heavy metals are defined by ground state. The emitted wavelengths are specific to a
PNS (and generally accepted) to include arsenic, barium, particular element and their intensity is proportional to the
cadmium, chromium, copper, lead, mercury, selenium, and concentration.
zinc (43). ICP-AES is a well-documented technique with multi-
Analysis requires minimum levels of detection in the element analytical capabilities and a suitable working range
range of parts per million (mg/L or mg/kg) and thus and levels of detection—parts per billion detection is
requires more sophisticated instrumental analysis. The possible for some elements (217). Operator skill levels are
assessment of low analytical levels poses certain challenges, intermediate between FAA and GFAA.
specifically when these elements are incorporated in a
matrix containing percent levels of an active ingredient.
5.3.2.3 ICP-MS
Instrumental techniques include AA spectroscopy, ICP-
atomic emission spectroscopy (ICP-AES), and ICP mass ICP-MS is the synergistic combination of an ICP ion gen-
spectroscopy (ICP-MS). eration and a quadrupole mass spectrometer. Ions are quan-
tified based on their mass/charge ratio. Ions of the selected
mass/charge ratio are directed at a detector that quantifies the
5.3.2.1 AA Spectroscopy (FAA and GFAA)
number of ions, further providing a sample concentration.
AA spectroscopy occurs when a ground state atom absorbs Applications are well documented and continue to grow
energy in the form of a specific wavelength and is elevated to rapidly. ICP-MS requires operator skill levels similar to those
an excited state. The amount of light absorbed at this wave- for ICP and GFAA.
length is proportional to the concentration of atoms of the ICP-MS combines multi-element detection capabilities
selected element. Either a flame (Flame Atomic Absorption with exceptionally low detection limits—parts per trillion
(FAA)) or electrothermal vaporization using a graphite fur- (ppt) levels are possible for some elements (217). An impor-
nace (Graphite Furnace Atomic Absorption (GFAA)) may be tant limitation is that samples with higher solids levels (i.e.,
employed to elevate the atom to an excited state. 0.2 percent) cannot be tolerated. However, given the
FAA is relatively simple to use and extensive applications extremely low levels of detection, samples may be diluted and
including those for snow and ice control products are avail- still provide results well below regulatory limits.
able. Excellent precision makes it a preferred technique for the
determination of major constituents (i.e., Ca, Mg, Na, and K)
5.3.2.4 Method Selection
and higher concentration analytes. GFAA techniques are well
documented, although not as completely as FAA techniques. Each of the techniques listed above for metals determina-
GFAA has exceptional detection limit capabilities, but a lim- tions is technically suitable for the measurement of metals
ited analytical working range. (and active ingredients) in snow and ice control products. A
Some metal ions in samples are present as strong com- skilled operator with knowledge of potential analytical
plexes—they are not easily decomposed to atoms in the interferences and methods of compensation is essential in

83
ensuring that performance-based objectives are met. The to as total Kjeldahl nitrogen (TKN). Nitrogen compounds are
main selection criteria include commonly reported as the nitrogen equivalent.
Test methods for determination of nitrogen compounds
• Detection limits required, can be referenced in APHA Standard Methods (67), and var-
• Equipment availability, and ious options are available for each nitrogen compound. It is
• Ability to meet desired quality objectives. the laboratory’s responsibility to ensure that the selected
methods meet accepted performance criteria for application
to snow and ice control materials.
5.3.3 Total Phosphorus
As discussed in Section 3.4.2.3, total nitrogen should be
Phosphorus can exist in various forms including the main index by which nitrogen content is evaluated. A
orthophosphate, which is readily soluble, and total phospho- separate evaluation of ammonia and nitrate content is war-
rus, which includes bound/combined phosphorus and ranted because ammonia is potentially toxic to aquatic life at
requires more aggressive analytical techniques such as hydrol- high concentrations, and nitrate is limited for drinking water
ysis using heat and acid before determination. Orthophos- supplies.
phates are thought to be more environmentally available,
although total phosphorus can degrade to orthophosphates.
5.3.5 Organic Matter (BOD and COD)
It is important when determining phosphate concentra-
tions that a total determination be conducted. PNS (43) ref- Biochemical oxygen demand (BOD) is an empirical test
erences APHA Standard Methods for the Examination of used to determine the relative oxygen requirements of a sam-
Water and Wastewater (67), which includes hydrolysis ple and provides indication of biodegradable organic matter.
methods employing perchloric acid, sulfuric/nitric acids, or The standard 5-Day BOD test has fixed incubation conditions
persulfate. of 20 ± 2°C for 5 days (67). It may be desirable in some cases
PNS methodologies include a dilution step before analysis to conduct the BOD test over longer periods of time and at
involving preparation of a 1-percent (v/v) test solution various temperatures to reflect biodegradability under alter-
(10 mL of liquid sample is acidified and diluted to 1,000 mL nate (i.e., field) conditions. The 5-day BOD test procedure
with ASTM D 1193 Type II water) before phosphorus deter- referenced in APHA Standard Methods (67) has widespread
mination (43). Analytical results are reported as mg/L of acceptance and is commonly conducted at most environ-
phosphorus on the 1-percent solution. Unless a thorough mental laboratories. PNS references this method as well (43).
understanding of the dilution procedure and the sample spe- Chemical oxygen demand (COD) is used to measure the
cific gravity is known, it is difficult to relate the determined oxygen equivalent of the organic matter content of a sample
result to phosphorus concentration in the snow and ice con- that is susceptible to a strong chemical oxidant (67). COD can
trol product. be conducted relatively quickly and can be empirically related
Section 3.4.2.3 discusses how the environmental charac- to BOD. COD values should always be equal to or greater than
teristics of snow and ice control materials should be evaluated BOD. The COD test procedures referenced in APHA Stan-
primarily in terms of total phosphorus, rather than specific dard Methods (67) have widespread acceptance and are com-
phosphorus fractions. monly conducted at most environmental laboratories. PNS
references these methods as well (43).
5.3.4 Nitrogen
5.3.6 pH
Forms of nitrogen present in snow and ice control prod-
ucts include pH is the negative log of the hydrogen ion concentration in
solution and represents the intensity of the acidic (low pH) or
• Nitrite (NO2 ), basic (high pH) character of a solution. Standard test meth-
• Nitrate (NO3 ), ods involve the use of a calibrated electronic pH meter. These
• Ammonia (NH3), devices determine hydrogen ion activity by potentiometric
• Organic Nitrogen (TON), measurement using a standard hydrogen ion electrode and a
• Total Kjeldahl nitrogen (TKN), and reference electrode. pH measurement is relatively free from
• Total Nitrogen (TN). interferences from color, turbidity, colloidal matter, oxidants,
reductants, or high salinity, except for sodium error at
Nitrite and nitrate are commonly determined together and pH > 10 (67). Most standard methods for pH measurements
are referred to as “nitrate and nitrite.” Ammonia and organic of liquids are relatively similar. Two examples include ASTM
nitrogen are commonly determined together and are referred D 1293 and those listed in APHA Standard Methods.

84
PNS (43) references ASTM D 1293 for the measurement of • LC50, or the median lethal concentration, is an estimate of
pH, but requires a sample dilution of one part chemical and the sample concentration that causes 50-percent mortality
four parts distilled water before measurement. of the test organism during the specified period.
• The IC25 or IC50, or the inhibition concentration, is an esti-
5.3.7 Cyanide mate of the sample concentration that causes a 25-percent
or 50-percent reduction in growth or reproduction during
Cyanide (CN) compounds exist as simple cyanides, where a specified period.
the CN ion is soluble and biologically available, or as com-
plexed cyanide compounds, where the CN ion is sparingly Test duration designates whether the measured endpoint
or almost insoluble and not biologically available. Methods of applies to acute or chronic toxicological effects. The U.S. EPA
analysis typically involve a digestion process (using acids, and Environment Canada recognize the 7-day fathead min-
reducing agents, and/or ultraviolet light) to convert cyanide now test as a chronic bioassay for freshwater fish whereas the
compounds into cyanide gas, which is distilled and adsorbed 96-hour test is recognized as an acute test. It has been judged
into a basic (NaOH) solution, from which a CN determina- that the 7-day test is sufficient in length to evaluate chronic
tion can be made. toxicity. In addition, this test assesses the life stage of the rep-
Test methods to determine cyanide can be referenced in resentative fish considered most sensitive to adverse effects
APHA Standard Methods (67). The chemistry of cyanide from chemicals and provides endpoints for growth and
analysis is complex and subject to numerous interferences. It survival. By conducting a 7-day test, one can determine the
is the laboratory’s responsibility to ensure that the selected 96-hr LC50 and 7-day LC50 as well as a 7-day EC50 on growth,
methods meet accepted performance criteria for application in addition to other endpoints, including LC25, EC25, lowest
to snow and ice control materials. observable effect concentration (LOEC) and no observable
effect concentration (NOEC).
The assays shown in Table 5-1 provide chronic toxicity
5.3.8 Aquatic Toxicity
endpoints for fish, invertebrates, and algae. These tests are
The literature provides generalized protocols for conduct- recognized as chronic tests by both Environment Canada and
ing ecological toxicity tests of various chemicals and materi- the U.S. EPA and are commonly used to examine sub-lethal
als. Methods specific to snow and ice control materials do not effects to aquatic organisms in programs such as Environ-
exist. Both Environment Canada and the U.S. EPA detail ment Canada’s Environmental Effects Monitoring, and the
methods for assessing toxicity to aquatic organisms (see U.S. EPA’s NPDES.
Table 5-1). These procedures involve exposure of a test organ- The fathead minnow chronic test is an early life stage test
ism to various material concentrations under controlled that measures sub-lethal effects at the life stage of the organism
conditions for a specified period. Based on the observed considered most sensitive to adverse effects from chemicals.
effects, a toxicity test endpoint value can be derived as follows: The Ceriodaphnia dubia test represents a lifecycle test where
Table 5-1. Freshwater toxicity test methods.
Test Method Standard Reference Method

Fathead Minnow (or Rainbow Trout) US Environmental Protection Agency, 1993, Methods for Measuring
96-hour Survival Test the Acute Toxicity of Effluents and Receiving Water to Freshwater and
Marine Organisms, 4th Edition, EPA/600/4-90/027F
Fathead Minnow 7-day Survival and US Environmental Protection Agency, 1994, Short-Tern Methods for
Growth Test Estimating the Chronic Toxicity of Effluents and Receiving Water to
Freshwater Organisms, 3rd Edition, EPA/600/4-91/002
Ceriodaphnia dubia 3-brood Survival US Environmental Protection Agency, 1994, Short-Tern Methods for
and Reproduction Test Estimating the Chronic Toxicity of Effluents and Receiving Water to
Selenastrum capricornutum 72-hour Environment Canada 1992/1997 amendments. Biological Test Method:
(Environment Canada) Growth Test Growth Inhibition Test Using the Freshwater Alga Selenastrum
capricornutum Report EPS1/RM/25.
or
Selenastrum capricornutum 96-hour US Environmental Protection Agency, 1994, Short-Term Methods for
(US EPA) Growth Test Estimating the Chronic Toxicity of Effluents and Receiving Water to

85
three reproductive cycles are represented. This test measures Although this test provides a fairly quick assessment of the
effects on survival and reproduction to aquatic invertebrates. freezing and thawing resistance of the concrete specimen, it is
The algae test is a chronic test that measures effects on growth of limited value in evaluating snow and ice control materials.
and survival to aquatic plants over several life cycles. Several issues have to be considered when evaluating these
The tests proposed involve representative organisms from materials that are not addressed in this test method.
three trophic levels in the aquatic environment—vertebrates, The concrete mix does not contain any air entrainment,
invertebrates, and plants. These three tests should be suffi- making it naturally susceptible to freezing and thawing dam-
cient to examine the potential for adverse effects to a range of age. Thus, most damage is most likely caused by the freezing
organisms in aqueous environments. and thawing action. The damage caused by the deicing mate-
rial is more likely a function of how well it can protect the
sample from freezing, rather than a result of the aggressive-
5.4 Evaluation of Concrete-Related
ness of the material. This may explain why some materials
Test Methods
perform better than others. Regardless of the deicing material
Several standard test procedures have been developed to eval- used, the small samples probably will freeze completely dur-
uate the performance of reinforced concrete exposed to snow ing the long exposure at 17.8oC.
and ice control materials. The test methods applicable to con- The concrete mix proportions use an unrealistically high
crete and reinforcing steel are outlined in SHRP-H-332, Hand- w/c ratio of 0.51, resulting in a relatively poor-quality con-
book of Test Methods for Evaluating Chemical Deicers, as follows: crete compared with that in highway structures. For concrete
exposed to freezing and thawing as well as snow and ice con-
• SHRP H-205.8, Test Method for Rapid Evaluation of trol materials, a much lower w/c ratio is typically used or even
Effects of Deicing Chemicals on Concrete; required by applicable standards.
• SHRP H-205.9, Test Method for Evaluation of Scaling Considering the long freeze cycles with the short thaw
Effects of Deicing Chemicals on Concrete Surfaces; and cycles, it is very unlikely that any chemical attack of the con-
• SHRP H-205.12, Test Method for Evaluation of Corrosive crete matrix would be significant. Therefore, the relatively low
Effects of Deicing Chemicals on Steel Reinforcement in number of cycles cannot be used to estimate the long-term
Concrete effects of a snow and ice control material. Chemical attacks
will take much longer to cause damage. In addition, these
These test methods are typically based on ASTM standards. reactions will continue during the warmer temperature times
The individual test procedures are discussed below. of the year after the snow and ice control material has per-
meated the concrete.
Overall, this test method is a test for freezing and thawing
5.4.1 SHRP H-205.8, Test Method for Rapid
resistance of the concrete. To obtain realistic results for the
Evaluation of Effects of Deicing
effects of snow and ice control materials, more representative
Chemicals on Concrete
conditions should be considered. The use of air-entrained
The objective of this test method is the evaluation of the concrete is now the standard for concrete exposed to
effects of snow and ice control materials and freezing and freeze/thaw conditions. The number of cycles should be
thawing cycling on the structural integrity of concrete. This increased, and the thaw periods extended. Considering that
test method uses small cylindrical concrete samples—1.5 freeze/thaw resistance is not the property that is to be evalu-
inches in diameter and 1.875 inches in length. ated, the use of higher freezing temperatures and shorter
The test specifies the concrete mix proportioning, casting, freeze cycles may also be considered.
and curing of the samples. No air entrainment is used in the
concrete. The samples are placed on sponges in a sealed con-
5.4.2 SHRP H-205.9, Test Method for
tainer exposed to the specified snow and ice control test solu-
Evaluation of Scaling Effects of Deicing
tion. The samples are subjected to ten freezing and thawing
Chemicals on Concrete Surfaces
cycles by placing them in a freezer at –17.8 oC for 16 to 18 hours
followed by thawing for 6 to 8 hours. The temperature and time This test method is used to evaluate scaling effects on hor-
parameters are as in ASTM C 672-98, Standard Test Method for izontal concrete surfaces exposed to freezing and thawing
Scaling Resistance of Concrete Surfaces Exposed to Deicing cycles in the presence of deicing chemicals. The effects are
Chemicals. The condition of each sample is evaluated by dry- evaluated qualitatively by visual examination and quantita-
ing and weighing the largest intact part of each specimen after tively by weight loss. Concrete samples used in this test are
deteriorated concrete has been removed. The test method used 152 mm 305 mm 76 mm in size and are contained in a
deionized water and NaCl solutions as control conditions. well holding the test solution. Similar to SHRP H-205.8, the

86
mix proportions, casting, and curing are specified. The con- Exposed to Chloride Environments, and involves the ponding of
crete does not contain air entrainment. It is recommended mortar or concrete prisms (280 mm 152 mm 114 mm)
that deionized water and NaCl solution be used as control with a deicing solution.
samples. The same concerns for air entrainment and w/c ratio One deformed #4 bar (12.7-mm diameter) is embedded 19
arise as for SHRP H-205.8. mm from the top surface and two bars are embedded 25 mm
After the samples have been cured, they are ponded with from the bottom of the prism. The top and bottom bars are
deicer solution and exposed to the same conditions as in connected with a resistor to form a macrocell, across which
SHRP H-205.8. The samples are exposed to five freezing and the potential difference can be measured. The macrocell cur-
thawing cycles and evaluated. Scaled material is washed and rent between the top and bottom bars can be calculated using
scraped from the sample surface, filtered, dried, and Ohm’s law. The samples are ponded with solution 3.8 mm in
weighed. The surface is visually examined and rated in depth, and the voltage is measured at monthly intervals. As
accordance with ASTM C 672-98, Standard Test Method for chlorides penetrate the sample and reach the top bar, it will
Scaling Resistance of Concrete Surfaces Exposed to Deicing begin to corrode, acting as an anode whereas the bottom bars
Chemicals as follows: then act as cathodes. The magnitude of the macrocell current
provides information about the corrosion rate of the top bar.
Rating Condition of surface This test measures two properties: the permeability of the
concrete to chloride ions, and corrosion rate. Using the same
0 No scaling
concrete properties, one can evaluate the diffusion properties
1 Very slight scaling (3.2 mm) depth,
and corrosiveness of different snow and ice control materials.
2 Slight to moderate scaling
Depending on the quality of the concrete, time to corro-
3 Moderate scaling (some aggregate visible)
sion initiation may be fairly long. However, long-term tests
4 Moderate to severe scaling
typically provide the most meaningful results. Installation of
5 Severe scaling (coarse aggregate visible over entire
bars at various depths may provide additional information
surface)
about the diffusion of chloride ions. The relative long-term
exposure also allows the evaluation of the effects of snow and
After evaluation, the samples can be subjected to subsequent
ice control materials on the concrete matrix—possible chem-
cycles, with evaluations being carried out after each set of five
ical reactions between the chemical and the cement and/or
freeze/thaw cycles.
aggregates.
With the exception of using non-air-entrained concrete,
As discussed for SHRP H-205.9, ponding tests are limited
this test is more realistic in evaluating various snow and ice
in their applicability to real structures. Spray application of
control materials. Considering that ASTM C 672-98 allows
the solution may be a more realistic test procedure than
for the use of air-entrained concrete, it would be more realis-
ponding.
tic for comparison with real structures. Also, the use of air-
entrainment would eliminate additional deterioration as a
result of freeze/thaw action. 5.4.4 Summary of Concrete-Related
To obtain meaningful results, the test should be carried for Test Methods
as many freeze/thaw cycles as possible. Some deterioration
may require several years to establish. Longer thaw cycles 5.4.4.1 General
should be considered because chemical reactions and ion dif-
As for most laboratory test procedures, there are limita-
fusion are favored by higher temperature levels.
tions in these test standards when making predictions for field
Another problem with ponding tests is the exclusion of
applications. It is difficult to make service life predictions for
oxygen from the concrete surface. Most real structures are
20, 30, 50, or even 100 years on the basis of several days or
subjected not only to freezing and thawing but also to wetting
months of laboratory testing. It is well known that chlorides
and drying.
penetrate concrete, but it is more difficult to predict the rate
at which this occurs. Changes in concrete technology and
snow and ice control systems will influence the service life of
5.4.3 SHRP H-205.12, Test Method for
structures.
Evaluation of Corrosive Effects of
It may take several years to make conclusive statements
Deicing Chemicals on Steel
about the effects of a particular snow and ice control material
Reinforcement in Concrete
from field use. To allow extrapolation from laboratory data to
This test method is based on ASTM G 109-99, Standard Test long-term field performance, one has to attempt to develop
Method for Determining the Effects of Chemical Admixtures on tests that resemble real application and climate conditions.
the Corrosion of Embedded Steel Reinforcement in Concrete Simply ponding samples is not sufficient. Accelerated tests

87
can be considered if exposure conditions are realistic and only There is also parallel research into test procedures being
when the number of cycles is increased to represent real- sponsored by AASHTO:
world conditions. Any long-term study should be comple-
mented by full-scale field studies if at all possible. • TP64-2003, Provisional Standard Test Method for Predicting
Considering the advancement in concrete technology, the Chloride Penetration of Hydraulic Cement Concrete by the
SHRP and other related test methods need to be modified to Rapid Migration Procedure and
incorporate mix designs developed for the intended ice con- • TP26-94, Method for the Relative Permeability of Concrete
trol chemical exposure. Air entrainment and SCMs are com- by Surface Air Flow.
monly used, and test methods must accommodate those.
Ideally, test samples should be prepared using the concrete A few highway departments are already using some of
actually being used in a particular region rather than the these tests. They will be contacted in Phase II about their
lower quality concrete now designated in some of the test pro- experience.
cedures. For example, the effects of SCMs in improving con-
crete durability generally and reduction in ion diffusion 5.5 Evaluation of Atmospheric
specifically here are well-documented (138). Corrosion Testing
The study team believes that SHRP H-205.8 is of limited
Laboratory corrosion tests typically are used to attempt to
value for the purposes of this study. It measures freezing and
predict the corrosion behaviors of metals versus various cor-
thawing scaling resistance only. SHRP H-205-12 is the most
rodants when service history is lacking or unavailable and
relevant test for the effect of snow and ice control chemicals time, budget, or suitable field test site constraints do not allow
on concrete itself and, more important, concrete rebar corro- field or test track testing. Experience has shown that all met-
sion. It has the disadvantages of long test times and incom- als and alloys do not respond alike to the many factors that
plete representation of field exposure. control corrosion and that “accelerated” corrosion tests give
indicative results only or may even be entirely misleading
5.4.4.2 Test Methods Under Development (218, 219).
Laboratory corrosion tests are particularly useful for
The study team is aware that substantial relevant research comparing materials and environments. They allow the
work is under way. Much of that work relates to the transport application of quality control measures and can provide
phenomena of ions into concrete. One problem with this insight into corrosion mechanisms. Corrosion under natu-
work, when considering ions other than chlorides, is that ral conditions can take long periods to materialize; thus,
there are no standard test methods for profiling the diffusion most laboratory corrosion tests are designed to accelerate
of such ions with depth of penetration into concrete. the corrosion processes. Test results must therefore be used
Following are some ASTM test methods in advanced stages with some caution because they may bear little or no rela-
of development or recently introduced: tionship to field use (196, 208, 215, 218, 220). Corrosion test
results typically depend on both the test method used and
• C 1556, Standard Test Method for Determining the Apparent the subject metals tested. Results can vary dramatically from
Chloride Coefficient of Cementitious Mixtures by Bulk Dif- test to test and thus are not comparable (175, 196, 202, 206,
fusion. This method provides the diffusion coefficient 208). Standardized laboratory corrosion tests are useful
techniques to compare various chemical compounds such as
needed for modeling and will be used in Phase II to provide
inhibitors or snow and ice control chemicals, but typically
relative diffusion values for the various ice control chemi-
they do not predict performance in real-world conditions
cals.
(191, 218, 220). Laboratory corrosion tests can basically be
• Standard Method for Indication of Concrete’s Ionic Conduc-
broken down into five general classes: immersion testing,
tivity. This test uses the RCP equipment to determine con- poultice testing, salt-spray testing, electrochemical testing,
ductivity, which is a fundamental part of corrosion. Such and test track testing.
corrosion current flow is greatly influenced by the presence
of diffused compounds and is again relevant to the detri-
mental action of ice control chemicals. 5.5.1 Immersion Testing
• Field Rate of Absorption. This test is similar to the ISAT test Immersion testing, probably the most common corrosion
used in Europe. It may find value here given its ability to test method used, involves immersion of the sample within a
quickly determine, in situ and non-destructively, the sus- liquid. There can be numerous variations in the test condi-
ceptibility of an existing concrete to ion penetration. tions such as temperature, agitation rate, and atmosphere.
• Rate of Sorptivity. Sorptivity is the rate of absorption of Immersion tests can entail the following (174, 187, 188,
liquids. 195–198, 206, 208, 215, 218, 221-229):

88
• Total immersion (specimen completely immersed in the that the test is conducted at a single inhibitor concentration
solution); and is an immersion test. The PNS test is not suitable for
• Alternate immersion (specimen immersed in the solution evaluating the effect of chemicals on metals (e.g., aluminum,
for a period and then removed from the solution and magnesium, or stainless steel) (187) or galvanic corrosion.
allowed to dry);
• Partial immersion (lower portion of the specimen is
immersed in the solution with the upper half in air or 5.5.1.2 SAE J2334 – Cosmetic Corrosion Lab Test
vapor); and This corrosion test method was developed for determin-
• Numerous variations.
ing cosmetic corrosion performance of coating systems,
substrate, process, or design (187, 230, 231). It can be used
All immersion testing can normally be carried out with as a validation tool as well as a development tool. The test
oxygen contact or with inert gas to exclude oxygen contact. uses ten standard steel sheet materials painted with a full
Immersion of test specimens in a corrosive medium automotive paint system consistent with standard practices
appears to be a simple test, but, to ensure reproducibility of used in the automotive industry. The test can be performed
test results, numerous test conditions must be controlled either manually or automatically, depending on the equip-
(e.g., temperature, humidity, chemical concentration and vol- ment available. Testing for 80 cycles by either method repro-
ume, time, surface area, flow rates, and number of samples). duces results obtained in the field after 5 years of vehicle
Immersion corrosion tests basically all follow or are mod- exposure in the Snow Belt area of the United States and
ifications of ASTM G 31, “Standard Practice for Laboratory Canada.
Immersion Corrosion Testing of Metals” (218), or NACE The test procedure consists of an application of a salt
Standard TM-01-69, “Laboratory Corrosion Testing of Met- solution (0.5 percent NaCl, 0.1 percent CaCl2, and 0.075
als for the Process Industries” (219). These standard prac- percent NaHCO3) by dip, fog, or spray for 15 minutes at
tices describe accepted procedures for and factors that ambient conditions; a drying period of 17 hours and 45
influence laboratory immersion corrosion tests, particularly minutes at 60°C and 50-percent relative humidity; and then
mass loss tests. The factors include specimen preparation, a 6-hour exposure at 50°C with a 100-percent relative
apparatus, test conditions, methods of cleaning specimens, humidity. For automatic operation, this is repeated daily for
evaluation of results, and calculation and reporting of cor- 80 cycles; for manual operation, the samples are allowed to
rosion rates. These practices also emphasize the importance dry at 60°C and 50-percent relative humidity on weekends
of recording all pertinent data and provide a checklist for and holidays. This test method could be classed as either an
reporting test data. ASTM G 31 is based on the NACE stan- intermittent immersion test or an intermittent salt-spray or
dard with modifications to relate more directly to practices. fog test, depending on the method of salt application (i.e.,
Specific corrosion testing protocols are described in the dip, fog, or spray).
following sections. The tests are monitored with weight loss bare steel
coupons, and the coated samples are evaluated by scribe creep
back values or corrosion rate (weight loss) measurements.
5.5.1.1 PNS (NACE Standard TM-01-69 as Modified
This test procedure is a field-correlated test for motor vehicle
by the Pacific Northwest States)
coating systems on steel substructures. It has an excellent cor-
This procedure is an intermittent immersion test used to relation with results obtained in 5-year tests on road-driven
evaluate snow and ice control chemicals for qualification cars in the Snow Belt area of the United States and Canada.
(43). The PNS test uses 1/2-inch flat steel washers for each If corrosion mechanisms other than cosmetic or general
deicer test solution. In each test, a single washer is exposed to corrosion are to be examined using this test, field correlation
the 3-percent snow and ice control chemical solution, 3-per- must be established.
cent sodium chloride solution, and distilled water. The test
consists of 72 cycles of intermittent exposure with 10 min-
5.5.1.3 SHRP H-205.7 “Test Method for Evaluation
utes in the liquid and 50 minutes air drying per cycle. For PNS
of Corrosive Effects of Deicing Chemicals
acceptance, the chemical must have a corrosion weight loss of
on Metals”
at least 70 percent less than the sodium chloride standard
solution. The PNS test has shown significant variability— This SHRP test method was developed to evaluate the
probably because of the lack of control of test conditions corrosive effects of deicing chemicals on metal substrates
such as temperature and humidity. It is unlikely that the PNS (215, 232). This test method was also intended to evaluate
test shows any direct correlation with field exposure, given the effectiveness of corrosion-inhibiting additives to snow

89
and ice control chemicals and the effects of other aqueous, of these have been revised to conform with most of the
near-neutral pH solutions. The test method is a weight loss details of the ASTM specifications, but they still incorporate
immersion corrosion test with air aspiration into the several statements relating to practices that experience has
solution for deicing chemicals and inhibitors using deion- shown to be desirable or beneficial for achieving reliable,
ized water and 3-percent sodium chloride master solutions reproducible results and maximum correlation among labo-
as a comparison of the relative effectiveness. It is specified ratories (174). The following are typical salt-spray (fog) cor-
as a matrix test with duplicate exposures ranging from 1 to rosion test procedures. Most are similar, but have different
8 weeks and inhibitor concentration, less than optimum, operating parameters.
optimum, and greater than optimum. It gives no guide-
lines as to the acceptable corrosion behavior for the tested
materials. 5.5.3.1 ASTM B 117 “Standard Method of
Salt-Spray (Fog) Testing” (224)
5.5.2 Poultice Corrosion Testing This is the most commonly used salt-spray (fog) testing for
inorganic and organic coatings. It is not suitable for testing
This poultice corrosion test was devised to simulate poul- metallic (e.g., zinc) coatings (235, 236) or localized corrosion
tice and crevice corrosion on automobiles and the results are (e.g., crevice and poultice corrosion). The test conditions are
reported in “Development of Poultice Corrosion Tests for Auto- as follows:
mobiles” (198). The poultice corrosion testing procedure
consists of two types of test methods: (1) partially inserting • Exposure time—8 to 3,000 hours, depending on product
rectangular sheet metal samples into washed sand doped or type of coating;
with various deicing chemical solutions; or (2) forming • Test media—5-percent sodium chloride solution with a
U-shaped channels with sealed ends from metal sheets and pH of 6.5 to 7.2; and
then filling the channels with washed sand doped with vari- • Test temperature—35°C (95°F) in the exposure zone of the
ous deicing chemical solutions. The test samples are exposed closed cabinet.
to a cyclic testing procedure that consists of a 22-hour expo-
sure in a humidity chamber at 95-percent relative humidity
and 40°C followed by a 2-hour exposure at ambient condi- 5.5.3.2 ASTM G 85 “Standard Practice for Modified
tions. The test samples are evaluated for corrosion after var- Salt-Spray (Fog) Testing Annex 1 – Acetic
ious exposure times (up to 16 weeks), by weight loss Acid – Salt-Spray (Fog) Testing” (226)
measurements and measured thickness loss. Given that the
corrosion tends to be highly localized at the air-sand inter- This test is also used for testing inorganic and organic coat-
face, the measured thickness loss is indicative of localized ings, but is particularly applicable to the study or testing of
corrosion rates such as would occur from poultice corrosion decorative chromium plate and cadmium plate on steel or
or crevice corrosion. zinc die castings and for the evaluation of the quality of a
product. Typical test conditions are as follows:
5.5.3 Salt-Spray Testing • Exposure time—can be as brief as 16 hours, but is typically

144 to 240 hours or longer;
Salt-spray testing has been used for approximately 100 years
• Test media—5-percent sodium chloride with pH adjusted
as an accelerated corrosion test to determine the comparative
corrosion resistance of various metals and alloys and the corro- to 3.1 to 3.3 by adding acetic acid; and
• Test temperature—35°C (95°F) in the exposure zone of the
sion protection of organic and inorganic coatings on metallic
substrates. There are numerous variations (174, 181, 186, 187, closed cabinet.
207, 221, 222, 224–226, 233–235). The principal objective of
salt-spray (fog) testing is to provide a rapid procedure for com-
5.5.3.3 ASTM B 368 “Standard Practice for Copper-
paring the performance of materials and coatings. Salt-spray
Accelerated Acetic Acid – Salt-Spray (Fog)
testing is an acceptable tool for evaluating the uniformity of
Testing (CASS Test)” (225)
thickness and degree of porosity of metallic (e.g., zinc) and non-
metallic protective coatings. In most, if not all cases, there is lit- This test is primarily used for the rapid testing of decora-
tle correlation with service performance. tive plating on steel and zinc die castings, but is also useful in
Many government agencies and automotive companies the testing of anodized, chromated, or phosphated alu-
have developed their own standards and procedures. Many minum. The test conditions are as follows:

90
• Exposure time—ranges from 6 to 720 hours; motor vehicles to controlled humidity, soaking and drying
• Test media—5-percent sodium chloride with addition of (corrosion booth), salt mist or spray, salt splash troughs, mud
0.26 gram per liter CuCl2⋅2H2O and pH adjusted to 3.1 to and gravel roads, grit and gravel troughs and mileage accumu-
3.3 with acetic acid; and lation (234, 238). The corrosion test results from test track tests
• Test temperature—49°C (120°F) in the exposure zone of have generally shown good correlation with customer motor
the closed cabinet. vehicles inspection and field corrosion measurements using
coupons mounted on motor vehicles (234).
5.5.4 Electrochemical Testing
5.5.6 General Comments Regarding
Corrosion is an electrochemical process and thus electro-
Laboratory Corrosion Tests
chemical techniques can be used as rapid procedures to eval-
uate and study most, if not all, forms of corrosion. Corrosion Laboratory corrosion tests are accelerated corrosion tests
requires an anode and a cathode in electrical contact with an and, therefore, may have little or no correlation with field
ionic conduction path through an electrolyte. Typical elec- experience (191, 215, 220–222, 234–236, 239). Most of the
trochemical procedures and tests are described in references accelerated corrosion test procedures commonly accepted by
(174, 190, 221, 227–229, 237). the automotive industry are designed for cosmetic corrosion
Linear polarization resistance (LPR) is a common electro- of steel and coated steel (237). Cosmetic corrosion is typically
chemical technique that measures instantaneous corrosion exterior corrosion that occurs as a result of the breakdown or
rates. It has been shown to correlate well in near-neutral solu- damage to a coating system. This type of corrosion does not
tions with weight-loss measurements for a wide variety of normally affect function or structural strength, but does com-
metals and corrosive media. Unfortunately, most corrosion of promise appearance.
motor vehicles occurs in moist or humid environments with Laboratory corrosion tests may be applicable to some
only a thin film of moisture present. Thus, LPR is not suitable degree to all metals, but they may not give valid comparisons
and/or measurements do not typically correlate with service even within the same metal systems (223). In addition, rank-
conditions and are thus unreliable. ings of snow and ice control chemicals will vary with the cor-
LPR measurements could be used as a rapid test method to rosion test procedures used and the metals and coatings
compare the corrosivity of various chemical deicers and tested (187). The tests typically are designed to determine the
inhibitors under different concentrations, temperatures, comparative corrosiveness of snow and ice control chemicals
inhibitor concentrations, or length of exposure time. LPR relative to a particular sodium chloride concentration under
would also be suitable for evaluating the effect on inhibitors controlled and reproducible conditions that are not typical of
with roadside dilution. Many types of inhibitors can acceler- field conditions (215).
ate localized corrosion if they are not present at sufficient Caution should be exercised for accelerated laboratory cor-
levels to protect all of the metal surfaces. Protected surfaces rosion tests because these tests are based on numerous
become cathodic to unprotected surfaces and if the cathodic- assumptions and arbitrary testing conditions. Corrosion data
to-anodic ratio is large, very rapid corrosion can occur on the developed using bulk solutions do not necessarily correlate
anodic areas. with corrosion in humid or moist environments such as
LPR testing is referenced in: occur on motor vehicles exposed to snow and ice control
chemicals.
• ASTM G 3, “Standard Practice for Conventions Applicable Weight-loss measurements are widely used for evaluating
to Electrochemical Measurements in Corrosion Testing”; corrosion resistance. Weight-loss is a suitable method for gen-
• ASTM G 96, “Standard Guide for On-Line Monitoring of eral corrosion, but it is not suitable for the most detrimental
Corrosion in Plant Equipment (Electrical and Electro- forms of motor vehicle corrosion (e.g., localized corrosion
chemical Methods)”; and such as pitting, crevice corrosion and under-deposit attack
• ASTM G 102,“Standard Practice for Calculating of Corro- (poultice corrosion), and galvanic corrosion).
sion Rates and Related Information from Electrochemical Immersion corrosion tests such as the PNS test are rapid
Measurements.” and provide an acceptance ranking for snow and ice control
chemicals based on a single sample (typically steel) and sin-
gle concentrations of chemical and inhibitor. Therefore, test
5.5.5 Test Track Testing
results do not necessarily bear any relationship to field
Test track testing is basically a form of accelerated field tests. conditions. In addition, the PNS tests appear to have inter-
Test tracks are designed to simulate field corrosion conditions, laboratory and intra-laboratory variations and are not
but in an accelerated time frame. They typically subject the suitable for metals other than steel. The validity of the PNS

91
test to predict uniform or general corrosion field conditions forms of localized corrosion by using the deicing chemicals
could be improved by running duplicate samples, controlling and inhibitors for the test media, instead of the normal salt
temperature and humidity, and running the test at a number solutions in environmental chambers with programmed dry-
of dilution ratios to try and simulate field conditions. These ing periods.
changes to the PNS test procedure will still not provide cor- Electrochemical corrosion tests can be used to evaluate
relation with field conditions because of localized corrosion, general corrosion of snow and ice control chemicals and cor-
galvanic corrosion, or corrosion of metals other than steel. rosion inhibitors over various concentrations; it is much
Salt-spray (fog) tests are primarily for the evaluation of more difficult, but not impossible, to devise a test program to
organic coatings on steel. They might be adapted to evaluate evaluate localized corrosion and galvanic corrosion.

92
CHAPTER 6
General Assessment of Information
6.1 General Conclusion effects to human health. Routine snow and ice control prac-
tices do not pose a significant risk to roadside microbial pop-
All snow and ice control materials have some potential for ulations outside what would already normally be considered
affecting the receiving environment and/or infrastructure. a disturbed environment. Human health concerns associated
The magnitude of the effect depends on a wide range of con- with handling of snow and ice control materials are associ-
ditions, including types of materials, quantities being applied, ated with very high doses and any real concerns are mitigated
transport pathways, exposure pathways, exposure duration, through application of proper safety measures (e.g., protec-
chemical-specific impacts, and other site-specific character- tive clothing and ventilation) that should already be in place.
istics. Assessment of the information gathered through this In addition, contamination of drinking water supplies to a
study has indicated where potential areas of impairment degree considered an impairment for a raw water supply
related to snow and ice control materials are most likely. could occur under some circumstances, especially in the case
Relevant product categories include NaCl, CaCl2, MgCl2, of individual wells very close to a point of routine application
acetates, organic biomass products, and abrasives. These for deicing materials.
material types represent the typical cross section of the pres-
ent market. Potential impairments associated with these
materials can also be applied to many of the available blended 6.1.2 Infrastructure Corrosion
materials.
6.1.2.1 Atmospheric Corrosion
Atmospheric corrosion processes related to snow and ice
6.1.1 Natural Receiving Environment
control chemicals are complex and it is difficult to provide
Table 6-1 summarizes the potential environmental impair- generalized evaluations. Much of the literature reviewed pro-
ment, including the estimated magnitude of effect and a vides information that depends on specific conditions; and
description of the contributing factors. The estimated mag- study conclusions tend to be wide-ranging and, in some cases,
nitude of effect has been based on the assessment and con- contradict one another. Discrepancies are related to the spe-
clusions developed in the referenced sections. cific conditions under which the assessment was carried out,
The most significant areas of potential impairment including type of snow and ice control material; presence, type
include water quality/aquatic life, air quality, vegetation, and concentration of corrosion inhibitor; alloy materials
soils, and animals. They also are the general areas of envi- being assessed; corrosion type being monitored; and specific
ronmental concern indicated in the survey of agency com- test method/conditions used. Despite the weakness in specific
ments (refer to Table 2-5). The exception is soil impacts, information, some broad assessments can be developed:
which have been included based on assessment of technical
information. Direct effects on soils are generally not a con- • Chloride-based snow and ice control chemicals are the
cern, but snow and ice control material can influence soil most corrosive to metals with the hygroscopic chlorides of
structure, and the presence of these materials in soil can fur- magnesium and calcium being the most aggressive because
ther affect water systems and vegetation. of the longer time of wetness.
Environmental areas reviewed in this study and found not • Acetate-based and organic snow and ice control materials
to be of significant concern include effects to soil biota and tend to cause significantly less corrosion than chloride-

93
Table 6-1. Generalized potential environmental impairment related to common snow and ice control chemicals.
Environmental Road Salt Calcium Chloride Magnesium Chloride Acetates Organic Biomass
Abrasives
Impact (NaCl) (CaCl2) (MgCl2) (CMA and KA) Products
Water Quality/ Moderate: Moderate: Moderate: High: High: High:
Aquatic Life Excessive chloride Excessive chloride Excessive chloride Organic content Organic matter leading Turbidity; increased
(Section 3.4) loading, metal loading; heavy metal loading; heavy metal leading to oxygen to oxygen demand; sedimentation.
contaminants; contamination. contamination. demand. nutrient enrichment by
ferrocyanide additives. phosphorus and
nitrogen; heavy
metals.
Air Quality Low: Low: Low: Low: Low: High:

(Section 3.8) Leads to reduced Leads to reduced Leads to reduced Leads to reduced Leads to reduced Fine particulate
abrasive use. abrasive use. abrasive use. abrasive use. abrasive use. degrades air quality.
Soils Moderate/High: Low/Moderate: Low/Moderate: Low/Moderate: Low: Low:

(Section 3.5) Sodium accumulation Improves soil Improves soil Improves soil Probably little or no Probably little or no
breaks down soil structure; increases structure; increases structure; increases effect; limited effect.
structure and permeability; potential permeability; potential permeability; potential information available.
decreases permeability for metals for metals for metals
and soil stability; mobilization. mobilization. mobilization.
potential for metals
mobilization.
Vegetation High: High: High: Low: Low: Low:

(Section 3.7) Spray causes foliage Spray causes foliage Spray causes foliage Little or no adverse Probably little or no Probably little or no
damage; osmotic damage; osmotic damage; osmotic effect; osmotic stress effect. effect.
stress harms roots; stress harms roots; stress harms roots; at high levels.
chloride toxicosis. chloride toxicosis. chloride toxicosis.
Animals Low: Low: Low: Low: Low: Low:

(Section 3.9) Sodium linked to salt Probably little or no Probably little or no Probably little or no Probably little or no Probably little or no
toxicosis and vehicle effect. effect. effect. effect; limited toxicity effect.
kills; magnitude information available.
unclear.
Conclusions based on the assessment of information reviewed in this study.
based chemicals, but can still accelerate corrosion by • The enabling and/or acceleration of corrosion of concrete
increasing the conductivity of the moisture. reinforcing (rebar) resulting from chloride ion penetra-
• Corrosion related to a given snow and ice control material tion. This effect is primarily related to bridges; however,
may show various responses between product concentra- there is a possibility of a secondary risk in concrete pave-
tions, metal types, and even between alloys within the same ments with doweled joints or continuous reinforcement.
metal group. • Reactions of snow and ice control materials with the
• Snow and ice control chemical corrosion inhibitors tested cement paste, resulting in deterioration of the cement
in the laboratory can show significant reductions in the matrix. This effect is primarily a risk for concrete pave-
corrosion rate of the metals being tested, but they may ments and the amount of deterioration is directly related
show little or no inhibiting effect on other metals (even to the quality of concrete. Evidence suggests that magne-
those in the same metal group) and may, in effect, acceler- sium reacts chemically with the cement paste in a process
ate corrosion of some metals.
that reduces the concrete strength.
• To develop any meaningful general corrosion rate data for
• All snow and ice control chemicals by definition of their
snow and ice control chemicals and corrosion inhibitors,
action increase the number of freezing and thawing
tests must be run over a wide range of concentrations and
cycles on the concrete in service. This can result in dete-
also under a number of ambient conditions to simulate
field conditions. rioration in the form of freezing and thawing and/or
scaling.
6.1.2.2 Concrete Corrosion

6.1.3 Other Infrastructure
Snow and ice control chemicals can have detrimental
effects on concrete infrastructure. These can be separated into Chloride-based snow and ice control products can affect
three general categories: the operation of railway traffic control signaling and power

94
distribution lines. This is related to their ability to become The decision tool is the most complex component of the
conductive, and in general, effects are related to extreme con- material section process and combines product rankings that
ditions. In the case of railway traffic control signaling, this reflect the material’s potential for impairment with agency
may involve pooling of brines across tracks; for power distri- objectives weighted by importance. A guidance tool for agen-
bution lines, it may involve power lines close to elevated cies was developed in Phase II of this study and includes a list
bridges and overpasses. However, the information is weak of potential operational issues and areas of impairment, along
regarding the frequency and magnitude of effect. with general recommendations to assist agencies in assigning
weights to decision categories.
Table 6-2 lists the four decision categories that form the
6.2 Draft Material Selection framework for the decision tool: cost, performance, natural
Framework environment, and corrosion. These categories balance oper-
ational concerns (“value” related to cost and performance)
6.2.1 Approach and Assumptions
with potential consequences of use (“costs” related to envi-
A process for selecting preferred snow and ice control ronmental and corrosion impacts). Also indicated is the
materials for generic application situations was developed average weighting given to categories as indicated from the
considering responses from the agency survey (see Section survey of agency comments (see Section 2.4.4.1). Environ-
2.3) and information on product characteristics, environmental and corrosion concerns currently are given much
mental effects, and material effects, as discussed in Chapters lower agency importance than cost and performance.
2 through 4. When selecting a snow and ice control material, Although these priorities may not be ideal, it is nonetheless
a series of steps and factors is considered. Figure 6-1 illus- the current reality.
trates the selection process. The general process is straight- Table 6-2 also lists subcategories that contribute to the
forward; however, certain components (e.g., the “decision development of natural environment and corrosion decision
tool”) are more involved. This complexity makes it difficult categories. These were selected based on agency survey results
to develop a model that incorporates “all” possible cases. and study team experience. The divisions are relatively
The framework proposed here addresses the most common straightforward and provide a reasonably thorough represen-
and most significant factors involved in selecting snow and tation of the higher level decision category.
ice control products. Development of similar subcategories for cost and per-
Before selecting a snow and ice control material, applica- formance offers the following challenges:
tion methods must be selected by the agency. Application
methods have considerable influence on level of service, • The cost of snow and ice control involves direct costs (i.e.,
and significant research has been applied toward develop- purchase price) and indirect costs (i.e., operational costs
ing guidelines for selecting snow and ice control applica- and capital costs). Direct and indirect costs can be influ-
tion methods (6). A challenge in developing a material enced by factors outside the scope of information gathered
selection process is to maintain independence from the in this study. Additionally, indirect costs are related more
application method selection processes, while realizing that to application strategy/policies than material selection. For
significant interrelations exist. For example, application this study, cost will represent purchase price per kilogram
methods influence volumes of materials released to the of solid material as provided by each agency.
environment and, therefore, influence the potential for • Performance is a broad term that may have numerous
environmental impact. When general application methods measures depending on conditions being assessed. For this
are selected before materials, the influence the application study, general performance will be represented by temper-
method can have on material selection can be carried ature performance assessed through application of eutec-
through the material selection process. tic temperatures using the Performance Index description
Once application methods are agreed on, an agency will in Section 2.6.1.
choose a list of candidate snow and ice control products.
A matrix of significant properties will exist or, in the case The result of the decision tool is a numerical product score
of newer products, be developed, and then preliminary that can be used to compare selected products objectively.
qualification will be applied through comparison with Product selection occurs after a review of other agency consid-
a specification. Materials meeting specification require- erations that may include, but not be limited to, concerns with
ments will be “qualified” for further assessment using the product availability (e.g., delivery times and quantities), quality
decision tool. testing results, and equipment availability.

