Aashto Codes2

Evaluation of Performance Graded Asphalt
Binder Equipment and Testing Protocol
By
John P. Zaniewski
Michael E. Pumphrey
Asphalt Technology Program
Department of Civil and Environmental Engineering
Morgantown, West Virginia
April 2004
i
NOTICE
The contents of this report reflect the views of the authors who are responsible for
the facts and the accuracy of the data presented herein. The contents do not necessarily
reflect the official views or policies of the State or the Federal Highway Administration.
This report does not constitute a standard, specification, or regulation. Trade or
manufacturer names which may appear herein are cited only because they are considered
essential to the objectives of this report. The United States Government and the State of
West Virginia do not endorse products or manufacturers. This report is prepared for the
West Virginia Department of Transportation, Division of Highways, in cooperation with
the US Department of Transportation, Federal Highway Administration.
ii
Technical Report Documentation Page
1. Report No. 2. Government 3. Recipient's catalog No.

Accociation No.
4. Title and Subtitle 5. Report Date April, 2004
Evaluation of Performance Graded Asphalt 6. Performing Organization Code
Binder Equipment and Testing Protocol
7. Author(s) 8. Performing Organization Report No.

John P. Zaniewski, Michael E. Pumphrey
9. Performing Organization Name and Address 10. Work Unit No. (TRAIS)
Asphalt Technology Program
Department of Civil and Environmental 11. Contract or Grant No.
Engineering
West Virginia University
P.O. Box 6103
Morgantown, WV 26506-6103
12. Sponsoring Agency Name and Address 13. Type of Report and Period Covered
West Virginia Division of Highways
1900 Washington St. East 14. Sponsoring Agency Code
Charleston, WV 25305
15. Supplementary Notes
Performed in Cooperation with the U.S. Department of Transportation - Federal Highway
Administration
16. Abstract The testing and grading of asphalt cement has progressively cha nged to meet
the demanding needs for quality hot mix asphalt pavements. From October 1987 through
March 1993, the Strategic Highway Research Program (SHRP) conducted a $50 million
research effort to improve asphalt pavement technology, resulting in the Performance Graded
asphalt binder tests and specifications.
The objective of this research was to critically evaluate the SHRP performance graded
testing equipment purchased by the Asphalt Technology Laboratory at West Virginia
University: Rolling Thin Film Oven, Pressure Aging Vessel, Rotational Viscometer, Dynamic
Shear Rheometer, and the Bending Beam Rheometer. A testing program was established to
evaluate the equipment performance and related test methods and to develop a confidence in
the testing protocol and equipment performance such that the new equipment could be used
for continued research and industry testing. Sources of equipment and operator error were
identified and documented. The underlying theories of the equipment and performance graded
specifications were consolidated and documented as well.
.
17. Key Words 18. Distribution Statement
Performance Grade binder tests
19. Security Classif. (of 20. Security Classif. 21. No. Of Pages 22. Price
this report) (of this page) 100
Unclassified Unclassified
Form DOT F 1700.7 (8-72) Reproduction of completed page authorized
iii
Table of Contents
List of Figures iv
List of Tables v
Chapter 1 Introduction 1
1.1 History of Asphalt Binder Grading Methods 1
1.2 Problem Statement 3
1.3 Research Objectives 3
1.4 Scope and Limitations 4
1.5 Organization of the Report 5
Chapter 2 Literature Review and Background 6
2.1 Introduction 6
2.2 Asphalt Rheology 6
2.3 Penetration Grading System Specifications 9
2.4 Viscosity Grading System Specifications 10
2.5 Performance Graded Binder Specifications 13
2.6 Rolling Thin Film Oven 16
2.7 Pressure Aging Vessel 19
2.8 Rotational Viscometer 22
2.9 Dynamic Shear Rheometer 28
2.10 Bending Beam Rheometer 44
2.11 Summary 54
Chapter 3 Research Approach 55
3.1 Introduction 55
iv
3.2 Testing Program Preparation 55
3.1 Testing Program 58
Chapter 4 Test Results Analysis and Equipment Evaluation 67
4.1 Introduction 67
4.2 Rolling Thin Film Oven 67
4.3 Pressure Aging Vessel 68
4.4 Rotational Viscometer 69
4.5 Dynamic Shear Rheometer 87
4.6 Bending Beam Rheometer 92
Chapter 5 Conclusion and Recommendations 97
5.1 Conclusions 97
5.2 Recommendations 99
References 101
LIST OF FIGURES
Figure 2.1 Mechanical Response of Elastic, Viscous, and Viscoelastic Materials.............7
Figure 2.2 Despatch Rolling Thin Film Oven ..................................................................17
Data Collection ..................................................................................................................17
Figure 2.3 ATS Pressure Aging Vessel ............................................................................20
Figure 2.4 Typical Equiviscosity Curve for Asphalt Binder ............................................23
Figure 2.5 Brookfield DV-III Rotational Viscometer.......................................................25
Figure 2.6 Bohlin Instruments DSR II Mechanical Dynamic Shear Rheometer..............30
Figure 2.7 Configuration and Load Cycle of Dynamic Shear Rheometer........................34

v
Figure 2.8 Phase Angles for Elastic and Viscous Materials .............................................37
Figure 2.9 Relationship Between Complex Modulus and Phase Angle ...........................39
Figure 2.10 Graphical Description of Linear Viscoelastic Region...................................40
Figure 2.11 ATS Bending Beam Rheometer ....................................................................47
Figure 2.12 Graphical Definition of Creep Rate, m-value ................................................51
Figure 2.13 Development of Time-Temperature Shift Factors ........................................52
Figure 3.1 Testing Program Summary..............................................................................60
Figure 3.2 Testing Program Summary – continued ..........................................................61
Figure 4.4 RV Temperature Sweep Test 17 Equiviscosity Curve ....................................84
LIST OF TABLES
Table 2.1 ASTM D946 Requirements for Penetration Graded Asphalt Cements ............10
Table 2.2 ASTM D3381 Requirements for Viscosity Graded Asphalt Cements .............12