95
Figure 6-1. Flowchart for materials selection process.
The following sections describe the material selection exist or be developed. This will consist of a series of tables that
processes in greater detail. cross-references a material with a measured parameter. In
cases where non-tested attributes are a concern, a descriptive
rating will be applied in place of a numerical test result. Appli-
6.2.2 Candidate Product Selection and
cation methods will be given consideration here, because they
Matrix of Significant Properties
have significant bearing on the volumes of materials released
As an initial step in product selection, an agency will select to the environment.
a list of candidate snow and ice control products. For each The approach here is to characterize each material on the
product, an associated matrix of significant properties will basis of components that affect cost, performance, the natu-

96
Table 6-2. Decision tool framework. erate a “short list” of products for detailed assessment quickly.
Phase II of this study involved developing a purchase specifi-
Decision Category Weighting1 Decision Subcategory cation for agencies to use in this step.
Cost 45
Agency Input Required 6.2.4 Decision Tool
Performance 33
The decision tool combines the product rankings devel-
Aquatic oped in this study and agency objectives for several decision
Air
categories and subcategories (see Table 6-2). The product is a
Natural Environment 12 Vegetation
Soil numerical score for a material/application selection that can
Animals be compared among materials. The decision tool consists of
Vehicles
Infrastructure
a relatively simple software application.
Corrosion 10
Concrete
Concrete Reinforcing
1 6.2.4.1 Product Rankings
Agency Survey, Section 2.3.4.1
The study team developed product rankings based on the
technological appraisal of information gathered and assessed
ral receiving environment, and corrosion. A matrix of sig- in this study and based on agency input. This takes the form
nificant properties was developed during Phase II of this of a matrix embedded in the decision tool that provides a
study. The information included in the “matrix of significant numerical product ranking for each product/subcategory
properties” will also be used in the decision tool to develop combination. Five levels of numerical ranking (1 = poor; 2 =
product rankings (see Section 6.2.4.1). poor/moderate; 3 = moderate; 4 = moderate/good; 5 = good)
provide sufficient numerical range to separate “poor” and
“good” product performance effectively during subsequent
6.2.3 Preliminary Qualification
modeling (decision tool), without being overly complex.
Agencies may choose minimum qualifications that singu- In the case of natural environment and corrosion, the study
larly affect the selection of a product. An example is a pur- team developed product rankings at the “subcategory” level
chase specification addressing product cost, toxicity, that were based on a material’s potential for impairment.
corrosion, or other policies important to the agency. This These were similar to the evaluations prepared in Section 6.1
process allows agencies to reject unsuitable products and gen- (Tables 6-1 and 6-3), but were supplemented by the informa-
Table 6-3. Generalized potential corrosion impairment related to common snow and ice control chemicals.
Environmental Road Salt Calcium Chloride Magnesium Chloride Acetates Organic Biomass Abrasives
Impact (NaCl) (CaCl2) (MgCl2) (CMA and KA) Products
Atmospheric High: High: High: Low/moderate: Low: Low:
Corrosion to Metals Will initiate and Will initiate and Will initiate and Potential to initiate Potential to initiate Probably little or no
(General) accelerate corrosion. accelerate corrosion; accelerate corrosion; and accelerate and accelerate effect
higher potential for higher potential for corrosion due to corrosion due to
corrosion related to corrosion related to elevated conductivity. elevated conductivity
hygroscopic properties hygroscopic properties claims of mitigation of
corrosion require
further evaluation.
Concrete Matrix Low/moderate: Low/moderate: Moderate/high: Moderate/high: Low: Low:

(Section 4.1) Will exacerbate Will exacerbate Will exacerbate Will exacerbate Probably little or no Probably little or no
scaling; low risk of scaling; low risk of scaling; risk of paste scaling; risk of paste effect. effect
paste attack. paste attack. deterioration from deterioration from
magnesium reactions. magnesium reactions
Concrete Reinforcing High: High: High: Low: Low: Low:

(Section 4.1) Will initiate corrosion Will initiate corrosion Will initiate corrosion Probably little or no Probably little or no No Effect
of rebar. of rebar. of rebar, evidence effect. effect; claims of
suggests MgCl 2 has mitigation of corrosion
highest potential for require further
corrosion of chloride evaluation
products
Conclusions based on the assessment of information reviewed in this study.

97
tion from the testing program and matrix of significant prop- 6.2.4.2 Agency Objectives
erties developed in Phase II of this study. For example, one
material may pose a significant corrosion risk to concrete Each agency will have a unique set of concerns specific to
reinforcing and be assigned a score of “1 = poor.” Alterna- their conditions. For example, in certain regions, legislation
tively, a second material posing lower risk may be assigned a may require protection of a particular aquatic environment;
score of “4 = moderate/good.” in other regions, corrosion of infrastructure may be a con-
Cost and performance rankings are applied at the “cate- cern. It must also be recognized that there is a high degree of
gory” level and are developed based on user input. For exam- reluctance to depart from traditional materials and methods
ple, an agency will provide purchase cost information, because of political, personal, or other reasons. The goal here
because these costs can be significantly affected by trans- is to have the agency express policies, objectives, operations,
portation costs, and are not similar for all agencies. To assess and other conditions though numerical weighting of the cat-
performance requirements, agencies will provide information egories and subcategories listed in Table 6-2. Agencies will
on the pavement temperatures (i.e., lowest 85th percentile) apply their criteria on two levels:
expected for the region. Many agencies do not routinely expe-
rience extreme cold temperatures and, therefore, would 1. Agencies will assign weights to decision “categories”—
receive limited additional performance benefit from low- cost, performance, natural environment, and corrosion.
temperature-performance snow and ice control materials. This level of weighting reflects general policy concerns.
To illustrate the approach, Table 6-4 provides an example of 2. Agencies will assign weights to “subcategories.” This rep-
preliminary product rankings for two potential product/ resents objectives and concerns at the operational level. A
application strategies: (1) Abrasive/NaCl deicing; and (2) MgCl2 guidance tool was developed in Phase II of this study and
anti-icing. These illustrate the categories and subcategories includes a list of potential operational issues and areas of
used and the aspects of the product considered to formulate impairment, along with general recommendations.
a product ranking. These do not reflect the actual rankings Table 6-5 provides current thinking on the general areas
developed during Phase II of this study, but they are reason- an agency will need to consider. To ensure this process is
able approximations based on the current knowledge. practical at the operational level, the requested informa-
Table 6-4. Proposed product rankings for abrasive/NaCl deicing vs. MgCl2 anti-icing.
Product
Ranking 1
(1=poor; 5=
Category Subcategory Description excellent)
Abrasive/ MgCl2
NaCl Anti-
Deicing icing
Cost - User input (example): cost of MgCl2 > Abrasive/NaCl 5 3
User input (example): general operational low temperature -
Performance - 4 5
4ºC; MgCl2 marginally better than Abrasive/NaCl
Abrasives have high potential to impact aquatic
Aquatic 1 3
Environment environments; MgCl2 is a moderate concern
Air Abrasives have high potential to impact air quality 1 5
Both NaCl and MgCl2 can impact vegetation; risk for anti-
Vegetation 5 4
icing is lower due to lower volumes of chemical release
NaCl can impact soil structure, both NaCl and MgCl2 can
Soil 4 2
potentially mobilize metals
Both NaCl have low (unclear) potential to impact species,
Animals 4 5
potential roadkills related to NaCl attraction
Both NaCl and MgCl2 increase corrosion; MgCl2 may be
Vehicles 2 1
Corrosion more aggressive due to its hygroscopic properties
Metal Both NaCl and MgCl2 increase corrosion; MgCl2 may be
2 1
Infrastructure more aggressive due to its hygroscopic properties
Concrete Matrix Evidence suggests MgCl2 may impact the concrete matrix 4 2
Concrete Both NaCl and MgCl2 increase corrosion; evidence suggests
2 1
Reinforcing MgCl2 may initiate corrosion sooner
1
Rankings based on Tables 6-1 and 6-3, except for Cost and Performance, which illustrate “examples” of agency input.
“ – ” indicates not applicable

98
Table 6-5. Operational objectives and concerns.
Decision
Agency Inputs and Areas of Potential Impairment
Subcategories
Purchase Cost 1. Agency to input cost for candidate products
Temperature
2. Agency to input mean winter temperature
Performance
3. Site-specific; regional stream standards; or antidegradation criteria
4. Road density per unit watershed area
5. Presence of standing or slowly moving waters near application area
Aquatic 6. Locations of unusual sensitivity (endangered or threatened species,
spawning habit)
7. Shallow drinking water wells (downstream) overlain by permeable
soils
Air 8. Poor air dispersion
9. Protected habitats within 30m of roadway

Vegetation
10. Aesthetic objectives
11. Potential for changes in soil structure (ESP >13)
Soil
12. Protected habitats within 30 m of the roadway
Animals 13. Protected habitats within 30 m of the roadway
Vehicles 14. Assumed constant for all agencies
15. Structural steel infrastructure
Steel Infrastructure
16. Pipelines along roadway, on bridge structure
Concrete 17. Aged/poor quality concrete
18. Reinforced concrete decking
19. Aged/poor quality concrete
tion must not be overly detailed, yet adequately address the (“category” level weightings); and operational objectives (“sub-
most significant areas of concern. If an agency has a con- category” level weightings), as follows:
cern not explicitly considered in the decision tool, agency
staff will be able to address this through adjustment of 1. Product rankings are as shown in Table 6-4.
assigned weights, as appropriate. Using the proposed 2. Current agency weightings for the decision “categories”
methods as a guideline, an agency will be able to develop are as shown in Table 6-2 (i.e., cost = 45 percent; perform-
weightings specific to the circumstances. ance = 33 percent; natural environment = 12 percent; and
The agency will also define other critical information at corrosion = 10 percent).
this step, such as the road sections of particular concern. It 3. Current agency weighting for natural environment and
is anticipated that the decision tool can be applied to road corrosion decision “subcategories” are represented in
areas ranging from small specific sites, such as bridges, to Table 6-6.
an entire maintenance region.
Table 6-7 shows the mathematical steps involved in com-
bining product rankings, policy objectives (categories), and
6.2.4.3 Decision Tool Model operational objectives (subcategories). The final result is a
The decision tool mathematically combines product product score. With the information illustrated in this
rankings and agency objectives for general categories and example, the final product score for Abrasive/NaCl deicing
provides a product score that can be compared with other is 68, which is higher than the MgCl2 anti-icing score of 64.
Although the resulting scores are similar, based on the infor-
products. The process is best illustrated using a case
mation in this example, Abrasive/NaCl deicing best meets
example:
agency objectives.
Example:
6.2.5 Product Selection
An agency has traditionally conducted abrasive/NaCl deicing
and is considering switching to MgCl2 anti-icing. The agency has The decision tool provides a numerical product score for
provided input to generate product rankings; policy objectives snow and ice control materials assessed; however, it is not

99
Table 6-6. Example: Agency objectives for decision subcategories.
Subcategory
Decision Weighting (%)
Subcategory Agency Objectives
Category Max. = 100% per
Decision Category
Primary concern; low recharge, protected water bodies
Aquatic 50
Natural Environment adjacent to large portions of maintenance area
Air Minor concern 10
Vegetation Aesthetic concern 20
Soil Not a concern 10
Animals Not a concern 10
Vehicles Concerns from public and trucking industry 40

Corrosion
Infrastructure Minimal concern 10
Concrete Matrix Minimal concern 10

Concrete
Aged bridge inventory 40
Reinforcing
necessary that an agency selects the highest ranked product. numerous contributing factors. These may include types of
The decision tool modeling process is not ideal, and final materials, quantities being applied, transport pathways,
product scores are best estimates based on the input infor- exposure pathways, exposure duration, receptor sensitivity,
mation. As illustrated in the example (Section 6.2.4.3), it is and other site-specific environmental characteristics. It is,
expected that scores will be similar among products. In such therefore, not possible to account for all conditions when
cases, agencies may allow other considerations to influence developing a product selection tool. At best, generalized
their product selection. These may include, but not be lim- assessments based on the most common and most signifi-
ited to, concerns with product availability (e.g., delivery cant factors can be made.
times and quantities), quality testing results, and equipment Many of the studies reviewed made conclusions based on
availability. information provided through laboratory studies or adapted
from studies conducted for purposes other than snow and
6.3 Information Gaps ice control, and limited field correlations were available.
Additionally, many materials have had a relatively short his-
It was expected that the findings of the information assess- tory of use (less than 10 to 15 years), and it is unclear what
ment conducted through this study would identify gaps in the potential for long-term adverse effects are over decade
information. Technologies and products employed for snow time scales. Addressing this will require much more infor-
and ice control are continually evolving. As well, many of the mation based on field studies than currently is available.
effects surrounding snow and ice control are highly complex Although the main types of impairment seem to have been
and variable, making development of any definitive conclu- identified, the actual documentation and modeling of these
sions difficult. General information gaps are included in this types of impairment under specific field conditions are not
section and those gaps specific to the natural environment available.
and infrastructure are included in Sections 6.3.1 and 6.3.2
respectively.
6.3.1 Information Gaps—Natural
The factors that affect the transport of snow and ice con-
Environment
trol materials and their ability to affect the receiving envi-
ronment are extremely complex. In general, the magnitude The scientific and technical literature supporting an
of impact is related to product-specific properties and dose evaluation of snow and ice control materials as related to
at the environmental or infrastructure receptor. Prediction natural environments is relatively weak in certain areas,
of dose at the receptor is complex and can be affected by given the scope and magnitude of road applications. Exam-

Table 6-7. Decision process example – Abrasive/NaCl deicing vs. MgCl2 anti-icing.
Mat er i al : Ab r as i v e & NaCl Ap p l i c at i o n St r at eg y : Dei c i n g
Decision Subcategory Level Decision Category Level
Product Ranking (1=poor;
1 Score (Maximum =
5= excellent) Subcategory User Weight % Score (1-5) Category User Weight %2
100)
5. 00 Cost 45
2.00 Performance 33
1 Aquatic 50 68
1 Air 10
5 Vegetation 20 2.40 Natural Environment 12
4 Soil 10
4 Animals 10
2 Vehicles 40
2 In frastructure 10
2.20 Corrosion 10
4 Concrete 10
2 Concrete Reinforcing 40
Material: MgCl2 Application Strategy: Anti Icing

Decision Subcategory Level Decision Category Level
Product Ranking (1=poor;
1 Score (Maximum =
5= excellent) Subcategory User Weight % Score (1-5) Category User Weight %2
100)
3.00 Cost 45
4.00 Performance 33
3 Aquatic 50 64
5 Air 10
4 Vegetation 20 3.50 Natural Environment 12
2 Soil 10
5 Animals 10
1 Vehicles 40
1 In frastructure 10
1.10 Corrosion 10
2 Concrete 10
1 Concrete Reinforcing 40
1 2
Maximum = 100 per category (i.e. natural environment or corrosion); Maximum = 100 (i.e. sum of cost, performance, natural environment, and corrosion)
101
ples of information gaps related to receptors in the natural 6.3.2 Information Gaps—Infrastructure
environment include
6.3.2.1 Atmospheric Corrosion
• When considering aquatic toxicity, only a handful of Atmospheric corrosion processes related to snow and ice
bioassay studies are available for snow and ice control control chemicals are extremely complex and it is difficult to
materials. provide generalized evaluations. The following information
• It is difficult to judge the degree of variability between lots gaps exist:
for such materials with respect to toxicity or other envi-
ronmental effects such as nutrient content, or to produce • Comparison of the reviewed literature shows wide-ranging
statistically robust information on the toxicity or other conclusions that in some cases contradict one another.
environmentally relevant characteristics of snow and ice • Many laboratory study conclusions are based on very
control materials in various categories. Thus, much more specific assessment conditions, including type and con-
extensive chemical and bioassay testing is warranted in centration of snow and ice material; presence, type and
order to make the environmental evaluation of snow and concentration of corrosion inhibitor; type of alloy materi-
ice control materials more robust. als; corrosion type being monitored; and specific test
• Information related to environmental monitoring in method/conditions used.
conjunction with application of snow and ice control • The literature is not decisive in providing relative corrosion
materials is weak. To be relevant to the direct evaluation
rates between snow and ice control material for various
of snow and ice control materials, such monitoring needs
metal types, and even between alloys within the same metal
to be designed to answer specific questions. Examples of
group.
relevant questions include the following: for specific cli-
• The effectiveness of common corrosion inhibitors outside
matic conditions and application rates, what are the
of controlled laboratory conditions (i.e., under field condi-
characteristic rates of dilution for application materials
tions, with dilution) is not fully supported in the literature.
after they leave the roadway? What are the conditions
• Although the main types of impairment related to snow and
under which stagnant or slowly moving waters are sig-
ice control materials seem to have been identified, actual
nificantly impaired by application materials? What kinds
documentation and modeling of these types of impairment
of assessment or monitoring methods would be appro-
under specific field conditions are not available.
priate for routine use as insurance against environmental
impairment?
• Metals mobilization in soils has been identified as a The variability and uncertainty surrounding atmospheric
potential concern for roadside soils; however, limited field corrosion processes is, therefore, considered a significant
evidence is available linking metals mobilization in soils weakness in information.
to applications of snow and ice control materials.
• Heavy metal ions can be displaced by specific ions such as 6.3.2.2 Concrete Reinforcing and Concrete Paste
calcium and magnesium. These metal ions may bioaccu-
mulate in the food chain; however, information is lacking The following information gaps exist with respect to effects
in public literature relating this potential effect to snow and of snow and ice control materials on concrete:
ice control materials.
• The literature indicates that reductions in soil stability and • Many of the studies reviewed through this assessment have
potential for soil erosion are concerns related to sodium made conclusions based on information provided through
applications, but limited field evidence is available. laboratory studies. These studies often used mortar and
• When considering soil microbial communities, limited lit- cement paste or used poorer quality concrete than is typi-
erature was found directly assessing exposures and effects cally used for infrastructure applications.
of these snow and ice control materials to soil microbial • The information indicates that magnesium can affect the
communities. concrete matrix, but the degree of deterioration risk is
• There is no conclusive field evidence demonstrating uncertain.
whether or not snow and ice control materials increase ani- • There is insufficient evidence to support the claimed effec-
mal deaths though toxicosis or through vehicle collisions. tiveness of common corrosion inhibitors (typically organic
• Radioactivity has been linked to well brines associated with biomass) added to snow and ice control chemicals in pre-
oil and gas production; however it is unknown what levels venting corrosion in reinforcing steel. These products have
can generally be expected from similarly obtained brines not been designed to penetrate concrete and will likely have
and mined material. short-lived duration because of biodegradation.

102
• Limited information is available that assesses the effects of formate) were used too infrequently for roadway applications
cations on the pore structure of concrete paste. The poten- to warrant any consideration for analytical testing. It was also
tial effects are twofold: (1) cations can, to some degree, unlikely that these materials would be used beyond very lim-
affect the durability of concrete pore structure, which can ited applications given their significant costs when compared
affect strength and chloride ion penetration; and (2) with more common products. Regardless, these materials
cations can affect the pH of the concrete pore solution, could still be considered at a later date using the framework
which may depassivate the reinforcing at the rebar level. established in Phase II of this study.
• Although the main types of impairment seem to have been Samples were obtained through cooperation with produc-
identified, the actual documentation of these types of ers and suppliers only. Collection of samples from user
impairment under specific field conditions is not available. sources was avoided so as to address issues related to possible
Currently, the history of use for many snow and ice control contamination or mishandling.
products is less than 15 years. It may take many more years To develop a list of significant properties for common
before effects related to their use are evident. snow and ice control materials, an analytical testing pro-
gram was proposed for chemical analysis, aquatic toxicity,
concrete corrosion, and atmospheric corrosion. Materials
6.3.2.3 Other Infrastructure
were selected to represent a cross section of the present
Evidence supports the suggestion that snow and ice con- market in terms of product types, additives, and vendors.
trol materials, specifically chloride-based materials, can Chemical products (Table 6-8) were selected for each of the
influence the proper operation of railway traffic control sig- proposed test programs. Most products represented com-
naling and power distribution lines. Although processes for mercially available products identified in the agency survey
impairment have been identified, actual documentation is (Section 2.4.1), although manufacturers’ names have not
not available. been disclosed. Products listed in Table 6-8 represent
the current situation and therefore depend on materials
6.3.3 Information Gaps—Materials availability.
Up to 50 samples were tested for chemical analysis. Efforts
Many products available for snow and ice control are pro- were reduced for aquatic toxicity, atmospheric corrosion, and
prietary, making it difficult to obtain specific test informa- concrete corrosion testing. For these groups, efforts focused
tion. Significant variability may exist in these products in on CaCl2 and MgCl2. These selected products represent the
terms of additive types and percentages. This has not been most-used snow and ice control materials and corrosion
well-documented in the literature. inhibitor combinations as well as most of the available prod-
ucts. KA and a CMA-KA blend were included in Table 6-8
6.4 Revised Work Plan because of their widespread use on automated bridge snow
and ice control systems. The use of these chemicals on bridges
The Phase II work plan covered Tasks 7 through 13 of the presents unique corrosion and environmental issues given
NCHRP Project 06-16 research plan. that many bridges are over waterways.
Commercial product blends are continually evolving and,
6.4.1 Task 7 (Analytical Testing Program) in many cases, product-specific test data become obsolete
when formulations are modified. This is especially true
6.4.1.1 Product Selection where inhibitors and additives are present. It was, therefore,
Agency surveys determined common snow and ice control not the purpose of this study to provide a comprehensive
materials were found to include one or more of the following classification of all available products. Proposed testing
base products: efforts were to provide an initial data set of commonly used
products. This study was not intended to determine variance
• Chloride salts—NaCl, MgCl2, CaCl2, and various propri- in each manufacturer’s production. To conduct such an
etary blends; effort with an acceptable level of statistical significance
• Acetates—calcium magnesium acetate; and would be beyond the scope of this study. Concerns of prod-
• Organic Products—typically agriculturally derived biomass. uct variability could, however, be addressed indirectly
through the application of the quality assurance program,
Subsequent testing efforts focused on these materials. although this would require continued management
Other products (e.g., glycols, alcohols, K2CO3, and potassium through a parent organization.

103
Table 6-8. Proposed test samples.
Testing Effort
Chemical Description Chemical Aquatic Concrete Atmospheric
Analysis Toxicity Corrosion Corrosion
(6.4.1.2) (6.4.1.3) (6.4.1.4) (6.4.1.5)
CaCl2 Liquid X X X X
Corrosion Inhibited Liquid X X X X
Corrosion Inhibited Liquid X
Solid Pellet X
Solid Flake X
Solid Flake X
MgCl2 Liquid X X X X
Solid Flake X
Solid Crystal X
NaCl Solid (ASTM Spec) X X X X
Solid X
Corrosion Inhibited Solid X X X X
Corrosion Inhibited Solid X
Blended Mined X X X X
Chloride Well Brine Liquid X X
Corrosion Inhibited Solid X X X X
CMA Manufactured Solid X X X X
KA Manufactured Liquid X X X X
CMAK Blended Liquid X X X
Organic Manufactured Liquid (Biomass) X
Manufactured Liquid (Biomass) X
Totals 43 15 15 15
6.4.1.2 Chemical Analysis chloride would be found in potassium acetate). However, it

was proposed that all samples requiring chemical analysis be
Chemical analyses were conducted to characterize the snow
subjected to the same suite of testing, except where this was
and ice control products and to determine the presence of any
physically impossible. This approach was taken to ensure com-
trace chemicals that might occur as impurities. It was expected
pleteness of the data set and to account for cases where unex-
that up to 50 test samples (Table 6-8) would be analyzed as part
pected analytes were shown to be present. Table 6-9 indicates
of the chemical testing program. These samples represented
the product types and associated chemical analyses.
different base chemicals, mix ratios, and additives.
Each test sample was subjected to a range of applicable ana-
6.4.1.3 Aquatic Toxicity Testing
lytical test parameters. The study team recognized that not all
sample types needed to be tested for all chemical analysis Aquatic test procedures were proposed for 15 snow and ice
parameters (i.e., it was unlikely that appreciable amounts of control products (Table 6-8). Standard test methods as indi-

104
Table 6-9. Chemical analysis.
Test Parameters Method

Reference
Metals
• Arsenic, Barium, Cadmium, Chromium, Copper, Iron, Lead, APHA
Manganese, Mercury, Molybdenum, Nickel, Selenium, Silver, Zinc
Cations
APHA
• Sodium, Calcium, Magnesium, Potassium
Phosphorus
APHA
• Total Phosphorus, Total Soluble Phosphorus
Nitrogen
APHA
• Nitrate & Nitrite, Ammonia, Total Kjeldahl Nitrogen (TKN)
Chloride APHA
Biochemical Oxygen Demand (BOD) APHA
Chemical Oxygen Demand (COD) APHA
pH APHA
Cyanide, Ferrocyanide APHA
Specific Gravity ASTM D 1429
APHA – American Public Health Association (67)
ASTM – American Society for Testing and Materials
cated in Table 6-10 were followed and represented chronic By conducting a 7-day test, one can determine the 96-hour
exposure assessments to various levels of aquatic biota, LC50 and 7-day LC50 as well as a 7-day EC50 on growth, in addi-
including vertebrates, invertebrates, and algae. tion to other endpoints, including LC25, EC25, lowest observ-
Chronic exposure tests were selected over more com- able effect concentration (LOEC), and no observable effect
monly used acute test procedures (i.e., Rainbow trout 96- concentration (NOEC).
hour LC50) because they provided more information related
to potential long-term impacts. Both Environment Canada
6.4.1.4 Concrete Reinforcing Corrosion
and the U.S. EPA recognize the 7-day fathead minnow test as
a chronic bioassay for freshwater fish whereas the 96-hour The properties of the concrete that required evaluation,
test is recognized as an acute test. It has been judged that the and the gaps in our current ability to conduct the evaluation,
7-day test is sufficient in length to evaluate chronic toxicity. are outlined in Section 6.3.2.2. It was proposed that a mini-
In addition, this test assesses the life stage of the representa- mum of 15 snow and ice control products be selected for con-
tive fish that is considered to be most sensitive to adverse crete corrosion testing (Table 6-8). It was beyond the scope of
effects from chemicals and provides endpoints for growth this study to develop completely new test methods. It was
and survival. intended to take existing methods, possibly combinations or
Table 6-10. Aquatic toxicity testing.

Fathead Minnow 7-day Survival and Growth U.S. EPA, 1994, Short-Term Methods for Estimating the
Test (will also include acute toxicity data for Chronic Toxicity of Effluents and Receiving Water to
Fathead Minnow 96-hour Survival Test) Freshwater Organisms, 3rd Edition, EPA/600/4-91/002
Ceriodaphnia dubia 3-brood Survival and U.S. EPA, 1994, Short-Term Methods for Estimating the
Reproduction Test Chronic Toxicity of Effluents and Receiving Water to
Selenastrum capricornutum 96-hour (US EPA) U.S. EPA, 1994, Short-Term Methods for Estimating the
Growth Test Chronic Toxicity of Effluents and Receiving Water to

105
modifications of those discussed in Section 5.4, and apply rosion tests have shown significant variability between test
them to the needs here. Results were in relative performance coupons used and between laboratories. The validity of the
only. In this regard, a single ice control chemical, presumably immersion tests to predict uniform or general corrosion field
NaCl at some established dosage, was selected as a control for conditions could be improved by running duplicate samples,
reference of others. It was expected that test sample concen- controlling temperature and humidity, and running the test
trations would be selected to match routine application expo- at a number of dilution ratios to try and simulate field condi-
sure concentrations. tions. These changes still did not provide correlation with
As indicated in Section 5.4.4.1, it was necessary to define a field conditions because of localized corrosion, galvanic cor-
standard concrete (not mortar) mixture proportion from rosion, or corrosion of metals other than steel.
which evaluation of the relative performance of ice control Salt-spray (fog) tests are primarily for the evaluation of
chemicals could be made. It should have been representative organic coatings on steel. They could be adapted to evaluate
of the quality of the concrete used in transportation works. forms of localized corrosion by using snow and ice control
As an initial effort in this regard, the AASHTO specification chemicals and inhibitors for the test media instead of the nor-
for such concretes was used. Concrete exposed to snow and mal salt solutions in environmental chambers with pro-
ice control chemicals proposed for testing were to have grammed drying periods. However, drawbacks include
method development (particularly for tests of solutions
• Air entrainment producing adequate bubble spacing fac- exceeding 5-percent concentration) and duration of time
tors to ACI 318 requirements; and required for testing. Salt-spray (fog) tests are quality control
• A maximum w/c ratio of 0.45. tests for coatings and not a viable tool for evaluating degra-
dation of infrastructure.
A test method was designed to evaluate, simultaneously, The literature and study team experience indicated that
the susceptibility of a surface-applied chemical’s potential laboratory poultice corrosion tests using various dilutions of
to cause deterioration of the paste and/or corrosion of the chemicals and inhibitors would provide good correlation
rebar. The test was similar to ASTM C 1556 (refer to Sec- with field experience for crevice corrosion, poultice corro-
tion 5.4.4.2) but included wetting and drying cycles. In sion, pitting, and galvanic corrosion of the various metals and
addition to the assessment of the rebar corrosion, ion dif- combinations of metals used in motor vehicles and metal
fusion profiles were obtained and the two aspects corre- bridge structures. Despite limited evidence of correlation
lated. Again, opinion on the method was sought from other with various forms of localized attack, the poultice test had
researchers. drawbacks—it was not a standardized (and therefore not
The study team had informal discussions with researchers widely applied) test; moreover, comparatively long exposure
assessing the effects of snow and ice control chemicals on times would be necessary.
concrete and concrete reinforcing. Detailed discussions were Rapid evaluation of the corrosive properties of snow and
held with these to obtain the benefit of their experience on ice control chemicals and inhibitors could best be carried out
test method development. Such discussions also avoided using electrochemical techniques (i.e., linear polarization
duplication of effort between this program and existing resistance (LPR) testing). Electrochemical corrosion tests
programs. could be used to evaluate general corrosion of snow and ice
control chemicals and corrosion inhibitors over various con-
centrations. The effectiveness of a corrosion inhibitor is influ-
6.4.1.5 Atmospheric Corrosion Testing enced by concentration (i.e., as affected by roadway dilution).
Various options are available to assess atmospheric corro- In part this also addressed manufacturer corrosion claims.
sion; however most accelerated laboratory tests do not corre- Further benefits of LPR were that it
late with field experience, particularly where the predominant
form of degradation is localized corrosion and galvanic cor- • Was a mature, proven technology and had been used exten-
rosion. The following represented the development of the sively to field-monitor corrosion of equipment and infra-
proposed test program. structure;
Immersion corrosion tests, such as the PNS corrosion test • Was considered suitable for evaluating the effect of
(43), are rapid tests that provide an acceptance ranking for inhibitors with roadside dilution;
snow and ice control chemicals based on a single sample, typ- • Had good correlation with weight loss measurements for
ically steel, and single concentrations of chemical and near-neutral pH solutions; and
inhibitor. However, test results do not necessarily bear any • Was considered to be most well adapted to aqueous
relationship to field conditions. In addition, immersion cor- corrosion.

106
As with the drawbacks of immersion and salt-spray corro- 2. Diluted to 10 percent (w/w) of the “as applied” product.
sion testing, LPR had limited applicability with respect to This represented the initial stages of roadway dilution.
evaluating localized corrosion and galvanic corrosion, but the 3. Diluted to 1 percent (w/w) of the “as applied” product.
benefits of this technique outweighed these. This represented later stages of roadway dilution.
The study team proposed to conduct LPR measurements
in accordance with the criteria established in ASTM G 96. The proposed test program would produce 225 individual
The test program would comply with the conventions determinations (15 samples 5 alloys 3 concentrations).
acknowledged in ASTM G 3 while the calculation of corro- Expected results would provide general corrosion compar-
sion rates would be carried out per ASTM G 102. Measure- isons of metals, dilution ratios, chemicals, and inhibitors.
ments would be conducted on five commonly available
alloys: 6.4.1.6 Abrasive Testing
1. Steel is used in the following functions: reinforcing bar, The original work plan accounted for testing of 15 abrasive
structural steel bridges, motor vehicle frames, bumpers, products using gradation and abrasion tests. From an envi-
brake line fittings, body panels, general roadside infra- ronmental perspective, the current body of information was
structure, guard rails, sign posts, and lamp standards. A well developed (i.e., abrasive use can degrade air quality and
wide range of steel grades is used, but most will have sim- lead to turbidity and sedimentation concerns for aquatic sys-
ilar corrosion rates, except for weathering steels. A stan- tems). The agency survey (Section 2.4) also indicated that
dard structural steel such as ASTM A 36 will be used for most material was obtained through an extremely wide range
testing. of regional sources, making it extremely difficult to obtain a
2. Wrought aluminum is used in the following functions: good representation. Current information was sufficient to
warrant no further testing of abrasive materials.
motor vehicle frames, diesel fuel tanks, body panels, road-
side infrastructure, building cladding, signage and so
forth. Wide ranges of aluminum alloys are used and have 6.4.2 Task 8 (Matrix of Significant
varying degrees of corrosion resistance. Aluminum alloy Properties)
6061-T6 will be used for testing because it is one of the
The matrix consisted of a series of tables that cross-
most common alloys used.
referenced a specific base product or blended material with a
3. Cast aluminum is used in motor vehicle wheel rims, trans-
measured parameter as developed through Task 7. In cases
mission cases, engines, and so forth. Cast aluminum alloy
where non-tested attributes were a concern, efforts were
356.0 will be used for testing because it is one of the most
made to include a descriptive rating that could be applied in
common aluminum alloys used. place of a numerical test result. The matrix was to display the
4. Cast magnesium is used primarily for motor vehicle wheel test results and information that would ultimately be applied
rims and transmission housings. Die cast magnesium alloy to the development of the decision tool. It was difficult to
AM60A-F samples will be used for testing because it is one describe the exact formatting and organization of these
of the most common alloys used for wheel rims. tables, except that efforts would be made to ensure that the
5. Brass is typically used for brake line fittings and for elec- tables were understandable and organized logically. A simple
trical contacts. A typical leaded free-machining brass such conceptualization of the matrix is indicated in Table 6-11.
as copper alloy C360 will be used for testing.
It was proposed that a minimum of 15 snow and ice con- 6.4.3 Task 9 (Purchase Specification)
trol products be selected as candidates for corrosion testing The developed purchase specifications probably would be
(Table 6-8). Testing for each sample/alloy combination would similar to those used by the Pacific Northwest Snowfighters
be conducted at three concentration ranges that represented (PNS) (43), modified by the findings of this project. These
snow and ice control material concentrations expected under would include recommended specifications for product
typical roadway exposure conditions: qualification based on environmental issues and product
properties. This process had been successful and had been
1. As applied, no dilution. For solid products, solutions used as a template by other jurisdictions. One area needing
would be prepared approximating liquid equivalents. For review was the maximum acceptable product concentrations
example, solid NaCl would be prepared as an approximate for certain contaminants developed by PNS that were based
21-percent (w/w) solution. on U.S. EPA water quality criteria and a “rule-of-thumb”

107
Table 6-11. Matrix of significant properties.
Product
Test Parameter
Product A Product B Product C Etc.
Metals
Cyanide
Chloride
Biochemical Oxygen Demand (BOD)
Phosphorus
Nitrogen
Toxicity
Atmospheric Corrosion Attribute
Concrete Corrosion Attributes
Etc.
hundredfold expected dilution after roadway application. It comparison tool. The final output is a product score that can
was not the purpose of the purchase specification to dictate be compared with other materials for a given set of conditions.
regulations—ideally, sufficient information related to prod-
ucts and environmental-related factors would be provided
6.4.4.3 Generate Guidance Document
by the decision tool (Task 10) to allow managers to assess
product risks. A straightforward user manual that explains the rationale
for the use of the software and potential limitations of the
6.4.4 Task 10 (Decision Tool Guidelines) software and provides step-by-step instructions for using the
software and ways of interpreting the results accompanies the
Task 10 involved the development of guidelines that incor- software. This incorporates a guidance tool that will assist
porated all the components of the decision framework (Sec- agencies in developing objectives and applying significance
tion 6.2); the matrix of significant properties (Task 8, Section though weighting to issues such as cost, performance, envi-
6.4.2); and the purchase specification (Task 9, Section 6.4.3). ronmental impacts, and corrosion impacts. This process was
Of these tasks, the development of the decision framework described in detail in Section 6.2.4.2.
into a practical decision tool involved the following tasks.
6.4.4.4 Quality Assurance Program

6.4.4.1 Develop Product Rankings
The quality assurance program takes the general form of a
Products were ranked based on their ability to impair an suite of qualifying tests and submittals, allowing the agency to
area of concern. These included product-specific environ- “qualify” and characterize the product before use. This is fol-
mental attributes known to present possible problems or lowed by procedures to determine suitability, along with sam-
advantages for snow and ice control. It quantified those that pling and random testing to ensure product quality.
reasonably yielded to testing (parameter matrix) and also Quality control test requirements focus on chemical analy-
addressed non-tested attributes and how they should be con- sis parameters. Chemical analysis provides a great deal of
sidered. The project team developed a set list of rankings that valuable information on a product’s potential environmental
incorporated the matrix of significant properties with factors
impact at a price that is not restrictive when compared with
identified in the agency inventory. A score was given on a scale
common corrosion and toxicity testing. However, it is sug-
of 1 to 5, where a higher score designated less potential for
gested that toxicity and corrosion testing be conducted at a
impact. This process was described in detail in Section 6.2.4.1.
reduced frequency or as product concerns dictate.
6.4.4.2 Incorporate Decision Tool into Software 6.4.5 Phase II Reporting

The decision tool was anticipated to consist of a relatively 6.4.5.1 Task 11 (Draft Task 10 Report and Review)
simple software application prepared through use of a com-
mercial spreadsheet program such as Microsoft ExcelTM. The A draft report incorporating the elements of Task 10 (Sec-
tool includes rankings of products and specific conditions tion 6.4.4) was prepared and submitted for review to the
developed through the study work and allows the user to NCHRP and to a project advisory panel established by the
assign weights for a number of categories of criteria that are study team that included industry experts representing DOT
important in selection of the product. This is essentially a maintenance managers from U.S. states and Canadian

108
provinces, the PNS, and AASHTO’s Winter Maintenance Pol- 6.4.5.3 Task 13 (Final Report)
icy Coordinating Committee. Researchers involved in related
A final report incorporating review revisions was submit-
issues were also included. This document provided sufficient
ted to the NCHRP and included
supporting information for reference and use by transporta-
tion managers so that they could make informed decisions
about user-specific requirements. • Documentation of the entire research effort;
• A stand-alone guidance document for the selection of
snow and ice control products that incorporates the
6.4.5.2 Task 12 (Compilation of Review Process)
review;
The results of the review steps were compiled and a tech- • An Executive Summary outlining the research results; and
nical memorandum was submitted to the NCHRP indicating • Recommendations for research required for refining the
how comments were to be addressed where possible. established guidelines.

109
CHAPTER 7
Analytical Program Overview
The analytical test program for NCHRP Project 06-16 was Both solid and liquid snow and ice control materials were
designed to help address gaps in current information and assessed in the analytical test program. Liquid materials were
establish a matrix of significant properties for common snow tested as received. Solid materials were dissolved in deionized
and ice control chemicals in the areas of environmental water at room temperature at concentrations equivalent to
chemistry, aquatic toxicity, concrete corrosion, and atmos- common liquid counterparts (Table 7-2). This approach
pheric corrosion. Producers and suppliers of snow and ice allowed testing of solids and liquids using the same test meth-
control chemicals provided 42 products for testing (Table 7-1). ods and further allowed direct comparison of data between
These materials represented the current U.S. and Canadian matrix types. Where insoluble solid material existed, addi-
market in terms of product type (e.g., NaCl, MgCl2, CaCl2, tional efforts were made to dissolve this material; these efforts
CMA, and KA), manufacturers, regional sources and addi- included extended mixing and further addition of water.
tives. Commercial products continue to evolve, especially Where insoluble material remained, only the liquid compo-
with respect to additives, and variability can exist between nent was tested. Water hydration was unknown for the three
product batches. The data and information presented in this solid CaCl2 products tested and the sample dilution scheme
report provide a snapshot of the materials available for use in underestimated the target concentration by about 25 percent.
2004. Abrasives were not included in the test program because These samples underwent chemical analysis only, and test
material sources are too numerous and varied. Testing efforts data were in the same general range as for liquids. Samples
are also listed in Table 7-1. All 42 products received full chem- were transferred into test bottles and logged into a laboratory
ical analysis. Fifteen materials representing each material type information management system to track sample analysis
and common additive combination were selected for aquatic steps, quality control efforts, and data management tasks. The
toxicity, concrete corrosion, and atmospheric corrosion following chapters describe and assess test results:
assessments. Conducting a large-scale analytical assessment
on a wide range of snow and ice control materials using the • Chapter 8, Chemical Analysis
same test methods provides data that can be directly com- • Chapter 9, Aquatic Toxicity Testing
pared for trends and significant material attributes can be • Chapter 10, Concrete Reinforcing Testing
identified. • Chapter 11, Atmospheric Corrosion Testing

110
Table 7-1. Analytical program testing effort.
Analytical Program
Lab
Base Chemical Matrix Additive1 Aquatic Concrete Atmospheric
ID Chemistry
Toxicity Corrosion Corrosion
1 NaCl Liquid - X
2 NaCl Liquid OMB X X X X
3 NaCl Liquid OMB X X X X
4 NaCl Solid - X
6 NaCl Solid - X X X X
7 NaCl Solid - X
8 NaCl Solid OMB X
9 NaCl Solid OMB X
10 NaCl Solid P X X X X
11 NaCl Solid P X
12 NaCl Solid Unknown X
13 CaCl2 Liquid - X X X X
14 CaCl2 Liquid - X
15 CaCl2 Liquid - X
16 CaCl2 Liquid - X
17 CaCl2 Liquid OMB X X X X
18 CaCl2 Liquid OMB X
19 CaCl2 Liquid OMB X
20 CaCl2 Liquid C, A X X X X
21 CaCl2 Liquid OMB X X X X
22 CaCl2 Solid - X
23 CaCl2 Solid - X
24 CaCl2 Solid - X
25 MgCl2 Liquid - X
26 MgCl2 Liquid - X X X X
27 MgCl2 Liquid - X
28 MgCl2 Liquid OMB X
29 MgCl2 Liquid OMB, TEA X X X X
32 MgCl2 Liquid OMB X X X X
34 Blended Chloride Liquid OMB X
35 Blended Chloride Liquid OMB X
36 Blended Chloride Solid - X
37 Blended Chloride Solid OMB, P X X X X
38 Blended Chloride Solid - X
40 CMA Liquid - X X X X
41 CMAK Liquid - X X X X
42 KA Liquid - X X X X
43 KA + OMB Liquid KA, OMB blend X
44 NAAC Solid - X
1
OMB – Organic Matter from Biomass, P – phosphate, TEA – triethanolamine, C – polyhydroxycarboxylate, A –
polyalkoxylated amine
‘ - ’ indicates not present
Table 7-2. Solid snow and ice control

material dilution scheme.
Solid Material Type Concentration Target
(% w/w)
NaCl 23 %
Blended Chlorides 30 %
CaCl2 30 %
NAAC 30 %

111
CHAPTER 8
Chemical Analysis
The chemical analysis program provided data for chemi- quality criteria, site-specific factors, and assumptions about
cal attributes of environmental concern so that trends and snow and ice control material transport in the field.
significant material attributes could be identified. All 42 Analytical data is presented in milligrams per liter (mg/L),
snow and ice control materials were subjected to the analy- which is the mass-to-liquid equivalent of parts per million.
tical test parameters shown in Table 8-1. Test parameters This convention was selected based on the processes used to
were selected based on a literature assessment of known develop the purchase specification GMC values. These units
potential environmental concerns conducted during Phase I can be converted to the mass-to-mass equivalent of parts per
of this study. Standard methods employed for testing were million (mg/kg) by applying a specific gravity conversion.
referenced from “Standard Methods for the Examination of
Water and Wastewater,” 20th Edition, 1998, American Public
8.1 Organic Matter (BOD and COD)
Health Association, American Waterworks Association,
Water Environment Federation (APHA). Organic matter in snow and ice control materials was
APHA contains well-recognized standard environmental assessed using Chemical Oxygen Demand (COD) and Bio-
test methods commonly used at environmental laboratories chemical Oxygen Demand (BOD) analyses (Table 8-2). COD
throughout the United States and Canada. For some param- measures the oxygen consumed after chemical oxidation;
eters, APHA lists multiple analytical techniques. Those shown BOD measures oxygen consumed after biological oxidation.
in Table 8-1 were selected based on suitability for the test By this convention, COD is greater than or equal to BOD. The
matrix; existing quality test programs, such as the PNS pro- standard BOD test determines oxygen consumed over a 5-day
gram; and the testing laboratory’s experience conducting period at 20°C. If a material is slowly biodegradable, higher
similar testing. values could be obtained for longer test periods.
Chemical test data is tabulated in Tables 8-2 through 8-15. Data shows that the highest organic matter levels are asso-
Data is organized based on analyte level. Detection limits ciated with acetate materials (e.g., CMA and KA). BOD val-
indicate limits of quantitation. APHA tests used for testing are ues ranged from 110,000 to 310,000 mg/L, and COD values
designed for water and wastewater matrixes, whereas snow ranged from 180,000 to 550,000 mg/L for these materials.
and ice control chemicals are concentrated salts. Where dilu- Liquid products containing organic matter from biomass
tions were necessary before testing and where interferences (OMB) as an additive also showed high organic matter levels.
existed, detection limit adjustments were necessary. If an ana- BOD values ranged from 1,000 to 120,000 mg/L, and COD
lyte was detected, it was not necessarily a cause for concern. If values ranged from 6,800 to 230,000 mg/L for these materi-
an analyte was not detected, it was not a cause for concern if als. The wide range in levels probably is related to the con-
the detection limit was below the threshold of adverse effects, centration of additives in the blended material.
but it could be a concern if the detection limit was near or For solid snow and ice control materials containing OMB
above this threshold. additives, BOD values for test solutions were generally less
Guidelines for Maximum Concentration (GMCs) developed than 2,400 mg/L, and COD values were less than 10,000 mg/L.
through the NCHRP Project 06-16 purchase specification will Lower values probably are related to the lower proportion of
be referenced throughout chemical analysis data discussions to OMB product added to solid materials than to liquids.
identify where areas of environmental concern may exist. These A high BOD/COD ratio may indicate more readily
values have been developed based on existing U.S. EPA water biodegradable materials. Conversely, a lower BOD/COD

112
Table 8-1. Chemical analysis and standard methods. 8.3 Phosphorus

Test Parameters Method Based On: Section Both orthophosphate (reported as phosphorus) and total
Organic Matter
• Biochemical Oxygen Demand (BOD) • APHA 5210 B phosphorus were determined (Table 8-4). Orthophosphate is
8.1
• Chemical Oxygen Demand (COD) • APHA 5220 D
highly available for uptake by plants, while total phosphorus
Nitrogen includes complexed and organic phosphorus, which can be
• Nitrate & Nitrite • APHA 4500-NO3- H
• Ammonia • APHA–NH3 F 8.2 mobilized after breakdown through biological processes.
• Total Kjeldahl Nitrogen (TKN) • APHA-Norg B When assessing potential for environmental impairment,
Phosphorus evaluation should be in terms of total phosphorus loading,
• Total Phosphorus • APHA 4500-P B, E
• Total Soluble Phosphorus • APHA 4500-P E
8.3 rather than specific phosphorus fractions.
Three materials containing phosphate-based corrosion
Cyanide (total) • APHA 4500-CN- C
8.4 inhibitors showed high phosphorus levels of 460, 690, and
Heavy Metals 780 mg/L. Materials with OMB additives also showed ele-
• Arsenic, Cadmium, Chromium, • APHA 3125
Copper, Lead, Mercury, Nickel, 8.5 vated phosphorus levels, although the spread in values was
Selenium, Silver, Zinc
broad, ranging from <2 to 690 mg/L. This probably is related
APHA – Standard Methods for the Examination of Water and Wastewater, 20 th
to the proportion of OMB additives in the blend and/or the
Edition, 1998, American Public Health Association, American Waterworks level of phosphorus in the OMB materials themselves. Cer-
Association, Water Environment Federation
tain acetate products (i.e., one KA source and a NAAC mate-
rial) also had somewhat elevated phosphorus levels of 160
ratio may indicate materials that are not biodegradable
and 340 mg/L.
or are slowly biodegradable. The BOD/COD ratio was
From an environmental standpoint, the most restrictive
between 56 and 76 percent for acetate materials and
purchase specification GMC for total phosphorus is 3.5 mg/L.
between 4 and 63 percent for materials containing organic
Values above this may be cause for concern for some aquatic
additives.
systems. The general conclusion for phosphorus is that each
snow and ice control material type provided by a specific ven-
8.2 Nitrogen dor must be tested if phosphorus is to be evaluated as a com-
Analysis was conducted for various forms of nitrogen, ponent of that material.
including ammonia, nitrate and nitrite, and total Kjeldahl
nitrogen (TKN) (Table 8-3). The total nitrogen composite 8.4 Cyanide
(sum of ammonia, nitrate and nitrite, and TKN) is the main
index by which nitrogen content should be evaluated from an Iron-cyanide compounds are commonly added to solid
environmental standpoint. Ammonia and nitrate content is NaCl as an anti-caking agent. Evidence suggests these materi-
important as well because ammonia can be toxic to aquatic als have limited bioavailability under natural field conditions;
life at high concentrations, and nitrite can affect drinking however, readily applicable standard methods for the direct
water supplies. measurement of these compounds do not exist. Therefore, as
Concentrations of total nitrogen showed little evidence of a conservative measure of potential environmental impact,
pattern across the different types of snow and ice control total cyanide content was measured (Table 8-5).
materials. Elevated nitrogen levels showed a direct relation- From an environmental standpoint, the most restrictive
ship to additive type, particularly OMB additives; however, purchase specification GMC for cyanide is 3.5 mg/L. Total
the spread of values for these materials is broad. The same was cyanide was detected in seven samples. Of these, two solid
true for ammonia and nitrate, although only a few materials NaCl products (17 and 18 mg/L) and one with blended chlo-
had high concentrations of these substances. ride product (9.4 mg/L) were above the GMC value.
From an environmental standpoint, the most restrictive
purchase specification GMC for ammonia is 3,000 mg/L and
for total nitrogen is 75 mg/L. None of the materials tested 8.5 Heavy Metals
exceeded the ammonia GMC, whereas about one-half of the 8.5.1 Arsenic
materials exceeded the total nitrogen GMC. Values above the
GMC levels may be cause for concern for some aquatic sys- All samples tested for arsenic were below the reported
tems. The general conclusion for nitrogen is that each snow detection limit of 2 mg/L, and below the most restrictive pur-
and ice control material type provided by a specific vendor chase specification GMC value of 5 mg/L (Table 8-6). The
must be tested if nitrogen is to be evaluated as a component data suggests that arsenic was not a significant attribute of
of that material. environmental concern.

113
Table 8-2. Organic matter (BOD and COD) analysis data.
BOD COD
Lab ID Matrix Material Type Additive 1
(mg/L) (mg/L)
0406064-43 Liquid Acetate OMB 310000 550000
0406064-42 Liquid Acetate 220000 290000
0406064-41 Liquid Acetate 180000 280000
0406064-31 Liquid MgCl2 OMB 36000 250000
OMB,
0406064-29 Liquid MgCl2 TEA 120000 230000
0406064-40 Liquid Acetate 160000 220000
0406064-44 Solid Acetate 110000 180000
0406064-03 Liquid NaCl OMB 86000 160000
0406064-28 Liquid MgCl2 OMB 100000 160000
0406064-02 Liquid NaCl OMB 35000 76000
0406064-17 Liquid CaCl2 OMB 5700 66000
0406064-18 Liquid CaCl2 OMB 21000 43000
0406064-19 Liquid CaCl2 OMB 4200 38000
0406064-35 Liquid Blended Chloride OMB 17000 37000
0406064-20 Liquid CaCl2 C, A 7400 30000
0406064-21 Liquid CaCl2 OMB 2400 30000
0406064-33 Liquid MgCl2 OMB 11000 27000
0406064-25 Liquid CaCl2 4200 21000
0406064-30 Liquid MgCl2 OMB 1100 20000
0406064-32 Liquid MgCl2 OMB 670 19000
0406064-26 Liquid MgCl2 <10 12000
0406064-13 Liquid CaCl2 <10 11000
0406064-14 Liquid CaCl2 <10 11000
0406064-16 Liquid CaCl2 <10 9900
0406064-37 Solid Blended Chloride OMB, P 2400 9500
0406064-27 Liquid MgCl2 <10 8500
0406064-24 Solid CaCl2 <10 8100
0406064-08 Solid NaCl OMB 910 7000
0406064-09 Solid NaCl OMB 940 6800
0406064-15 Liquid CaCl2 <10 6800
0406064-11 Solid NaCl P <10 6200
0406064-23 Solid CaCl2 <10 6200
0406064-01 Liquid NaCl <10 5900
0406064-12 Solid NaCl Unknown <10 5600
0406064-07 Solid NaCl <10 5300
0406064-10 Solid NaCl P <10 4900
0406064-04 Solid NaCl <10 4800
0406064-22 Solid CaCl2 <10 4700
0406064-06 Solid NaCl <10 <4000
0406064-36 Solid Blended Chloride <5 <4000
0406064-38 Solid Blended Chloride <5 <4000
1
Additive: OMB – Organic Matter from Biomass
P – Phosphate
TEA – triethanolamine
C – polyhydroxycarboxylate
A – polyalkoxylated amine

114
Table 8-3. Nitrogen analysis data.
Nitrate
TN-
Ammonia & TKN
Lab ID Matrix Material Type Additive 1 calc
(mg/L) Nitrite (mg/L)
(mg/L)
(mg/L)
0406064-31 Liquid MgCl2 OMB 4 <5.0 8900 8904
0406064-03 Liquid NaCl OMB 150 <5.0 6350 6500
0406064-17 Liquid CaCl2 OMB 2200 <5.0 3860 6060
0406064-29 Liquid MgCl2 OMB, TEA <5 <5.0 3000 3000
0406064-11 Solid NaCl P 1400 <5.0 1550 2950
0406064-18 Liquid CaCl2 OMB 44 94 2040 2178
0406064-33 Liquid MgCl2 OMB 450 <5.0 1290 1740
0406064-34 Liquid Blended Chloride OMB 280 320 933 1533
0406064-10 Solid NaCl P 710 <5.0 699 1409
0406064-19 Liquid CaCl2 OMB 12 <5.0 1220 1232
0406064-21 Liquid CaCl2 OMB 16 260 773 1049
0406064-35 Liquid Blended Chloride OMB 14 <5.0 889 903
0406064-09 Solid NaCl OMB 330 <5.0 502 832
0406064-30 Liquid MgCl2 OMB 35 <5.0 620 655
0406064-32 Liquid MgCl2 OMB 15 <5.0 629 644
0406064-25 Liquid CaCl2 1 <5.0 589 590
0406064-16 Liquid CaCl2 150 <5.0 136 286
0406064-02 Liquid NaCl OMB 50 <5.0 126 176
0406064-20 Liquid CaCl2 C, A 2 <5.0 155 157
0406064-28 Liquid MgCl2 OMB 2 <5.0 89.8 91.8
0406064-26 Liquid MgCl2 2 <5.0 73 75
0406064-37 Solid Blended Chloride OMB, P <0.2 <5.0 53.7 53.7
0406064-38 Solid Blended Chloride 2 <5.0 25.6 27.6
0406064-14 Liquid CaCl2 14 <5.0 11.3 25.3
0406064-24 Solid CaCl2 9 <5.0 7.24 16.24
0406064-41 Liquid Acetate 4 <5.0 10.5 14.5
0406064-01 Liquid NaCl 11 <5.0 3.25 14.25
0406064-40 Liquid Acetate <0.2 <5.0 10 10
0406064-13 Liquid CaCl2 6 <5.0 2.86 8.86
0406064-27 Liquid MgCl2 0.6 <5.0 8.14 8.74
0406064-42 Liquid Acetate 0.3 <5.0 7.6 7.9
0406064-08 Solid NaCl OMB <0.2 <5.0 7.36 7.4
0406064-07 Solid NaCl 3 <5.0 3.65 6.65
0406064-12 Solid NaCl Unknown 3 <5.0 3.47 6.5
0406064-23 Solid CaCl2 3 <5.0 2.29 5.29
0406064-15 Liquid CaCl2 2 <5.0 1.6 3.6
0406064-06 Solid NaCl 2 <5.0 0.83 2.83
0406064-43 Liquid Acetate OMB <0.2 <5.0 <0.50 <5.0
0406064-44 Solid Acetate 0.4 <5.0 <0.50 <5.0
0406064-36 Solid Blended Chloride 1 <5.0 1.54 2.54
0406064-22 Solid CaCl2 1 <5.0 <0.50 1
0406064-04 Solid NaCl 0.2 <5.0 <0.50 0.2
1
P – Phosphate

115
Table 8-4. Phosphorus analysis data.
Ortho- Total
Lab ID Matrix Material Type Additive 1 Phosphate as Phosphorus
P (mg/L) (mg/L)
0406064-10 Solid NaCl P 780 780
0406064-37 Solid Blended Chloride OMB, P 98 690
0406064-33 Liquid MgCl2 OMB 23 540
0406064-11 Solid NaCl P 440 460
0406064-34 Liquid Blended Chloride OMB 5.6 430
0406064-09 Solid NaCl OMB 370 420
0406064-42 Liquid Acetate 340 370
0406064-21 Liquid CaCl2 OMB 230 350
0406064-44 Solid Acetate 160 190
0406064-31 Liquid MgCl2 OMB <20 110
0406064-30 Liquid MgCl2 OMB 22 76
0406064-29 Liquid MgCl2 OMB, TEA <2.0 56
0406064-32 Liquid MgCl2 OMB 34 52
0406064-03 Liquid NaCl OMB <20 50
0406064-17 Liquid CaCl2 OMB <20 44
0406064-28 Liquid MgCl2 OMB 4.8 28
0406064-43 Liquid Acetate OMB <2.0 19
0406064-02 Liquid NaCl OMB <2.0 14
0406064-40 Liquid Acetate <2.0 13
0406064-41 Liquid Acetate 2.5 11
0406064-20 Liquid CaCl2 C, A <10 10
0406064-38 Solid Blended Chloride <2.0 8.7
0406064-25 Liquid CaCl2 <2.0 7.2
0406064-36 Solid Blended Chloride <2.0 6.5
0406064-18 Liquid CaCl2 OMB <2.0 3.4
0406064-26 Liquid MgCl2 <2.0 3.2
0406064-04 Solid NaCl <2.0 2.5
0406064-01 Liquid NaCl <2.0 <2.0
0406064-06 Solid NaCl <2.0 <2.0
0406064-07 Solid NaCl <2.0 <2.0
0406064-08 Solid NaCl OMB <2.0 <2.0
0406064-12 Solid NaCl Unknown <2.0 <2.0
0406064-13 Liquid CaCl2 <2.0 <2.0
0406064-14 Liquid CaCl2 <2.0 <2.0
0406064-15 Liquid CaCl2 <2.0 <2.0
0406064-16 Liquid CaCl2 <2.0 <2.0
0406064-19 Liquid CaCl2 OMB <2.0 <2.0
0406064-22 Solid CaCl2 <2.0 <2.0
0406064-23 Solid CaCl2 <2.0 <2.0
0406064-24 Solid CaCl2 <2.0 <2.0
0406064-27 Liquid MgCl2 <2.0 <2.0
1
P – Phosphate
8.5.2 Cadmium 8.5.3 Chromium

Cadmium was detected in seven samples, although values Chromium was common in most materials and was
were low and near the detection limit of 0.02 mg/L (Table 8- detected in all but two samples. Chromium levels ranged
7). The most restrictive purchase specification GMC for cad- between 0.05 mg/L and 1.4 mg/L (Table 8-8). These levels
mium is 0.12 mg/L for NaCl and KA products and 2.5 mg/L are below the most restrictive purchase specification GMC
for CaCl2, MgCl2, and CMA. The highest three samples (i.e., for chromium of 5.5 mg/L and are not a significant envi-
0.09, 0.06, and 0.04 mg/L) were associated with CaCl2 from ronmental concern.
one source; however, from an environmental protection The data suggests a relationship between chromium levels
standpoint these levels were not a concern. and OMB additive materials. Twelve of the sixteen high

116
Table 8-5. Cyanide analysis data. Table 8-6. Arsenic analysis data.
Cyanide Arsenic
Lab ID Matrix Material Type Additive 1 Lab ID Matrix Material Type Additive 1
(mg/L) (mg/L)
0406064-12 Solid NaCl Unknown 18 0406064-01 Liquid NaCl <2.0
0406064-07 Solid NaCl 17 0406064-02 Liquid NaCl OMB <2.0
0406064-36 Solid Blended Chloride 9.4
0406064-03 Liquid NaCl OMB <2.0
0406064-35 Liquid Blended Chloride OMB 0.94
0406064-04 Solid NaCl <2.0
0406064-40 Liquid Acetate 0.41
0406064-34 Liquid Blended Chloride OMB 0.38 0406064-06 Solid NaCl <2.0
0406064-02 Liquid NaCl OMB 0.32 0406064-07 Solid NaCl <2.0
0406064-01 Liquid NaCl <0.15 0406064-08 Solid NaCl OMB <2.0
0406064-03 Liquid NaCl OMB <0.15 0406064-09 Solid NaCl OMB <2.0
0406064-04 Solid NaCl <0.15 0406064-10 Solid NaCl P <2.0
0406064-06 Solid NaCl <0.15 0406064-11 Solid NaCl P <2.0
0406064-08 Solid NaCl OMB <0.15 0406064-12 Solid NaCl Unknown <2.0
0406064-09 Solid NaCl OMB <0.15 0406064-13 Liquid CaCl2 <2.0
0406064-10 Solid NaCl P <0.15 0406064-14 Liquid CaCl2 <2.0
0406064-11 Solid NaCl P <0.15
0406064-15 Liquid CaCl2 <2.0
0406064-13 Liquid CaCl2 <0.15
0406064-16 Liquid CaCl2 <2.0
0406064-14 Liquid CaCl2 <0.15
0406064-15 Liquid CaCl2 <0.15 0406064-17 Liquid CaCl2 OMB <2.0
0406064-16 Liquid CaCl2 <0.15 0406064-18 Liquid CaCl2 OMB <2.0
0406064-17 Liquid CaCl2 OMB <0.15 0406064-19 Liquid CaCl2 OMB <2.0
0406064-18 Liquid CaCl2 OMB <0.15 0406064-20 Liquid CaCl2 C, A <2.0
0406064-19 Liquid CaCl2 OMB <0.15 0406064-21 Liquid CaCl2 OMB <2.0
0406064-20 Liquid CaCl2 C, A <0.15 0406064-22 Solid CaCl2 <2.0
0406064-21 Liquid CaCl2 OB <0.15 0406064-23 Solid CaCl2 <2.0
0406064-22 Solid CaCl2 <0.15 0406064-24 Solid CaCl2 <2.0
0406064-23 Solid CaCl2 <0.15 0406064-25 Liquid CaCl2 <2.0
0406064-24 Solid CaCl2 <0.15
0406064-26 Liquid MgCl2 <2.0
0406064-25 Liquid CaCl2 <0.15
0406064-27 Liquid MgCl2 <2.0
0406064-26 Liquid MgCl2 <0.15
0406064-27 Liquid MgCl2 <0.15 0406064-28 Liquid MgCl2 OMB <2.0
0406064-28 Liquid MgCl2 OMB <0.15 0406064-29 Liquid MgCl2 OMB, TEA <2.0
0406064-29 Liquid MgCl2 OMB, TEA <0.15 0406064-30 Liquid MgCl2 OMB <2.0
0406064-30 Liquid MgCl2 OMB <0.15 0406064-31 Liquid MgCl2 OMB <2.0
0406064-33 Liquid MgCl2 OMB <0.15 0406064-34 Liquid Blended Chloride OMB <2.0
0406064-37 Solid Blended Chloride OMB, P <0.15 0406064-35 Liquid Blended Chloride OMB <2.0
0406064-38 Solid Blended Chloride <0.15 0406064-36 Solid Blended Chloride <2.0
0406064-41 Liquid Acetate <0.15
0406064-37 Solid Blended Chloride OMB, P <2.0
0406064-38 Solid Blended Chloride <2.0
0406064-43 Liquid Acetate OMB <0.15
0406064-44 Solid Acetate <0.15 0406064-40 Liquid Acetate <2.0
1
P – Phosphate
TEA – triethanolamine 0406064-43 Liquid Acetate OMB <2.0
C – polyhydroxycarboxylate 0406064-44 Solid Acetate <2.0
A – polyalkoxylated amine 1
P – Phosphate
chromium values were associated with materials containing TEA – triethanolamine
OMB materials; however, other OMB-containing materials A – polyalkoxylated amine
had much lower chromium values, suggesting the concentra-
tion depends on the specific material. and additives and higher values probably depend more on
material source than type.
8.5.4 Copper
8.5.5 Lead
Copper was detected in 26 of 42 samples and ranged
between the detection limit of 0.1 mg/L and 0.91 mg/L (Table All samples displayed lead levels below the reported ana-
8-9). The most restrictive purchase specification GMC for lytical detection limit of 0.5 mg/L (Table 8-10). The most
copper is 4.5 mg/L for NaCl and KA products and 15 mg/L for restrictive purchase specification GMC for lead is 1.3 mg/L
CaCl2, MgCl2, and CMA. Test sample levels were not consid- for NaCl and KA products, and 5.5 for CaCl2, MgCl2, and
ered to be of significant environmental concern. The data CMA. Based on the data, lead was not a significant attribute
indicates no obvious trends associated with material types of environmental concern for the samples tested.

117
Table 8-7. Cadmium analysis data.
Cadmium
Lab ID Matrix Material Type Additive 1
(mg/L)
0406064-15 Liquid CaCl2 0.09
0406064-20 Liquid CaCl2 C, A 0.06
0406064-22 Solid CaCl2 0.04
0406064-09 Solid NaCl OMB 0.02
0406064-10 Solid NaCl P 0.02
0406064-11 Solid NaCl P 0.02
0406064-43 Liquid Acetate OMB 0.02
0406064-01 Liquid NaCl <0.02
0406064-04 Solid NaCl <0.02
0406064-06 Solid NaCl <0.02
0406064-07 Solid NaCl <0.02
0406064-08 Solid NaCl OMB <0.02
0406064-12 Solid NaCl Unknown <0.02
0406064-13 Liquid CaCl2 <0.02
0406064-14 Liquid CaCl2 <0.02
0406064-16 Liquid CaCl2 <0.02
0406064-17 Liquid CaCl2 OMB <0.02
0406064-18 Liquid CaCl2 OMB <0.02
0406064-19 Liquid CaCl2 OMB <0.02
0406064-21 Liquid CaCl2 OMB <0.02
0406064-23 Solid CaCl2 <0.02
0406064-24 Solid CaCl2 <0.02
0406064-25 Liquid CaCl2 <0.02
0406064-26 Liquid MgCl2 <0.02
0406064-27 Liquid MgCl2 <0.02
0406064-28 Liquid MgCl2 OMB <0.02
0406064-29 Liquid MgCl2 OMB, TEA <0.02
0406064-30 Liquid MgCl2 OMB <0.02
0406064-31 Liquid MgCl2 OMB <0.02
0406064-32 Liquid MgCl2 OMB <0.02
0406064-33 Liquid MgCl2 OMB <0.02
0406064-34 Liquid Blended Chloride OMB <0.02
0406064-35 Liquid Blended Chloride OMB <0.02
0406064-37 Solid Blended Chloride OMB, P <0.02
0406064-44 Solid Acetate <0.02
1
P – Phosphate
8.5.6 Mercury 8.5.7 Nickel

Of the samples tested, mercury was detected in one Nickel was detected in two samples, although levels were
MgCl2 sample at 0.02 mg/L (Table 8-11). This value was near the detection limi of 1 mg/L (Table 8-12). These levels
well below the most restrictive purchase specification GMC were well below the most restrictive purchase specification
value of 0.39 mg/L. Based on the data, mercury was not a GMC value of 26 mg/L. Based on the data, nickel was not a
significant attribute of environmental concern for the sam- significant attribute of environmental concern for the sam-
ples tested. ples tested.

118
Table 8-8. Chromium analysis data. Table 8-9. Copper analysis data.
Chromium Copper
(mg/L) (mg/L)
0406064-43 Liquid Acetate OMB 1.40 0406064-44 Solid Acetate 0.91
0406064-29 Liquid MgCl2 OMB, TEA 1.10 0406064-17 Liquid CaCl2 OMB 0.68
0406064-28 Liquid MgCl2 OMB 0.91
0406064-44 Solid Acetate 0.76
0406064-16 Liquid CaCl2 0.51
0406064-18 Liquid CaCl2 OMB 0.68
0406064-17 Liquid CaCl2 OMB 0.64 0406064-32 Liquid MgCl2 OMB 0.35
0406064-33 Liquid MgCl2 OMB 0.58 0406064-37 Solid Blended Chloride OMB, P 0.34
0406064-26 Liquid MgCl2 0.55 0406064-34 Liquid Blended Chloride OMB 0.31
0406064-30 Liquid MgCl2 OMB 0.52 0406064-36 Solid Blended Chloride 0.31
0406064-25 Liquid CaCl2 0.46 0406064-40 Liquid Acetate 0.31
0406064-19 Liquid CaCl2 OMB 0.45 0406064-25 Liquid CaCl2 0.28
0406064-03 Liquid NaCl OMB 0.43 0406064-21 Liquid CaCl2 OMB 0.26
0406064-20 Liquid CaCl2 C, A 0.40 0406064-30 Liquid MgCl2 OMB 0.23
0406064-40 Liquid Acetate 0.40 0406064-41 Liquid Acetate 0.23
0406064-32 Liquid MgCl2 OMB 0.37 0406064-18 Liquid CaCl2 OMB 0.22
0406064-27 Liquid MgCl2 0.30 0406064-06 Solid NaCl 0.21
0406064-20 Liquid CaCl2 C, A 0.20
0406064-22 Solid CaCl2 0.28
0406064-26 Liquid MgCl2 0.20
0406064-10 Solid NaCl P 0.27
OMB,
0406064-29 Liquid MgCl2 TEA 0.20
0406064-15 Liquid CaCl2 0.26
0406064-14 Liquid CaCl2 0.25
0406064-35 Liquid Blended Chloride OMB 0.20
0406064-35 Liquid Blended Chloride OMB 0.25 0406064-09 Solid NaCl OMB 0.17
0406064-06 Solid NaCl 0.24 0406064-27 Liquid MgCl2 0.17
0406064-34 Liquid Blended Chloride OMB 0.24 0406064-28 Liquid MgCl2 OMB 0.16
0406064-13 Liquid CaCl2 OMB 0.23 0406064-19 Liquid CaCl2 OMB 0.12
0406064-09 Solid NaCl 0.21 0406064-38 Solid Blended Chloride 0.12
0406064-38 Solid Blended Chloride 0.21 0406064-07 Solid NaCl 0.10
0406064-02 Liquid NaCl OB 0.20 0406064-01 Liquid NaCl <0.10
0406064-23 Solid CaCl2 0.19 0406064-02 Liquid NaCl OMB <0.10
0406064-24 Solid CaCl2 0.19 0406064-03 Liquid NaCl OMB <0.10
0406064-12 Solid NaCl Unknown 0.17 0406064-04 Solid NaCl <0.10
0406064-16 Liquid CaCl2 0.14 0406064-08 Solid NaCl OMB <0.10
0406064-07 Solid NaCl 0.13
0406064-10 Solid NaCl P <0.10
0406064-08 Solid NaCl OMB 0.13
0406064-11 Solid NaCl P <0.10
0406064-04 Solid NaCl 0.12
0406064-11 Solid NaCl P 0.12
0406064-12 Solid NaCl Unknown <0.10
0406064-37 Solid Blended Chloride OMB, P 0.11 0406064-13 Liquid CaCl2 <0.10
0406064-21 Liquid CaCl2 OMB 0.09 0406064-14 Liquid CaCl2 <0.10
0406064-01 Liquid NaCl <0.05 0406064-15 Liquid CaCl2 <0.10
0406064-36 Solid Blended Chloride <0.05 0406064-22 Solid CaCl2 <0.10
1 0406064-23 Solid CaCl2 <0.10
P – Phosphate
0406064-24 Solid CaCl2 <0.10
TEA – triethanolamine 0406064-42 Liquid Acetate <0.10
C – polyhydroxycarboxylate 0406064-43 Liquid Acetate OMB <0.10
A – polyalkoxylated amine 1
P – Phosphate
8.5.8 Selenium 8.5.9 Silver

Selenium was detected in 10 samples, at levels between 1 Silver was detected in one sample at the detection limit of
and 1.6 mg/L (Table 8-13). These levels were below the most 0.1 mg/L (Table 8-14). The most restrictive purchase specifi-
restrictive purchase specification GMC value of 2.5 mg/L cation GMC for silver is 0.32 mg/L for NaCl and KA products
and were thus not considered a significant environmental and 3.5 mg/L for CaCl2, MgCl2, and CMA. Based on the data,
concern. Some caution should be exercised with the conclu- silver was not a significant attribute of environmental con-
sion in that the detection limit capability for selenium was cern for the samples tested; however, detection limit capabil-
near limits for environmental concern. ity for silver was near limits for environmental concern.

119
Table 8-10. Lead analysis data. Table 8-11. Mercury analysis data.
Lead Mercury
(mg/L) (mg/L)
0406064-01 Liquid NaCl <0.5 0406064-31 Liquid MgCl2 OBM 0.02
0406064-02 Liquid NaCl OMB <0.5 0406064-01 Liquid NaCl <0.01
0406064-03 Liquid NaCl OMB <0.5 0406064-02 Liquid NaCl OBM <0.01
0406064-04 Solid NaCl <0.5 0406064-03 Liquid NaCl OBM <0.01
0406064-06 Solid NaCl <0.5 0406064-04 Solid NaCl <0.01
0406064-07 Solid NaCl <0.5 0406064-06 Solid NaCl <0.01
0406064-08 Solid NaCl OMB <0.5 0406064-07 Solid NaCl <0.01
0406064-09 Solid NaCl OMB <0.5 0406064-08 Solid NaCl OBM <0.01
0406064-10 Solid NaCl P <0.5 0406064-09 Solid NaCl OBM <0.01
0406064-11 Solid NaCl P <0.5 0406064-10 Solid NaCl P <0.01
0406064-12 Solid NaCl Unknown <0.5 0406064-11 Solid NaCl P <0.01
0406064-13 Liquid CaCl2 <0.5 0406064-12 Solid NaCl Unknown <0.01
0406064-14 Liquid CaCl2 <0.5 0406064-13 Liquid CaCl2 <0.01
0406064-17 Liquid CaCl2 OMB <0.5
0406064-16 Liquid CaCl2 <0.01
0406064-17 Liquid CaCl2 OBM <0.01
0406064-18 Liquid CaCl2 OBM <0.01
0406064-20 Liquid CaCl2 C, A <0.5
0406064-19 Liquid CaCl2 OBM <0.01
0406064-20 Liquid CaCl2 C, A <0.01
0406064-22 Solid CaCl2 <0.5
0406064-21 Liquid CaCl2 OBM <0.01
0406064-23 Solid CaCl2 <0.5
0406064-22 Solid CaCl2 <0.01
0406064-24 Solid CaCl2 <0.5
0406064-23 Solid CaCl2 <0.01
0406064-25 Liquid CaCl2 <0.5
0406064-24 Solid CaCl2 <0.01
0406064-26 Liquid MgCl2 <0.5
0406064-25 Liquid CaCl2 <0.01
0406064-27 Liquid MgCl2 <0.5
0406064-26 Liquid MgCl2 <0.01
0406064-28 Liquid MgCl2 OMB <0.5
0406064-29 Liquid MgCl2 OMB, TEA <0.5 0406064-27 Liquid MgCl2 <0.01
0406064-30 Liquid MgCl2 OMB <0.5 0406064-28 Liquid MgCl2 OBM <0.01
0406064-31 Liquid MgCl2 OMB <0.5 0406064-29 Liquid MgCl2 OBM, TEA <0.01
0406064-34 Liquid Blended Chloride OMB <0.5 0406064-33 Liquid MgCl2 OBM <0.01
0406064-35 Liquid Blended Chloride OMB <0.5 0406064-34 Liquid Blended Chloride OBM <0.01
0406064-36 Solid Blended Chloride <0.5 0406064-35 Liquid Blended Chloride OBM <0.01
0406064-37 Solid Blended Chloride OMB, P <0.5 0406064-36 Solid Blended Chloride <0.01
0406064-38 Solid Blended Chloride <0.5 0406064-37 Solid Blended Chloride OBM, P <0.01
0406064-40 Liquid Acetate <0.5 0406064-38 Solid Blended Chloride <0.01
0406064-41 Liquid Acetate <0.5 0406064-40 Liquid Acetate <0.01
0406064-42 Liquid Acetate <0.5 0406064-41 Liquid Acetate <0.01
0406064-43 Liquid Acetate OMB <0.5 0406064-42 Liquid Acetate <0.01
0406064-44 Solid Acetate <0.5 0406064-43 Liquid Acetate OBM <0.01
1 0406064-44 Solid Acetate <0.01
1
P – Phosphate Additive: OMB – Organic Matter from Biomass
TEA – triethanolamine P – Phosphate
C – polyhydroxycarboxylate TEA – triethanolamine
A – polyalkoxylated amine C – polyhydroxycarboxylate
8.5.10 Zinc 8.6 Chemical Analysis Summary

Of the materials tested, zinc was detected in all but five The following general conclusions can be made from the
samples and levels ranged from the detection limit of 1 mg/L chemical analysis program:
to 20.6 mg/L (Table 8-15). The most restrictive purchase • High organic matter concentrations are associated with
specification GMC for zinc is 59 mg/L for NaCl and KA acetate materials and with materials containing OMB. For
products and 190 mg/L for CaCl2, MgCl2, and CMA. Based the latter, levels depended on the proportion of OMB in the
on the data, zinc was not a significant attribute of environ- blend. Solids with OMB generally had much lower organic
mental concern. matter levels than liquid counterparts.

120
Table 8-12. Nickel analysis data. Table 8-13. Selenium analysis data.
Nickel Selenium
(mg/L) (mg/L)
0406064-31 Liquid MgCl2 OMB 3 0406064-25 Liquid CaCl2 1.6
0406064-03 Liquid NaCl OMB 1 0406064-19 Liquid CaCl2 OMB 1.5
0406064-01 Liquid NaCl <1 0406064-26 Liquid MgCl2 1.3
0406064-02 Liquid NaCl OMB <1 0406064-14 Liquid CaCl2 1.2
0406064-04 Solid NaCl <1 0406064-32 Liquid MgCl2 OMB 1.2
0406064-06 Solid NaCl <1 0406064-34 Liquid Blended Chloride OMB 1.2
0406064-07 Solid NaCl <1 0406064-16 Liquid CaCl2 1.1
0406064-08 Solid NaCl OMB <1 0406064-27 Liquid MgCl2 1
0406064-09 Solid NaCl OMB <1 0406064-29 Liquid MgCl2 OMB, TEA 1
0406064-10 Solid NaCl P <1 0406064-30 Liquid MgCl2 OMB 1
0406064-11 Solid NaCl P <1 0406064-01 Liquid NaCl <1.0
0406064-12 Solid NaCl Unknown <1 0406064-02 Liquid NaCl OMB <1.0
0406064-13 Liquid CaCl2 <1 0406064-03 Liquid NaCl OMB <1.0
0406064-14 Liquid CaCl2 <1 0406064-04 Solid NaCl <1.0
0406064-17 Liquid CaCl2 OMB <1 0406064-08 Solid NaCl OMB <1.0
0406064-18 Liquid CaCl2 OMB <1 0406064-09 Solid NaCl OMB <1.0
0406064-19 Liquid CaCl2 OMB <1 0406064-10 Solid NaCl P <1.0
0406064-20 Liquid CaCl2 C, A <1 0406064-11 Solid NaCl P <1.0
0406064-21 Liquid CaCl2 OMB <1 0406064-12 Solid NaCl Unknown <1.0
0406064-22 Solid CaCl2 <1 0406064-13 Liquid CaCl2 <1.0
0406064-23 Solid CaCl2 <1 0406064-15 Liquid CaCl2 <1.0
0406064-24 Solid CaCl2 <1 0406064-17 Liquid CaCl2 OMB <1.0
0406064-25 Liquid CaCl2 <1 0406064-18 Liquid CaCl2 OMB <1.0
0406064-26 Liquid MgCl2 <1 0406064-20 Liquid CaCl2 C, A <1.0
0406064-27 Liquid MgCl2 <1 0406064-21 Liquid CaCl2 OMB <1.0
0406064-28 Liquid MgCl2 OMB <1 0406064-22 Solid CaCl2 <1.0
0406064-29 Liquid MgCl2 OMB, TEA <1 0406064-23 Solid CaCl2 <1.0
0406064-30 Liquid MgCl2 OMB <1 0406064-24 Solid CaCl2 <1.0
0406064-32 Liquid MgCl2 OMB <1 0406064-28 Liquid MgCl2 OMB <1.0
0406064-33 Liquid MgCl2 OMB <1 0406064-31 Liquid MgCl2 OMB <1.0
0406064-34 Liquid Blended Chloride OMB <1 0406064-33 Liquid MgCl2 OMB <1.0
0406064-35 Liquid Blended Chloride OMB <1 0406064-35 Liquid Blended Chloride OMB <1.0
0406064-36 Solid Blended Chloride <1 0406064-36 Solid Blended Chloride <1.0
0406064-37 Solid Blended Chloride OMB, P <1 0406064-37 Solid Blended Chloride OMB, P <1.0
0406064-38 Solid Blended Chloride <1 0406064-38 Solid Blended Chloride <1.0
0406064-40 Liquid Acetate <1 0406064-40 Liquid Acetate <1.0
0406064-43 Liquid Acetate OMB <1 0406064-43 Liquid Acetate OMB <1.0
0406064-44 Solid Acetate <1 0406064-44 Solid Acetate <1.0
1 1
Additive: OMB – Organic Matter from Biomass Additive: OMB – Organic Matter from Biomass
P – Phosphate P – Phosphate
TEA – triethanolamine TEA – triethanolamine
C – polyhydroxycarboxylate C – polyhydroxycarboxylate
A – polyalkoxylated amine A – polyalkoxylated amine
• Nutrient (nitrogen and phosphorus) levels found • Cyanide could be an environmental concern for materials
in many of the materials tested can be considered a poten- such as solid NaCl that routinely contain iron-cyanide
tial environmental concern. Nutrient levels had a strong anti-caking agents. More research and information is nec-
relationship to the presence of OMB materials in the essary to assess the bio-availability of iron-cyanide materi-
product blend. Higher nutrient levels were also present in als under field conditions conclusively.
other material types. The range in nutrient levels was • Overall, metals levels were below levels considered to be a
broad, suggesting that each material should be assessed significant environmental concern. The literature does
independently if nutrients are to be evaluated. indicate that higher metals levels have routinely been

121
Table 8-14. Silver analysis data. Table 8-15. Zinc analysis data.
Silver Zinc
(mg/L) (mg/L)
0406064-01 Liquid NaCl <0.1 0406064-16 Liquid CaCl2 20.6
0406064-02 Liquid NaCl OMB <0.1 0406064-31 Liquid MgCl2 OMB 15.4
0406064-03 Liquid NaCl OMB <0.1 0406064-03 Liquid NaCl OMB 6.6
0406064-04 Solid NaCl <0.1 0406064-35 Liquid Blended Chloride OMB 5
0406064-06 Solid NaCl <0.1 0406064-33 Liquid MgCl2 OMB 4.7
0406064-07 Solid NaCl <0.1 0406064-17 Liquid CaCl2 OMB 2.8
0406064-08 Solid NaCl OMB <0.1 0406064-18 Liquid CaCl2 OMB 2.6
0406064-09 Solid NaCl OMB <0.1 0406064-30 Liquid MgCl2 OMB 2.6
0406064-10 Solid NaCl P <0.1 0406064-32 Liquid MgCl2 OMB 2.4
0406064-11 Solid NaCl P <0.1 0406064-06 Solid NaCl 2.3
0406064-12 Solid NaCl Unknown <0.1 0406064-26 Liquid MgCl2 2.3
0406064-13 Liquid CaCl2 <0.1 0406064-20 Liquid CaCl2 C, A 2.2
0406064-14 Liquid CaCl2 <0.1 0406064-41 Liquid Acetate 2.1
0406064-15 Liquid CaCl2 <0.1 0406064-37 Solid Blended Chloride OMB, P 2
0406064-16 Liquid CaCl2 <0.1 0406064-40 Liquid Acetate 2
0406064-17 Liquid CaCl2 OMB <0.1 0406064-28 Liquid MgCl2 OMB 1.9
0406064-18 Liquid CaCl2 OMB <0.1 0406064-34 Liquid Blended Chloride OMB 1.9
0406064-19 Liquid CaCl2 OMB <0.1 0406064-25 Liquid CaCl2 1.8
0406064-20 Liquid CaCl2 C, A <0.1 0406064-09 Solid NaCl OMB 1.7
0406064-21 Liquid CaCl2 OMB <0.1 0406064-29 Liquid MgCl2 OMB, TEA 1.7
0406064-22 Solid CaCl2 <0.1 0406064-44 Solid Acetate 1.7
0406064-23 Solid CaCl2 <0.1 0406064-11 Solid NaCl P 1.6
0406064-24 Solid CaCl2 <0.1 0406064-27 Liquid MgCl2 1.6
0406064-25 Liquid CaCl2 <0.1 0406064-38 Solid Blended Chloride 1.6
0406064-26 Liquid MgCl2 <0.1 0406064-43 Liquid Acetate OMB 1.6
0406064-27 Liquid MgCl2 <0.1 0406064-01 Liquid NaCl 1.5
0406064-28 Liquid MgCl2 OMB <0.1 0406064-13 Liquid CaCl2 1.4
0406064-30 Liquid MgCl2 OMB <0.1 0406064-15 Liquid CaCl2 1.3
0406064-31 Liquid MgCl2 OMB <0.1 0406064-19 Liquid CaCl2 OMB 1.3
0406064-32 Liquid MgCl2 OMB <0.1 0406064-04 Solid NaCl 1.2
0406064-33 Liquid MgCl2 OMB <0.1 0406064-08 Solid NaCl OMB 1.2
0406064-34 Liquid Blended Chloride OMB <0.1 0406064-10 Solid NaCl P 1.1
0406064-35 Liquid Blended Chloride OMB <0.1 0406064-12 Solid NaCl Unknown 1.1
0406064-36 Solid Blended Chloride <0.1 0406064-22 Solid CaCl2 1.1
0406064-37 Solid Blended Chloride OMB, P <0.1 0406064-42 Liquid Acetate 1.1
0406064-38 Solid Blended Chloride <0.1 0406064-02 Liquid NaCl OMB 1
0406064-40 Liquid Acetate <0.1 0406064-14 Liquid CaCl2 1
0406064-41 Liquid Acetate <0.1 0406064-07 Solid NaCl <1.0
0406064-42 Liquid Acetate <0.1 0406064-21 Liquid CaCl2 OMB <1.0
0406064-43 Liquid Acetate OMB <0.1 0406064-23 Solid CaCl2 <1.0
0406064-44 Solid Acetate <0.1 0406064-24 Solid CaCl2 <1.0
0406064-29 Liquid MgCl2 OMB, TEA 0.1 0406064-36 Solid Blended Chloride <1.0
1 1
Additive: OMB – Organic Matter from Biomass Additive: OMB – Organic Matter from Biomass
P – Phosphate P – Phosphate
TEA – triethanolamine TEA – triethanolamine
C – polyhydroxycarboxylate C – polyhydroxycarboxylate
A – polyalkoxylated amine A – polyalkoxylated amine
observed in snow and ice control materials. Continual suggest that it is important to continue to monitor chemi-
monitoring of snow and ice materials received in the field cal attributes of snow and ice control materials through a
over a longer time would provide a more realistic assess- quality assurance plan. The Purchase Specification pro-
ment of metals levels. vided in the attached Guidebook provides further guidance
• The variability in the data between products and the gen- on quality monitoring plans.
eral lack of strong trends outside of those already listed

122
CHAPTER 9
Aquatic Toxicity Testing
Aquatic toxicity testing was conducted on 15 snow and ice tolerated as well as other cations associated with snow and ice
control products following U.S. EPA standard test methods control materials. Verification will, however, require further
for chronic exposure to various levels of aquatic biota, research, given that only two of the materials tested contained
including vertebrates, invertebrates, and algae (Table 9-1). potassium as a major ingredient. Another interesting aspect
Chronic exposure tests were selected over more commonly of Table 9-2 is the relatively small separation between thresh-
used acute test procedures because chronic exposure tests olds involving acute mortality and chronic suppression of
provide more information related to potential long-term growth.
effects. Both the U.S. EPA and Environment Canada recog- Selenastrum capricornutum, a green alga often used as a test
nize these tests to be sufficient in length to evaluate chronic organism, proved to be substantially more sensitive than
effects. For fathead minnows, by conducting a 7-day test, one fathead minnows to most snow and ice control materials. For
can also determine the 96-hour LC50, as well as a 7-day IC50 Selenastrum, there was no detectable difference between
and IC25, on growth. Aquatic toxicity test data is presented in potassium-containing materials and other materials, in
Tables 9-2 through 9-4. The data shows the concentration of contrast to the results for fathead minnows.
material required to achieve a desired endpoint (e.g., LC50 Ceriodaphnia dubia, a water flea known to have low toxi-
and IC50) over the test duration. Lower values represent cological tolerance, showed tolerances lower than those
higher toxicity levels (i.e., less material is required to achieve of Selenastrum and much lower than those of the fathead
the test endpoint). minnow. Ceriodaphnia, like Selenastrum, was not especially
The extensive toxicity testing reported in Tables 9-2 sensitive to potassium-containing snow and ice control
through 9-4 provided data for comparisons across snow and materials.
ice control material types, comparisons for a given material It might be expected that snow and ice control materials of
type, and comparisons of sensitivity among different kinds of a certain class (e.g., NaCl) would show essentially the same
organisms. In addition, the tables can be used to show what thresholds of tolerance for a given kind of organism. This is
fraction of the snow and ice control material might be con- not always the case, as shown by the test data. Differences
sidered environmentally damaging at a given dilution ratio between materials are partly a matter of variance in the test-
on the roadway (e.g., 500:1). ing methods, but undoubtedly extend beyond this to include
The fathead minnow was more tolerant than Ceriodaphnia effects of additives and contaminants that may differ from
and Selenastrum, especially for chloride-based snow and ice one type of material to another.
control materials. Although there is considerable variation From the viewpoint of applications in the field, it is
among chloride materials, there is no evidence that the asso- important to know how the tolerance thresholds shown in
ciated cation (e.g., sodium, calcium, or magnesium) makes a Tables 9-2 through 9-4 compare with the expected dosage
difference in toxicity. CMA falls within the same range of that organisms would receive in the field. Although the con-
thresholds as chloride-based deicers. CMAK and KA, how- centrations of snow and ice control materials at the time of
ever, produce surprisingly low toxicity thresholds for fathead application vary, the roadway dilution of 500:1 can be
minnows. This may be an important finding, given that low applied to a solid. The threshold for solids is shown in
toxicity thresholds (i.e., higher toxicity) for CMAK and KA Figure 9-1. 500:1 dilutions for liquids would produce a line
have not been reported in the literature, nor would they be that falls below the one for solids. Materials with toxicity
expected. The implication of the data is that potassium is not thresholds above the 500:1 dilution line would, according to

123

Fathead Minnow 7-day Survival and Growth Test (will also U.S. EPA, 1994, Short-Term Methods for Estimating the Chronic Toxicity
include acute toxicity data for Fathead Minnow 96-hour Survival of Effluents and Receiving Water to Freshwater Organisms, 3 rd Edition,
Test) EPA/600/4-91/002
Ceriodaphnia dubia 3-brood Survival and Reproduction Test U.S. EPA, 1994, Short-Term Methods for Estimating the Chronic Toxicity
of Effluents and Receiving Water to Freshwater Organisms, 3 rd Edition,
EPA/600/4-91/002
Selenastrum capricornutum 96-hour (US EPA) Growth Test U.S. EPA, 1994, Short-Term Methods for Estimating the Chronic Toxicity
of Effluents and Receiving Water to Freshwater Organisms, 3rd Edition,
EPA/600/4-91/002
Table 9-2. Fathead minnow toxicity testing endpoints (mg/L) (95% confidence limits).
Endpoint Concentration (mg/ L)

Lab Base 1
Matrix Additive Fathead Minno w 96-h Fathead Minno w 7-d Fathead Minno w 7-d Fathead Minno w 7-d
ID Chemical
LC502 LC502 IC25 (Gro wth) 3 IC50 (Gro wth) 4
2 NaCl Liquid OMB 18,415 (16,951, 20,202) 15,324 (3,201, 20,958) 8,574 (3,810, 15,349) 13,545 (N/A)
3 Liquid OMB 11,691 (10,465, 12,822) 10,927 (9769, 12,149) 5 9 2 3 (5 0 2 6 , 6 6 8 1 ) 8663 (7323, 10,303)
6 Solid - 8 2 8 7 (8 0 0 2 , 8 5 9 3 ) 7598 (7057, 7914) 3 3 5 2 (6 2 0 , 5 1 2 6 ) 5 6 0 6 (4 2 1 3 , 6 4 3 5 )
10 Solid P 4 5 9 8 (4 4 4 0 , 4 8 4 1 ) 4 3 9 8 (4 0 7 9 , 4 6 5 9 ) 2071 (989,3375) 3857 (N/A)
13 CaCl2 Liquid - 13,893 (10,672, 16,368) 10,656 (8246, 13,720) 7 7 0 7 (6 8 0 5 , 8 7 5 2 ) 10574 (9457, 12,278)
17 Liquid OMB 11,609 (10,779, 13,378) 10,971 (10,486, 12,002) 5 8 1 0 (4 5 4 8 , 6 5 8 1 ) 8 4 3 4 (7 5 9 9 , 9 0 0 4 )
20 Liquid C, A 79.9 (N/A) 48.6 (36.9, 98.5) >25 >25
21 Liquid OMB 11002 (9648, 12,666) 9 3 1 5 (8 7 8 7 , 9 9 0 4 ) 7 1 9 5 (6 2 1 2 , 8 2 2 1 ) 9 9 4 6 (8 6 3 7 , 1 1 3 4 0 )
26 Mg Cl2 Liquid - 3 6 9 1 (3 4 8 0 , 3 8 8 3 ) 3 6 3 9 (3 4 1 7 , 3 8 5 2 ) 2631 (N/A) >3038 (N/A)

29 Liquid OMB, TEA 6 0 0 0 (4 6 9 0 , 6 5 7 5 ) 5 6 6 4 (4 3 1 6 , 6 7 5 5 ) >2400 >2400
32 Liquid OMB 3 5 2 8 (3 3 0 6 , 3 8 0 3 ) 3 3 2 6 (3 0 2 3 , 3 6 5 0 ) 1 8 8 7 (1 4 9 4 , 2 3 9 1 ) 3045 (N/A)
37 Blended Chlor ide Solid OMB, P 7692 (6924, 10,377) 6 4 0 2 (5 7 3 5 , 7 1 2 7 ) 3 6 6 6 (2 8 9 7 , 4 4 4 1 ) 5388 (N/A)
40 CMA Liquid - 22,093 (15,182, 27,767) 12,586 (11,771, 13,283) 6 3 3 7 (4 5 4 1 , 7 7 4 5 ) 9493 (7022, 12,152)
41 CMAK Liquid - 2 7 1 1 (2 5 9 3 , 2 9 3 0 ) 2 5 9 6 (2 4 5 9 , 2 8 9 1 ) 1 3 0 8 (8 8 4 , 1 7 6 2 ) 2143 (N/A)
42 KA Liquid - 1 0 3 3 (9 6 3 , 1 0 8 5 ) 9 0 7 (8 3 5 , 9 2 6 ) >788 >788
1
OMB – Org anic Matter from Biomass, P – Phosphate, TEA – triethanolam ine, C – polyhydroxycarboxylate, A – polyalkoxylated amine
2
LC50 = concentration at which 50% of the population exposed would surviv e
3
IC25 = concentration which would cau se a 25% inhibition in repro duction or g rowth
4
IC50 = concentration which would cau se a 50% inhibition in repro duction or g rowth
‘ - ’ indicates not p resent

124
Table 9-3. Selenastrum capricornutum toxicity testing endpoints (mg/L) (95% confidence limits).
Lab Base Endpoint Concentration (mg/L)

Matrix Additive1
ID Chemical Selenastrum 96-h IC252 (Growth) Selenastrum 96-h IC503 (Growth)
2 NaCl Liquid OMB 224 (206, 253) 2100 (1335, 2741)

3 Liquid OMB <747 <747
6 Solid - 1439 (764, 1715) 2956 (1302, 4115)

10 Solid P 1306 (867, 1417) 1990 (1694, 2194)
13 CaCl2 Liquid - 1295 (803, 9814) 18993 (9707, 27314)

17 Liquid OMB 928 (242, 1111) 1489 (1148, 4123)
20 Liquid C, A 23.9 (6.5, 49.4) 50.5 (17.1, 67.5)

21 Liquid OMB 569 (374, 889) 1146 (910, 1447)
26 MgCl2 Liquid - 1031 (0, 3371) 3365 (370, 5424)
29 Liquid OMB, TEA 623 (0, 964) 1206 (907, 1424)

32 Liquid OMB 941 (69, 1158) 1593 (1115, 2812)
37 Blended Chloride Solid OMB, P 1184 (697, 1493) 1892 (1557, 2075)
40 CMA Liquid - 1001 (57.8, 2273) 3443 (1666, 5346)

41 CMAK Liquid - 1326 (0, 1920) 2480 (1884, 2865)
42 KA Liquid - 2373 (1640, 2660) 3535 (3128, 3789)
1
OB – Organic Matter from Biomass, P – Phosphate, TEA – triethanolamine, C – polyhydroxycarboxylate, A – polyalkoxylated amine
2
IC25 = concentration which would cause a 25% inhibition in reproduction or growth
3
the toxicity data, be unlikely to harm the representative melt following application probably would not extend for a
organisms used in these tests. long period. The suitability of acute or chronic criteria would
Figure 9-1 makes clear that a dilution to 500:1 would be depend on the location and type of application.
insufficient to prevent nonlethal chronic effects of most snow Although it appears likely that the most sensitive organ-
and ice control materials on the most sensitive organisms isms would be affected in some way at dilutions that could
(i.e., Selenastrum and Ceriodaphnia), although other organ- be expected at the margin of a roadway for many snow and
isms might not be affected at these concentrations (i.e., fat- ice control materials, it is also true that dilutions greater
head minnow). Dilution at the margin of a roadway varies, than 500:1 can be expected within a relatively short distance
but 500:1 would be close to the amount of dilution for many beyond the roadway. Therefore, the toxicity data suggest
applications. that there could be an impact zone relatively close to the
Figure 9-2 shows that these lethal effects under acute expo- roadway, bounded by dilutions that offset acute or chronic
sures at a dilution of 500:1 would affect sensitive organisms toxicities close to the roadway. Dilution rates characteristic
for a number of snow and ice control materials but not for for a specific site could be determined easily from domi-
others. Acute exposure might be the most meaningful guide nant marker substances in a given type of material (e.g.,
for field applications, given that transport of materials with chloride).

125
Table 9-4. Ceriodaphnia dubia toxicity testing endpoints (mg/L) (95% confidence limits).
Endpoint Concentration (mg/L)

Lab Base
Matrix Additive1 3-brood IC253 3-brood IC504
ID Chemical 3-brood LC50 2
(Reproduction) (Reproduction)
2 NaCl Liquid OMB 3889 (3461, 4330) 564 (464, 809) 1659 (1325, 2098)
3 Liquid OMB 180 (160. 253) 107 (90.6, 132) 159 (130, 215)
6 Solid - 2349 (2187, 2555) 659 (371, 1153) 1598 (738, 2204)
10 Solid P 2353 (2130, 3118) 1082 (829, 1227) 1561 (1356, 1777)
13 CaCl2 Liquid - 2937 (1872, 3740) <570 1885 (1535, 2196)
17 Liquid OMB 997 (724, 1300) 246 (184, 295) 367 (327, 400)
20 Liquid C, A 380 (252, 440) 76.0 (55.6, 118.6) 170 (132, 232)
21 Liquid OMB 2040 (1248, 3326) 258 (171, 362) 900 (754, 1022)
26 MgCl2 Liquid - 3202 (3013, 4234) 272 (175, 439) 1250 (526, 1885)
29 Liquid OMB, TEA 1983 (1525, 2529) 697 (374, 862) 1306 (1076, 1546)
32 Liquid OMB 3358 (3149, 4025) 442 (321, 555) 1804, 1283, 2131)
37 Blended Chloride Solid OMB, P 372 (118, 754) <118 <118
40 CMA Liquid - 6311 (414, 8651) 305 (73.8, 723) 1030 (660, 1236)
41 CMAK Liquid - 4961 (N/A) 435 (285, 597) 1033 (718, 1614)
42 KA Liquid - 1994 (N/A) <25 <25
1
OMB – Organic Matter from Biomass, P – Phosphate, TEA – triethanolamine, C – polyhydroxycarboxylate, A – polyalkoxylated amine
2
LC50 = concentration at which 50% of the population exposed would survive
3
4

126
Figure 9-1. Sub-lethal (chronic) aquatic toxicity.

Summary of the chronic sub-lethal bioassay data on fathead minnow, Selenastrum, and Ceriodaphnia. For each
species, each source is assigned a rank and the threshold concentrations corresponding to the ranks are shown in the
graph. The data used in this graph are for chronic toxicity involving growth or reproduction (not mortality). The
solid line near the bottom of the graph shows the concentration corresponding to a 500:1 dilution of a snow and ice
control chemical. Points falling well above the line are unlikely to present any danger of toxicity for the combina-
tions of species and snow and ice control materials in this test.
Figure 9-2. Lethal (acute) aquatic toxicity.

Summary of the acute lethal bioassay data on fathead minnow, and Ceriodaphnia (no data avail-
able for Selenastrum). For each species, each source is assigned a rank and the threshold concen-
trations corresponding to the ranks are shown in the graph. The data used in this graph are for
acute toxicity involving mortality thresholds. The solid line near the bottom of the graph shows
the concentration corresponding to a 500:1 dilution of a snow and ice control chemical. Points
falling well above the line are unlikely to present any danger of toxicity for the combinations of
species and snow and ice control materials in this test.

127
CHAPTER 10
Effects of Snow and Ice Control Materials

on Reinforcing Steel Corrosion
10.1 Introduction analytical program involved the relative evaluation of corro-

sion rates of concrete reinforcing steel under exposure to 15
In addition to environmental impacts, snow and ice con- common snow and ice control materials, including those
trol materials can be detrimental to the durability of concrete with corrosion inhibitors. The methodology was based on
structures (e.g., bridges, concrete pavements, and parking ASTM G 109 ponding tests, modified to represent field-
garages). The effects of these materials on concrete can be applied material concentrations. In addition, chloride ion
twofold—either by attacking the concrete itself or by pene- diffusion properties were tested for chloride-based snow and
trating the concrete cover and corroding the reinforcing steel. ice control materials using ASTM C 1556. Testing was initi-
Deterioration of the concrete was not included in the scope ated in June of 2004 and results to June of 2006 are discussed.
of this project because a study by the South Dakota DOT is in
progress (Project SD2002-01). That study focuses on the
10.2 Background
effects of magnesium chloride and other concentrated salt
solutions on pavement and structural Portland cement Snow and ice control materials are known to affect con-
concrete. Considering that the South Dakota study is not crete structures—either through corrosion of reinforcing
addressing reinforcing steel corrosion, a laboratory testing steel or deterioration of the concrete paste. Corrosion of
program was included in this project. reinforcing steel has typically been the primary deterioration
Corrosion effects related to snow and ice control materials mechanism and has been linked to use of chloride-based
are a recognized and costly consequence of use. A U.S. FHWA snow and ice control materials (e.g., NaCl, CaCl2, and
report conducted in 1999 on “Corrosion Costs and Preven- MgCl2). NaCl has been used for more than 50 years and asso-
tion Strategies in the United States” estimated annual dollar ciated long-term effects to reinforced concrete are fairly well
effect of corrosion on steel highway bridge components to be understood. Although there has been extensive research into
$2 billion for maintenance and the cost of capital for bridge the use of alternative materials (e.g., MgCl2 and CaCl2), there
substructures and superstructures, and approximately $500 is limited data about their long-term effects.
million for maintenance painting on steel bridges. Concrete is a composite material consisting of stone and
Sodium chloride has been used for roadway snow and ice sand held together in a matrix of hydrated cement paste. The
control for more than 50 years. In recent years, alternative two major components of the hydrated cement paste are cal-
products such as magnesium chloride (MgCl2), calcium cium-silicate-hydrate (C-S-H) and calcium hydroxide
chloride (CaCl2), calcium magnesium acetate (CMA), and (Ca(OH)2). The C-S-H phase provides the bond strength of
potassium acetate (KA) have gained popularity. Manufac- concrete. The Ca(OH)2 buffers the pH of the pore solution
turers also blend snow and ice control materials with various to about 12.5. At this pH, reinforcing steel embedded in con-
materials intended to inhibit corrosion, including phos- crete is protected against corrosion by a “passive” layer (i.e.,
phates, amines, and agriculturally derived organic matter chemical coating) that forms on its surface. Degradation of
from biomass (OMB). the passive layer initiates active corrosion and can be caused
The objective was to determine the rate at which snow and by the presence of chloride ions. The typical threshold value
ice control materials penetrate the concrete cover and initi- (i.e., concentration) for chloride in concrete to initiate cor-
ate active corrosion of the reinforcing steel as well as to deter- rosion of black steel is 0.4 percent by weight of cement, but
mine corrosion rates once corrosion has been initiated. The values as low as 0.15 percent may be considered dangerous.

128
The threshold value also depends somewhat on the pH of the Table 10-1. Concrete mixture proportions.
concrete pore solution at the reinforcing steel level. Both
MgCl2 and CaCl2 cause a drop in pH—not typically observed Cement content (Type I) 405 kg/m3
to the same extent with NaCl. This suggests that the chloride w/c 0.37
threshold for NaCl may be somewhat higher than for MgCl2
Plastic air content 6.4 %
and CaCl2. However, this argument requires that the cation
diffuses through the concrete at a rate similar to that of Slump 140 mm
chloride. If the diffusion rate of the cation is lower than for
43.2 MPa @28 days
chloride, the pH effects will not necessarily influence the Compressive Strength 48.3 MPa @ 56 days
corrosion initiation threshold value. 50.2 MPa @ 90 days
Considering that chloride-induced corrosion is independ- 3116 Coulombs (moderate) @ 28 days
ent of the associated cation, there is no reason to consider any Rapid Chloride Ion
2611 Coulombs (moderate) @ 90 days
Permeability (ASTM C1202)
relative differences in corrosivity between various chloride
salts (i.e., the steel does not differentiate between the source
of chloride). However, the amount of chlorides that will pen- significantly different with more severe deterioration being
etrate the concrete and their rate of penetration do affect cor- observed in laboratory studies. Therefore, it was decided to
rosion, and evidence suggests this may be related to the use concrete representative of actual structures. A standard
associated salt cation. concrete (not mortar) mix design was used to evaluate the
Chloride is a relatively small, negatively charged anion that relative performance of snow and ice control materials. As an
can diffuse through concrete as a result of concentration gra- initial effort in this regard, the AASHTO specification for
dients. Concrete quality, pore structure, and micro-cracks will such concrete was used. Given the variability in the type, use,
influence diffusion rates. Evidence also suggests that chloride and availability of supplementary cementing materials (e.g.,
ion diffusion depends on the type of salt used. fly ash, slag, and silica fume), it was decided to omit these
materials from the concrete mixture.
Properties of the concrete used in this study are presented
10.3 Laboratory Program
in Table 10-1. All samples were cured for 28 days in a stan-
The objective was to assess relative effects of various snow dard moist curing room at 100-percent relative humidity.
and ice control materials on concrete reinforcing steel, Hardened concrete samples were analyzed in accordance
including those with corrosion inhibitor additives, for their with ASTM C 114 for acid-soluble chloride ion content to
potential to initiate corrosion in concrete reinforcing. It was determine total chloride ion background. Total chloride ion
beyond the scope of the work to develop completely new test content in the concrete was determined at 0.004 percent by
methods; therefore, two existing standard test methods were mass of concrete.
used:
10.3.2 Snow and Ice Control Materials
• Macrocell corrosion/ponding testing to assess time to
corrosion was conducted in accordance with ASTM G The snow and ice control materials used for assessment
109-99a (2005) Standard Test Method for Determining the were selected based on currently available common materi-
Effects of Chemical Admixtures on the Corrosion of Embed- als. Although the overall project has a broader selection of
ded Steel Reinforcement in Concrete Exposed to Chloride snow and ice control materials, 15 products were selected that
Environment. represent the current U.S. and Canadian market in terms of
• Chloride-ion diffusion testing was conducted in accor- product type (i.e., e.g., NaCl, MgCl2, CaCl2, CMA, and KA),
dance with ASTM C 1556-04, Standard Test Method for and common additive combinations were selected for this
Determining the Apparent Chloride Diffusion Coefficient of assessment (Table 10-2). The same materials were also used
Cementitious Mixtures by Bulk Diffusion. to study atmospheric corrosion effects (Chapter 5). Primar-
ily chloride-based materials were selected because these are
most commonly used for snow and ice control on public
roads, primarily because of cost. Also, chloride-based materi-
10.3.1 Concrete Properties
als, when compared with other materials, are more likely to
Many previous laboratory studies used mortar or cement cause corrosion of the reinforcing steel.
paste samples instead of concrete, thus not representing the The benefit of conducting a large-scale analytical assess-
actual material used in structures. Review of the literature ment on a wide range of snow and ice control materials using
has shown that observations in the field and laboratory were the same test methods was that a direct comparison for trends

129
Table 10-2. Snow and ice control material description.

Approximate Salt
Lab ID Base Chemical1 Matrix Additive2
Concentration %
2 NaCl Liquid 23 OMB
3 NaCl Liquid 23 OMB
6 NaCl Solid 23 -
10 NaCl Solid 23 P
13 CaCl2 Liquid 32 -
17 CaCl2 Liquid 32 OMB
20 CaCl2 Liquid 32 C, A
21 CaCl2 Liquid 32 OMB
26 MgCl2 Liquid 29 -
29 MgCl2 Liquid 29 OMB, TEA
32 MgCl2 Liquid 29 OMB
37 Blended Chloride Solid 30 OMB, P
40 CMA Liquid 25 -
41 CMAK Liquid 37.5 -
42 KA Liquid 50 -
1 NaCl - Sodium Chloride, CaCl2 - Calcium Chloride, MgCl2 - Magnesium Chloride, CMA -
Calcium Magnesium Acetate, KA - Potassium Acetate, Blended Chloride - mixture of NaCl, MgCl2,
and CaCl2.
2 OMB – Organic Matter from Biomass, P – phosphate, TEA – triethanolamine, C – polyhydroxycar-
boxylate, A –polyalkoxylated amine
could be made. Commercial products continue to evolve, experiment. Triplicate samples were prepared for all 15 snow
especially with respect to additives, and variability can exist and ice control materials, which included three distilled water
between product batches. control samples, resulting in 48 specimens. Figure 10-2 shows
Liquid materials were tested as received. Solid materials the setup of all samples.
were dissolved in deionized water at room temperature at Samples were ponded with snow and ice control materials
concentrations approximately equivalent to common liquid for 2 weeks followed by 2 weeks of drying per ASTM specifi-
counterparts. These concentrations are typical for use in cations. Macrocell currents were measured every 2 weeks
roadway applications. This approach was conservative and when the samples were filled and drained, and half-cell
provided an unbiased representation of field exposure. potential values (ASTM C 876) were measured monthly when
draining the samples. Half-cell potentials provided informa-
tion about the probability of corrosion as shown in Table 10-3
10.3.3 ASTM G 109 Exposure
and generally provided an indication when corrosion was
The ASTM G 109 ponding test was selected to provide a initiated. The main deviation from the specified ASTM pro-
relative assessment of the potential of various snow and ice cedure was the solution concentration. Although ASTM G
control materials to initiate corrosion by means of diffusion 109 specifies a 3-percent NaCl solution, much higher con-
of chlorides or other ionic species through the concrete. Once centrations were used to represent actual field conditions.
active corrosion was initiated, the macrocell corrosion cur- The 3-percent NaCl concentration is typically used to assess
rent (i.e., corrosion rate) could be measured between the the performance of different types of concretes whereas the
anode and cathode bars of the specimens. The effects of the objective of this study was to assess different types of solu-
snow and ice control materials on the cement and concrete tions with identical concrete specimens. Macrocell currents
properties could also be assessed. However, the samples were were calculated by measuring the potential between the
not subject to freezing and thawing cycles and, as such, the bottom two bars (cathode) and the top bar (anode) across a
combined effects of freezing and thawing and snow and ice 100-ohm resistor and using Ohm’s law:
control materials could not be assessed.
Concrete test specimens were prepared in accordance with I E/R
ASTM G 109 with a clear concrete cover to the top bar
(anode) of 25 mm. Figure 10-1 is a schematic illustration and Where
photograph of the specimens. In addition, for each material R resistance in ohms (100-ohm resistor installed)
an additional sample without reinforcing bars was cast to E potential in volts
determine chloride penetration at the end of the exposure I current in amperes

130
Figure 10-1. ASTM G 109 specimen illustration.

131
research has assessed diffusion coefficients of various con-

crete types and has assumed that diffusion rates depend pri-
marily on concrete properties and have generally not
considered the possibility that diffusion may be affected by
the associated salt cation. Concrete cylinders were used for
diffusion testing in accordance with ASTM C 1556 to obtain
a relative assessment of chloride diffusion properties of dif-
ferent chloride-based snow and ice control materials. Dupli-
cate samples were prepared with MgCl2 (~29 percent), CaCl2
(~32 percent), and NaCl (~23 percent) products only,
because the primary objective was to determine chloride ion
diffusion through the concrete. Continual concrete specimen
exposure to the snow and ice control materials began in Sep-
tember of 2004 and one set of each duplicate sample was
tested for chloride penetration in May of 2006. The second
set will be left for future testing to assess further chloride
penetration. Concrete powder samples were collected at 1- to
3-mm increments using a 25-mm diamond core bit and a
milling machine. Samples were removed by drilling without
any cooling liquids.
10.4 Results and Discussion

After 2 years of ponding, corrosion has not been initiated
for any of the snow and ice control materials—ions have not
yet diffused through 25 mm of concrete cover causing active
Figure 10-2. ASTM G 109 specimen setup. corrosion of the top reinforcing bar. Thus, it is not possible
to comment on the relative differences in chloride ion diffu-
sion and corrosion activity of the various chloride-based
Table 10-3. Probability of corrosion
(ASTM C 876). materials based on the ASTM G 109 ponding samples. Sim-
ilarly, non-chloride materials have not corroded the rein-
Ecorr vs. (mVCSE*) Probability of Corrosion forcing steel.
More positive than -200 mV 10% Figure 10-3 shows half-cell potential values for one sam-
ple each of all 15 materials for approximately 2 years of
-200 to -350 mV Uncertain
exposure. Potential values for replicate samples for each
More negative than -350 mV 90% material were essentially identical, with differences in
potentials being typically less than 10 mVCSE. Half-cell
*CSE – Copper/copper-sulfate reference electrode potential values are in the passive (less than 10-percent
probability) and transitional ranges for all materials when
compared with ASTM C 876 specifications shown in
Table 10-3. Initial negative values for some materials are
Samples were cast in April of 2004 and exposure of the most likely associated with the passivation of the reinforc-
G109 samples to snow and ice control material solutions ing steel. Over time, these values have stabilized in the
started in June of 2004. passive and transitional ranges, indicating that active
corrosion has not been initiated.
For clarification, Figure 10-4 shows half-cell potentials for
10.3.4 Chloride Diffusion—ASTM C 1556
one sample of each type of snow and ice control material, as
In addition to ponding samples, the ASTM C 1556 well as one control sample (water exposure). Half-cell poten-
method was selected to assess the relative diffusion rate of tials are comparable for all snow and ice control materials
several chloride-based snow and ice control materials. Most with no significant difference in values.

132
Figure 10-3. Half-cell potential values.
Figure 10-4. Half-cell potential values after 2 years’ exposure (ASTM G 109 specimens).

133
Potentials across the 100-ohm resistor remained at 0 V 10.5 Conclusions and

throughout the 2-year exposure period, thus no macrocell Recommendations
currents were recorded. This suggests that chlorides have
not yet reached the top steel and corrosion has not been After approximately 2 years of exposure to various
initiated. chloride-based and non-chloride snow and ice control mate-
Over time, the half-cell potentials for these materials have rials, active corrosion has not been initiated for ponding of
stabilized. Visual observation of the samples did not identify ASTM G 109 samples using concrete with a clear cover over
any deterioration of the concrete at the exposed surface. the top bar of 25 mm. ASTM C 1556 samples suggest that
Chloride penetration for the ASTM C 1556 samples was chlorides have migrated through the concrete to a depth
measured to a depth of 18 mm by drilling concrete powder beyond 18 mm; however, concentrations at a depth of 25 mm
samples. Acid-soluble chloride concentrations were very remain below the threshold to initiate active corrosion.
high at the exposed concrete surface. The highest chloride These results are generally supported by observations made
concentration at 18 mm depth of 1.02 percent by mass of for real structures where corrosion initiation normally takes
cement was measured for the CaCl2 sample and the lowest several years. Corrosion-related damage in shorter periods is
concentration at the same depth was measured for the often attributed to cracking of the concrete cover. The rela-
MgCl2 sample (Figure 10-5). Generally, at 18-mm depth, tively slow process of chloride migration through the concrete
the acid-soluble chloride concentrations are above the provides the service life for reinforced concrete structures.
threshold concentration to initiate corrosion. There Generally, good-quality concrete and control of cracking is
appears to be some evidence of inconsistent results with essential to slow penetration of chlorides. The intent of this
some higher concentrations at 18 mm than at smaller study was to use a concrete mix design representative of con-
depths. This is most likely a result of large aggregate parti- crete used in the field rather than laboratory conditions that
cles in the specimens. would accelerate the testing process. As such, longer exposure
Results shown in Figure 10-5 suggest that chloride ion times will be required to obtain results showing corrosion ini-
concentrations at a depth of 25 mm have not yet reached the tiation and measure corrosion rates once initiated.
corrosion initiation threshold. The duplicate samples will be Chloride diffusion testing (ASTM C 1556) shows chloride
tested at the same exposure age that corrosion initiation is penetration for NaCl, CaCl2 and MgCl2 to a depth of at least
observed. Initial results indicate that CaCl2 and NaCl are 18 mm, with CaCl2 showing the highest chloride concentra-
migrating faster that MgCl2 and blended chlorides. tion at that depth. The lowest concentration was measured for
Although similar results were reported in other studies for the MgCl2 sample, which is contrary to findings in the litera-
CaCl2, the slower penetration of MgCl2 is contrary to previ- ture. Longer term exposure and comparison of the diffusion
ous studies. and ponding tests will provide more conclusive results.
Figure 10-5. Chloride ion concentrations after 20 months of exposure.

134
It is recommended that the testing continue so as to deter- between the top bar (anode) and the bottom bars (cathode)
mine time to initiation of active corrosion. Based on the dif- as well as microcell corrosion rates using the linear polariza-
fusion experiments, it is anticipated that this may require at tion resistance technique.
least one more year of exposure. Once initiated, corrosion Autopsy of the samples after evidence of corrosion damage
rates should be monitored to determine the corrosivity of the has become obvious (i.e., corrosion staining and cracking) is
different snow and ice control materials. Corrosion rate also recommended. It is anticipated that the overall experi-
measurements should include macrocell corrosion rates ment should continue for an additional 2 to 4 years.

135
CHAPTER 11
Atmospheric Corrosion Testing
11.1 Introduction Furthermore, the relevance of laboratory-based corrosion

rate determinations is deeply rooted in the relationship of the
In conjunction with a larger study relating to the chemistry test environment to the service environment and does not
and toxicity of snow and ice control materials and corrosion of necessarily represent conditions that exist in the field
concrete from snow and ice control materials, electrochemical environment.
corrosion testing of 15 snow and ice control products of varied Electrochemical techniques, such as Linear Polarization
concentrations with five engineered alloys was completed. The Resistance (LPR), can also be used to assess corrosion rates.
objective of the study was to conduct relative assessments of Given that LPR is a mature, rapid, and sensitive corrosion rate
snow and ice control materials corrosivity on various alloys on measurement technique, LPR augmented by weight loss
the basis of both dilution and chemical family type. Embedded determinations was used for the study. The test program
in the general test scheme, it was anticipated that the effect of involved developing a rapid and reliable test protocol that was
corrosion inhibitor additives could be assessed. more robust than traditional electrochemical techniques. Of
Corrosion, an inherently slow process, is often initiated particular benefit was the potential suitability of LPR for
after an incubation period, may be comparatively rapid after evaluating the corrosiveness of inhibited snow and ice control
onset, and tends to slow over time. Various standard methods materials and the effect of roadside dilution.
are available to assess atmospheric corrosion, including Fifteen snow and ice control material/inhibitor combinations
immersions tests, poultice corrosion tests, and salt-spray tests. were selected for the study and are described in Table 11-1.
Generally, the most accurate corrosion tests are weight-loss These materials represent common snow and ice control mate-
measurements used to estimate an average penetration rate, rial chemical type and additive combinations being used in the
although they are very time-intensive (i.e., weeks, months, or United States and Canada for roadway snow and ice control.
years in duration). Testing of each snow and ice control material was conducted on
The accuracy of weight loss measurements depends five commonly available alloys representative of materials used
significantly on the summation of effects over time (essen- in highway infrastructure and vehicles (Table 11-2).
tially an integration). Corrosion penetration estimations To represent snow and ice control material concentrations
based on weight loss can seriously underestimate the expected in the field (i.e., application concentration through
corrosion penetration caused by localized processes (e.g., dilution), LPR assessments were conducted on liquid snow
pitting, cracking, and crevice corrosion). In addition, bias and ice control materials at 100-percent (v/v) (as received),
in corrosion test results may result from the following 10-percent (v/v), and 1-percent (v/v) concentrations for each
factors: snow and ice control material with each alloy. Solid snow and
ice control materials used in this test program were prepared
• Inappropriate exposure time to represent corrosion condi- into solutions at concentrations equivalent to common liquid
tions accurately; counterparts (refer to Chapter 7).
• Film formation whereby precipitates or complexes form on Corrosion rate data was reviewed for test method robustness,
the test specimen that may inhibit corrosion; and trends were assessed that related to alloy type, snow and ice
• Inadequate corrosion product removal or excessive metal control material type, test solution concentration, and the effec-
removal before final weighing of test specimens; and tiveness of corrosion inhibitors. Conclusions were summarized
• Depletion of corrodant. and recommendations for further research have been made.

136
Table 11-1. Snow and ice control material occur. The measured corrosion current will be indicative of
description. the rate (kinetics) of the reaction and may be readily related
to current density and corrosion rate.
Lab ID Chemical Type Matrix Additive1
The ratio of a small potential shift of a corroding metal sur-
2 NaCl Liquid OMB
face to the current required (ΔI) to maintain this potential
3 NaCl Liquid OMB
displacement (ΔE) is called the polarization resistance and is
6 NaCl Solid -
10 NaCl Solid P
inversely proportional to corrosion current (Icorr).
13 CaCl2 Liquid -
17 CaCl2 Liquid OMB ΔE I
α
20 CaCl2 Liquid C, A ΔI Icorr
21 CaCl2 Liquid OMB
26 MgCl2 Liquid - The potentiostat controls the potential between the work-
29 MgCl2 Liquid OMB, ing electrode and the reference electrode while it measures the
TEA current between the working electrode and the counter (aux-
32 MgCl2 Liquid OMB
iliary) electrode. Although the step interval for LPR testing
37 Blended Chloride Solid OMB, P
may be varied, the method should be considered one of
40 CMA Liquid -
pseudo-instantaneous measurement.
41 CMAK Liquid -
42 KAc Liquid -
1
OMB = Organic Matter from Biomass, P = Phosphate, TEA = 11.2.3 Weight Loss
Triethanolomine,
C = Polyhydroxycarboxylate, A = Polyalkoxylated amine To supplement LPR, measurements of electrode mass were
‘ - ’ indicates not present made before and after each test. In the absence of localized
corrosion, weight loss may be used as a principal measure of
Table 11-2. Alloy description.
uniform corrosion. In this study, estimates of average pene-
tration by weight loss were completed as a quality control for
Alloy Type Designation comparison with the LPR data.
Structural Steel ASTM A 36 Instances whereby negative corrosion rates are suggested
Wrought Aluminum Alloy AA6061-T6 by the weight loss data may be indicative of corrosion prod-
Cast Aluminum Alloy AA356.2 uct retention on the specimens or errors in weighing. Alter-
Free Machining Brass CDA 360 nately, the formation of durable complexes (not corrosion
Magnesium Alloy AZ91C-T6
products) may occur on the electrode surface which, when
not removed by cleaning, cause the electrodes to gain weight.
11.2 Technology Overview
11.3 Test Method
11.2.1 Electrochemistry
Corrosion rate determinations were made during the
Metallic corrosion is the consequence of a redox reaction period of August 2004 to April 2005. The general measure-
whereby oxidation of the anode liberates electrons and reduc- ment process involved immersing test electrodes made from
tion at the cathode consumes electrons. Oxidation is always the various alloys into snow and ice control material solutions
accompanied by reduction in an electrochemical cell. Electrical and taking LPR measurements over time. All LPR measure-
potential is the driving force for a redox reaction and is an effec- ments were completed using a Metal Samples MS1500L
tive measure of the thermodynamic (energetic) state of the sys- handheld measurement and data collection device configured
tem. Electron flow (i.e., current) is the result of a redox reaction for Three-Electrode Linear Polarization. Additionally, elec-
and is an effective measure of the kinetics or rate of the corro- trode mass was recorded before and after each test.
sion reaction. Electrochemical potential measurements may be The general test protocol is described in Appendix A, which
considered an indication of system equilibrium. is available for download from the TRB website. Of specific
interest to the conduct of the testing was
11.2.2 LPR
• Configuration of the MS1500L to complete cathodic polar-
The essence of a potentiostatic test such as LPR is the ization (cathodic control is considered more likely to be
alteration of a system electrical potential in some systematic linear than anodic control).
manner and the measurement of current response. Corrosion • The use of 15-minute LPR test cycle times for all diluted
current cannot be measured directly. The change in potential (1- and 10-percent) solution measurements and 3-minute
of an electrode (polarization) causes a redox reaction to LPR test cycle times for all testing of concentrated solu-

137
tions. Systems exhibiting passive behavior and lower material systems tested was relatively constant over the dura-
corrosion rates mandate longer cycle times. tion of the test (24 or 72 hours).
• Testing of coupons in 1- and 10-percent concentration There are various possible methods for assessing replicate
solutions at intervals of 3, 6, 24, 48, and 72 hours (five corrosion rate data. The practice of characterizing corrosion
readings) following immersion; coupons immersed in rates using average values was applied to the replicate data
100-percent solutions were tested at 1, 3, 6, and 24 hours and is consistent with ASTM G 16, Standard Guide for
(four readings) following immersion. Applying Statistics to Analysis of Corrosion Data. In corrosion
• Surveillance of electrodes for pitting corrosion (corrosion testing, average corrosion rates are frequently used in charac-
current measurement is surface-area-dependent; pitting terizing corrosion rates, particularly where replicate meas-
renders surface area unreliable). urements are as likely to overestimate the experimental rate
• The MS1500L uses different working electrode surface as to underestimate it. For data that span several orders of
areas for different alloys. Surface area is varied in propor- magnitude, the median is less sensitive to extreme values and
tion to the ratio of equivalent weight of the corroding provides a stable statistic that reflects the center of the distri-
metal to corroding metal density rather than input of the bution. When considered together, the mean and median
constant K (related to Tafel slope). provide an indication of the influence of extreme values in a
• Surface preparation by ultrasonic cleaning followed by data set. If the two measures of central tendency are close to
abrasion. one another, extreme values in the data set are not considered
influential. Such comparison was considered for analysis of
Uniformity of the sample preparation and data collection data samples from the population linked by alloy type and
were considered paramount to develop a basis of comparison snow and ice control materials concentration.
in the data sets. The data collected are described in the “Cor- Relative standard deviation (RSD) values were calculated
rosion Test Work Sheet and Summary” (refer to Appendix B, for the replicate values for each alloy-snow and ice control
which is available for download from the TRB website). materials-concentration system (Appendix C, which is avail-
able for download from the TRB website). Average RSD values
for each alloy-concentration system are shown in Table 11-3.
11.4 Test Program Evaluation RSD values indicate the relative precision (i.e., measurement
Before evaluating data for specific trends (related to alloy variability) that exists for the replicate measurements made.
type, snow and ice control materials type, snow and ice control Overall, the discrete corrosion rate values obtained for replicate
materials concentration, and corrosion inhibitor performance), measurements were within about one standard deviation
the overall robustness of the test program was evaluated. This (100-percent RSD) of the average value. RSD values of 15 per-
involved various statistical assessments of the data and com- cent are considered excellent, and RSD values of 50 percent are
parison of LPR corrosion rate data with corrosion rate data considered good. Extremely high RSD values were typically
obtained through supplementary weight loss measurements on associated with low corrosion rate values as test method
the test electrodes. The data set is presented in Appendixes C uncertainty increases at the lower limits of the corrosion test.
and D, which are available for download from the TRB website. Table 11-4 summarizes statistical values calculated for each
Replicate LPR measurements were made over time for each alloy tested. Average, minimum, and maximum corrosion rate
snow and ice control material/concentration/alloy system. A values indicate the range of values obtained for each alloy/
minimum of four LPR readings (concentrated snow and ice concentration system. Median values and average corrosion
control materials at 1, 3, 6, and 24 hours) or five LPR readings rates integrated over time were also calculated and are generally
(diluted snow and ice control materials at 3, 6, 24, 48, and 72 consistent with average corrosion rate values. Plots of cumula-
hours) were conducted. Comparison of corrosion potential tive probability versus corrosion rate (measured using a loga-
measurements at the outset and finish of the test generally rithmic scale) and linear probability versus corrosion rate
showed minimal changes in most instances, suggesting that (measured using a linear scale) were prepared to graphically
the measured corrosion rate of the alloy/snow and ice control characterize the pooled corrosion data as to symmetry.
Table 11-3. Average relative standard deviations for each alloy and solution concentration system.
Average (%) (minimum (%), maximum(%)) RSD Values for each Test Solution
Alloy Concentration
100% Solution 10% Solution 1% Solution
Structural Steel, ASTM A 36 64% (0, 200) 68% (17, 137) 60% (8, 196)
Wrought Aluminum Alloy, AA6061-T6 17% (3, 86) 26% (3, 70) 108% (68, 163)
Cast Aluminum Alloy, AA356.2 28% (2, 119) 34% (3, 108) 110% (51, 170)
Free Machining Brass, CDA 360 62% (2, 115) 50% ( 11, 134) 100% (10, 258)
Magnesium Alloy, AZ91C-T6 54% (15, 104) 50% (17, 83) 55% ( 9, 91)

Table 11-4. Univariate descriptive statistics corrosion rate by alloy and snow and ice control material concentration.
Corrosion Rate – MPY

SICM Average of Population Solution Distribution
Alloy Population Population Aberrant
Concentration Max. Min. Rate by Area Standard ID Characteristic
Mean Median Value*
Under Curve Deviation
AA6061-T6 1% 0.87 0.01 0.215 0.110 0.158 0.276 - Likely log normal
10% 52.23 0.05* 29.74 34.40 29.30 15.49 0.05 40 Appears normal
100% 94.88 2.27 59.57 66.23 57.19 25.06 2.27 13 Likely normal
82.2 10
AA356.2 1% 7.27* 0.01 0.663 0.06 0.27 1.850 7.27 20 Likely log normal
10% 79.07 0.05* 28.08 26.09 26.65 23.00 0.05 40 Likely normal
100% 80.41 0.00 46.64 52.30 47.36 30.93 Likely normal
A36 1% 16.96* 0.02* 3.20 2.21 2.17 4.02 0.02 41 Inconclusive
0.03 42
16.96 29
10% 7.38 0.02 2.49 2.38 2.14 2.00 - Likely normal
100% 9.06 0.06 2.32 1.58 2.25 2.79 0.06 40 Likely log normal
0.06 41
CDA 360 10% 13.82 0.09 3.23 1.41 2.94 4.19 - Likely log normal
10% 97.54 0.13 28.72 24.28 30.38 31.46 - More likely normal
100% 85.57 0.20 14.30 3.39 12.31 24.60 - Likely log normal
AZ91C-T6 1% 19.54 0.67 5.28 3.10 4.62 5.34 - Likely log normal
10% 79.74 4.98 36.48 36.98 36.51 27.22 - ** Likely log normal
100% 124.38 5.39 57.39 57.17 78.13 38.74 - *** Likely log normal
MPY is equivalent to 0.001 inch per year.

* Aberrant Value – anomalous or extreme data value
** Magnesium alloy exposure to 100% SICM 3, 17, 20, 21, 26, 29 and 32 exhibited corrosion rates exceeding 200 MPY
*** Magnesium alloy exposure to 10% SICM 3, 6 and 21 exhibited corrosion rates exceeding 200 MPY
“ - ” indicates no value observed
Note: Individual corrosion rates were derived from the average of four or five LPR measurements.
139
Quality control was conducted by comparing LPR corro- percent and 10-percent concentration levels, except for
sion rate determinations with corrosion rates determined by magnesium alloy at the 100-percent concentration level.
weight loss measurement on the test electrodes (refer to
Table 11-5 and Appendix D, which is available for download
11.5 Data Assessment
from the TRB website). To determine equivalency between
the two corrosion rate test methods, a two-tailed “t-test” was Complete LPR corrosion rate data for each alloy type, snow
conducted at the 95-percent confidence interval (refer to and ice control materials type, and concentration level are tab-
ASTM G 16, Clause 7). Populations (i.e., all possible items ulated in Appendix C, which is available for download from
possessing one or more common characteristics) are often the TRB website, including replicate LPR measurements,
too large to be studied adequately. This is particularly true maximum and minimum values, averages values, standard
when the population is conceptual as it is in an engineering deviation values, and relative standard deviation values. Data
experiment or when the collection of the data requires alter- summaries of average values are represented both graphically
ation of items such as corrosion of the electrodes. Despite the and in tabular form in Figures 11-1 through 11-5.
complexity of the test matrix, data was compiled into sample Trend analysis was conducted using average LPR corrosion
groupings based on alloy type and snow and ice control mate- rate values. The population is diverse and difficult to study
rial solution concentration. This allowed reasonable group- adequately, but the testing did provide comparative analysis
ings of adequate size to allow performance of the statistical of a comprehensive matrix of snow and ice control materials.
test. Equivalency analysis indicated that It was also important to consider measurement precision and
variability as expressed by RSD when comparing LPR values.
• In approximately 53 percent (8 of 15) of the cases exam- Overall, standard deviation values for replicate LPR measure-
ined, the corrosion rate determined by LPR was similar to ments were of the same order of magnitude as average values.
the corrosion rate determined by weight loss. Therefore, trends in average LPR data had to exist at a
• The strongest correlations between the two tests occurred resolution of at least an order of magnitude or greater to be
where LPR corrosion rate values were low. This included considered significant.
measurements made for carbon steel and measurements Overall, mean corrosion rate increased with increasing snow
made for 1-percent solution concentrations. and ice control material concentration for all alloys except for
• There tended to be poor correlation between the two the brass (CDA 360) alloy where the intermediate (10-percent)
tests where LPR measurements were high. Weight loss snow and ice control materials concentration was most
measurements underestimated LPR measurements at 100- corrosive. Corrosion rates for carbon steel (A36) also did not
Table 11-5. Comparison of corrosion rate by LPR versus weight loss.
Solution Weight Null

Alloy LPR t-stat Tc Remark
Concentration Loss Hypothesis
AA356.2 100 46.63 3.56 5.30 2.14 Accept Significant differences
exist
10 29.71 1.69 4.23 2.16 Accept Significant differences
exist
1 0.64 0.96 -0.27 2.14 Reject Negligible difference
A36 100 2.18 2.30 -0.19 2.14 Reject Negligible difference
10 2.61 2.43 0.25 2.16 Reject Negligible difference
AA6061-T6 100 59.58 4.29 8.27 2.14 Accept Significant differences
exist
exist
1 0.21 0.29 -0.13 2.14 Reject Negligible difference
CDA 360 100 14.28 0.60 2.13 2.14 Reject Statistic indicates
similarity
exist
exist
AZ91C-T6 100 57.39 91.47 -0.96 2.16 Reject Negligible difference
10 33.40 10.94 3.36 2.16 Accept Differences exist
Null Hypothesis: A difference between LPR and weight loss exists; Alternative Hypothesis: No difference between LPR and weight
loss; Alpha = 0.05; Two tailed test

140
demonstrate overall increasing corrosion rates with increasing corrosion rate levels, variability in measurement can be
snow and ice control material concentration. In this instance, greater or equal to the test method resolution, making
corrosion rates were lower than for the other alloys and any trend identification difficult.
trends were likely obscured by measurement variability at these • Chloride-based snow and ice control materials (2, 3, 6,
levels. The comparative arrangement of alloys which experi- 10, 13, 17, 20, 21, 26, 29, 32, and 37) had roughly equiv-
enced the least to greatest corrosion rates were as follows: alent average corrosion rates at each concentration level
(100, 10, and 1 percent). This data suggests little differ-
• Carbon steel (A36) Least corrosion ences in corrosivity of various chloride-based snow and
• Brass (CDA 360) ice control materials.
• Cast aluminum (AA356.2) • Corrosion rates for acetate-based snow and ice control
• Wrought aluminum (AA6061-T6) materials (40, 41, 42) were much lower than those for chlo-
• Magnesium (AZ91C-T6) Greatest corrosion ride-based snow and ice control materials. Observed values
bordered on the minimum practical detection ability of
The following sections describe key observations and test instrumentation.
trends in the data for each metal alloy tested as they pertain
to snow and ice control material type, concentration level,
11.5.2 Wrought Aluminum, AA6061-T6
and additive ability to inhibit corrosion.
Refer to Figure 11-2 for data.
11.5.1 Structural Steel, A36
• With the exception of the uninhibited CaCl2 solution (13),
Refer to Figure 11-1 for data. the chloride-based snow and ice control materials (2, 3, 6,
10, 17, 20, 21, 26, 29, 32, and 37) had roughly equivalent
• Corrosion rates measured for structural steel (A36) were average corrosion rates at each concentration level (100,
lower than for the other alloy materials tested. At lower 10, and 1 percent). This suggests little difference in the
Solution
Figure 11-1. Summary of average corrosion rates (MPY) for carbon steel A36.

141
Figure 11-2. Summary of average corrosion rates (MPY) for wrought aluminum alloy AA6061-T6.
corrosivity of various chloride-based snow and ice control 11.5.3 Cast Aluminum AA356.2
materials on wrought aluminum.
• Uninhibited CaCl2 (13) had a comparatively low corrosion Refer to Figure 11-3 for data.
rate at the 100-percent concentration level when com-
pared with inhibited CaCl2 snow and ice control materials • NaCl-based snow and ice control materials with corrosion
(17, 20, 21) and other chloride-based snow and ice con- inhibitors (2, 3, 10) demonstrated corrosion rates similar to
trol materials. However, corrosion rates were consistent the NaCl snow and ice control materials without a corrosion
between inhibited and uninhibited CaCl2 snow and ice inhibitor (6) at 100-, 10-, and 1-percent concentration lev-
control materials at 10- and 1-percent concentration levels. The NaCl snow and ice control materials containing a
els. Similar trends were observed for cast aluminum phosphate inhibitor (10) demonstrated the lowest corrosion
(AA356.2). A potential explanation may be that the 100- rate at 100- and 10-percent concentration levels; however
percent solution was sufficiently aggressive that a corro- the magnitude of this difference borders on the precision
sion product barrier was formed on the measurement limits of the test and thus cannot be conclusively attributed
electrodes that prevented further corrosion and corrosion to corrosion performance or measurement variability.
measurements. • Uninhibited CaCl2 (13) had a comparatively low corrosion
• Acetate-based snow and ice control materials containing rate at the 100-percent concentration level when compared
potassium (CMAK (41) and KA (42)) demonstrated cor- with inhibited CaCl2 snow and ice control materials (17, 20,
rosion rates similar to chloride-based snow and ice control 21) and other chloride-based snow and ice control materi-
materials at all concentration levels. Conversely, the CMA als. Corrosion rates were consistent between inhibited and
snow and ice control materials (40) had comparatively uninhibited CaCl2 snow and ice control materials at 10- and
lower corrosion rates. A similar trend was also observed 1-percent concentration levels. Similar trends were observed
for cast aluminum (AA356.2) and is contrary to general for wrought aluminum (AA6061-T6). A potential explana-
opinion that all acetate products are equally non-corrosive tion may be that the 100-percent solution was sufficiently
to metals. aggressive that a corrosion product barrier was formed on

142
Figure 11-3. Summary of average corrosion rates (MPY) for cast aluminum alloy AA356.2.
the measurement electrodes that prevented further corro- 11.5.4 Free Machining Brass, CDA 360
sion and corrosion measurements.
Refer to Figure 11-4 for data.
• Corrosion rates for MgCl2-based snow and ice control
materials (26, 29, 32) varied greatly among test solutions.
• Brass is recognized for its ability to resist electrochemical
Product 32 contained an OMB inhibitor and displayed a
low corrosion rate at the 100-percent concentration level, corrosion because of the formation of a relatively thin film
but had a corrosion rate consistent with uninhibited MgCl2 of corrosion product, though this ability can be influenced
(26) at the 10-percent concentration level. Product 29 con- by other chemical interactions that may exist in the corro-
tained an OMB/TEA corrosion inhibitor and displayed a sion system.
corrosion rate similar to uninhibited MgCl2 (26) at the 100- • For 12 of the 15 snow and ice control materials tested, the
percent concentration level, but a much lower corrosion corrosion rate was greater at the 10-percent concentration
rate at the 10-percent concentration level. A possible expla- level than at the 100-percent concentration level (6, 10, 13,
nation for these observations may be that corrosion 17, 20, 21, 26, 32, 37, 40, 41, 42). This is contrary to the
inhibitor effectiveness depends on snow and ice control trends observed for the other alloys tested, where corrosion
material and/or inhibitor concentration. rates increased with concentration.
• Acetate-based snow and ice control materials containing • For NaCl-based snow and ice control materials, the two
potassium (CMAK (41) and KA (42)) demonstrated corro- products with OMB-type additives (2, 3) demonstrated
sion rates similar to chloride-based snow and ice control comparatively low corrosion rates at the 10-percent concen-
materials for all test concentrations. Conversely, the CMA tration level. However, a similar trend was not observed at
snow and ice control materials (40) had comparatively the 100-percent concentration level. A similar corrosion
lower corrosion rates. A similar trend was also observed for reduction effect was observed for CaCl2 snow and ice con-
wrought aluminum (AA6061-T6) and is contrary to trol materials containing OMB corrosion additives (17, 21).
general opinion that all acetate products are equally non- • The MgCl2-based snow and ice control material containing
corrosive to metals. an OMB additive (32) did not demonstrate low corrosion

143
Figure 11-4. Summary of average corrosion rates (MPY) for free matching brass 360.
rates as was observed for similarly blended NaCl-based and underestimate the true corrosion rate in these exposures.
CaCl2-based snow and ice control materials. However, the Estimates of corrosion rate by area under curve (integra-
MgCl2 snow and ice control material containing TEA (29) tion) are considered more meaningful in the 100- and
demonstrated a reduced corrosion rate at the 100-percent 10-percent concentration levels.
concentration level and a very low corrosion rate at the • Chloride-based snow and ice control materials (2, 3, 6, 10,
10-percent concentration level. 13, 17, 20, 21, 26, 29, 32, 37) had roughly equivalent
• Acetate snow and ice control materials generated corrosion corrosion rates at each concentration level (100-, 10-, and
rates that were similar and/or higher than chloride-based 1-percent). This suggests little difference in the corrosivity
snow and ice control materials for all concentration levels of various chloride-based snow and ice control materials
tested. on magnesium alloy based purely on electrochemical cor-
rosion processes.
• Acetate-based snow and ice control materials demon-
11.5.5 Magnesium Alloy, AZ91C-T6
strated corrosion rates that were slightly lower than the
Refer to Figure 11-5 for data. chloride-based snow and ice control materials. However,
from the standpoint of potential impairment from corro-
• Magnesium is an extremely reactive metal, and higher cor- sion, acetate snow and ice control materials would still be
rosion rates were measured than for the other alloys tested considered corrosive.
in the study. Given the exposure of the magnesium alloy
to 100-percent concentration levels for snow and ice 11.6 Conclusions
control materials 3, 17, 20, 21, 26, 29, and 32, in addition
to 10-percent concentration levels for snow and ice con- The study of commercially significant alloys and 15 snow
trol materials 3, 6, and 21, resulted in corrosion rates and ice control materials was consistent with the following
exceeding 200 MPY (the MS1500L maximum capacity), conclusions based on the average LPR corrosion rate data
the mean and median corrosion rates for magnesium alloy obtained through this study:

144
Figure 11-5. Summary of average corrosion rates (MPY) for magnesium alloy AZ91C-T6.
1. LPR is a viable corrosion rate measurement technique 4. Overall, chloride-based snow and ice control materials
to relatively assess corrosion rates of various snow and (i.e., NaCl, CaCl2, MgCl2, and blended chlorides) displayed
ice control materials. Average corrosion rates may not be similar corrosion rates for each metal type tested at each
the most precise indicators of central tendency, particu- concentration level. This suggests little difference in cor-
larly for asymmetrical distributions. Comparison of rosion rates among chloride-based snow and ice control
the mean and median provide an indication of the materials. Field conditions along with other snow and ice
influence of extreme values in the data set (similar control material properties (e.g., hygroscopicity) may
means and medians suggest negligible extreme value contribute more to the relative corrosiveness of these
influence). materials in the service environment. Other corrosion
2. Overall, the structural steel alloy (A36) demonstrated the mechanisms not emulated by aqueous testing may be
lowest corrosion rate values while the magnesium alloy significant.
(AZ91C-T6) demonstrated the highest corrosion rate 5. Acetate materials (i.e., CMA, CMAK, and KA) demonstrated
values. Wrought aluminum alloy (AA6061-T6), cast alu- a range of LPR corrosion rates for the alloys assessed:
minum alloy (AA356.2), and brass alloy (CDA 360) • The lowest corrosion rates were for structural steel (A36).
demonstrated intermediate corrosion rate values. Observed values bordered on the minimum practical detec-
3. With the exception of structural steel (A36) and brass tion ability of test instrumentation for all acetate types.
(CDA 360), corrosion rates increased with increasing • Acetates were equally or slightly less corrosive than
snow and ice control material concentration. Structural chlorides to magnesium (AZ91C-T6).
steel corrosion rates were sufficiently low that identifica- • Acetates were equally or slightly more corrosive than
tion of trends may have been obscured by measurement chlorides to brass (CDA 360).
variability at these levels. Brass corrosion rates were • For aluminum alloys (AA356.2 and AA6061-T6),
highest when snow and ice control materials were at the acetates containing potassium demonstrated corrosion
10-percent concentration level. rates equal to chloride-based snow and ice control mate-

145
rials. CMA demonstrated a much lower corrosion rate 1. Use an LPR corrosion rate meter (i.e., electrochemical meas-
for these alloys. urement) calibrated to read metal loss directly. Corrosion
rate measurement capacity exceeding 200 MPY is preferred
6. The effectiveness of corrosion inhibitors was varied and
for reactive metals, particularly magnesium alloys.
poorly quantified. In some cases, these materials provided 2. Confine testing to 10-percent snow and ice control mate-
no detectable corrosion improvement; in other cases, cor- rial concentrations. 100-percent snow and ice control
rosion improvement appeared to depend on the concen- material solution concentrations are not present for long
tration level of the test solution. This suggests that, in the service environment because of dilution and excess
although some benefit may be realized from the use of reactants at these concentrations manifest atypical
these materials, the overall effectiveness under field condi- corrosion rates for the short term. One-percent snow and
tions (including dilution) and after exposure to numerous ice control material solution concentrations developed
potential alloy types is questionable. corrosion rates that were essentially unresolvable.
3. Test snow and ice control material/alloy combinations
in triplicate; assess corrosion rate with a single LPR
11.7 Proposed Testing Program
measurement at a maximum test time of 24 hours.
Based on the study team’s experience with the snow and ice 4. Consider confining future studies to the most commer-
control material corrosion test program thus far, refinements of cially relevant alloys (i.e., structural steel, wrought alu-
the initial test parameters are recommended as follows: minum, and cast aluminum).

146
References
1. Ice and Snow Technologies, LLC. Proactive Guide to Snow and Ice 16. Chang, W., et al. Environmentally Sensitive Sanding and
Control—A Guide for Winter Maintenance Personnel, 2001. Deicing Practices, prepared for the Colorado Department of
2. Ketcham, S. A., et al. Manual of Practice for an Effective Anti-Icing Transportation by the ESSD Research Group of the Department
Program—A Guide for Highway Winter Maintenance Personnel, of Civil Engineering of the University of Colorado at Denver,
U.S. Army Cold Regions Research and Engineering Laboratory 1994.
Corps of Engineers for the Federal Highway Administration, 17. Transportation Research Board. Highway Deicing Comparing Salt
Hanover, New Hampshire, 1996. and Calcium Magnesium Acetate, Special Report 235. Transporta-
3. Environment Canada and Health Canada. Canadian Environmental tion Research Board, Committee on the Comparative Costs of
Protection Act, 1999, Priority Substances List Assessment Report Road Rock Salt and Calcium Magnesium Acetate (CMA) for Highway
Salts, 2000. Deicing, Washington DC, 1991.
4. Transportation Association of Canada (TAC). Code of Practice— 18. Colwill, D. M., J. R. Thompson, and A. J. Rutter. An Assessment of
Design of Road Maintenance Yards, 1999. the Condition of Shrubs Along Motorways, Report 1061, U.K.
5. Williams, D. Past and Current Practices of Winter Maintenance at Transport and Road Research Laboratory, Crowthorne,
the Montana Department of Transportation (MDT), Montana Berleshire, England, 1992.
Department of Transportation, 2001. 19. Jones, P. H., et al. Environmental Impact of Road Salting—State of
6. Blackburn, R. R., et al. Guidelines for Snow and Ice Control Materials the Art, Ontario Ministry of Transportation Technical Report
and Methods, NCHRP Project No. 6-13. National Cooperative RR237, 1986.
Highway Research Program, Transportation Research Board, 20. Novotny, V., et al. Urban and Highway Snowmelt: Minimizing the
National Research Council, 2003. Impact on Receiving Water, Water Environment Research Founda-
7. Wisconsin Transportation Information Center. Wisconsin Trans- tion, Project 94-IRM-2, 1999.
portation Bulletin No. 6: Using Salt and Sand for Winter Road 21. Morin, D., et al. Impacts Evaluation of Road Salts Loads on Soils and
Maintenance, 1996. Surface Waters, Report Submitted to the Environmental Resource
8. Technology Transfer Center, University of New Hampshire. Manual Group for Road Salts, Environment Canada, Commercial Chemi-
of Practice for Anti-Icing of Local Roads, 1996. cals Evaluation Branch, 2000.
9. Fonnesbech, J. K. Ice Control Technology With 20% Brine on High- 22. Berthouex, P. M. and G. A. Prior. “A Study of Salt Pollution of Soil
ways, Highway Maintenance Division, County of Funen, Den- by Highway Salting.” Highway Research Record 193, HRB, National
mark, 2003. Research Council, Washington D.C. (1967): 8–21.
10. Ontario Ministry of Transportation. Memorandum: Pilot Imple- 23. Hofstra, G. and D. W. Smith. “The Effects of Road Deicing Salt on
mentation of Best Practices for Pre-Wetting and Variable Salt the Levels of Ions in Roadside Soils in Southern Ontario.” Journal
Application Rates, 2001. of Environmental Management 19 (1984): 261–71.
11. Kuemmel, D. A. and Q. Bari. Benefit-Cost Comparison of Salt-Only 24. Hutchinson, F. E. and B. E. Olson. “The Relationship of Road Salt
Versus Salt-Abrasive Mixtures Used in Winter Highway Maintenance Applications to Sodium and Chloride Ion Levels in the Soil Bor-
in the United States, Marquette University, Milwaukee, WI, 1996. dering the Major Highways.” Highway Research Record 193,
12. Nixon, Wilfrid A. The Use of Abrasives in Winter Maintenance: National Research Council, Washington DC (1967).
Final Report of Project TR 434, Iowa Institute of Hydraulic Research 25. McBean, E. and S. Al-Nassri. “Migration Pattern of De-Icing Salts
Technical Report No. 416, Iowa Department of Transportation and from Roads.” Journal of Environmental Management, no. 25 (1987):
the Iowa Highway Research Board, 2001. 231–38.
13. Chang, N.-Y. and B. Brady. Cost of Sanding, CDOT-DTD-R-2002- 26. McBean, E. A. “Roadside Pollutant Concentrations Arising From
5. Colorado Department of Transportation, 2002. Bare Pavement Policies.” Proc. Int. Conf. on Urban Hydrology under
14. Ice and Snow Technologies, LLC. Cypress Bowl Liquid Chemical Wintry Conditions, Narvik, Norway (1990): 19–21.
Case Study, the Insurance Corporation of British Columbia, 2001. 27. Eppard, R. A., et al. “Effects of Deicing Salt on Overstory Vegeta-
15. Breen, B. D. Anti Icing Success Fuels Expansion of the Program in tion in the Lake Tahoe Basin.” Transportation Research Record
Idaho, Idaho Department of Transportation, 2001. 1352 (1992).

147
28. Cain, N. P. et al. Review of the Effects of NaCl and Other Road Salts 46. U.S. EPA. “TENORM Sources: Summary Table.” Web page,
on Terrestrial Vegetation in Canada, Environment Canada, Com- [accessed November 2003]. Available at http://www.epa.gov/
mercial Chemicals Evaluation Branch, 2001. radiation/tenorm/sources_table.htm.
29. Kelsey, P. and R. G. Hootman. “Deicing Salt and Effects on Vege- 47. NYSDEC. An Investigation of Naturally Occurring Radioactive
tation along Highways. Case Study: Deicing Salt Deposition on the Materials (NORM) in Oil and Gas Wells in New York State, New
Morton Arboretum.” Chemical Deicers and the Environment, York State Department of Environmental Conservation, Division
253–81. Chelsea, Michigan: Lewis Publishers, 1992. of Soil and Hazardous Materials, 1999.
30. Chung, Y. G.“Roadside Air Pollution.” Chinnok 4, no. 1 (1981): 4-5. 48. Gould, G. A. and D. L. Grant. Cases and Materials on Water Law.
31. Organization for Economic Co-operation and Development St. Paul, MN: West Publishing Co, 1995.
(OECD) Scientific Expert Group. Road Transport Research: Cur- 49. Cappelow, C. C, et al. The Evaluation Procedures for Deicing Chem-
tailing Usage of De-Icing Agents in Winter Maintenance, Paris, 1989. icals and Improved NaCl, SHRP-H-205. Strategic Highway
32. Transportation Research Board. Special Report 235: Highway Research Program, Washington DC, 1993.
Deicing—Comparing NaCl and CMA. National Research Council, 50. Manahan, S. E. Environmental Chemistry. Boca Raton FL: Lewis
Washington DC, 1991. Publishers, 1994.
33. Granato, G. E. and K.P. Smith. Estimating Concentrations of Road- 51. Waters, T. F. “Sediment in Streams, Sources, Biological Effects and
Salt Constituents in Highway-Runoff from Measurements of Specific Control Monograph No. 7.”American Fisheries Society, Bethesda,
Conductance, Water Resources Investigation Report 99-4077. U.S. MD, 1995.
Department of the Interior, U.S. Geological Survey, Massachusetts 52. Lewis, W. M. Jr. Studies of Environmental Effects of Magnesium
Highways Department, Northborough, Massachusetts, 1999. Chloride Deicer in Colorado, Report No. CDOT-DTD-R-99-10.
34. Mussato, B. T. Assessment of Environmental Impacts of Liquid Colorado Department of Transportation, Research Branch, 1999.
Magnesium Chloride from Anti Icing, Insurance Corporation of 53. Wetzel, R. G. Limnology: Lake and River Ecosystems. San Diego,
British Columbia, 2001. CA: Academic Press, 2001.
35. Lewis, W. Magnesium Chloride Deicer: A Literature Review with 54. Meybeck, M. and R. Helmer.“The Quality of Rivers: From Pristine
Emphasis on the State of Colorado, CDOT-CTI-95-5. Colorado Stage to Global Pollution.” Palaeogeography, Palaeoclimatology,
Department of Transportation, 1997. Palaeoecology 75 (1989): 283–309.
36. Horner, R. R. Environmental Monitoring and Evaluation of 55. Hem, J. D. Study and Interpretation of the Chemical Characteristics
Calcium Magnesium Acetate (CMA), National Cooperative High- of Natural Water, 2nd Edition, Geological Survey Water-Supply
way Research Program 305, 1988. Paper 1473. A Review of Chemical, Geological, and Hydrologic Prin-
37. Ecological Planning and Toxicology Inc. A Critical Review of ciples and Processes That Control the Composition of Natural Water,
Methods for Developing Ecological Soil Quality Guidelines and With Methods for Studying and Interpreting Chemical Analyses, US
Criteria, American Petroleum Institute Biomonitoring Task Force, Government Printing Office, Washington DC, 1975.
Washington DC, 1999. 56. Dickman, M. D. and M. B. Gochnaur. “Impact of Sodium Chlo-
38. U.S. EPA. Contaminant Candidate List Preliminary Regulatory Deter- ride on the Microbiota of a Small Stream.” Environmental Pollu-
mination Support Document for Sodium, EPA 815-R-01-014, 2001. tion 17 (1978): 109–26.
39. Environment Canada, Ecoplans Ltd. “List of Canadian Programs, 57. Feick, G., R. A. Horne, and D. Yeaple. “Release of Mercury from
Initiatives, and Guidelines Related to Road Salts.” Web page, June Contaminated Freshwater Sediments by the Runoff of Road Deic-
2001. Available at http://www.ec.gc.ca/noop/roadsalt/en/list.htm. ing Salt.” Science 175 (1972): 1142–43.
40. Environment Canada. “Risk Management Strategies for Road 58. Cowan, J. A. The Biological Chemistry of Magnesium. New York,
Salts.” Web page, Available at http://www.ec.gc.ca/nopp/roadsalt/ NY: VHC, 1995.
en/rms.htm. 59. Brady, N. C. and R. Weil. The Nature and Properties of Soils. Upper
41. Bright, D. A. and J. Addison. Derivation of Matrix Soil Standards Saddle River, NJ: Prentice Hall, 1996.
for Salt under the British Columbia Contaminated Sites Regula- 60. Stumm, W. and J. J. Morgan. Aquatic Chemistry: Chemical Equi-
tion. Report to the British Columbia Ministry of Water, Land libria and Rates in Natural Waters. New York, NY: John Wiley &
and Air Protection, Ministry of Transportation and Highways, Sons, Inc, 1996.
British Columbia Buildings Corporation and Canadian Associ- 61. Cheng, K. C. and T. F. Guthrie. Liquid Road Deicing Environment
ation of Petroleum Producers, 2002. Impact, Insurance Corporation of British Columbia, 1998.
42. Canadian Council of Ministers of the Environment.“Summary of 62. Moran, V. M., L. A. Weinberger, and L. W. Abron. A Comparison of
Existing Canadian Environmental Quality Guidelines.” Web page, Conventional and Alternative Deicers: An Environmental Impact
2002. Available at http://www.ccme.ca/assets/pdf/e1_062.pdf. Assessment, 341–61. Chelsea, Michigan: Lewis Publishers, 1992.
43. Pacific Northwest Snowfighters. Pacific Northwest Snowfighters 63. Bubeck, R. C., et al. “Runoff of Deicing Salt: Effect in Irondequoit
Snow and Ice Control Chemical Products Specifications and Test Bay, Rochester New York.” Science 172 (1971): 1128–32.
Protocols for the PNS Association of British Columbia, Idaho, 64. Diment, W. H., R. C. Bubeck, and B. L Ceck. Some Effects of Deic-
Montana, Oregon, and Washington, 2002. ing Salton of Irondequoit Bay and Its Drainage Basin, Highway
44. Mussato, B. T. and T. Guthrie. Anti Icers—Chemical Analysis and Research Board 425: 23–24, 1973.
Toxicity Test Results, Insurance Corporation of British Columbia, 65. Kelly, Underwood Hansen. Influence of Highway Deicing Salts
2000. Upon Water Quality in Chocolate Lake, Halifax, Nova Scotia, Nova
45. Otton, J. K., S. Asher-Bolinder, D. E. Owen, and L. Hall. Effects of Scotia Department of the Environment, Halifax, NS, 1976.
Produced Waters at Oilfield Production Sites on the Osage Indian 66. Thurman, E. M. Organic Geochemistry of Natural Waters. Dor-
Reservation, Northeastern Oklahoma, Open File Report 97-28. U.S. drecht, the Netherlands: Martinus Nijhoff/Dr W. Junk Publishers,
Department of the Interior, U.S. Geological Survey, 1997. 1985.

148
67. APHA. American Waterworks Association Water Environment 88. Westing, A. H. “Plants and Salt in the Roadside Environment.”
Federation. Standard Methods for the Examination of Water and Phytopathology (1969): 1174–81.
Wastewater. 20th Edition ed. Washington DC: American Public 89. Bohn, J, B. L. McNeal, and G. A. O’Conner. Soil Chemistry. New
Health Association, 1998. York: Wiley-Interscience, 1985.
68. Lewis, W. M. Jr. Preliminary Environmental Evaluation of Calibre 90. Brandt, G. H.“Potential Impacts of Sodium Chloride and Calcium
M1000 Deicer for Use in Colorado. CDOT Report No. 200-12. Chloride De-Icing on Roadside Soils and Plants.” Highway
Colorado Department of Transportation, 2000. Research Record 425 (1973): 52–66.
69. Fischel, M. Evaluation for Selected Deicers Based on a Review of the 91. Winters, G. R., J. Gidley, and H. Hunt. Environmental Evaluation
Literature, Colorado Department of Transportation Research of Calcium Magnesium Acetate (CMA), California Department of
Branch, 2001. Transportation, CA, 1985.
70. Meeussen, J. C. L., M. G. Keizer, and F. A. M. DeHaan. “Chemical 92. Amrhein, C., et al. “Trace Metal Solubility in Soils and Waters
Stability and Decomposition Rate of Iron Cyanide Complexes in Receiving Deicing Salts.” Journal of Environmental Quality 23, no.
Soil Solutions.” Environmental Science and Technology 26 (1992): 2 (1994): 219–27.
511–16. 93. Erhart, E. and W. Hartl. “Effects of Potassium Carbonate as an
71. Dodds, W. K. Freshwater Ecology: Concepts and Environmental Alternative De-Icer on Ground Vegetation and Soil.” Ann. App.
Applications. San Diego, CA: Academic Press, 2002. Biol 136 (2000): 281–89.
72. Kalff, J. Limnology: Inland Water Ecosystems. Upper Saddle River, 94. Amrhein, C., J. E. Strong, and P. A. Mosher. “Effect of Deicing Salt
NJ: Prentice Hall, 2001. on Metal and Organic Matter Mobilization in Roadside Soils.”
73. Welch, E. B. Ecological Effects of Wastewater: Applied Limnology and Environmental Science and Technology 26 (1992): 703–9.
Pollutant Effects. London, England: Chaman & Hall, 1992. 95. Klecka, G. M., C. L. Carpenter, and B. D. Landenberger.“Biodegra-
74. Levine, S. N., and D. W. Schindler. “Radiochemical Analysis of dation of Aircraft Deicing Fluids in Soil at Low Temperatures.”
Orthophosphate Concentrations and Seasonal Changes in the Flux Ecotoxicology and Environmental Safety 25 (1993): 280–295.
of Orthophosphate to Seston in Canadian Shield Lakes.” Canadian 96. Prince Edward Island, Agriculture and Forest. “Urea Fertilizer in
Journal of Fisheries and Aquatic Sciences 37 (1980): 479–87. Crop Production.” Web page [accessed 2003]. Available at
75. OECD (Organization for Economic Co-operation and Develop- http://www.gov.pe.ca/af/agweb/index.php3?number=71253.
ment). Eutrophication of Waters: Monitoring, Assessment and 97. Oberts, G. L.“Pollutants Associated with Sand and Salt Applied to
Control, OECD, Paris, France, 1982. Roads in Minnesota.” Water Resources Bulletin 22, no. 3 (1986):
76. Dodds, W. K. and E. B. Welch. “Establishing Nutrient Criteria in 479–83.
Streams.” Journal of the North American Benthological Society 19 98. Olson, A. C. and T. Ohno. “Determination of Free Cyanide Levels
(2000): 186–96. in Surface and Ground Waters Affected by Highway Salt Storage
77. Kaushal, S. and W. M. Lewis Jr. “Patterns in the Chemical Frac- Facilities in Maine.” FOCUS Conf. on Eastern Regional Ground-
tionation of Organic Nitrogen in Rocky Mountain Streams.” water Issues.
Ecosystems 6 (2003): 483–92. 99. Butler, B. J. and J. Addison. “Biological Effects of Road Salts on
78. U.S. EPA. Ambient Water Quality Criteria for Ammonia—1984, EPA Soil, Report Submitted to the Environment Canada CEPA Prior-
440/5-85-001. USEPA, Office of Water, Washington DC, 1985. ity Substances List.” Environment Canada, Commercial Chemi-
79. Lewis, W. M. Jr. “Yield of Nitrogen from Minimally Disturbed cals Evaluation Branch (2000).
Watersheds of the United States.” Biogeochemistry 57/58 (2002): 100. Gunter, M. and B. Wilke. “Effects of De-Icing on Soil Enzyme
375–85. Activity.” Water, Air and Soil Pollution (1982).
80. U.S. EPA. 1999 Update of Ambient Water Quality Criteria for 101. Transport Canada, Airport Group Safety and Technical Services.
Ammonia, EPA-822-R-99-014. USEPA, Office of Water 4304, Environmental Impact from the Use of Potassium Acetate (KA) at
Washington DC, 1999. Halifax International Airport, TP 12286E, 1994.
81. Resource Concepts Inc. Survey of Alternative Road Deicers 102. French, D. W. “Boulevard Trees are Damaged by Salt Applied to
Prepared for the State of Nevada Department of Transportation Streets.” Minnesota Farm and Home Science 16, no. 2 (1959).
and the State of California Department of Transportation, 1992. 103. Sucoff, E. “Effects of Deicing Salts on Woody Vegetation along
82. Elliot, H. A. and J. H. Linn. Journal of Environmental Quality 16, Minnesota Roads.” Minnesota Highways Department, Investiga-
no. 3 (1987). tion 636, Final Report, St. Paul Minnesota, 1975.
83. Amrhein, C. and J. E. Strong.“The Effect of Deicing Salts on Trace 104. Lacasse, N. L. and A. E. Rich. “Maple Decline in New Hampshire.”
Metal Mobility in Roadside Soils.” Journal of Environmental Phytopathology 54 (1964): 1071–75.
Quality 19, no. 4 (1990). 105. Bowers, M. C. and J. H. Hesterberg. “Environmental Implication
84. Public Sector Consultants, Inc. The Use of Selected Deicing Mate- for Highway Deicing Agents on White Pine in Marquette County,
rials on Michigan Roads: Environmental and Economic Impacts, Michigan.” The Michigan Botanist 15 (1976): 77–89.
Michigan Department of Transportation, Michigan, 1993. 106. Scharpf, R. F. and M. Srago. Conifer Damage and Death Associated
85. Doner, H. E.“Chloride as a Factor in Mobilities of Ni(II), Cu(II), and with the Use of Deicing Salt in the Lake Tahoe Basin of California
Cd(II) in Soil.” Soil Science Society of America 42 (1978): 882–85. and Nevada, Forest Pest Control Technical Report 1, California
86. Pilon, E., and K. W. F. Howard. “Contamination of Subsurface Region, USDA Forest Service, Sacramento, 1975.
Waters by Road Deicing Chemicals.” Water Pollution Research 107. Gidley, J. The Environmental Impact of Highway Deicing, Publication
Journal of Canada 22, no. 1 (1987): 157–72. 33 (1990): 20–48.
87. Carter, M. R. for Canadian Society of Soil Science. Soil Sampling 108. Kliejunas, J., M. Marosy and J. Pronos.“Conifer Damage and Mor-
and Methods of Analysis. Boca Raton, Florida: Lewis Publishers, tality Associated With Highway Deicing and Snow Removal in the
1993. Lake Tahoe Area.” 1989.

149
109. Backman, L. and L. Folkeson. “Influence of Deicing Salt on Vege- 128. Fraser, D.“Sightings of Moose, Deer and Bears on Roads in North-
tation, Groundwater and Soil along Two Highways in Sweden.”4th ern Ontario.” The Wildlife Society Bulletin 7, no. 3 (1979): 181–84.
International Symposium for Snow Removal and Ice Control Tech- 129. Fraser, D.G. Fish and Wildlife Review, Ontario Ministry of Natural
nology, Washington DC: National Academy Press, 1996. Resources 17, No. 4 (1978).
110. Fostad, O. and P. A. Pefersen. “Vitality, Variation and Causes of 130. McClain, T. Motor Vehicle-Deer Crashes in Wisconsin in 2002,
Decline of Trees in Oslo Center (Norway).” Journal of Arboricul- Wisconsin Department of Transportation, 2002.
ture 23, no. 4 (1997): 155–65. 131. Hiatt, G. F. S., et al. “Calcium Magnesium Acetate: Comparative
111. Hanes, R. E., C. W. Zelazney, and R. E. Blaser. Effects of Deicing Salts Toxicity Tests and Industrial Hygiene Site Investigation.” Trans-
On Water Quality and Biota—Literature Review and Recom- portation Research Record 1157 (1989).
mended Research, Highway Research Board, Report 91, Washing- 132. Rosenberg, A., C. M. Hansson, and C. Andrade. “Mechanisms of
ton DC, 1970. Corrosion of Steel in Concrete.” 285–313. American Ceramic
112. Marshall, J. P. “Evaluation for Vegetative Damage along Nevada Society, 1989.
Highways in the Lake Tahoe Basin.” Forest Pest Management, State 133. ACI. “ACI Committee 222R-96 Corrosion of Metals in Concrete.”
and Private Forestry, Intermountain Region, USDA Report No. 84–3 American Concrete Institute, 1996.
(1984). 134. Raupach, M. “Chloride-Induced Macrocell Corrosion of Steel in
113. Davis, G., S. Krannitz, and M. Goldstein. A Reconnaissance Study Concrete—Theoretical Background and Practical Consequences.”
of Roadside Tree Injury and Decline at 17 Sites in Interior British Construction and Building Materials 10, no. 5 (1996): 329–38.
Columbia, Roadside Tree Injury Committee, Victoria BC, 1992. 135. Newton, C. J. and J. M. Sykes. “The Effect of Salt Additions on the
114. Blomqvist, G. De-Icing Salt and Roadside Environment—Strategies Alkalinity of Ca(OH)2 Solutions.” Cement and Concrete Research
for Impact Analyses, Swedish National Road and Transport 17 (1987): 765–76.
Research Institute, 2001. 136. Deja, J. and G. Loj. “Effects of Cations Occurring in the Chloride
115. Eppard, R. A., et al. “Effects of Deicing Salt on Overstory Vegeta- Solutions on the Corrosion Resistance of Slag Cementitious Mate-
tion in the Lake Tahoe Basin.” Transportation Research Record 1352 rials.” R. N. Swamy, 603–620, 1999.
(1992). 137. Kondo, R., M. Satake, and H. Ushiyama.“Diffusion of Various Ions
116. California Fertilizer Association. Western Fertilizer Handbook. into Hardened Portland Cement.” 41–43, Cement Association of
Dansville, Illinois: The Interstate Printers and Publishers Inc., Japan, 1974.
1985. 138. Deja, J. and G. Loj. “Effects of Cations Occurring in the Chloride
117. Tisdale, S. L. and W. L. Nelson. Soil Fertility and Fertilizers, Solutions on the Corrosion Resistance of Slag Cementitious Mate-
MacMillan Co., New York, 1966. rials.” R. N. Swamy, 603–620, 1999.
118. Environment Canada. Calcium Chloride: Environmental and Tech- 139. Kondo, R., M. Satake, and H. Ushiyama.“Diffusion of Various Ions
nical Information for Problem Spills, Environment Canada, Protec- into Hardened Portland Cement.” 41–43, Cement Association of
tion Programs Directorate, Technical Services Branch, 1984. Japan, 1974.
119. Paul, R., M. Rocer, and R. Ipens. “Influence of Winter De-Icing 140. Callahan, M. R. “Deicing Salt Corrosion with and without
With Calcium Chloride on Sorbus, Acer, Tilia and Plantanus.” The Inhibitors.” Transportation Research Record 1211 (1990): 12–17.
Science of the Total Environment 59 (1987): 277–82. 141. Nielsen, E. P. and M. R. Geiker. “Chloride Diffusion in Partially
120. David Nairne and Associates. Potential Impact of Calcium Chlo- Saturated Cementitious Material.” Cement and Concrete Research
ride: An Environmental Assessment as a Dust Control Agent, Yukon 33 (2003): 133–38.
Highways Maintenance Branch Project 2404, 1992. 142. Mends, N. and P. Carter. “Economic Impact of Magnesium Chlo-
121. Bogemans, J. L., J. M. Neirinckx, and Stassart. “Effects of Deicers ride Anti-Icing on Montana Bridges.” 739–744, Canadian Society
Sodium Chloride and Calcium Chloride on Spruce (Picea Abies).” for Civil Engineering, 2002.
Plant and Soil 120, no. 2 (1989): 203–11. 143. Williams D. “Rebuttal to Report by Mends and Carter.” 2003.
122. McLaren Environmental Engineering. “Dust-Off Environmental 144. Moukwa, M. “Characteristics of the Attack of Cement Paste by
Safety Report for Leslie Salt Company.” Leslie Salt Company, MgSO4 and MgCl2 From Pore Structure Measurements.” Cement
Newark, CA, 1989. and Concrete Research 20 (1990): 148–58.
123. Nolan Davis. A Comparison of the Effects of Urea, Potassium 145. Cody, R. D., A. M. Cody, P. G. Spry, and G.-L. Gan. “Experimental
Acetate, Calcium Magnesium Acetate and Sodium Formate Runway Deterioration of Highway Concrete by Chloride Deicing Salts.”
Deicers on the Environment, Transport Canada Report TP12285E, Environmental and Engineering Geoscience 2, no. 4 (1996): 575–88.
Ottawa, Ontario, 1994. 146. Arya, C., N. R. Buenfeld, and J. B. Newman. “Factors Influencing
124. Audesirk, G. and T. Audesirk. Biology, Life on Earth. New York: Chloride-Binding in Concrete.” Cement and Concrete Research 20,
MacMillan Publishing Co., 1993. no. 2 (1990): 291–300.
125. Boulder (City of), Colorado. Boulder Transportation Master Plan, 147. Salcedo, R. N. and W. Jensen. “Corrosivity Tests Pit Road Salt vs.
Chapter 9, Mobile Source Air Pollution, City of Boulder Colorado, CMA.” Public Works 118, no. 11 (1987): 58–61.
1996. 148. McCrum, R. L. “Corrosion Evaluation of Calcium Magnesium
126. U.S. EPA.“Denver’s Road to Attainment for Six Criteria Air Pollu- Acetate (CMA), Salt (NaCl), and CMA/Salt Solutions.”Corrosion 89.
tants.” Web page, [accessed November 2003]. Available at 149. Ihs, A. and K. Gustafson. “Test and Evaluation of Calcium Mag-
http://www.epa.gov/region8/air/history.html. nesium Acetate-Sodium Chloride Mixtures in Sweden.” Snow and
127. Brownlee, L. Mineau P. and A. Baril. Road Salts and Wildlife—An Ice Control Technology—Fourth International Symposium, Trans-
Assessment of the Risk, Report Submitted to the Environmental portation Research Board, National Research Council, 1997.
Resource Group for Road Salts, Environment Canada, Commercial 150. Callahan, M. R. “Deicing Salt Corrosion with and without
Chemicals Evaluation Branch, 2000. Inhibitors.” Transportation Research Record 1211 (1990): 12–17.

150
151. Locke, C. E., et al. “A Study of Corrosion Properties of a New 171. Murray, D. and U. Ernst. An Economic Analysis of the Environmen-
Deicer, Calcium Magnesium Acetate.” Transportation Research tal Impact of Highway Deicing, Office of Research and Develop-
Record 1113 (1988). ment, Municipal Environmental Research Laboratory,
152. Kennelley, K. J. and C. E. Locke. “Electrochemical Behaviour of Environmental Protection Agency, Cincinnati, 1976.
Steel in Calcium Magnesium Acetate.” Corrosion 46, no. 11 (1990): 172. Menzies T.R. “National Cost of Motor Vehicle Corrosion from
888–95. Deicing Salts.” Corrosion/91, Houston, TX: NACE International,
153. Ushirode, W. M., J. T. Hinatsu, and F. R. Foulkes. “Voltammetric 1991.
Behaviour of Iron in Cement Part IV: Effect of Acetate and Urea 173. Koch G.H., et al. “Corrosion Cost and Preventive Strategies in the
Additions.” Journal of Applied Electrochemistry 22, no. 3 (1992): United States.” Report No. FHWA-RD-01-156. US Department of
224–29. Transportation, Federal Highway Administration, Office of Infra-
154. Chollar, B. H. and Y. P. Virmani. “Effects of Calcium Magnesium structure Research and Development, 2002.
Acetate on Reinforced Steel Concrete.” Public Roads 51, no. 4 174. ASM International.“Metals Handbook Ninth Edition, Volume 13,
(1988): 113–15. Corrosion.” Materials Park, Ohio: ASM International, 1987.
155. Committee on the Comparative Costs of Rock Salt and Calcium
175. Soepenberg E.N. “History of Materials Technology for the Auto-
Magnesium Acetate (CMA) for Highway Deicing. Highway
motive Industry—A Continuing Environment-Materials-Energy
Deicing—Comparing Salt and Calcium Magnesium Acetate, Spe-
Debate.” Corrosion/93, Houston, TX: NACE International, 1993.
cial Report 235. Transportation Research Board, National
176. Rendahl B. and S. Hedlund. “The Influence of Road Deicing Salts
Research Council, Washington, DC, 1991.
on Motor Vehicle Corrosion.” Corrosion/91, Houston, TX: NACE
156. Slick, D. Effects of CMA on Pavements and Motor Vehicles, FHWA-
RD-87-037. US Department of Transportation, FHWA, 1987. International, 1991.
157. Pigeon, M. and R. Pleau. Durability of Concrete in Cold Climates. 177. Kruger J. “Corrosion Mechanisms Relevant to the Corrosion of
Modern Concrete Technology Series. New York, NY: E & FN Automobiles.” Corrosion/91, Houston, TX: NACE International,
Spon, 1995. 1991.
158. Marchand, J., E. J. Sellevod, and M. Pigeon. “The Deicer Salt Scal- 178. Bryant A.W., et al. “US Automotive Corrosion Trends at Five and
ing Deterioration of Concrete—An Overview.” 1–46, 1994. Six Years”. Automotive Corrosion and Protection, p. 3-1 to 3-17, ed.
159. Lee, H., et al. “Effects of Various Deicing Chemicals on Pavement Robert Baboian, Houston, TX: NACE International, 1992.
Concrete Deterioration.” 151–55, 2000. 179. Nakazawa M., Y. Miyoshi, and D. Davidson.“Corrosion Investiga-
160. Balasubramanian, T. M., et al. “Deterioration of Concrete in tion from 10-Year Field Exposure Vehicles.” Corrosion/95, Hous-
Industrial Environments.” 1533–38, Central Electrochemical ton, TX: NACE International, 1995.
Institute, Karaikudi, India, 1987. 180. Simpson M.W., et al. “License-Plate Cosmetic Corrosion Tests of
161. Beaudoin, J. J., S. Catinaud, and J. Marchand. “Volume Stability of Automotive Coated Steel Sheet.” Corrosion/98, Houston, TX:
Calcium Hydroxide in Aggressive Solutions.” Cement and Concrete NACE International, 1998.
Research 31 (2001): 149–51. 181. Fujita S.“Perforation Corrosion and Its Mechanism of Galvanized
162. Cody, R. D., et al. “The Role of Magnesium in Concrete Deterio- Steel Sheets on Vehicles”. Corrosion/98, Houston, TX: NACE
ration.” Final Report—Iowa DOT HR-355. The Iowa Highway International, 1998.
Research Board, 1994. 182. ASM International. “Metals Handbook Ninth Edition Volume 13
163. Wakeley, L. D., et al. “Geochemical Stability of Cement-Based Corrosion.” Corrosion in the Automotive Industry, 1011–18. Mate-
Composites in Magnesium Brines.” 333–50, International Cement rials Park, Ohio: ASM International, 1987.
Microscopy Association, 1992. 183. Baboian R. “Chemistry and Corrosivity of the Automotive Envi-
164. Newton, C. J. and J. M. Sykes. “The Effect of Salt Additions on the ronment.” Automotive Corrosion and Protection, p. 6-1 to 6-19, ed.
Alkalinity of Ca(OH)2 Solutions.” Cement and Concrete Research Robert Baboian, Houston, TX: NACE International, 1992.
17 (1987): 765–76. 184. NACE International. “NACE Basic Corrosion Course.” Houston,
165. Helmy, I. M., A. A. Amer, and H. El-Didamony. “Chemical Attack TX: NACE International, 1971.
on Hardened Pastes of Blended Cements—Part 1: Attack of Chlo-
185. Committee on the Comparative Costs of Rock Salt and Calcium
ride Solutions.” Zement-Kalk-Gips 44, no. 1 (1991): 46–50.
Magnesium Acetate (CMA) for Highway Deicing. “Highway
166. Rechenberg, W. and H.-M. Sylla. “The Effect of Magnesium on
Deicing—Comparing Salt and Calcium Magnesium Acetate,”
Concrete.” ZKG International 49, no. 1 (1996): 44–56.
Special Report 235. Transportation Research Board, National
167. Peterson, O. “Chemical Effects on Cement Mortar of Calcium
Research Council, Washington, D.C., 1991.
Magnesium Acetate as Deicing Salt.” Cement and Concrete
Research 25, no. 3 (1995): 617–26. 186. Smith G.D., J. R. Crum, and H. L. Flower.“High Temperature Cor-
168. Santagata, M. C. and M. Collepardi. “The Effect of CMA Deicers rosion Performance of Automotive Coupling Alloys.” Corrosion/
on Concrete Properties.” Cement and Concrete Research 30 (2000): 98, Houston, TX: NACE International, 1998.
1389–94. 187. Xi Y. and Z. Xie. Corrosion Effects of Magnesium Chloride and
169. Pianca, F., K. Carter, and H. Sedlak. A Comparison of Concrete Scal- Sodium Chloride on Automobile Components, No. CDOT-DTD-R-
ing Resistance Caused by Calcium Magnesium Acetate and Sodium 2002-4. Colorado Department of Transportation, Research
Chloride in Laboratory Tests, Ontario Ministry of Transportation Branch, Denver, CO, 2002.
and Communications, 1987. 188. Safford C.M, and T. G. Sanders. “Deicers: Melting Away Winter
170. Manning, D. G. and L. W. Crowder. A Comparative Field Study of Woes.” Snow & Ice Manager (2002): 5–11.
Calcium Magnesium Acetate and Rock Salt During the Winter of 189. Chappelow C.C., et al.“Handbook of Test Methods for Evaluating
1986–1987, Ontario Ministry of Transportation and Communi- Chemical Deicers,” SHRP-H-332. Strategic Highway Research Pro-
cations, 1987. gram, National Research Council, Washington, DC, 1992.

151
190. Koefod R.S. “Mechanism of Action of Corrosion-Inhibited High- 212. U.S. Department of Transportation, Federal Highway Adminis-
way Deicers.” Transportation Research Record 1490: p. 75–82. tration Highway Rail Grade Crossing Technical Working Group.
191. Washington State Department of Transportation. “WSDOT Salt Guidance on Traffic Control Devices at Highway-Rail Grade
Pilot Project Draft Report,” 2003. Crossings, 2002.
192. Nixon W.A. and A. D. Williams. “A Guide for Selecting Anti-Icing 213. Federal Highway Administration. Warning Time Requirements at
Chemicals, Version 1.0”, IIHR Technical Report No. 420. IIHR- Railway-Highway Grade Crossings With Active Traffic Control,
Hydroscience and Engineering, College of Engineering, the Report No. FHWA SA-91-007, Washington, DC, 1991.
University of Iowa, 2001. 214. Ontario Hydro and CEA, 1997. “The Presentation of Pole Fires:
193. Fink E., et al. “Motor Carrier Concerns with Corrosion and The Utility Guide to Reducing the Incidence of Pole Fires.”
Vehicle Damage, Transcript of Panel Discussion.” Undated. Web page, [accessed October 2003]. Available at http://www.
194. Ihs A. “Test and Evaluation of Alternative De-Icing Methods and alexanderpublications.com.
Materials in Sweden.” Sweden-Japan Workshop on Winter Road 215. Chappelow, C. C, et al. Handbook of Test Methods for Evaluating
Management & Pavement, December 11–13, 2000, ed. Kent Chemical Deicers, SHRP-H-332. Strategic Highway Research
Gustafson, Linkoping, Sweden: Swedish National Road and Trans- Program, Washington, DC, 1992.
port Research Institute, 2001. 216. CRC Press Inc, Weast R. C. Editor. CRC Handbook of Chemistry
195. Suga H. and T. Suzuki. “Development of New Chloride Group and Physics. Boca Raton, Florida: CRC Press Inc, 1980.
Deicing/Anti-Freezing Agents.” 217. Perkin Elmer. Web page [accessed September 2003]. Available at
196. McCrum R. L.“Corrosion and Alternate Deicers, Interim Report,” http://las.perkinelmer.com.
Research Report No. R-1315. Michigan Department of Trans- 218. ASTM.“ASTM G-31 Standard Practice for Laboratory Immersion
portation, Lansing, Michigan, 1991. Corrosion Testing of Metals.” Philadelphia, PA: ASTM.
197. Ireland D. T. “Chemical Deicers and the Environment.” Chapter 219. NACE Technical Practices Committees.“NACE Standard TM-01-
14, New Salt Based Highway Deicers with Improved Anti-Corrosive 69 Test Method—Laboratory Corrosion Testing of Metals for the
Properties. 403-20. Boca Raton: Lewis Publishers, 1992. Process Industries.” Houston, TX: NACE International.
198. Baboian R. and R. C. Turcotte. “Development of Poultice Corro- 220. Dean S.W. “Automobile Atmospheric Corrosion—Testing and
sion Tests for Automobiles.” Materials Performance, Vol. 24, no. 12 Evaluation.” Materials Performance Volume 30 (1991): 48–57.
221. Moran J.P. and M. W. Egbert. “Galvanic Corrosion of Aluminum
(1985): p. 13–17.
Auto Body Sheet Coupled to Steel.” Corrosion/95, Houston, TX:
199. Public Sector Consultants, Inc. “The Use of Selected Deicing
NACE International, 1995.
Materials on Michigan Roads: Environmental and Economic
222. Snodgrass J.S. and J. R. Weir. “Corrosion Test Selection for Alu-
Impacts.” Public Sector Consultants Inc., 1993.
minum Autobody Sheet with Chromium-Free Pretreatments.”
200. McCrum R.L. “Corrosion Evaluation of Calcium Magnesium
Corrosion/95, Houston, TX: NACE International, 1995.
Acetate (CMA), Salt (NaCl), and CMA/Salt Solutions.” Corrosion/
223. ASTM. “ASTM G44 Standard Practice for Evaluating Stress Cor-
89, Houston, TX: NACE International, 1989.
rosion Cracking Resistance of Metals and Alloys by Alternate
201. ASM International. “ASM Handbook, Volume 13A, Corrosion:
Immersion in 3.5% Sodium Chloride Solution.” Philadelphia, PA:
Fundamentals, Testing, and Protection.” Materials Park, Ohio:
ASTM, 1994.
ASM International, 2003.
224. ASTM. “ASTM B 117 Standard Practice for Operating Salt Spray
202. Uhlig H.H. The Corrosion Handbook. 1948. (Fog) Testing Apparatus.” ASTM, Philadelphia, PA: ASTM.
203. Uhlig’s Corrosion Handbook, Second Edition. New York: John 225. ASTM. “ASTM B 368 Standard Method for Copper-Accelerated
Wiley and Sons Inc., 2000. Acetic Acid-Salt Spray (Fog) Testing (CASS Test).”Philadelphia,
204. “Winter Maintenance: Snow and Ice Removal.” University of PA: ASTM.
Michigan, 2003. 226. ASTM. “ASTM G 85 Standard Practice for Modified Salt Spray
205. Williams D.“Past and Current Practices of Winter Maintenance at (Fog) Testing.” Philadelphia, PA: ASTM, 1985.
the Montana Department of Transportation (MDT),” 2001. 227. ASTM. “ASTM G 96 Standard Practice for Conventions Applica-
206. Fischel, M. “Evaluation of Selected Deicers Based on a Review of ble to Electrochemical Measurements in Corrosion Testing.”
the Literature.” CDOT-DTD-R-2001-15. Colorado Department of Philadelphia, PA: ASTM.
Transportation, 2001. 228. ASTM. “ASTM G96 Standard Guide for On-Line Monitoring of
207. Ihs A. and K. Gustafson. “Test and Evaluation of Calcium Corrosion in Plant Equipment (Electrical and Electrochemical
Magnesium Acetate-Sodium Chloride Mixtures in Sweden.” Snow Methods).” Philadelphia, PA: ASTM, 1990.
Removal and Ice Control Technology—Fourth International Sym- 229. ASTM. “ASTM G 102 Standard Practice for Calculation of Corro-
posium, Reno, Nevada, August 11–16, 1996, Washington, DC: sion Rates and Related Information from Electrochemical Mea-
Transportation Research Board, National Research Council, 1997. surements.” Philadelphia, PA: ASTM.
208. McCrum R.L. “Chemical Deicers and the Environment.” Chapter 230. SAE International. “SAE J2334 Cosmetic Corrosion Lab Test.”
13, Corrosion and Alternate Deicers, 363–402. Boca Raton: Lewis Warrendale, PA: SAE International.
Publishers, 1992. 231. Lutze F.W.,et al.“The Effects of Temperature and Salt Composition on
209. Corrosion Inhibited Calcium Chloride, Liquid Hard on Ice, Easy on the Speed of the SAE J2334 Cyclic Corrosion Test.”ENROCORR 2000:
Corrosion. General Chemical Corporation, Undated. Past Success—Future Challenges, September 10–14, 2000, London.
210. Setting the Record Straight: Calcium Chloride vs. Magnesium 232. “Test Method for Evaluation of Corrosive Effects of Deicing
Chloride. Harrow, Ontario: Undated. Chemicals on Metals.” SHRP H-205.7. Strategic Highway
211. Mori N., et al.“Snowy and Ice Road Countermeasures Using Anti- Research Program, National Research Council, Washington,
Freezing Admixture.” DC, 1992.

152
233. Townsend H.E. “Behavior of Painted Steel and Aluminum Sheet 237. Gao G. “In-Situ Electrochemical Study of Corrosion of Steel and
in Laboratory Corrosion Tests.” Corrosion/95, Houston, TX: Aluminum/Steel Couples during Cyclic Corrosion Test.” Corrosion
NACE International, 1995. 98, Houston, TX: NACE International, 1998.
234. Whitton G.G. and J. W. Sprys. “Vehicle Component Corrosion- 238. Schumacher W.A., G. Hook, and L. L. Piepho. “Corrosion Testing
Resistant Coatings and Test Methods Comparison.” Corrosion/95, of Complete Vehicles at General Motors Proving Ground.” Auto-
Houston, TX: NACE International, 1995. motive Corrosion and Protection, p. 17-1 to 17-11, ed. Robert
235. Romanchik D. “Engineers Say “Salt Spray Test Is Bogus.” Newton, Baboian, Houston, TX: NACE International, 1992.
MA: Test and Measurement World, 2000. 239. Dean S.W. “Atmospheric Corrosion—Testing and Evaluation.”
236. Dunham K. “Private Communication.” American Galvanizers Automotive Corrosion and Protection, p. 12-1 to 12-12, ed. Robert
Association, 2002. Baboian, Houston, TX: NACE International, 1992.

Guidebook to the Decision

Tool, Purchase Specification,
and Quality Assurance
Monitoring Program

CONTENTS
G-1 Chapter 1 General Overview

G-1 1.1 Introduction
G-1 1.2 Material Selection Process Framework
G-4 Chapter 2 Decision Tool
G-4 2.1 Policy Objectives
G-4 2.2 Operational Objectives
G-4 2.3 Snow and Ice Control Strategy and Material Selection
G-5 2.3.1 Strategies
G-6 2.3.2 Material Selection
G-8 2.4 Maintenance Area Information
G-11 2.4.1 Aquatic Life Worksheet
G-11 2.4.2 Drinking Water from Groundwater Worksheet
G-12 2.4.3 Air Quality Worksheet
G-12 2.4.4 Vegetation Worksheet
G-12 2.4.5 Soil Worksheet
G-12 2.4.6 Terrestrial Fauna Worksheet
G-12 2.4.7 Vehicles Worksheet
G-12 2.4.8 Metal Infrastructure Worksheet
G-12 2.4.9 Concrete Matrix Worksheet
G-13 2.4.10 Concrete Reinforcing Worksheet
G-13 2.5 Identified Cost
G-13 2.6 Performance
G-14 2.6.1 Temperature Information
G-15 2.6.2 Performance Measured by Melting Potential
G-18 2.7 Decision Tool Application
G-27 2.8 Software
G-30 CHAPTER 3 Purchase Specification
G-30 3.1 Chemical Specifications
G-34 3.2 Other Specifications
G-34 3.2.1 Physical Specifications—Chemical Products
G-36 3.2.2 Physical Specifications—Abrasives
G-36 3.2.3 Toxicity Specifications
G-37 3.2.4 Corrosion Specifications
G-38 CHAPTER 4 Quality Assurance Monitoring Program
G-38 4.1 Documentation
G-38 4.2 Field QA Procedure
G-38 4.3 Sample Collection Procedure
G-39 4.4 Laboratory Testing Effort
G-39 4.5 Standard Test Methods
G-40 4.5.1 Sample Preparation

G-40 4.5.2 Active Ingredient

G-40 4.5.3 Metals
G-41 4.5.4 Total Phosphorus
G-41 4.5.5 Nitrogen
G-41 4.5.6 Cyanide
G-42 4.5.7 Organic Matter (BOD and COD)

G-1
CHAPTER 1
General Overview
1.1 Introduction Project 06-16 and provides a practical snow and ice control
material selection process (decision tool), including sup-
Every year, considerable quantities of snow and ice control porting technical information and contributing thought
products are applied to highways, and environmental and processes.
regulatory agencies have questioned the environmental
effects of these products. Transportation agencies are asked
1.2 Material Selection Process
to use “environmentally-friendly” or less toxic alternatives
Framework
wherever possible, but there is no commonly accepted guid-
ance for determining which products meet these criteria. When selecting a snow and ice control material, the factors
The traditional use of roadsalt has been prohibited in some affecting the decision processes can be numerous and com-
locations, leaving highway agencies uncertain about how plex. The framework provided here is relatively straightfor-
traffic safety can be maintained in winter conditions. ward and addresses the most common and significant items.
Studies of the most common chemical alternatives— The product is a practical tool that generates a numerical
sodium chloride (salt, NaCl), magnesium chloride (MgCl2), evaluation that can be used to compare snow and ice control
calcium chloride (CaCl2), and acetate-based materials—have materials. Although the research team has established the pri-
focused on performance and cost under various weather mary steps in the decision process, user input is a key element
conditions without evaluating their relative effects on the to ensure that maintenance area conditions are represented.
environment. Several new chemical preparations, including Users will gather and input a range of information associated
some that are proprietary formulations, have entered the with materials and maintenance area conditions. Even
market as snow and ice control chemicals for use by trans- though some information may be unfamiliar or considered
portation agencies, but there is limited information about irrelevant to agency conditions, a “complete” data set is nec-
their environmental impacts. essary so that relative rankings can be developed for materi-
There is a need for rational decision-making guidelines to als. A secondary benefit is that users will research and develop
assist winter maintenance managers in selecting the most knowledge in new areas, further increasing their understand-
appropriate snow and ice control materials for the conditions ing of the elements involved in snow and ice control material
in their jurisdictions. A transportation agency must consider a selection.
range of factors when assessing snow and ice control materials The overall process for selecting snow and ice control
for use. These include performance, cost, potential for the materials is illustrated in Figure 1-1. Each agency will have
material to impair the natural receiving environment, and unique objectives and priorities that will affect selections.
potential for the material to impair infrastructure. Each agency These objectives and priorities will be expressed through
will have unique objectives and conditions that influence the percentage weightings of the categories and subcategories
importance of each of these factors in the decision process. that form the framework for the snow and ice control mate-
The overall objective of this project is to develop guidelines rial selection process (Table 1-1). Decision categories
for selection of snow and ice control materials based on their include cost, performance, potential to impair the natural
properties, together with common site-specific conditions environment, and potential to impair infrastructure. These
near roadways on which these products would be used. This balance economic value related to cost and performance
report expands on the information developed through the with potential consequences of use related to environmental
information assessment conducted for Phase I of NCHRP and corrosion impacts.

G-2
Figure 1-1. Decision tool process flow chart.
Decision subcategories that support the natural environ- trol material type, numerical rankings were developed. These
ment and infrastructure decision categories are relatively rankings are based on a material’s potential to elicit an effect on
straightforward and represent the most significant compo- a receptor, as developed from a comprehensive information
nents contributing to the higher level decision category. review and analytical test results. These values are applied where
To quantify risks of impairment for environmental and infra- sensitive receptors exist in the maintenance area as identified
structure subcategories resulting from each snow and ice con- through user-completed receptor inventory worksheets.
Table 1-1. Decision tool framework.
Decision Category Decision Subcategory

(Policy Objectives) (Operational Objectives)
Purchase Price Not Applicable
Performance (as Melting Potential (MP)) Not Applicable
Aquatic Life
Drinking Water
Air
Natural Environment
Vegetation
Soil
Terrestrial Animals
Vehicles
Metals Infrastructure
Infrastructure
Concrete

G-3
When assessing relative cost and performance of snow and For each decision category, a weighted score is generated.
ice control materials, many factors are influential and consid- These are summed for each material to produce a numerical
erable research continues to be done here. The decision tool product score that is used to compare snow and ice control
uses the following approaches: material types objectively. Once a material type has been
selected, various sources may exist. Considerable variation in
• Cost is based on the purchase price of the dry active certain attributes can exist among materials from different
ingredient. Active ingredients include the five primary sources, or between batches from the same source. A purchase
snow and ice control chemicals: NaCl, MgCl2, CaCl2, specification containing concentration-based guidelines for
CMA, and KA. Section 2.5 provides further discussion of key attributes will allow transportation agencies to screen out
cost quantification. unsuitable products (Chapter 3). The purchase specification
• Performance is based on melting potential derived from is supported by a quality assurance monitoring program to
phase curve information and expected winter temperature assess received materials and by standard test methods to
in the maintenance region. Section 2.6 provides further measure these attributes (Chapter 4).
discussion of performance quantification. Product selection occurs after a review of the purchase
specification and supplemental considerations that may
These approaches may not be ideal; however, they provide a include, but are not limited, to product availability (e.g.,
standard, practical method that reasonably represents the delivery times and quantities), quality testing results, and
dominant component for each decision category. equipment availability.

G-4
CHAPTER 2
Decision Tool
The following sections detail the development of the snow monitor and communicate the effectiveness of future policy
and ice control material selection tool and discuss how it is to changes.
be used, including required inputs. Application notes are
included in the text and identify key steps in the decision tool
2.2 Operational Objectives
process.
Application Note 2 – Operational Objectives
2.1 Policy Objectives
Application Note 1 – Agency Objectives Identify operational objectives for decision subcategories
through numerical weighting.
Identify policy objectives for decision categories through
Natural environment and infrastructure decision cate-
numerical weighting. gories are supported by decision subcategories representing
the most significant components contributing to the higher
Transportation agencies have unique objectives and level decision criteria (Table 2-2). Users will apply percent
concerns specific to their maintenance areas that affect weightings to these according to agency priorities. Subcate-
selection of snow and ice control materials. Users will gory (operational) weightings enable users to prioritize areas
express these objectives through percent weightings for the of greatest concern. The potential extent or severity of dam-
decision categories shown in Table 2-1. These categories age is greatest for aquatic ecosystems, rather than for other
represent the most significant items influencing the snow environmental subcategories—an exception may be for abra-
and ice control material selection process. When combined, sives where road dust contributes to air quality concerns.
purchase price and performance represent economic value Therefore, it is recommended that the aquatic life subcate-
in the snow and ice control material (i.e., performance per gory not be assigned a weighting less than 25 percent, unless
unit cost). Natural environment and infrastructure cate- reasonable information suggests otherwise.
gories represent potential consequences of use and long-
term costs. Decision makers must balance value with
consequences. As an example, agencies wishing to reduce
2.3 Snow and Ice Control Strategy
environmental concerns by using alternate materials
and Material Selection
should ensure that cost and performance objectives are Significant research has been applied toward developing
also met. guidelines for selecting snow and ice control application
Assignment of weights to the decision categories should be methods (i.e., NCHRP Project 06-13). Application methods
conducted at the beginning of the snow and ice control mate- can influence volumes of materials used, which can influ-
rial selection process by management staff familiar with ence performance and the potential for negative impacts. A
agency policy decisions. By setting clear objectives at the onset challenge when developing a material selection process is to
of the decision process, managers can demonstrate trans- ensure that potential impairments related to a material type
parency in their decision criteria and establish a baseline to are considered independently of application methods.

G-5
Table 2-1. Policy objectives.
Decision Category Decision Weighting (%)

1
Purchase Price
Performance as Melting Potential (MP)2
Sum = 100 %
Natural Environment
Infrastructure
1
The Decision Tool uses purchase price per dry unit of active ingredient, including shipping as a cost
benchmark. This is a practical approach and yields a consistent and unbiased assessment of the
dominant cost element (refer to Section 2.5 for further discussion).
2
Given that the primary objective of winter maintenance is controlling snow and ice, it is recommended
that performance weighting not be assigned a value below 25%.
Table 2-2. Operational objectives.
Decision Category Decision Subcategory

Decision Weighting (%)
(Policy Objectives) (Operational Objectives)
Aquatic Life1
Drinking Water
Air Quality
Natural Environment Sum = 100 %
Vegetation
Soil
Animals
Vehicles
Metal Infrastructure
Infrastructure Sum = 100 %
Concrete Corrosion
Concrete Degradation
1
It is recommended that aquatic life weighting not be assigned a value below 25%
Establishing the application method before material selec- dry chemicals range from 25 kg per lane kilometer (100 lbs.
tion normalizes application method-based influences per lane mile) to 100 kg per lane kilometer (400 lbs. per
between materials, allowing the materials to be assessed lane mile), and sometimes more, depending on conditions.
based on their specific properties. Deicing typically involves the application of solid NaCl
and, as with any dry application, traffic action can result in
2.3.1 Strategies a significant loss of material to the roadside environment.
• Anti-icing, a proactive strategy, involves application of
Application Note 3 – Application Method Selection snow and ice control materials before a snow, ice, or frost
event. This strategy delays or can prevent precipitation
Identify the primary application method for the material from bonding (i.e., freezing) with the pavement surface.
When only a delay in the bonding of the ice is achieved, it
evaluation. weakens the bonds that form, allowing easier removal than
with deicing. Anti-icing applications are not limited to liq-
Most winter maintenance practices follow one or a combi- uid chemicals. Properly timed applications of either dry or
nation of the snow and ice control strategies shown below. It pre-wetted chemicals can achieve the goal of preventing
is assumed that mechanical removal of snow and ice is used bond formation if these materials can be held in place and
as appropriate within all strategies. Further information on not removed by traffic action.
the selection and use of snow and ice control strategies is pre- • Abrasive use has long been a strategy of choice by many
sented in NCHRP Report 526. agencies because it is a very visible and assumed to be a low-
cost approach to managing pavement friction. However,
• Deicing, a reactive strategy, involves application of a chem- when abrasives are placed “dry” on the road surface (i.e.,
ical on top of a layer of snow, ice, or frost already bonded without significant pre-wetting) they provide, at best, a very
to the surface of the pavement. Typical application rates for short-term increase in road surface friction. As roadway

G-6
traffic levels increase, and vehicle speeds are greater than attributes are addressed though the purchase specification
30 mph (50 kph), any benefit from abrasive use diminishes. (Chapter 3).
• Pre-wetting is injecting or spraying a liquid chemical Table 2-4 lists common snow and ice control strategy and
on solid chemicals or abrasives so as to enhance their material combinations. Materials are denoted as being pri-
effectiveness. It is a viable and desirable alternative to dry mary or secondary. For example, when pre-wetting NaCl with
application of solid materials. Pre-wetting may be con- liquid MgCl2, NaCl is the primary product, and MgCl2 is the
ducted to support anti-icing, deicing, or abrasive use. secondary product. Using typical application rates, secondary
Using standard application rates, the pre-wet liquid has materials contribute little to total material loading. Therefore
little contribution to the overall chemical loading. only primary materials are considered in the decision tool
For example, when solid NaCl is pre-wet with liquid process. Organic matter from biomass (OMB) materials are
CaCl2 at a rate of 10 gallons per ton, the chloride contri- generally not used alone for snow and ice control and will
bution from CaCl2 is less than 0.2 percent of total typically exist as an additive or a secondary product to the
chloride loading. other chemical materials.
2.3.2.1 Blended Materials

2.3.2 Material Selection
Blends of snow and ice control materials are common and
Application Note 4 – Snow and Ice Control Materials are generally prepared with specific goals in mind, such as
Selection corrosion inhibition or performance enhancement. Com-
mercial blends of chloride-based chemicals generally contain
between 5 and 25 percent of an OMB material, although
Select the specific material for evaluation, in addition to OMB levels up to 50 percent are also commercially available.
Blended snow and ice control materials pose certain chal-
standard products.
lenges when assessing potentials for impairment to the envi-
ronment and infrastructure, because it is not always clear
Several types of materials are used for snow and ice control, what concentrations are present or which materials are active
and new products continue to emerge in the marketplace. ingredients. The decision tool defines standard active ingre-
However, most materials used for roadway snow and ice con- dients for snow and ice control as being
trol contain or will continue to contain one or more of the
material types shown in Table 2-3. The potential for impair- • Sodium Chloride (NaCl);
ment to the natural environment or infrastructure is linked • Calcium Chloride (CaCl2);
to the properties (i.e., components and attributes) of each • Magnesium Chloride (MgCl2);
snow and ice control material. The decision tool addresses • Calcium Magnesium Acetate (CMA);
effects from the primary components and attributes of snow • Potassium Acetate (KA); and
and ice control material types. Secondary components and • Abrasives.
Table 2-3. Primary components and secondary attributes of

selected snow and ice control materials.
Snow and Ice Primary Components Secondary Components
Material Type
Control Material and Attributes and Attributes
Heavy metals, Cyanide (CN),
Sodium Chloride (NaCl) Na, Cl
Nutrients (N, P)
Chloride Salts
Calcium Chloride (CaCl2) Ca, Cl Heavy metals, Nutrients (N, P)
Magnesium Chloride (MgCl2) Mg, Cl Heavy metals, Nutrients (N, P)
Calcium Magnesium Acetate Ca, Mg, Organic Matter

Acetates Heavy metals, Nutrients (N, P)
(CMA) (acetate) as BOD1
K, Organic Matter (acetate)
Potassium Acetate (KA) Heavy metals, Nutrients (N, P)
as BOD1
Organic Matter from Biomass
Organic Products Organic Matter as BOD1 Heavy metals, Nutrients (N, P)
(OMB), Agricultural By Products
Particulate – Air Quality
Abrasive Abrasives (PM-10) Heavy metals, P
Sediments – Water Quality
1
Biochemical Oxygen Demand

G-7
Table 2-4. Common snow and ice control strategy & material combinations.
Acetates Organic
Strategy NaCl CaCl2 MgCl2 Matter from Abrasives
CMA KA CMAK
Biomass
Solid Only P P P1 P - - - -
Deicing
Liquid Only P P P P P P S -
Pre-wet Solid P, S S S S S S S -
1 1
Solid Only P P P P -
Anti-Icing
Liquid Only P P P P P P S -
Pre-wet Solid P, S S S S S S S -
Abrasive Abrasive Only - - - - - - - P

Use
Pre-wet abrasive S S S S S S S P
P = denotes a primary material
S = denotes a secondary product material
1
Not Commonly Used
‘ - ’ denotes not applicable
These materials are the primary snow and ice control mate- case, it is recommended that further verification be conducted
rials, as identified through a comprehensive literature review by a laboratory using approved methods (refer to Chapter 4).
and industry survey conducted in Phase I of this study. These Table 2-5 provides molecular weights and primary com-
materials also have a proven ability to melt ice under field ponent proportions for standard snow and ice control mate-
conditions. Recent information suggests that OMB materials rials. Chloride-based materials (NaCl, MgCl2, and CaCl2)
may also contribute to melting performance, but information have chloride concentrations within a close relative range
is limited. Any measurable performance benefits of OMB (61- to 74-percent chloride) implying that chloride-related
materials should be assessed on a case-by-case basis through effects for these materials would generally be of the same
development of material-specific phase curves. Abrasives order of magnitude. The following example shows how to cal-
have no snow and ice melting potential, but are an established culate the chloride content of a blend that is 75-percent NaCl
winter maintenance strategy for many agencies. and 25-percent MgCl2:
For commercially available blends, it is necessary for suppli-
ers to provide complete, accurate, and independently devel- Chloride
oped data that identifies product components. Preliminary Material % Chloride % of Mixture Contribution
qualification may include basic physical and chemical data and
NaCl 61 % 75 % 0.75 × 0.61 = 45 %
MSDS information. Where transportation agencies prepare
MgCl2 74 % 25 % 0.25 × 0.74 = 19 %
blended products in house, chemical and physical properties
can be calculated through proportional summation. In either Total Chloride 64 %
Table 2-5. Molecular weights and constituent proportions of five

standard snow and ice control materials.
Ion Na Mg Ca K Cl Acetate
Material Molecular
22.99 24.31 40.08 39.01 35.45 59.04
Weight
NaCl 58.44 39 % - - - 61 % -
MgCl2 95.21 - 26 % - - 74 % -
CaCl2 110.99 - - 36 % - 64 % -
KA 98.14 - - - 40 % - 60 %
1 1 1
CMA 1
90 - 15 % 15 % - - 60 %1
1
Concentration may vary depending on manufacturing processes.

G-8
This mixture contains 64-percent chloride, which is within The transport process for snow and ice control materials
the 61- to 74-percent chloride range typically exhibited by the from the roadway involves movement of liquid (through
chloride-based products. Similar principles can be applied to splashing, spraying, runoff, surface water movement, and
other components and blended materials. groundwater movement) and movement of dry material or
residue mobilized or resuspended by traffic action and wind.
Although it is important to understand all possible transport
processes, they have little practical application to a material
2.4 Maintenance Area Information
selection tool that is broadly applied. It is necessary to simplify
Application Note 5 – Maintenance Area the transport processes into the following generalizations
developed through a comprehensive technical assessment of
information reviewed during Phase I of this study:
Identify and briefly describe the roadway section or region
• Assessment of numerous studies investigating snow and ice
undergoing assessment. Inventory potential maintenance
control material movement over roadside soil (from
area receptors using Tables 2-6 through 2-15 and develop splash, spray, and runoff) shows that the effects on soil con-
centrations typically drop quickly beyond the first 10- to
potential for impairment scores specific to receptors and 20-m-wide strip of land bordering the roadway or along
materials. paths of overland flow.
• Many studies have attempted to model the concentrations of
direct water runoff; however, findings are highly variable and
The decision tool can assess maintenance area sizes rang- depend on site-specific factors, including precipitation
ing from small areas with specific concerns (e.g., bridges or amounts, first flush effects, soil infiltration, hydraulic con-
environmentally sensitive areas) to entire maintenance ductivity, and dilution effects. A conservative estimate is that
regions. Users must clearly define maintenance area limits in snow and ice control materials will be diluted 500-fold at the
order to set the constraints for establishing an inventory of point they leave the roadway. In certain instances, lower
site-specific considerations. dilution rates are possible, although dilution will increase
For impairment to occur, snow and ice control materials substantially as the distance from the roadway increases.
must contact a receptor in a concentration significant enough • Air quality effects can be further reaching than effects on
to elicit a negative response. Predicting transport processes soil, water, vegetation, and animals and should be assessed
from the roadway and identifying all possible receptors in a on a local to regional level.
maintenance area can be challenging and some broad
assumptions are necessary. The decision tool applies a conservative approach to iden-
Many historical problems related to snow and ice control tify potential at-risk receptors in the maintenance area. This
materials are associated with chemical overuse that resulted involves a series of checklists used to inventory common and
in higher than necessary loadings to the environment. Subse- significant areas of concern for the environmental and infra-
quent assessments provided in this decision tool are based on structure subcategories. The inventory should be conducted
normal use of snow and ice control products. The following by individuals with a broad understanding of the conditions
references provide suggestions for normal application rates of that exist in the maintenance region. Suitable individuals will
snow and ice control materials: include supervisory and management staff with knowledge of
snow and ice control operations and environmental and
• Ketcham, S. A., L. D. Minsk, R. R. Blackburn, and E. J. infrastructure conditions and concerns. This step may direct
Fleege. Manual of Practice for an Effective Anti-Icing Pro- further investigation, which if necessary, would require input
gram – A Guide for Highway Winter Maintenance Personnel, by additional professionals trained in the area of concern.
US Army Cold Regions Research and Engineering Labora- Tables 2-6 through 2-15 are practical worksheets that sum-
tory Corps of Engineers for the FHWA, Hanover, New marize common and significant maintenance area receptors
Hampshire, 1996. that may be impaired by snow and ice control materials.
• Blackburn, R. R., D. E. Amsler, S. E. Boselly, and A. D. These tables address the primary components and attributes
McElroy. Guidelines for Snow and Ice Control Materials and of snow and ice control materials (refer to Table 2-4).
Methods, NCHRP Project No. 06-13. National Cooperative Secondary components and attributes are addressed through
Highway Research Program, Transportation Research a purchase specification (Chapter 3).
Board, National Research Council, 2003. Each snow and ice control material will have a certain
• Salt Institute, The Snowfighters Handbook, July 1999. potential to impair an environmental or infrastructure

G-9
Table 2-6. Aquatic life—material/receptor potential for impairment

worksheet.
Exists?
Aquatic Life Conditions NaCl MgCl2 CaCl2 CMA KA OMB Abrasives
(Y/N)
Small stream with coarse substrate 4 4 4 2 2 1 1
Small stream with fine substrate 4 4 4 2 2 1 5
Intermediate stream with coarse substrate 5 5 5 3 3 2 1
Intermediate stream with fine substrate 5 5 5 3 3 2 5
Large stream with coarse substrate 5 5 5 4 4 3 3
Large stream with fine substrate 5 5 5 4 4 3 5
Wetland, nutrient poor 3 3 3 3 3 1 1
Wetland, nutrient rich 3 3 3 3 3 3 1
Small lake, nutrient poor 2 2 2 3 3 1 5
Small lake, nutrient rich 2 2 2 4 4 3 5
Intermediate lake, nutrient poor 4 4 4 4 4 1 5
Intermediate lake, nutrient rich 4 4 4 5 5 4 5
Large lake, nutrient poor 5 5 5 5 5 3 5
Large lake, nutrient rich 5 5 5 5 5 5 5
Sum of Selected Conditions
Average of Selected Conditions
Definition of terms:
• Stream Size: small stream: average width of water surface less than 10 yards under winter flow conditions;
intermediate stream: 10 to 30 yards; large stream: greater than 30 yards.
• Lake Size: small lake: < 10 acres; intermediate lake: 10 to 100 acres; large lake: > 100 acres.
• Wetland: aquatic body not defined as a lake or stream, i.e. marsh, ponds, etc.
• Substrate: coarse substrate: boulder, cobble, gravel; fine substrate: sand, clay, muck.
• Nutrient Levels: nutrient poor: total P below 15 µg/L, total N below 300 µg/L; nutrient rich: > 15 µg/L P,
>300 µg/L N.
Table 2-7. Drinking water from groundwater—material/receptor

potential for impairment worksheet.
Exists?
Drinking Water Conditions NaCl MgCl2 CaCl2 CMA KA OMB Abrasives
(Y/N)
Shallow ground water used for drinking 3 5 5 4 4 3 5

Table 2-8. Air quality—material/receptor potential for impairment worksheet.
Exists?
Air Quality Conditions NaCl MgCl2 CaCl2 CMA KA OMB Abrasives
(Y/N)
Existing air quality concerns relevant to

particulate matter (i.e. non-attainment 5 5 5 5 5 5 1
area)
Poor air dispersion, closed air shed 5 5 5 5 5 5 1
Table 2-9. Vegetation—material/receptor potential for impairment worksheet.
Exists?
Vegetation Conditions NaCl MgCl2 CaCl2 CMA KA OMB Abrasives
(Y/N)
Agricultural lands within 10m of the

1 2 2 4 4 5 4
roadway
Protected vegetation within 10m of the
1 2 2 4 4 5 4
roadway
Aesthetic objectives within 10m of the
1 2 2 4 4 5 4
roadway

G-10
Table 2-10. Soil quality—material/receptor potential for impairment

worksheet.
Exists?
Soil Quality Conditions NaCl MgCl2 CaCl2 CMA KA OMB Abrasives
(Y/N)
Poor soil structure within 10m of the roadway1 2 5 5 5 5 5 5

1
Exchangeable Sodium Percentage (ESP) >13
Table 2-11. Terrestrial fauna—material/receptor potential for impairment

worksheet.
Exists?
Animal Conditions NaCl MgCl2 CaCl2 CMA KA OMB Abrasives
(Y/N)
Potential for effects on terrestrial species 1 4 4 4 5 5 5 5

Road kill for mammals or birds is a concern 4 4 4 5 5 5 5
1
Includes sensitive mammals, birds, amphibians, reptiles, insects, spiders, etc.
Table 2-12. Vehicles—material/receptor potential for impairment worksheet.
Exists?
Vehicles NaCl MgCl2 CaCl2 CMA KA OMB Abrasives
(Y/N)
Vehicles 1 1 1 4 4 5 1
Table 2-13. Metal infrastructure—material/receptor potential for impairment

worksheet.
Exists?
Metal Infrastructure Conditions NaCl MgCl2 CaCl2 CMA KA OMB Abrasives
(Y/N)
Steel Bridges 1 1 1 4 4 5 5
Other metal infrastructure of concern
Table 2-14. Concrete matrix—material/receptor potential for impairment

worksheet.
Exists?
Concrete Degradation NaCl MgCl2 CaCl2 CMA KA OMB Abrasives
(Y/N)
Concrete pavement and structures 4 3 4 3 4 5 5


G-11
Table 2-15 Concrete reinforcing—material/receptor potential for

impairment worksheet.
Exists?
Concrete Reinforcing NaCl MgCl2 CaCl2 CMA KA OMB Abrasives
(Y/N)
Reinforcing Steel Bar 1 1 1 5 5 5 5

receptor. This is reflected by product/receptor-specific with a potentially adverse environmental effect tends to
numerical rankings developed based on a comprehensive increase as the size of the receiving water increases.
technical assessment of information conducted during • Abrasives block interstitial spaces within coarse substrates,
Phase I of this study and based on the results of analytical thus damaging habitat. This particular category of damage
testing conducted during Phase II of this study. Numerical is not an issue for streams with fine substrates. The same is
rankings correspond to the following conditions: true of wetlands.
• Organic matter and nutrients cause oxygen depletion or
• 1 to 2 – undesirable, eutrophication where potential dilution is low, but are
• 2 to 3 – potential for some impairment, much less likely to be harmful where dilution is high.
• 3 to 4 – potential for minor impairment, and Higher nutrient levels tend to be associated with OMB
• 4 to 5 – low likelihood of impairment. materials. Nutrient-poor water bodies have a lower
tolerance than nutrient-rich water bodies to additional
These values have undergone considerable development nutrient loadings.
and review. Where a user thinks that other information is • Nutrient-poor streams typically contain organisms that
more representative of the conditions in a maintenance area, are not tolerant of oxygen depletion. Nutrient-rich
these numerical rankings may be adjusted accordingly. Any streams are more likely to experience oxygen depletion,
adjustments must be technically defensible and conducted by and, therefore, typically contain organisms that are more
individuals with knowledge of the topic area. tolerant of oxygen depletion. Therefore, the tolerance for
Instructions for use of Tables 2-6 through 2-15 are as follows: oxygen-demanding substances, such as acetate and
OMB, is rated lower for nutrient-poor streams than for
1. A given roadway may have more than one type of poten- nutrient-rich streams, particularly if the streams are
tial receptor in each decision subcategory. Check the one small.
or more types of receptors that occur along a roadway • The probability that the chloride standard for aquatic life
under consideration for snow and ice control. will be exceeded is highest for standing waters of small vol-
2. Obtain a mean score for each material under considera- ume and lowest for flowing waters of large volume.
tion by summing the scores for the selected receptors and
dividing by the number of selected receptors. If no selec-
tions are made, a default average score of 5 applies. The
2.4.2 Drinking Water from Groundwater
lowest score indicates the least desirable choice.
Worksheet
3. For snow and ice control products blended from more
than one material, select the lowest (most conservative) The rationale for the numerical scores for drinking water
score applicable to either material in the blend. For a mate- obtained from groundwater is as follows:
rial composed of MgCl2 and OMB applied near a “small
coldwater stream with coarse substrate” (refer to Table 2-6) • Magnesium, calcium, and chloride are not associated with
the MgCl2 score is “4,” and the organic biomass score is “1.” impairment of human health.
Using this approach, the applied score is “1.” • Sodium in groundwater is a potential threat to persons
with cardiovascular impairment, although the threshold
concentration for health effects is relatively high.
2.4.1 Aquatic Life Worksheet
• Addition of organic carbon compounds (e.g., acetate or
The rationale for the numerical scores for aquatic life is as mixed organic matter from biomass) may cause deoxy-
follows: genation of groundwater, which in turn could cause
release of potentially harmful substances (e.g., heavy met-
• Small water bodies offer less dilution than large water als) into the groundwater and could cause taste and odor
bodies. Consequently, acceptability of any given material problems.

G-12
2.4.3 Air Quality Worksheet unclear whether snow and ice control materials increase
vehicle-animal collision rates. Some research supports this,
The use of abrasives for snow and ice control can degrade but most current information suggests other conditions con-
air quality through increases in airborne fine particulate tribute much more significantly to vehicle-animal collisions.
(i.e., PM-10). This is of greatest concern for areas with lim-
ited recharge capacity or where existing air quality issues
exist. Chemical materials are not known to result in similar 2.4.7 Vehicles Worksheet
concerns. Chloride-based snow and ice control materials are corro-
sive to metallic motor vehicle components. Numerous stud-
2.4.4 Vegetation Worksheet ies have attempted to rank these materials for corrosion but
complex contributing factors limited the development of
Effects associated with roadside vegetation impairment definitive conclusions. From the standpoint of developing
from snow and ice control materials are linked to extremely practical rankings, these materials have been ranked equally
high levels of ions contacting foliage and roots and damage high for their potential to cause corrosion. Information sug-
related to elevated chloride levels. Vegetation damage is of gests that acetate-based materials are less corrosive to steel
greatest concern in the primary material deposition zone than chlorides, but may lead to corrosion concerns for other
10 meters from the roadway. Although certain types of alloys. Although OMB materials are accepted as being non-
vegetation are more tolerant to these stresses than others, corrosive to metals, limited evidence suggests that these mate-
a conservative approach ensures the greatest level of rials, used as corrosion-inhibition additives, significantly
environmental protection. reduce corrosion rates under field conditions. Abrasives can
Chloride materials have the greatest potential to impair degrade coatings that protect metals from the corrosive
vegetation. Sodium can degrade soil structure, which can effects of snow and ice control materials.
impede water uptake and root growth. Evidence suggests that
acetate materials can also affect vegetation through osmotic
stress damage, although this does not occur to the same 2.4.8 Metal Infrastructure Worksheet
degree as for chlorides. OMB materials pose minimal concern Chloride-based snow and ice control materials are corro-
to vegetation when compared with inorganic salts. Abrasives sive to metal infrastructure components. Numerous studies
are not completely innocuous and can cause smothering have attempted to rank these materials for corrosion but
concerns in some instances. complex contributing factors limited the development of
definitive conclusions. From the standpoint of developing
2.4.5 Soil Worksheet practical rankings, these materials have been ranked equally
high for their potential to cause corrosion. Information
Snow and ice control chemicals can affect soils chemically suggests that acetate-based materials are less corrosive to steel
(by affecting concentrations of available ionic components in than chlorides, but may lead to corrosion concerns for other
the soil) and physically (by affecting soil structure). Chemical alloys. Although OMB materials are accepted as being non-
effects on the soil are primarily a vegetation-impact concern corrosive to metals, limited evidence suggests that these mate-
and are not significant when considering soil as the individ- rials, used as corrosion-inhibition additives, significantly
ual receptor. Elevated sodium levels can degrade soil structure reduce corrosion rates under field conditions.
by reducing soil particle size. This can reduce permeability
and create a higher potential for soil erosion. Sodium effects
2.4.9 Concrete Matrix Worksheet
are of greatest concern when poor soil conditions already
exist (Exchangeable Sodium Percentage (ESP) >13) within Considerable research is ongoing on the effects of snow
10 meters of the roadway. and ice control materials on concrete paste. To date, there
have been conflicting findings as to how the various materi-
als affect the durability of the concrete. Traditional snow and
2.4.6 Terrestrial Fauna Worksheet
ice control materials (e.g., NaCl and CaCl2) are known to have
Snow and ice control materials have been linked to reports some detrimental effects on concrete, primarily causing
of salt toxicosis from ingestion and increases in vehicle- scaling of the concrete surface, although no problems with
animal collisions. Some information suggests salt toxicosis the bulk properties of the concrete are known. With good-
from ingestion is a concern for small animals such as birds, quality concrete, scaling problems can be avoided. Some lab-
although most terrestrial animals have high salt tolerances, oratory studies have shown deterioration of the cement paste
especially when adequate drinking water is available. It is and bulk properties of the concrete by CaCl2; however, the

G-13
material is widely used and these findings do not appear to be difficult to predict with certainty. This is particularity true for
substantiated in field applications. indirect costs, which are closely related to application strategy
Laboratory studies with magnesium-based materials (e.g., influences (e.g., roadway retention, melting capacity at spe-
MgCl2 and CMA) have found chemical degradation of the cific temperatures, and applications required per given storm
concrete matrix as a result of magnesium reacting with the event). The decision tool uses purchase price per dry unit of
calcium-silica-hydrate (C-S-H) by replacing the calcium with active ingredient, including shipping, as a cost benchmark.
magnesium and, in time, resulting in strength-loss of the con- This is a practical approach and yields a consistent and unbi-
crete. Although there have been some reports of MgCl2 caus- ased assessment of the dominant cost element. It is necessary
ing deterioration of the concrete in the field, there does not for the decision tool user to obtain the purchase price and
appear to be a widespread problem. However, MgCl2 is used active ingredient concentration of standard materials (i.e.,
to a somewhat lesser extent. Considering the reactivity of NaCl, CaCl2, MgCl2, KA, CMA, and abrasives) and other
magnesium, long-term effects are possible and, given the candidate materials. Some standard materials may not be
findings of past studies, future work will be required. Close strong agency considerations, but the information is impor-
monitoring of structures where MgCl2 is used is also tant to ensure completeness of the decision tool assessment.
recommended. A secondary benefit is that users will research and develop
knowledge of new material types, further increasing their
understanding of the important elements involved in snow
2.4.10 Concrete Reinforcing Worksheet
and ice control material selection.
Chlorides enable and/or accelerate corrosion of concrete Snow and ice control materials can exist in various forms
reinforcing. Current evidence suggests that chloride associated (i.e., liquids and solids), concentrations, and blends. Cost
with magnesium may have higher chloride diffusion coeffi- must be expressed as dry weight of active ingredient. For
cients than NaCl or CaCl2. Considerable research is ongoing example, NaCl may be provided as a solid material, a blended
and, for practical purposes, all chloride materials have been solid material, a pure liquid, or a blended liquid. Examples for
ranked equally high in causing corrosion of the reinforcing calculating cost per unit of active ingredient are shown in
steel. This study includes an analytical research program inves- Table 2-16. Inert substances are defined as substances that do
tigating the relative chloride diffusion coefficients between not melt ice and primarily include insoluble substances that
chloride-based snow and ice control materials. After 1.5 years do not enter into solution.
of exposure testing, corrosion has not been initiated for any of Relative price rankings of products can be derived from
the materials. The rankings will remain based on the compre- purchase price information (Table 2-17). A high price rank-
hensive literature assessment, as this time of exposure can be ing represents the preferred product for this category.
considered short when viewed in terms of concrete durability
where current service life requirements in the order of 30 to
2.6 Performance
100 years are commonly specified. Acetate-based materials,
OMBs, and abrasives are not considered corrosive to concrete Measuring performance of snow and ice control chemicals
reinforcing. continues to be a topic of considerable research. Several
approaches for quantifying performance exist and many are
not easily quantifiable. The primary considerations affecting
2.5 Identified Cost snow and ice control material performance are as follows:
Application Note 6 – Price and Concentration
• Chemical characteristics of snow and ice control chemi-
cals, primarily the concentrations of active ingredients
Input price and active ingredient concentrations for stan- responsible for melting snow and ice affect performance.
Phase curves provide a simple method to correlate con-
dard snow and ice control materials, plus other materials centration with melting point.
• Physiochemical properties, such as solubility rates and
under consideration.
exothermic (i.e., heat-producing) properties, can also
affect melting rate and ice penetration rates, but these char-
This section provides a model to assess costs associated acteristics are difficult to quantify. Information related to
with snow and ice control materials. The complete cost of measuring these characteristics is available in SHRP
snow and ice control involves both direct costs (i.e., material H-332, Handbook of Test Methods for Evaluating Chemical
purchase price) and indirect costs (i.e., operational costs and Deicers, although this document does not address all types
capital costs). Costs are dynamic and are influenced by factors of snow and ice control materials.

G-14
Table 2-16. Example of cost per active ingredient calculation.

Material
Common
Form and Cost Per Active Ingredient Calculations
Examples
Type
Solid NaCl (Total material weight – inert substances and moisture) ÷ purchase price
Hydrated CaCl2-xH2O,
(Total material weight – inert substances and moisture) ÷ purchase price
Solid MgCl2-xH2O
NaCl, CaCl2,
Liquid (Total material weight – inert substances and moisture) ÷ purchase price
MgCl2, KA, CMA
Blend Varies (Total material weight – inert substances and moisture) ÷ purchase price
Note: materials suppliers can supply total moisture percentage and purity for solid materials (hydrated and non-
hydrated solids).
Table 2-17. Development of relative purchase price rankings.
NaCl CaCl2 MgCl2 KA CMA Abrasive

Purchase Price per Ton ($)1 $ 50 $ 130 $ 120 $ 1000 $ 1150 $ 80
Active Ingredient Concentration (%) 23 % 32 % 30 % 50 % 25 % NA
Purchase Price per Dry Pound ($/lb) $ 0.11 $ 0.20 $ 0.20 $ 1.00 $ 2.302 $ 0.04
Maximum Price ÷ Material Price 21 11 12 2 1 58
Price Ranking (Maximum 100) 37 20 20 4 2 100
1
Pricing was gathered from contracts published on the Internet and provides semi-realistic demonstration of relative
purchase price calculations and rankings.
2
Bold text indicates maximum purchase price per dry pound ($/Lb)
• Weather has considerable influence on material per- ingredient concentration are known, a melting potential
formance and includes precipitation (e.g., type, amount, using phase curve information can be used to assess materi-
and rate) and temperature. Although precipitation affects als. The following sections address how to apply temperature
performance of chemicals through dilution, application and active ingredient information to assess the relative per-
rates will vary directly with storm magnitude—more formance of snow and ice control materials.
chemical will be applied for high precipitation storms.
Temperature, on the other hand, can be correlated with
2.6.1 Temperature Information
chemical properties to provide a practical assessment of
performance. Application Note 7 – Temperature Information
• Application methods, including timing, application rates,
and equipment, influence overall snow and ice control
chemical performance significantly. However, when appli- Input the 85th percentile cold winter temperature.
cation methods are established before material selection,
these influences are normalized, allowing the materials to Pavement (or working) temperatures can be correlated
be assessed based on their specific properties. with a chemical’s phase properties to provide a practical
• Roadway factors, including grade, texture, and the pres- assessment of performance. Users will input expected cold
ence of surface contaminants, can affect snow and ice con- temperatures for a maintenance area. Historical pavement
trol material performance. These factors are not easily temperatures provide the best source of this information.
quantified over a large maintenance area, but will be the Historical air temperatures can also be used if pavement
same for each material assessment made. temperature data is not available. Temperature information
• Agency operational concerns (e.g., handling, storage, ease is generally obtained from Roadway Weather Information
of use, familiarity, and ease of delivery) can affect assess- Systems (RWIS), installed recording devices, purchased
ment of performance. weather information, or other sources the user consults for
decision-making.
Although numerous factors can affect snow and ice control The selected cold temperature should represent the 85th
material performance, the dominant performance compo- percentile of cold temperatures occurring in the maintenance
nent remains the chemical’s ability to melt ice at a specified area. Higher percentiles (90th or 95th) may be selected to
temperature. If the application temperature and the active represent higher levels of service objectives, while lower

G-15
Figure 2-1. Sample cumulative temperature plot.
percentiles (80th or 75th) may be selected to represent lower for a hypothetical chemical and indicate how this material
levels of service objectives. Consistent use of a single temper- performs at different concentrations with correlating freeze-
ature and service level objective throughout the decision tool points. Each point on the phase curve in Figure 2-2 is referred
process will ensure that all products will be evaluated equally. to as the eutectic temperature and represents a chemical con-
Figure 2-1 is a sample cumulative plot of historical, sub- centration and a corresponding freeze-point. The optimum
freezing temperatures. The temperature is 16°F or higher eutectic temperature is the lowest freezing temperature
85 percent of the time when the temperature is below freezing. achievable for material concentration and is typically the bot-
From another perspective, if a snow and ice control material is tom point (the “v”) on many phase curves. In this example,
effective at 16°F, it will theoretically work in 85% of expected the optimum eutectic point occurs at 0°F at a concentration
winter storm events. It is recommended that the user construct of 25 percent.
a chart(s) similar to that shown in Figure 2-1, ensuring that Standard phase curves for pure forms of NaCl, MgCl2, CaCl2,
each data point represents a consistent temperature measure- CMA, and KA exist in the scientific literature. Ice-melting
ment time period (e.g., hourly, daily, or weekly) and historical properties can deviate significantly if these materials contain
temperatures that are at or below the freeze-point. additional active ingredients, additives, or contaminants. For
impure or blended materials, phase curves developed by an
independent laboratory using standard test procedures are rec-
2.6.2 Performance Measured by Melting
ommended. Section 3.2.1.2 provides a standard method for
Potential
assessing the freezing point of snow and ice control materials.
Application Note 8 – Material-Specific Phase Information If independent verification is not available, the phase curve for
the primary active constituent may be used until additional
information can be obtained.
For non-standard or blended snow and ice control mate- The snow and ice melting capacity of snow and ice control
chemicals is based on their concentration and temperature in
rials, input material-specific phase curve data.
solution. As melting occurs, water is produced that dilutes the
original chemical concentration. As the solution concentra-
Short of extensive field and laboratory testing, the only tion reduces, the freezing point (i.e., ice/snow melting tem-
objective method of comparing snow and ice control mate- perature) changes. The mass of ice melted (i.e., water
rial performance is to assess materials based on their poten- produced) before refreezing occurs for a given chemical
tial to melt ice and snow as developed from phase (eutectic) application at a specified temperature can be used to assess
curve information. Figure 2-2 and Table 2-18 show phase data performance.

G-16
Figure 2-2. Phase/eutectic curve for hypothetical chemical X.
Melting potential (MP) is a reliable and accurate method ficient moisture be present). The melting performance assess-
to calculate the amount of ice and snow that can be melted at ment compares materials independent of application rate,
a given temperature and is represented mathematically by the and only materials in solution can effectively melt ice. For
following equation: solids, it is assumed that a saturated solution forms after
application. Use of any value for BC greater than the saturated
MP BC /EC 1 solution would bias the MP assessment toward materials
Where applied as a solid. Saturated solution estimations for standard
BC = Beginning Concentration (% w/w) materials are based on standard solubility information at 32°F
(0°C) (Table 2-19).
EC = Ending Concentration (% w/w)
Once snow and/or ice melting begins, the chemical will
1 = Constant representing the initial application.
dilute until it reaches a concentration where melting stops.
For liquid materials, BC is the applied concentration. For This concentration is referred to as the ending concentration
solid materials, BC is the saturated concentration (should suf- (EC) and is also referred to as “refreeze.”
Table 2-18. Phase/eutectic data for hypothetical chemical X.

Material Material
Freezing Freezing
Concentration Specific Gravity Concentration Specific Gravity
Temperature (F) Temperature (F)
(%) (%)
0 1 32 16 1.118 14
1 1.007 31 17 1.126 12.6
2 1.014 30 18 1.134 11.2
3 1.021 29 19 1.142 9.8
4 1.028 28 20 1.150 8.4
5 1.035 27 21 1.158 6.9
6 1.043 25.9 22 1.166 5.4
7 1.050 24.8 23 1.175 3.7
8 1.057 23.7 24 1.183 1.9
9 1.065 22.6 25 1.191 0
10 1.072 21.5 26 1.199 1
11 1.080 20.3 27 1.207 3
12 1.087 19.1 28 1.214 5
13 1.095 17.9 29 1.221 7
14 1.103 16.6 30 1.224 9
15 1.111 15.3

G-17
Table 2-19. Solubility products for common snow and ice control
materials.
Chemical Percent Solution at Saturation1
NaCl 26.28 % at 0°C (32°F)
CaCl2 36.70 % at 0°C (32°F)
MgCl2 33.96 % at 0°C (32°F)
CMA NA
KA (KC2H3O2) 68.40 % at 0°C (32°F)

1
CRC Handbook of Chemistry and Physics 85th edition.
Three examples are provided to illustrate MP calculations Example 3 – Melting Potential at 30° F
for Chemical X (from Table 2-18). This material has an opti-
mum eutectic point at a 25-percent concentration and a • Consider a 25-percent BC solution of Chemical X applied
freeze-point of 0°F. For these examples it is assumed that the at 30° F. This temperature corresponds to a 2-percent EC
roadway surface is either coated with ice or under continual solution.
snowfall. • Every 100 pounds of 25-percent BC solution contains 25
pounds of solid active ingredient. After dilution, the 25
Example 1 – Melting Potential at 27° F pounds of active ingredient generates a 2-percent EC solu-
tion with a total mass of 1,250 pounds. Therefore, 100
• Consider a 25-percent BC solution of Chemical X applied pounds of the 25-percent solution will melt 1,150 pounds
at a temperature of 27°F. This temperature corresponds to of ice and/or snow at 2° F. Using the information from the
a 5-percent EC solution. example:
• Every 100 pounds of 25-percent BC solution contains 25
pounds of solid active ingredient. After dilution, the 25 MP (30° F) = (25%/2%) 1 12.5 1 11.5
pounds of active ingredient generates a 5-percent EC
Summaries of data for these examples are presented in
solution with a total mass of 500 pounds. Therefore,
Table 2-20.
100 pounds of the 25-percent solution will melt 400 pounds
The MP formula can also be used to evaluate hypereutec-
of ice and/or snow at 27°F. Using the information from the
tic chemical solutions—solutions with concentrations in
example:
excess of the optimum eutectic concentration. Examples of
hypereutectic solutions are 30-percent magnesium chloride
MP (27°F) (25%/5%) 1 4 (optimum eutectic concentration = 21.6 percent) and 32-
percent calcium chloride (optimum eutectic concentration =
Example 2 – Melting Potential at 14° F 28.9 percent). When compared with lower concentration
solutions, hypereutectic solutions can withstand higher dilu-
• Consider a 25-percent BC solution of Chemical X applied tions, helping to prolong the time of effectiveness for a chem-
at a temperature of 14° F. This temperature corresponds to ical. However, hypereutectic solutions can have higher (i.e.,
a 16-percent EC solution. warmer) freeze-points than the optimum eutectic tempera-
• Every 100 pounds of 25-percent BC solution contains 25 ture and should be used with attention to temperatures. For
pounds of solid active ingredient. After dilution, the 25 example, the hypothetical chemical in Figure 2-2 supplied at
pounds of active ingredient generates a 16-percent EC a 30-percent concentration will freeze at 10°F, which is much
solution with a total mass of 160 pounds. Therefore, 100 warmer than the optimum eutectic point (25-percent con-
pounds of the 25-percent solution will melt 60 pounds of centration at 10°F).
ice and/or snow at 14° F. Using the information from the Melting potential represents an objective and easily calcu-
example: lated method to measure material performance at various
temperatures. The MP approach can be applied to blended
MP (14°F) (25%/16%) 1 1.6 1 0.6 materials using product-specific phase information. Table

G-18
Table 2-20. Melting potential examples.
Percent Mass of Total

Mass of
Active Active Solution
Water
Ingredient Ingredient Mass
Beginning
Concentration 25 % 25 lbs 100 lbs 75 lbs
Example 1 (BC)
Ending
Temperature = 27° F Concentration 5% 25 lbs 500 lbs 475 lbs
(EC)
MP = 4
Mass of Snow and Ice Melted (Water Produced) 400 lb
Beginning
Example 2 (BC)
Ending
Temperature = 14° F Concentration 16 % 25 lbs 160 lbs 135 lbs
(EC)
MP = 0.6
Mass of Snow and Ice Melted (Water Produced) 60 lbs
Beginning
Example 3 (BC)
Ending
Temperature = 30° F Concentration 2% 25 lbs 1250 lbs 1225 lbs
(EC)
MP = 11.5
Mass of Snow and Ice Melted (Water Produced) 1150 lbs
2-21 shows examples of MPs for five standard snow and ice • Users will input cost and concentration information for
control materials using phase curves from various commer- candidate material snow and ice chemicals.
cial sources. Phase curves from commercial sources can and • Cost will be represented by purchase price per dry unit of
will differ from the pure chemical data from sources such as active ingredient (Section 2.5). Active ingredients will
the CRC Handbook of Chemistry and Physics because of the include standard snow and ice control chemicals: NaCl,
presence of other species of chlorides and/or chemicals. This MgCl2, CaCl2, CMA, and KA.
highlights the necessity of using phase curves specific to the • Performance will be represented by melting potential
products being assessed, rather than depending on general- (Section 2.6), which is based on standard phase versus
ized curves. concentration data for standard active ingredients and
The numeric MP value can be easily applied to determine on a user-supplied expected cold winter temperature
relative performance of snow and ice control materials. MP (i.e., 85th percentile). Product-specific phase curve
and purchase price can be combined to generate the eco- information will be required for non-standard chemicals
nomic value of a snow and ice control material (i.e., MP per such as blends.
price per mass), which can be used to establish relative rank- • OMB materials are rated for environmental impact but are
ings of material types. not included in performance assessments. If a performance
assessment is desired, product-specific phase curve infor-
mation will be required.
2.7 Decision Tool Application
• Abrasives are rated for environmental impact, but are not
This decision tool is designed to assist users in selecting the included in performance assessments. Abrasives do not
most appropriate snow and ice control materials based on melt snow and ice.
user priorities and conditions. Decision tool steps and • Users will complete an inventory of the most common and
assumptions are summarized below: significant receptors that may be affected by snow and ice
control materials (Section 2.4). The individual environ-
• Users will prioritize agency policy objectives for cost, mental impacts assigned to each chemical are subject to
performance, environmental impacts, and infrastruc- modification, pending new information.
ture impacts. Users will further prioritize subcategory
objectives that support the environmental and infra- The following examples demonstrate the process and
structure objectives. results.

G-19
Table 2-21. Examples of melting potentials for standard commercial snow and ice
control materials.
NaCl Ca Cl2 Mg Cl2 KAc CMA
Temperature
BC (%) = 23 BC (%) = 32 BC (%) = 30 BC (%) = 50 BC (%) = 25
°F °C EC (%) MP EC (%) MP EC (%) MP EC (%) MP EC (%) MP
30 -1.1 1.9 11.3 2.5 11.7 2.1 13.4 5.0 9.0 5.0 4.0
29 -1.7 2.8 7.2 3.7 7.7 3.0 8.9 6.0 7.3 6.0 3.2
28 -2.2 3.7 5.2 4.8 5.7 3.9 6.7 7.0 6.1 7.0 2.6
27 -2.8 4.6 4.0 5.7 4.6 4.7 5.4 8.0 5.3 8.0 2.1
26 -3.3 5.4 3.2 5.7 4.6 5.3 4.7 9.0 4.6 9.0 1.8
25 -3.9 6.3 2.7 6.4 4.0 6.0 4.0 10.0 4.0 10.0 1.5
24 -4.4 7.1 2.2 7.0 3.6 6.7 3.5 11.0 3.5 11.0 1.3
23 -5.0 7.9 1.9 8.0 3.0 7.3 3.1 12.0 3.2 11.9 1.1
22 -5.6 8.7 1.7 9.0 2.6 8.1 2.7 13.0 2.8 12.7 1.0
21 -6.1 9.4 1.4 9.5 2.4 8.6 2.5 13.5 2.7 13.5 0.9
20 -6.7 10.1 1.3 10.0 2.2 9.0 2.3 14.0 2.6 14.4 0.7
19 -7.2 10.8 1.1 10.5 2.0 9.5 2.2 15.0 2.3 15.0 0.7
18 -7.8 11.5 1.0 11.0 1.9 10.0 2.0 16.0 2.1 15.8 0.6
17 -8.3 12.2 0.9 11.5 1.8 10.4 1.9 16.5 2.0 16.5 0.5
16 -8.9 12.8 0.8 12.0 1.7 10.8 1.8 17.0 1.9 17.2 0.5
15 -9.4 13.4 0.7 12.5 1.6 11.1 1.7 17.5 1.9 17.7 0.4
14 -10.0 14.1 0.6 13.0 1.5 11.4 1.6 18.0 1.8 18.3 0.4
13 -10.6 14.6 0.6 13.5 1.4 11.9 1.5 19.0 1.6 18.8 0.3
12 -11.1 15.2 0.5 14.0 1.3 12.3 1.4 19.5 1.6 19.4 0.3
11 -11.7 15.8 0.5 14.5 1.2 12.7 1.4 20.0 1.5 19.9 0.3
10 -12.2 16.3 0.4 15.3 1.1 13.1 1.3 20.5 1.4 20.6 0.2
9 -12.8 16.8 0.4 15.6 1.1 13.4 1.2 21.0 1.4 21.3 0.2
8 -13.3 17.3 0.3 16.0 1.0 13.7 1.2 21.5 1.3 21.8 0.1
7 -13.9 17.9 0.3 16.3 1.0 13.9 1.2 22.0 1.3 22.3 0.1
6 -14.4 18.3 0.3 16.7 0.9 14.2 1.1 22.5 1.2 22.8 0.1
5 -15.0 18.8 0.2 17.0 0.9 14.5 1.1 23.0 1.2 23.3 0.1
4 -15.6 19.3 0.2 17.3 0.8 14.8 1.0 23.5 1.1 23.8 0.1
3 -16.1 19.7 0.2 17.8 0.8 15.2 1.0 24.0 1.1 24.3 0.0
2 -16.7 20.2 0.1 18.2 0.8 15.5 0.9 24.5 1.0 24.8 0.0
1 -17.2 20.6 0.1 18.5 0.7 15.8 0.9 25.0 1.0 NA NA
0 -17.8 21.0 0.1 18.7 0.7 16.1 0.9 25.5 1.0 NA NA
-1 -18.3 21.4 0.1 19.0 0.7 16.3 0.8 25.9 0.9 NA NA
-2 -18.9 21.8 0.1 19.3 0.7 16.6 0.8 26.3 0.9 NA NA
-3 -19.4 22.2 0.0 19.7 0.6 16.8 0.8 27.7 0.8 NA NA
-4 -20.0 22.6 0.0 20.0 0.6 17.1 0.8 28.1 0.8 NA NA
-5 -20.6 22.9 0.0 20.3 0.6 17.4 0.7 28.6 0.8 NA NA
-6 -21.1 23.3 (0.0) 20.5 0.6 17.6 0.7 29.0 0.7 NA NA
-7 -21.7 NA NA 20.8 0.5 17.8 0.7 28.4 0.8 NA NA
-8 -22.2 NA NA 21.0 0.5 18.1 0.7 28.8 0.7 NA NA
-9 -22.8 NA NA 21.3 0.5 18.3 0.6 29.2 0.7 NA NA
-10 -23.3 NA NA 21.5 0.5 18.5 0.6 29.6 0.7 NA NA
-11 -23.9 NA NA 21.8 0.5 18.7 0.6 30.0 0.7 NA NA
-12 -24.4 NA NA 22.0 0.5 19.0 0.6 30.4 0.6 NA NA
-13 -25.0 NA NA 22.3 0.4 19.2 0.6 30.7 0.6 NA NA
-14 -25.6 NA NA 22.5 0.4 19.4 0.5 31.1 0.6 NA NA
-15 -26.1 NA NA 22.8 0.4 19.6 0.5 31.5 0.6 NA NA
-16 -26.7 NA NA 23.0 0.4 19.8 0.5 31.8 0.6 NA NA
-17 -27.2 NA NA 23.3 0.4 20.0 0.5 32.2 0.6 NA NA
-18 -27.8 NA NA 23.5 0.4 20.2 0.5 32.6 0.5 NA NA
-19 -28.3 NA NA 23.8 0.3 20.4 0.5 32.9 0.5 NA NA
-20 -28.9 NA NA 24.0 0.3 20.6 0.5 33.3 0.5 NA NA

G-20
Table 2-21. Examples of melting potentials for standard commercial snow and ice
control materials (Continued).
NaCl Ca Cl2 Mg Cl2 KAc CMA

Temperature
BC (%) = 23 BC (%) = 32 BC (%) = 30 BC (%) = 50 BC (%) = 25
°F °C EC (%) MP EC (%) MP EC (%) MP EC (%) MP EC (%) MP
-21 -29.4 NA NA 24.2 0.3 20.8 0.4 33.6 0.5 NA NA
-22 -30.0 NA NA 24.4 0.3 20.9 0.4 34.0 0.5 NA NA
-23 -30.6 NA NA 24.6 0.3 21.1 0.4 34.3 0.5 NA NA
-24 -31.1 NA NA 24.8 0.3 21.6 0.4 34.7 0.4 NA NA
-25 -31.7 NA NA 25.0 0.3 22.0 0.4 35.0 0.4 NA NA
-26 -32.2 NA NA 25.2 0.3 NA NA 35.3 0.4 NA NA
-27 -32.8 NA NA 25.3 0.3 NA NA 35.7 0.4 NA NA
-28 -33.3 NA NA 25.5 0.3 NA NA 36.0 0.4 NA NA
-29 -33.9 NA NA 25.7 0.2 NA NA 36.3 0.4 NA NA
-30 -34.4 NA NA 25.8 0.2 NA NA 36.7 0.4 NA NA
-31 -35.0 NA NA 26.0 0.2 NA NA 37.0 0.4 NA NA
-32 -35.6 NA NA 26.2 0.2 NA NA 37.3 0.3 NA NA
-33 -36.1 NA NA 26.3 0.2 NA NA 37.7 0.3 NA NA
-34 -36.7 NA NA 26.5 0.2 NA NA 38.0 0.3 NA NA
-35 -37.2 NA NA 26.6 0.2 NA NA 38.3 0.3 NA NA
-36 -37.8 NA NA 26.8 0.2 NA NA 38.7 0.3 NA NA
-37 -38.3 NA NA 26.9 0.2 NA NA 39.0 0.3 NA NA
-38 -38.9 NA NA 27.1 0.2 NA NA 39.3 0.3 NA NA
-39 -39.4 NA NA 27.2 0.2 NA NA 39.7 0.3 NA NA
-40 -40.0 NA NA 27.4 0.2 NA NA 40.0 0.3 NA NA
-41 -40.6 NA NA 27.5 0.2 NA NA 40.3 0.2 NA NA
-42 -41.1 NA NA 27.6 0.2 NA NA 40.6 0.2 NA NA
-43 -41.7 NA NA 27.8 0.2 NA NA 40.9 0.2 NA NA
-44 -42.2 NA NA 27.9 0.1 NA NA 41.3 0.2 NA NA
-45 -42.8 NA NA 28.1 0.1 NA NA 41.6 0.2 NA NA
-46 -43.3 NA NA 28.2 0.1 NA NA 41.9 0.2 NA NA
-47 -43.9 NA NA 28.3 0.1 NA NA 42.2 0.2 NA NA
-48 -44.4 NA NA 28.4 0.1 NA NA 42.5 0.2 NA NA
-49 -45.0 NA NA 28.6 0.1 NA NA 42.8 0.2 NA NA
-50 -45.6 NA NA 28.7 0.1 NA NA 43.1 0.2 NA NA
-51 -46.1 NA NA 28.7 0.1 NA NA 43.4 0.2 NA NA
-52 -46.7 NA NA 28.8 0.1 NA NA 43.8 0.1 NA NA
-53 -47.2 NA NA 28.9 0.1 NA NA 44.1 0.1 NA NA
-54 -47.8 NA NA 29.1 0.1 NA NA 44.4 0.1 NA NA
-55 -48.3 NA NA 29.2 0.1 NA NA 44.7 0.1 NA NA
-56 -48.9 NA NA 29.3 0.1 NA NA 45.0 0.1 NA NA
-57 -49.4 NA NA 29.4 0.1 NA NA 45.3 0.1 NA NA
-58 -50.0 NA NA 29.5 0.1 NA NA 45.5 0.1 NA NA
-59 -50.6 NA NA 29.6 0.1 NA NA 45.8 0.1 NA NA
-60 -51.1 NA NA NA NA NA NA 46.0 0.1 NA NA

G-21
Table 2-22. Three example agency policy weightings.
Example 3
Example 1 Example 2
(Section 2.7.1.3)
(Section 2.7.1.1) (Section 2.7.1.2)
Decision Category Environment
Survey Objective Equal Objective
Infrastructure
Weighting Weighting
Priority
Purchase Price 45 % 25 % 0%
Performance1 35 % 25 % 25 %
Environment 11 % 25 % 37.5 %
Infrastructure 9% 25 % 37.5 %
1 Given that the primary objective of winter maintenance is controlling snow and ice, it is recommended
that performance weighting not be assigned a value below 25%
Table 2-23. Example subcategory weightings.
Decision Category Decision Subcategory Survey Weightings

Aquatic Life1, 2 20 %
Drinking Water1 20 %
Air Quality 27 %
Environment
Vegetation 20 %
Soil 0%
Animals 13 %
Vehicles 22 %
Infrastructure Metal Infrastructure 28 %

Concrete 20 %
Concrete Reinforcing 30 %
1 Agency survey addressed water quality in its entirety and did not differentiate between aquatic life and
drinking water concerns. For this demonstration, equal weighting was assigned to each category.
2 This example is based on average responses for the survey conducted in Phase I of this study. In practice,
it is recommended that aquatic life weighting not be assigned a value below 25%.
This section provides three examples with differing policy • Cost and performance are summed into an “Economic
objectives to demonstrate the decision tool (Table 2-22). Objective.”
Example 1 shows agency priorities determined from the • Given that temperature influences melting potential,
agency survey conducted during Phase I of this study. The assessments have been conducted over temperatures rang-
values represent current averaged priorities in North America. ing from 30 to –20°F.
Example 2 shows equal priorities for all decision categories. • Environmental and infrastructure potentials for impair-
Example 3 describes where environmental and infrastructure ment are based on the maintenance area inventory condi-
concerns are highest. A practical example may include a steel tions represented in Tables 2-24 through 2-33.
bridge with reinforced concrete decking over a sensitive wet-
The decision tool provides a numerical product score that
land. Agency subcategory weightings for environment and
combines user priorities and conditions with a technical
infrastructure are based on agency survey results and will
assessment of potentials for impairment. Higher scores rep-
remain constant for the three examples (Table 2-23). The
resent preferred products, but it is not necessary that an
following conditions are also assumed for each of the three agency select the highest ranked product.
example assessments: Some materials have similar properties under certain con-
ditions and scores may have similar orders of magnitude. In
• Anti-icing is the selected snow and ice control strategy. such cases, considerations such as the following may influ-
• Pricing and concentration have been obtained for all stan- ence product selection:
dard snow and ice control products. The price values used
are semi-realistic examples obtained from contracts pub- • Quality-testing results compared with purchase specifi-
lished on the Internet. cations;

G-22
Table 2-24. Aquatic life—material/receptor potential for impairment

worksheet.
Exists?
Aquatic Life Conditions NaCl MgCl2 CaCl2 CMA KA OMB Abrasives
(Y/N)
Small stream with coarse substrate Y 4 4 4 2 2 1 1
Small stream with fine substrate N 4 4 4 2 2 1 5
Intermediate stream with coarse substrate N 5 5 5 3 3 2 1
Intermediate stream with fine substrate N 5 5 5 3 3 2 5
Large stream with coarse substrate N 5 5 5 4 4 3 3
Large stream with fine substrate N 5 5 5 4 4 3 5
Wetland, nutrient poor N 3 3 3 3 3 1 1
Wetland, nutrient rich N 3 3 3 3 3 3 1
Small lake, nutrient poor Y 2 2 2 3 3 1 5
Small lake, nutrient rich N 2 2 2 4 4 3 5
Intermediate lake, nutrient poor N 4 4 4 4 4 1 5
Intermediate lake, nutrient rich N 4 4 4 5 5 4 5
Large lake, nutrient poor N 5 5 5 5 5 3 5
Large lake, nutrient rich N 5 5 5 5 5 5 5
Sum of Selected Conditions 6 6 6 5 5 2 6
Average of Selected Conditions 3 3 3 2.5 2.5 1 3
• Stream Size: small stream: average width of water surface less than 10 yards under winter flow conditions;
intermediate stream: 10 to 30 yards; large stream: greater than 30 yards.
• Lake Size: small lake: < 10 acres; intermediate lake: 10 to 100 acres; large lake: > 100 acres.
• Wetland: aquatic body not defined as a lake or stream, i.e. marsh, ponds, etc.
• Substrate: coarse substrate: boulder, cobble, gravel; fine substrate: sand, clay, muck.
• Nutrient Levels: nutrient poor: total P below 15 µg/L, total N below 300 µg/L; nutrient rich: > 15 µg/L P,
>300 µg/L N.
Table 2-25. Drinking water from groundwater—material/receptor

potential for impairment worksheet.
Exists?
Drinking Water Conditions NaCl MgCl2 CaCl2 CMA KA OMB Abrasives
(Y/N)
Shallow ground water used for drinking Y 3 5 5 4 4 3 5

Average of Selected Conditions 3 5 5 4 4 3 5
Table 2-26. Air quality—material/receptor potential for impairment

worksheet.
Exists?
Air Quality Conditions NaCl MgCl2 CaCl2 CMA KA OMB Abrasives
(Y/N)
Existing air quality concerns relevant to

particulate matter (i.e. non-attainment Y 5 5 5 5 5 5 1
area)
Poor air dispersion, closed air shed Y 5 5 5 5 5 5 1
Table 2-27. Vegetation—material/receptor potential for impairment

worksheet.
Exists?
Vegetation Conditions NaCl MgCl2 CaCl2 CMA KA OMB Abrasives
(Y/N)
Agricultural lands within 10m of the roadway Y 1 2 2 4 4 5 4

Protected vegetation within 10m of the roadway N 1 2 2 4 4 5 4
Aesthetic objectives within 10m of the roadway Y 1 2 2 4 4 5 4

G-23
Table 2-28. Soil quality—material/receptor potential for impairment

worksheet.
Exists?
Soil Quality Conditions NaCl MgCl2 CaCl2 CMA KA OMB Abrasives
(Y/N)
Poor soil structure within 10m of the roadway1 N 2 5 5 5 5 5 5

1
Exchangeable Sodium Percentage (ESP) >13
Table 2-29. Terrestrial fauna—material/receptor potential for impairment

worksheet.
Exists?
Animal Conditions NaCl MgCl2 CaCl2 CMA KA OMB Abrasives
(Y/N)
Potential for effects on terrestrial species 1 N 4 4 4 5 5 5 5

Roadkill for mammals or birds is a concern N 4 4 4 5 5 5 5
1
Includes sensitive mammals, birds, amphibians, reptiles, insects, spiders, etc.
Table 2-30. Vehicles—material/receptor potential for impairment worksheet.
Exists?
Vehicles NaCl MgCl2 CaCl2 CMA KA OMB Abrasives
(Y/N)
Vehicles Y 1 1 1 4 4 5 1
Table 2-31. Metal infrastructure—material/receptor potential for

impairment worksheet.
Exists?
Metal Infrastructure Conditions NaCl MgCl2 CaCl2 CMA KA OMB Abrasives
(Y/N)
Steel Bridges Y 1 1 1 4 4 5 5
Other metal infrastructure of concern Y 1 1 1 4 4 5 5
Table 2-32. Concrete matrix—material/receptor potential for impairment

worksheet.
Exists?
Concrete Degradation NaCl MgCl2 CaCl2 CMA KA OMB Abrasives
(Y/N)
Concrete pavement and structures Y 4 3 3 3 4 5 5


G-24
Table 2-33. Concrete reinforcing—material/receptor potential

for impairment worksheet.
Exists?
Concrete Reinforcing NaCl MgCl2 CaCl2 CMA KA OMB Abrasives
(Y/N)
Reinforcing Steel Bar Y 1 1 1 5 5 5 5

• Customer service (e.g., product availability and delivery Figure 2-3. Based on the selected conditions, NaCl is the pre-
times); and ferred product between 32 and 12°F, after which MgCl2 and
• Other performance concerns (e.g., roadway slickness and CaCl2 are preferred.
equipment compatibility).
2.7.1.2 Decision Tool Example 2—Equal Priorities
2.7.1.1 Decision Tool Example 1—Survey Priorities
The calculation spreadsheet for this case is provided as
The calculation spreadsheet for this case is provided as Table 2-35; temperature-dependent scoring is shown in
Table 2-34; temperature-dependent scoring is shown in Figure 2-4. Based on the selected conditions, NaCl is the pre-
Table 2-34. Decision tool example 1—survey.

A) Economic Weighting: Purchase Price: 45%
Performance: 35%
Temperature (F): 15
Material Information Parameters NaCl CaCl2 MgCl2 KA CMA
Input Purchase Price per ton = $ 50.00 $ 130.00 $ 120.00 $ 1,000.00 $ 1,150.00
Input Material Concentration (%) = 23 30 28 50 25
Purchase Price per dry pound ($/lb) $ 0.11 $ 0.22 $ 0.21 $ 1.00 $ 2.30
Melting Potential (MP) 0.9 1.4 1.5 1.9 0.4
(MP)/(Purchase Price/lb) 7.9 6.5 7.1 1.9 0.2
A) Weighted "Economic" Score 80.0 65.7 72.3 18.9 1.8
B) Environmental Weighting: 11%
NaCl CaCl2 MgCl2 KA CMA

Operational Objectives (Maximum 100) Potential For Impairment (1=High, 5=Low)
Aquatic Life 20% 3 3 3 2.5 2.5
Drinking Water 20% 3 5 5 4 4
Air Quality 27% 5 5 5 5 5
Vegetation 20% 1 2 2 4 4
Soil 0% 5 5 5 5 5
Animals 13% 5 5 5 5 5
TOTAL SCORE 3.4 4.0 4.0 4.1 4.1
MAX SCORE 5.0 5.0 5.0 5.0 5.0
B) Weighted "Environment" Score 7.5 8.8 8.8 9.0 9.0
C) Infrastructure Weighting: 9%

Vehicles 22% 1 1 1 4 4
Metal Infrastructure 28% 1 1 1 4 4
Concrete 20% 4 3 3 3 4
Concrete Reinforcing 30% 1 1 1 5 5
TOTAL SCORE 1.6 1.4 1.4 4.1 4.3
MAX SCORE 5.0 5.0 5.0 5.0 5.0
C) Weighted "Infrastructure" Score 2.9 2.5 2.5 7.4 7.7
Final Material Scores

Final Material Scores = A) + B) + C) 90.4 77.0 83.6 35.3 18.5

G-25
Figure 2-3. Decision tool example 1—survey priorities.

G-26
Table 2-35. Decision tool example 2—equal priorities.

Performance: 25%
Temperature (F): 15
B) Environmental Weighting: 25%

Aquatic Life 20% 3 3 3 2.5 2.5
Soil 0% 5 5 5 5 5
Animals 13% 5 5 5 5 5
TOTAL SCORE 3.4 4.0 4.0 4.1 4.1
MAX SCORE 5.0 5.0 5.0 5.0 5.0
C) Infrastructure Weighting: 25%

Vehicles 22% 1 1 1 4 4
Concrete 20% 4 3 3 3 4
TOTAL SCORE 1.6 1.4 1.4 4.1 4.3
MAX SCORE 5.0 5.0 5.0 5.0 5.0


G-27
Figure 2-4. Decision tool example 2—equal priorities.
ferred product between 32 and 12°F, after which MgCl2 and and CMA lose performance benefits. Although KA is the pre-
CaCl2 are preferred. ferred product, CMA also scores well at warmer temperatures.
2.7.1.3 Decision Tool Example 3—Environment/ 2.8 Software

Infrastructure Priorities
A software program was developed to assist users with
The calculation spreadsheet for this case is provided applying the snow and ice control material decision tool. The
as Table 2-36; temperature-dependent scoring is shown in software was written in “C+” programming language using
Figure 2-5. Based on the selected conditions, all materials have Microsoft Visual Studio and requires a personal computer
similar scores until temperatures reach below 10°F and NaCl installed with Microsoft Windows 2000, Windows XP, or later

G-28
Table 2-36. Decision tool example 3—environmental and infrastructure

priority objectives.
Performance: 25%
Temperature (F): 15
B) Environmental Weighting: 37.5%

Aquatic Life 20% 3 3 3 2.5 2.5
Soil 0% 5 5 5 5 5
Animals 13% 5 5 5 5 5
TOTAL SCORE 3.4 4.0 4.0 4.1 4.1
MAX SCORE 5.0 5.0 5.0 5.0 5.0
C) Infrastructure Weighting: 37.5%

Vehicles 22% 1 1 1 4 4
Concrete 20% 4 3 3 3 4
TOTAL SCORE 1.6 1.4 1.4 4.1 4.3
MAX SCORE 5.0 5.0 5.0 5.0 5.0


G-29
Final Product Scores Based on Te mperature

Te mperature (F) NaCl CaCl2 MgCl2 KA CMA
30 62.5 57.2 59.2 63.8 63.5
25 62.5 60.7 60.8 66.3 63.8
20 62.5 58.5 59.7 66.5 63.6
15 62.5 61.0 63.1 67.4 63.6
10 59.5 61.4 65.5 68.3 63.4
5 52.2 60.8 65.5 68.3 63.2
0 46.6 60.7 65.5 68.5 TOO COLD
-5 38.2 59.9 65.5 68.1 TOO COLD
-10 TOO COLD 59.6 65.5 68.7 TOO COLD
-15 TOO COLD 58.8 65.5 68.8 TOO COLD
-20 TOO COLD 57.6 65.5 68.9 TOO COLD
Final Score vs. Temperature (F)
80.0
70.0
60.0
50.0
Final Score
40.0
30.0
20.0
10.0
0.0
30 25 20 15 10 5 0 -5 -10 15
-15 -20
Temperature (F)
Figure 2-5. Decision tool example 3—environmental and infrastructure priority

objectives.
operating systems. Some systems may require installation of score, based on input information, that can be printed on a
Microsoft.NET runtime (available from Microsoft). summary report with the input information.
The software follows a “wizard-type” format and prompts Appendix A, which is available for download from the
users to populate fields with key agency and maintenance area TRB website, shows screen captures from the Decision Tool
information. Measures are built in to address inappropriate software program. User instructions are included on each
or insufficient information. The result is a numerical product page.

G-30
CHAPTER 3
Purchase Specification
This purchase specification is one of the steps in the snow based on a comprehensive technical assessment of environ-
and ice control material selection process developed mental impacts associated with snow and ice control materi-
through NCHRP Project 06-16. Its purpose is to assist trans- als and should be used to supplement existing agency
portation agencies with mitigating impacts to the natural purchasing policies and documents.
environment by allowing the user to qualify and compare
snow and ice control materials based on key environmental
3.1 Chemical Specifications
attributes. By establishing recommended material accept-
ance criteria, transportation agencies can screen unsuitable Numerical criteria specific to snow and ice control materi-
products. The purchase specification is to be applied after als have been established based on
the agency has completed the NCHRP 06-16 decision tool
and a snow and ice control material type has been selected • U.S. EPA numerical water quality criteria for the protection
for use. of aquatic life and drinking water (based on EPA National
Aquatic systems are the primary environmental concern Recommended Water Quality Criteria: 2002, EPA-822-R-
associated with snow and ice control materials. The purchase 02-047, November 2002);
specification focuses on snow and ice control material chem- • Site-specific factors, specifically receiving water type, size,
ical attributes that can impair aquatic systems under normal and current nutrient levels; and
use but are not addressed through the decision tool. In the • The default value for expected roadway runoff dilution,
overall material selection process, the decision tool addresses which is estimated to be 500 times at the point the materi-
primary components and attributes (e.g., chlorides, cations, als leave the roadway—this may be adjusted to reflect any
and organic matter from biomass), whereas the purchase other assumed or measured dilution factor.
specification addresses secondary components and attrib-
utes (e.g., heavy metals, nutrients, and cyanide) (refer to Table 3-1 lists current U.S. EPA criteria for potential chem-
Table 2-3). ical toxins and nutrients of general concern in snow and ice
Existing U.S. EPA aquatic life and drinking water quality control materials. These attributes were developed based on
criteria are used to establish guidelines for maximum accept- an information assessment conducted through Phase I of this
able concentrations (i.e., specification limits) for snow and ice study. Other potential toxins not listed in the table may be
control material chemical attributes. The chemical attributes found under unusual circumstances and may be considered
selected for assessment are easily measured using standard in addition to those in the table. Table 3-1 includes chronic
environmental test methods. Guidance is also provided and acute criteria for the protection of freshwater aquatic life
should the agency wish to develop specifications independ- and drinking water criteria. For metal species affected by
ently for physical, toxicity, and corrosion properties of snow hardness, criteria have been provided for hardness values of
and ice control materials. Areas for consideration and associ- 100 mg/L and 400 mg/L.
ated test methods are referenced; however, no numerical The Guidelines for Maximum Concentrations (GMCs) of
criteria have been recommended. constituents in snow and ice control materials are listed in
Most transportation agencies have formal purchasing poli- Table 3-2. These concentrations were derived assuming that
cies and documents in place that specify purchasing terms the concentrated snow and ice control materials are diluted
and conditions. This purchase specification was developed 500 times at the point these materials leave the roadway.

G-31
Table 3-1. USEPA criteria used in calculation of guidelines

for maximum concentrations for chemical constituents of concern
in snow and ice control materials. Criteria are subject to change with
USEPA criteria.
Chemical Freshwater Aquatic Life3 Drinking

Chemical 6
Symbol Acute, μg/L Chronic, μg/L Water, μg/L
Arsenic As 340 1509 10
Cadmium1 Cd 3 0.25 5
Cadmium2 Cd 13 9.6 -
Chromium (III) Cr (III) 570 74 100
Chromium (IV) Cr (IV) 16 11 100
Copper1 Cu 13 9 -
Copper2 Cu 50 29 -
Lead1 Pb 65 2.5 15
Lead2 Pb 280 11 -
Mercury Hg 1.4 0.779 2
Nickel1 Ni 470 52 -
Nickel 2 Ni 1520 170 -
Selenium Se (25)7 5 -
Silver1 Ag 3.2 (0.64) 8 -
Silver2 Ag 35 (7) 8 -
Zinc1 Zn 117 118 -
Zinc2 Zn 379 382 -
Cyanide (Free) CN 22 5.2 140
Ammonia (Total)4 NH3 30,000 6000 -
Phosphorus (Total)5 P 15 7.0 -
Nitrogen (Total)5 N 300 150 -
1
Criterion is dependent on hardness; the value shown here is for a hardness of 100 mg/L.
2
Criterion is dependent on hardness; the value shown here is for a hardness of 400 mg/L
3
Concentrations are the potentially dissolved forms, except as noted. Source: http://www.epa.gov.
4
At pH 7.5, 5o C.
5
National criteria vary by region; see text for basis.
6
No value is shown for constituents having more restrictive limits for protection of aquatic life; metals
are as total for drinking water.
7
Acute value for selenium does not exist. Reported value estimated as 5 times the chronic value.
8
Chronic value for silver does not exist. Reported value estimated as 1/5th the acute value.
9
Routine consumption of fish taken from an affected site would reduce the limit to approximately 0.001
times the indicated chronic value.
Where this is not valid, the numbers could be adjusted maximum allowable concentrations for these metals are
accordingly. Furthermore, GMC values assume that no lethal highest when calcium and magnesium are major com-
or chronic effects will be allowed to extend beyond the mar- ponents of ice and snow control materials.
gin of the roadway. Small extensions beyond the roadway 2. A second category applies to snow and ice control chemi-
might allow considerably more dilution and could be judged cals that contain neither calcium nor magnesium as major
reasonable on a site-specific basis. constituents (i.e., NaCl, KA, agriculturally derived organic
The subsequent text describes the basis for the develop- materials). Such products lack any hardness-inducing
ment of the GMC values. Three general categories exist component in large quantities, which requires more
according to type of snow and ice control material. restrictive limits for metals whose toxicity is affected by
hardness.
1. The first category includes chemical products with a 3. The third category of snow and ice control materials in the
calcium or magnesium base (i.e., with calcium or table is abrasives. These materials are not soluble, but may
magnesium as major ingredients and not as incidental contain contaminants that are soluble. Abrasives obtained
contaminants: MgCl2, CaCl2, CMA). Calcium and mag- from industrial processes (e.g., metallurgical slag, mining
nesium are the two components of hardness for water. waste, and furnace slag.) can contain high toxicant levels.
Concentration limits (i.e., aquatic life criteria) for certain Table 3-2 reflects the assumption that abrasives to be
metals (i.e., Cd, Cu, Pb, Ni, Ag, and Zn) likely to be found tested will be leached in pH-adjusted (i.e., acidic) water
in snow and ice control materials depend on hardness. according to U.S. EPA Method 1311, Toxicity Characteris-
The presence of calcium and magnesium in solution tic Leaching Procedure (Revision 0, July 1992), which
(i.e., hardness) reduces the relative toxicity of a given incorporates a liquid-to-solid ratio of 20:1 (50 grams
concentration of any one of these metals. Therefore, the abrasive plus 1 liter extraction fluid).

Table 3-2 Guidelines for Maximum Concentrations (GMCs) of potential toxins and nutrients present in snow and ice control materials
consistent with the protection of aquatic life and use of untreated ground water for drinking by humans.
Other Inorganics Nutrients
2
Free Total Total 2
Total
Maximum Allowable Potentially Dissolved Metals and Metalloids, mg/L CN, NH3 , P,
N, mg/L
mg/L mg/L mg/L
Category of Material and Environmental Context As Cd Cr4 Cu Pb Hg1 Ni Se Ag1 Zn

Chemical with Ca or Mg base:
MgCl2 , CaCl2, CMA - Liquid chemical or 25% w/w solid
solution
Low nutrient small stream, wetland, pond, or small lake 75 4.8 5.5 15 5.5 0.39 84 2.5 3.5 190 2.6 3000 3.5 75
High nutrient small stream, wetland, pond, or small lake 75 4.8 5.5 15 5.5 0.39 84 2.5 3.5 190 2.6 3000 3.5 75
Low nutrient large or intermediate stream, lake, or wetland 170 6.6 8.0 24 140 0.70 760 13 18 190 11 15000 7.5 150
High nutrient large or intermediate stream, lake, or wetland 170 6.6 8.0 24 140 0.70 760 13 18 190 11 15000 7.5 150
Drinking water wells near roadway3 5 2.5 50 - 7.5 1.0 - - - - 70 - - -
Chemical lacking Mg or Ca base:

NaCl, KA - Liquid chemical or 25% w/w solid solution
Low nutrient small stream, wetland, pond, or small lake 75 0.13 5.5 4.5 1.3 0.39 26 2.5 0.32 59 2.6 3000 3.5 75
High nutrient small stream, wetland, pond, or small lake 75 0.13 5.5 4.5 1.3 0.39 26 2.5 0.32 59 2.6 3000 3.5 75
Low nutrient large or intermediate stream, lake, or wetland 170 1.5 8.0 6.7 32 0.70 235 13 1.6 59 11 15000 7.5 150
High nutrient large or intermediate stream, lake, or wetland 170 1.5 8.0 6.7 32 0.70 235 13 1.6 59 11 15000 7.5 150
Drinking water wells near roadway3 5 2.5 50 - 7.5 1.0 - - - - 70 - - -
Abrasive Leachate:
1:20 w/v leachate per EPA 1311 (TCLP)
Low nutrient small stream, wetland, pond, or small lake 0.75 0.0013 0.055 0.045 0.013 0.004 0.26 0.025 0.0032 0.59 0.026 30 0.035 0.75
High nutrient small stream, wetland, pond, or small lake 0.75 0.0013 0.055 0.045 0.013 0.004 0.26 0.025 0.0032 0.59 0.026 30 0.0.35 0.75
Low nutrient large or intermediate stream, lake, or wetland 1.7 0.015 0.080 0.067 0.32 0.007 2.4 0.13 0.016 0.59 0.11 150 0.075 1.5
High nutrient large or intermediate stream, lake, or wetland 1.7 0.015 0.080 0.067 0.32 0.007 2.4 0.13 0.016 0.59 0.11 150 0.075 1.5
Drinking water wells near roadway3 0.05 0.025 0.50 - 0.075 0.010 - - - - 0.70 - - -
1
Fish ingestion by humans will require much lower limits.
2
Note that ammonia limits (based on toxicity) will be overridden by total N limits (based on nutrient effects).
3
Total for metals and metalloids.
4
Values based on hexavalent chromium.
• Small stream: average width of water surface less than 10 yards under winter flow conditions; intermediate or large stream: greater than 10 yards.
• Small lake, pond, or wetland: < 10 acres; intermediate or large lake, pond, or wetland: ≥10 acres
Nutrient poor: total P below 15 μg/L, total N below 300 μg/L; nutrient rich: P > 15 μg/L , N >300 μg/L.
G-33
U.S. EPA 1311, a widely recognized standard test method, readily available. For this reason, acute rather than chronic
provides a conservative estimate of leachability of contami- criteria are used in computing maximum allowable concen-
nants contained on solid materials. Use of a less aggressive trations. As in the case of small aquatic environments, this
extraction fluid (e.g., deionized water) may be considered a category is subdivided into low-nutrient and high-nutrient
more suitable approach to simulating highway environments. waters, and the allowable concentrations of nutrients are
Within each of the three categories of snow and ice control determined by a rationale identical to that used for smaller
materials shown above, five possible types of environmental waters.
context are given Total nitrogen and total phosphorus may be evaluated in
a somewhat different way than other constituents listed in
1. Small aquatic environment, low nutrient levels; the tables. Unlike the other constituents shown in the table,
2. Small aquatic environment, high nutrient levels; total nitrogen and total phosphorus are nutrients, rather
3. Large or intermediate aquatic environment, low nutrient than toxins. The stimulating effects of nutrients on plant
levels; growth may be more tolerable in many instances than poten-
4. Large or intermediate aquatic environment, high nutrient tial lethal or sub-lethal effects associated with toxins.
levels; and Although all of the limits recommended in the tables are well
5. Drinking water wells near the roadway. justified from the viewpoint of preserving the integrity of
aquatic ecosystems, the limits for N and P may be viewed in
Small aquatic environments (i.e., streams, wetlands, ponds, a more flexible, site-specific manner than limits associated
or lakes) account for two of these categories. For purposes of with toxic substances.
constructing the table, it has been assumed that these envi- A final subcategory for each category of snow and ice con-
ronments offer minimal dilution, and that the maximum trol material involves the presence of drinking water wells
allowable concentrations for potential toxins should be set on near the roadway. In this case, limits for human health are
the basis of chronic toxicity criteria. The criteria are taken relevant and are taken directly from the U.S. EPA’s Safe
directly from the U.S. EPA’s current list of water-quality crite- Drinking Water Standards. Concentrations are listed only
ria for protection of aquatic life. For metals affected by hard- for those substances that have human health-related stan-
ness, hardness is assumed to be 100 mg/L, if calcium and dards that are near or below the aquatic life standards; in
magnesium are not major constituents of the snow and ice other cases, it is assumed that the more stringent aquatic life
control material, and 400 mg/L, if the snow and ice control criteria will also be adequate to meet the safe drinking water
material contains calcium and magnesium as major con- standards. Because drinking water standards are subject to
stituents. The actual hardness at the edge of the roadway change, maximum allowable concentrations should be
probably will exceed 400 mg/L in the latter case, but the ben- adjusted downward proportionately if they are made more
eficial effects of hardness in reducing the toxicity of heavy restrictive in the future. Drinking water wells should be
metals generally is assumed to be fully expressed at about 400 tested for contaminants, even if the table indicates that com-
mg/L hardness. For very soft waters (100 mg/L), maximum pliance with the maximum allowable limits will provide
allowable concentrations can be set somewhat lower than protection.
indicated in Table 3-2, unless calcium and magnesium are Maximum allowable concentrations for the constituents
added as major constituents. listed in Table 3-2 are computed on the assumption that a
The small aquatic environments fall into two categories: 500-times dilution occurs on the roadway (i.e., road surface
low-nutrient and high-nutrient. Critical concentrations for plus road margins and medians) during runoff. This is a con-
nutrients in snow and ice control materials are set on the servative estimate based on a technical assessment of infor-
assumption that concentrations of total phosphorus (P) and mation gathered in Phase I of this study. For liquid snow and
total nitrogen (N) in low-nutrient systems should not be ice control materials, the 500 to 1 factor is applied to U.S. EPA
augmented by more than about 50 percent in areas near the criteria to provide a material GMC that is directly applicable
roadway. For environments that have high-nutrient concen- to the liquid product. Where solid chemicals are applied, the
trations, the amount of nutrients that can be added without material should be tested in liquid form. A sample of the
substantially perturbing the ambient condition is assumed to material should be dissolved in an amount of water equal to
be equal to the threshold concentrations separating low- 3 times the mass of the solid (25-percent w/w solution). This
nutrient from high-nutrient conditions. allows for testing using the same standard methods as for liq-
Two additional subcategories for each type of snow and ice uid chemicals and corresponds approximately to the concen-
control material apply to aquatic environments that are of tration of liquid snow and ice control materials. The
intermediate to large size, as defined in the footnote to Table produced liquid then can be evaluated as a liquid using Table
3-2. In this case, it is assumed that a substantial dilution is 3-2. In the case of abrasives, the same overall dilution rate of

G-34
500 times still applies; however, two other factors require 3.2.1 Physical Specifications—Chemical
consideration: Products
1. The test solution being evaluated is a leachate extract 3.2.1.1 Specific Gravity
based on a 20:1 liquid-to-solid ratio. To provide compara-
bility with chemicals, the 500-times dilution factor must Specific gravity is a material’s density (i.e., mass/volume)
be reduced by 1/20th. compared with water. Many agencies use specific gravity to
2. Abrasives are generally applied at average application rates monitor percentage of active ingredient in a liquid snow and
5 times that of chemicals (generally between 2 to 8 times). ice control material using standard density versus tempera-
It is, therefore, necessary to reduce the 500-times dilution ture information. This approach is relevant only for pure
factor by 1/5th. liquid products or if product-specific data is available.
ASTM D 1429, Standard Test Methods for Specific Gravity of
Overall the factor applied to applicable U.S. EPA water quality Water and Brine, lists methods to determine specific gravity
to develop abrasive leachate Maximum Acceptable Concen- for brines.
trations is as follows:
500-times dilution factor 1/20 leachate factor (Item 1) 3.2.1.2 Freezing Point and Eutectic Temperature
1/5 times application factor (Item 2) 5 times
Eutectic temperature is the lowest freeze-point tempera-
If there is a site-specific basis for stricter or more lenient ture achievable for a given product in solution. Eutectic
dilution allowance, the limits shown in Table 3-2 can be temperature, by definition, applies to pure products; how-
adjusted proportionately. The same reasoning applies to ever, in practice (i.e., for commercially available blends and
cyanide, total ammonia, and nutrients. The toxicity of total so forth) a material-specific phase curve can be developed
ammonia to aquatic life is affected by pH and temperature. For over a range of material concentrations. This involves
the purposes of preparing Table 3-2, it is assumed that the measuring the freezing characteristics of a solution with a
temperature will be 5°C and the pH will be 7.5. Large devia- previously identified concentration. ASTM D 1177, Stan-
tions from these assumptions are unlikely in winter, but if they dard Test Method for Freezing Point of Aqueous Engine
occur, a proportionate adjustment can be made based on U.S. Coolants, was developed to determine the freezing point of
EPA ammonia criteria as a function of pH and temperature. engine coolants and can be applied to liquid snow and ice
Evaluation of the GMC values shown in Table 3-2 shows control materials.
there is generally not a large difference in magnitude of GMC
values within each material category and environmental con-
3.2.1.3 Settleable Solids
text. It is also likely that a maintenance area will include more
than one of the possible conditions listed in Table 3-2. To PNS has developed a method to determine percent Total
ensure that the information can be applied in a practical man- Settleable Solids and Percent Solids Passing a No. 10 Sieve after
ner, the most conservative values have been summarized in a liquid chemical is stored at a specified cold temperature
Table 3-3 and should be applied for general use. (17.8 ± 1°C for MgCl2 and 29 ± 1°C for CaCl2) without
agitation for 168 hours (i.e., 1 week). The test procedure is
3.2 Other Specifications designed to determine the potential for solids formation
under extreme-cold field temperatures. The formation of
Historically, purchase specifications for snow and ice con- solids in liquid products can impair the operation of applica-
trol materials have focused on physical properties, primarily tion equipment.
to address operational and performance concerns. In recent
years, specifications for corrosion properties have also been
3.2.1.4 Gradation
developed. Given the concern surrounding some specific
environmental issues, some agencies may wish to adopt Methods for determining particle size distribution of solid
aquatic toxicity specifications as well. This purchase specifi- chemical products are based on methods similar to ASTM C
cation does not directly address physical, toxicological, or cor- 136, Sieve Analysis of Fine and Coarse Aggregates. The follow-
rosion properties of snow and ice control materials, and no ing test methods for gradation exist for solid chemical snow
numerical criteria are suggested. However, areas of consider- and ice control products:
ation have been identified, along with recommended test
methods to characterize these properties. If an agency • ASTM D 632, Standard Specification for Sodium Chloride;
chooses to incorporate specifications for these properties, it is and
recommended that this be conducted under the guidance of • ASTM D 98 and ASTM D 345, Standard Specifications for
a professional with expertise in the subject area. Calcium Chloride.

Table 3-3. Summary of most conservative Guidelines for Maximum Concentrations (GMCs) for general purchase
specification application.
Other Inorganics Nutrients
Free Total Total
Maximum Allowable Potentially Dissolved Metals and Metalloids, Total N,
CN, NH3, P,
mg/L mg/L2
mg/L mg/L2 mg/L
Category of Material and Environmental
As Cd Cr4 Cu Pb Hg1 Ni Se Ag1 Zn
Context
Chemical with Ca or Mg base:

MgCl2, CaCl2, CMA - Liquid chemical or 5 4.8 5.5 15 5.5 0.39 84 2.5 3.5 190 2.6 3000 3.5 75
25% w/w solid solution
Liquid lacking Mg or Ca base: 5 0.12 5.5 4.5 7.5 0.39 26 2.5 .32 59 2.6 3000 3.5 75
NaCl, KA - Liquid chemical or 25% w/w solid solution
Abrasive Leachate: 0.05 0.0013 0.055 0.045 0.013 0.004 0.26 0.025 0.0032 0.59 0.026 30 0.035 0.75
1:20 w/v leachate per EPA 1311 (TCLP)
G-36
3.2.2 Physical Specifications—Abrasives Determining the Percentage of Fractured Particles in

Coarse Aggregate, may be used to assess fracture.
From the standpoint of a purchase specification, abrasives
must be considered separately from snow and ice control
chemicals. These materials have their own physical properties 3.2.2.3 Particle Size—Flat/Elongated
such as gradation, hardness, fracture, density, etc. which can
affect performance. Large flat or elongated particles (sometimes referred to as
“arrow heads”) can cause tire punctures and can be assessed
using ASTM D4791, Standard Test Method for Flat Particles,
3.2.2.1 Gradation and Particle Size Elongated Particles, or Flat and Elongated Particles in Coarse
Gradation refers to the particle sizes in an aggregate mate- Aggregate.
rial. Particle size, which is related to mass, is considered to
affect the ability of abrasives to remain on the roadway. How- 3.2.2.4 Mechanical Abrasion
ever, heavier particles are more likely than lighter particles to
cause windshield damage. When pre-wetting fine particles, The mechanical abrasion test, sometimes called the “LA
application equipment can become clogged. Fine particles Rattler,” is used to determine the resistance of particles to
also contribute to the formation of frozen lumps and/or mechanical abrasion. Higher values indicate softer aggre-
clumps in outdoor stockpiles. The following methods may be gates. Relatively soft abrasives have a greater potential to
used to assess particle size of abrasives: break down, which can lead to reduced performance. Traf-
fic action on soft abrasives can create large amounts of fine
• ASTM D 5444, Standard Test Method for Mechanical Size particles, which can lead to air quality issues. ASTM C-131,
Analysis of Extracted Aggregate (AASHTO T-27 is equiva- Standard Test Method for the Resistance to Degradation by
lent); Abrasion and Impact in the Los Angeles Machine
• ASTM C 142, Standard Test Method for Clay Lumps and (AASHTO T-96 is equivalent), may be used to assess
Friable Particles in Aggregates (AASHTO T-112 is equiva- mechanical abrasion.
lent); and
• ASTM D 2419, Standard Test Method for Sand Equivalent
3.2.3 Toxicity Specifications
Value of Soils and Fine Aggregate (AASHTO T-176 is
equivalent). Table 3-4 lists common aquatic toxicity test methods for
chronic exposure to test organisms representing aquatic
vertebrates (i.e., fathead minnow), aquatic invertebrates
3.2.2.2 Fracture
(i.e., Ceriodaphnia dubia) and aquatic vegetation/algae (i.e.,
Fracture refers to the number of manufactured fractures Selenastrum capricornutum). Other species of concern (i.e.,
on an aggregate particle. Abrasives with large numbers of threatened or endangered) can be tested as well using appro-
fractured faces are thought to remain on the pavement priate methodologies and under the guidance of appropriate
surface longer. ASTM D-5821, Standard Test Method for professionals.

Fathead Minnow 7-day Survival and US Environmental Protection Agency, 1994, Short-Term
Growth Test Methods for Estimating the Chronic Toxicity of Effluents
and Receiving Water to Freshwater Organisms, 3rd Edition,
EPA/600/4-91/002
Ceriodaphnia dubia 3-brood Survival and US Environmental Protection Agency, 1994, Short-Term
Reproduction Test Methods for Estimating the Chronic Toxicity of Effluents
EPA/600/4-91/002
Selenastrum capricornutum 96-hour (US US Environmental Protection Agency, 1994, Short-Term

EPA) Growth Test Methods for Estimating the Chronic Toxicity of Effluents
EPA/600/4-91/002

G-37
3.2.4 Corrosion Specifications assess corrosion potential that an agency may consider adopt-
ing if this is a significant area of concern. Key areas include
Corrosion properties of snow and ice control materials are atmospheric corrosion of exposed metals, concrete degrada-
related to the ionic constituents of snow and ice control mate- tion, and corrosion of concrete reinforcing. Laboratory test
rials. Chloride salts can initiate and accelerate corrosion in results should be viewed as relative corrosion values only and
metals. Many commercially available snow and ice control not as absolute indicators of corrosion performance. Most
products are modified with additives intended to inhibit cor- corrosion tests reflect a particular type of corrosion under
rosion. There is significant ongoing research on the effective- defined conditions and for a limited number of metals or
ness of corrosion inhibitors in the field. Users may wish to alloys. The conditions that exist on the roadway can be highly
specify a corrosion inhibitor, but should weigh the added cost variable in terms of weather, chemical application, and the
in relation to identified needs and current scientific literature presence of other contaminants not related to snow and ice
on the subject carefully. There are numerous methods to control chemicals.

G-38
CHAPTER 4
Quality Assurance Monitoring Program
The Quality Assurance (QA) program is a suite of qualifying 2. Visual inspection, a simple act, can sometimes detect obvi-
tests and submittals for the agency to characterize snow and ice ous problems with a product. This may involve assessing
control products before purchase or use. This section includes whether the delivered material looks similar to previous
procedures for collecting and testing received snow and ice con- products.
trol chemicals. Standard test methods used to quantify material
characteristics are also referenced. To ensure suitability of the
4.3 Sample Collection Procedure
products with agency purchase specifications, it is suggested
that QA procedures be included in contract documents. It is recommended that each delivered load be sampled.
The following sample collection procedures for snow and ice
control materials reflect procedures developed by the PNS
4.1 Documentation
in 2002:
When receiving snow and ice control materials, documen-
tation must be maintained to ensure sufficient traceability to 1. QA sampling should be conducted at the time of delivery,
the materials source. This allows tracking in the event of non- directly from the delivery vehicle. It is recommended that
performing materials. Information may include composite samples be collected by quantitatively combin-
ing discrete (i.e., grab) samples collected from several
• Material verification information including, as applicable, points in the material transfer process. If single grab sam-
active ingredient concentration, purity, and material spe- ples are preferred, it is recommended that these be col-
cific phase curve information. It is recommended that the lected midway through the material transfer process.
latter be prepared by an independent source. Exercise care to allow sufficient flushing of valves, fittings,
• Shipping data with certified weights and volumes shipped. and hoses to prevent potential cross contamination. Metal
• Lot numbers. All materials should be identified by produc- (specifically brass) fittings can be problematic if this step
tion and/or shipping lot numbers. is not performed.
• Material Safety Data Sheets (MSDS). All materials should 2. A replicate sample (i.e., two containers of the same mate-
be accompanied by an MSDS that provides sufficient safety rial) is recommended for each material sampled. One
information for handling and in case of an accident. replicate is for laboratory submission; the second replicate
is retained by the agency in case of dispute or if further
testing is required.
4.2 Field QA Procedure
3. The testing laboratory should be contacted to confirm
Simple field observations can help ensure product suit- required sample volume and container type required for
ability. It is recommended that each delivered load be sub- the requested testing. A clean quart/liter plastic bottle is
jected to the following recommended field QA procedures: generally sufficient to conduct a general chemical analysis
of liquid materials. For solid products, plastic bags or
1. Specific gravity per ASTM D 1429 (refer to Section 3.2.1.1) wide-mouth jars (i.e., plastic or glass) capable of holding
should be conducted on liquid materials. Comparing 1 pound (0.450 kg) are generally sufficient.
measured values with standard curves can provide initial 4. Clearly label each product sample with adequate infor-
assurance of active ingredient concentration. mation to ensure traceability. This will include but not be

G-39
limited to material type, product name, vendor, lot num- APHA methods are well-recognized and are commonly
ber, delivery location, lot number, date, and sampler. used at most environmental laboratories in the United States
and Canada. For some parameters, APHA references more
Samples of abrasives can be collected based on ASTM D 75. than one analytical technique that may be used. Trends in
analytical chemistry are toward performance-based method-
ologies (PBMs) with specific data quality objectives (DQOs).
4.4 Laboratory Testing Effort
It is the laboratory’s responsibility to ensure that perform-
Chemical analysis of snow and ice control materials can ance-based objectives are met for the sample matrix being
provide a great deal of information in a relatively short time. tested. When adapting APHA methods to snow and ice con-
Testing should correlate with agency purchase specification trol materials, the following considerations are necessary.
criteria and should include the following:
• Standard test methods for water and wastewater samples
• Active ingredient concentration (i.e., NaCl, MgCl2, CaCl2, primarily involve volumetric sample aliquots. Liquid snow
KA, and CMA); and ice control products are denser than water with spec-
• Metals (i.e., arsenic, cadmium, chromium [un-speciated], ify gravity values generally ranging from 1.15 to 1.35. It is
copper, lead, mercury, nickel, selenium, silver, and zinc)— common to report chemical data for snow and ice control
All should be tested in their potentially dissolved form; materials both volumetrically (e.g., mg/L, %w/v) and
• Chloride; gravimetrically (e.g. mg/kg, %w/w).
• BOD; To support the purchase specification (Section 3.1), data
• COD; must be reported as mg/L with the exception of active
• Phosphorus, Total; ingredient concentration, which should be reported based
• Nitrogen, Total; on weight (%w/w). Initial measurements may be taken vol-
• Ammonia; umetrically or, alternatively, weight-based measurements
• pH; and can be corrected using specific gravity values.
• Cyanide. • Units and basis of measurement should be clearly indicated.
Ambiguous units such as parts per million (ppm) or percent
The list can be modified to suit agency needs, but should (%) can cause confusion and should be clarified to reflect the
reflect the content of the agency’s purchase specification. basis of measurement. The following measurement units are
Where products are known not to contain significant preferred: milligrams per liter (mg/L), milligrams per kilo-
amounts of a constituent (e.g., for chloride in a CMA prod- gram (mg/kg), percent volume per volume (% v/v), percent
uct), it is still recommended that the full suite of testing be mass per volume (% w/v) or percent mass per mass (% w/w).
conducted to ensure completeness of the data set. • Reported Method Detection Limits (MDL) should be
Overall testing frequency should be determined by the sufficient to support purchase specification values with
agency. Where multiple loads of a single product are received, reasonable data confidence. It is recommended that MDLs
it is recommended that at least 10 percent of the received be equal to or less than 1/5th of the purchase specification
loads be tested, selected on a random basis. When multiple limit. If this is not possible using current and reasonably
product types are received, it is recommended that at least one available technologies, the reported result must, at mini-
test be conducted for each material. Additionally, any loads mum, be less than the reported detection limit or the data
suspected of being out of specification should be tested. Test must be qualified.
protocols are listed in Section 4.5. • Methods used for testing must go through sufficient
method validation to ensure that the method is fit for the
purpose. This will generally include an assessment of
4.5 Standard Test Methods
method detection limit, precision, and accuracy.
It is essential for a QA program to specify standard test • The laboratory should also use sufficient quality samples
methods to ensure data consistency among laboratories and with each batch analyzed to support the reported data. This
over time. Many chemical analysis methods used for snow will generally include where applicable
and ice control materials have been adapted from standard
environmental test methods for water and wastewater, most 1. Duplicates to monitor precision, conducted at a fre-
notably from Standard Methods for the Examination of Water quency of 5 to 10 percent with relative percent differ-
and Wastewater, 20th Edition, American Public Health Asso- ence (RPD) targets of less than 20 percent.
ciation, American Water Works Association, Water Environ- 2. Laboratory control samples (e.g., blank spikes, matrix
ment Federation, 1998, Washington DC (APHA). spikes, and reference materials) to assess recovery,

G-40
conducted at a frequency of 5 to 10 percent with recov- concentration is equivalent to the purchased concentration.
ery targets between 80 and 120 percent. Further, product concentrations are important to help
3. Blanks to assess contamination and interferences, con- guide material performance estimations. Active ingredient
ducted at a frequency of 5 to 10 percent with target concentrations may be estimated using specific gravity
acceptance levels of less than the reported detection limit. charts specific to each material; however, direct chemical
analysis of the cationic species (e.g., Ca, Mg, Na, K) fol-
• Test reports should include sample identification, test lowed by a molecular weight (MW) conversion to the salt
results, units of measurement, method detection limits, species is preferred. This approach assumes that all of the
analysis methods employed, and special notes if applicable. cation exists as the salt species. For example, Ca would be
measured to determine CaCl2 levels and Ca and Mg would
be measured to determine CMA concentrations. For chloride-
4.5.1 Sample Preparation
based chemicals, it can be helpful to balance cation con-
The following sections outline sample preparation steps for centrations with chloride concentrations as an additional
liquid chemical, solid chemical, and abrasive samples. check of data quality.
• NaCl (% w/w) Na (% w/w) 58.44 (MW NaCl)/22.99

4.5.1.1 Liquid Chemicals (MW Na);
Liquid snow and ice control chemicals are to be tested and • CaCl2 (%w/w) Ca (% w/w) 110.99 (MW CaCl2)/40.08
reported based on the received material in the units of mg/L, (MW Ca);
with the exception of active ingredient concentration, which • MgCl2 (%w/w) Mg (% w/w) 95.21 (MW MgCl2)/24.31
is to be converted to percent based on weight (% w/w). (MW Mg);
• KA (% w/w) K (% w/w) 98.14 (MW KA)/39.01
(MW K); and
4.5.1.2 Solid Chemicals • CMA (% w/w) [Ca (% w/w) 99.12 (MW CA)/40.08
Solid snow and ice control chemicals are to be diluted at (MW Ca)] [Mg (% w/w) 83.35 (MW MA)/24.31
room temperature with reagent-grade water (at minimum, (MW Mg)]
ASTM D 1193, Type II) to achieve a 25-percent w/w solution
(i.e., dissolve 250 grams of solid material in 750-grams of Simple wet chemistry techniques can be effective at meas-
reagent-grade water). Any insoluble materials should be noted. uring percent-level active ingredient concentrations common
When comparing test results with the purchase specification, with snow and ice control products. Standard titration pro-
test data is to be reported based on the prepared solution in cedures exist for calcium and magnesium determination:
mg/L, with the exception of active ingredient concentration, ASTM E 449 Standard Test Methods for Analysis of Calcium
which is to be converted to percent based on weight (% w/w). Chloride. Procedures are given for the determination of cal-
cium chloride, magnesium chloride, potassium chloride,
sodium chloride, and calcium hydroxide.
4.5.1.3 Abrasives Instrumental analysis procedures are more complex, but
Abrasives are to be prepared based on U.S. EPA Method are generally thought to be more specific and precise than wet
1311, Toxic Characteristic Leaching Procedure, U.S. EPA chemistry techniques. Ca, Mg, Na, and K are metals that can
SW-846, Test Methods for the Evaluation of Solid Waste, be detected by various instrumental procedures including
Physical/Chemical Methods, www.epa.gov/epaoswer/hazwaste/ atomic absorption spectroscopy, atomic emission spec-
test/sw846.htm. This is an extraction method that involves troscopy and mass spectroscopy. Refer to Section 4.5.3 for
a 1:20 (solid-to-liquid) leachate procedure. The leachate further details.
extract is to be tested directly, and test data is to be based on Units of measurement for active ingredient concentration
the leachate solution in units of mg/L. shall be percent based on weight (% w/w).
4.5.2 Active Ingredient 4.5.3 Metals

Active ingredient refers to the primary components of This purchase specification addresses the following heavy
the snow and ice control chemical. For present purposes, metals: arsenic, cadmium, chromium, copper, lead, mercury,
standard active ingredients include NaCl, CaCl2, MgCl2, nickel, selenium, silver, and zinc. These are to be tested in their
CMA, and KA. An accurate assessment of the active ingre- potentially dissolved form. Sophisticated instrumental analy-
dient is important to ensure that the received product sis is required to achieve the detection levels needed to meet

G-41
purchase specification limits. Instrumental techniques are not 4.5.5 Nitrogen

without drawbacks. Snow and ice control materials contain
high-percent levels of active ingredients that can interfere Various forms of nitrogen are present in snow and ice con-
with precise quantification of low-level analytes. Where high- trol products and include Nitrite (NO2 ) and Nitrate (NO3 ) ;
level dilutions are necessary, variances in measurement can be Ammonia (NH3); Organic Nitrogen (TON); Total Kjeldahl
greatly amplified after dilution factors are applied. nitrogen (TKN); and Total Nitrogen (TN). Total nitrogen
Instruments used for metals analysis include the following: analysis includes all forms of nitrogen and is the preferred
index to evaluate nitrogen content with respect to potential
• APHA 3111, Metals by Flame Atomic Absorption Spec-
environmental impairment. A separate evaluation of ammo-
troscopy (AA). Full range of metals analyses possible. nia is warranted because ammonia is potentially toxic to
• APHA 3112, Metals by Cold-Vapor Atomic Absorption
aquatic life at high concentrations.
(CVAA). Applicable to Hg analysis. Test methods to determine nitrogen compounds include
• APHA 3120, Metals by Inductively Coupled Plasma/Optical
various available options for each nitrogen compound:
Emission Spectroscopy (ICP/OES). Full range of metals
• APHA 4500-N A, B, and C;
analyses possible.
• APHA 4500-NH3 Nitrogen (Ammonia);
• APHA 3125, Metals by Inductively Coupled Plasma/Mass
Spectroscopy (ICP/MS). Full range of metals analyses • APHA 4500-NO2 Nitrogen (Nitrite);
possible. • APHA 4500-NO3 Nitrogen (Nitrate);
• APHA 4500-Norg Nitrogen (Organic)—This section includes
Each instrument technique listed above is technically suitable the standard Total Kjeldahl Nitrogen (TKN) methodology.
for the measurement of metals (and active ingredients) in
snow and ice control products. Some methods are better It is the responsibility of the analytical laboratory to ensure
suited to analyzing certain metals than others. A skilled that the selected methods meet accepted performance crite-
operator with knowledge of potential analytical interferences ria for application to snow and ice control materials.
and corrections is important in ensuring that performance
objectives are met.
4.5.6 Cyanide
Iron-cyanide compounds are commonly added to salt as
4.5.4 Total Phosphorus
anti-caking agents. Cyanide (CN) can exist as free cyanide,
Phosphorus can exist in various forms ranging from where the CN molecule is soluble and biologically available,
orthophosphate, which is readily soluble, to total phosphorus, or as complexed cyanide compounds (i.e., iron-cyanide com-
which also includes bound and combined phosphorus. pounds), where the CN molecule is sparingly soluble and not
Orthophosphates are thought to be more environmentally readily biologically available. Direct standard analytical tech-
available, although total phosphorus can degrade to niques are not available for iron-cyanide compounds; there-
orthophosphates through natural processes. Therefore snow fore, total cyanide analysis provides a conservative surrogate
and ice control materials should be evaluated primarily in method for their quantification.
terms of total phosphorus, rather than specific phosphorus Total cyanide analysis requires digestion using acids, reduc-
fractions. In order to degrade total phosphorus into a more ing agents, and/or ultraviolet light to convert cyanide com-
easily measurable form, APHA (67) includes the following pounds into cyanide gas (HCN), which is distilled and
total hydrolysis method: adsorbed into a basic solution, from which HCN can be
determined. The chemistry involved in cyanide analysis can
• APHA 4500-P B, Sample Preparation, describes procedures be subject to numerous interferences. The laboratory must
that employ perchloric acid, sulfuric/nitric acids, or per- ensure that methods used meet acceptable performance
sulfate digestion. criteria. Cyanide analysis should be based on APHA 4500-
CN C, Total Cyanide after Distillation. This procedure
After digestion, phosphorus may be determined colorimetri- details the acid digestion procedure used to liberate CN from
cally using the following methods: the sample and adsorb it into a basic NaOH solution.
The following methods detail measurement of the CN ion
• APHA 4500-P C, Vanadomolybdophosphoric Acid Colori- after digestion:
metric Method
• APHA 4500-P D, Stannous Chloride Method • APHA 4500-CN E, Colorimetric Method, and
• APHA 4500-P E, Ascorbic Acid Method • APHA 4500-CN F, Selective Ion Electrode Method.

G-42
The approach taken here—measurement of total of the QA Monitoring Program. BOD is an empirical test
cyanide—is conservative because cyanide from iron-cyanide used to determine the relative oxygen requirements of a sam-
complexes can become available in the natural environment. ple subjected to biochemical oxidation. The standard 5-Day
Other analytical methods (e.g., measurement of cyanide- BOD test has fixed incubation conditions of 20 ± 2°C. To
amenable chloride) may be a better indicator of readily avail- assess biodegradability under alternate (i.e., field) condi-
able cyanide, but practical challenges exist. Given the high tions, it may be desirable to conduct BOD testing over longer
potential organic content of many snow and ice control mate- periods and at various temperatures. The 5-day BOD test
rials, very high chlorination levels would be required to first procedure is conducted according to APHA 5210 B, 5-Day
oxidize all organic matter and then liberate the CN. In a prac- BOD Test.
tical sense, this may render any cyanide amenable to chlorine Chemical oxygen demand (COD) is used to measure the
potentially non-detectable. oxygen equivalent of the organic matter subjected to a strong
chemical oxidant. COD can be conducted relatively quickly
and can be empirically related to BOD. COD values will be
4.5.7 Organic Matter (BOD and COD)
equal to or greater than BOD. COD can be analyzed using the
Organic matter content influences snow and ice control following test methods:
material selection through the decision tool and is a recom-
mended component of the purchase specification. Organic • APHA 5220 B, Open Reflux Method
matter tests have environmental importance, are relatively • APHA 5220 C, Closed Reflux Method
simple to conduct, and are, therefore, recommended as part • APHA 5220 D, Closed Reflux Colorimetric Method

Abbreviations and acronyms used without definitions in TRB publications:

AAAE American Association of Airport Executives
AASHO American Association of State Highway Officials
AASHTO American Association of State Highway and Transportation Officials
ACI–NA Airports Council International–North America
ACRP Airport Cooperative Research Program
ADA Americans with Disabilities Act
APTA American Public Transportation Association
ASCE American Society of Civil Engineers
ASME American Society of Mechanical Engineers
ASTM American Society for Testing and Materials
ATA Air Transport Association
ATA American Trucking Associations
CTAA Community Transportation Association of America
CTBSSP Commercial Truck and Bus Safety Synthesis Program
DHS Department of Homeland Security
DOE Department of Energy
EPA Environmental Protection Agency
FAA Federal Aviation Administration
FHWA Federal Highway Administration
FMCSA Federal Motor Carrier Safety Administration
FRA Federal Railroad Administration
FTA Federal Transit Administration
IEEE Institute of Electrical and Electronics Engineers
ISTEA Intermodal Surface Transportation Efficiency Act of 1991
ITE Institute of Transportation Engineers
NASA National Aeronautics and Space Administration
NASAO National Association of State Aviation Officials
NCFRP National Cooperative Freight Research Program
NCHRP National Cooperative Highway Research Program
NHTSA National Highway Traffic Safety Administration
NTSB National Transportation Safety Board
SAE Society of Automotive Engineers
SAFETEA-LU Safe, Accountable, Flexible, Efficient Transportation Equity Act:
A Legacy for Users (2005)
TCRP Transit Cooperative Research Program
TEA-21 Transportation Equity Act for the 21st Century (1998)
TRB Transportation Research Board
TSA Transportation Security Administration
U.S.DOT United States Department of Transportation

De Icing Guidelines

Uploaded by

Copyright:

Available Formats

De Icing Guidelines

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

De Icing Guidelines

Uploaded by

Copyright:

Available Formats

THE NATIONAL ACADEMIES PRESS

This PDF is available at http://nap.edu/23178 SHARE

Guidelines for the Selection of Snow and Ice Control

194 pages | 8.5 x 11 | PAPERBACK

FIND RELATED TITLES

– Access to free PDF downloads of thousands of scientiﬁc reports

Copyright © National Academy of Sciences. All rights reserved.

NATIONAL COOPERATIVE HIGHWAY RESEARCH PROGRAM

NCHRP REPORT 577

LEVELTON CONSULTANTS LIMITED

TRANSPORTATION RESEARCH BOARD

Copyright National Academy of Sciences. All rights reserved.

NATIONAL COOPERATIVE HIGHWAY NCHRP REPORT 577

Published reports of the

NATIONAL COOPERATIVE HIGHWAY RESEARCH PROGRAM

Transportation Research Board

and can be ordered through the Internet at:

Copyright National Academy of Sciences. All rights reserved.

Copyright National Academy of Sciences. All rights reserved.

COOPERATIVE RESEARCH PROGRAMS

CRP STAFF FOR NCHRP REPORT 577

NCHRP PROJECT 6-16 PANEL

Copyright National Academy of Sciences. All rights reserved.

Copyright National Academy of Sciences. All rights reserved.

Copyright National Academy of Sciences. All rights reserved.

Copyright National Academy of Sciences. All rights reserved.

29 3.3 Components of Concern

Copyright National Academy of Sciences. All rights reserved.

71 4.1.4 Atmospheric Corrosion Costs

Copyright National Academy of Sciences. All rights reserved.

92 CHAPTER 6 General Assessment of Information

Copyright National Academy of Sciences. All rights reserved.

131 10.3.4 Chloride Diffusion—ASTM C 1556

Copyright National Academy of Sciences. All rights reserved.

Guidelines for the Selection of

NCHRP Project 06-16 Phase I

Copyright National Academy of Sciences. All rights reserved.

• The enabling and/or acceleration of corrosion of concrete reinforcing (rebar) resulting

Copyright National Academy of Sciences. All rights reserved.

amount of deterioration is directly related to the quality of concrete. Current evidence

NCHRP Project 06-16 Phase II

Copyright National Academy of Sciences. All rights reserved.

Copyright National Academy of Sciences. All rights reserved.

Copyright National Academy of Sciences. All rights reserved.

Copyright National Academy of Sciences. All rights reserved.

Quality Assurance Monitoring Program

Copyright National Academy of Sciences. All rights reserved.

Copyright National Academy of Sciences. All rights reserved.

Copyright National Academy of Sciences. All rights reserved.

Operational and Performance Considerations

Copyright National Academy of Sciences. All rights reserved.

Table 2-2. Chloride salts, general properties.

Copyright National Academy of Sciences. All rights reserved.

Table 2-3. Organic products, general properties.

Other Organic Glycols Not Available Not

Copyright National Academy of Sciences. All rights reserved.

Copyright National Academy of Sciences. All rights reserved.

Table 2-4. Snow and ice control material preference.

1st 2nd 3rd 4th 5th 6th

NaCl solid 57% (16) 18% (5) 4% (1) 0 0 0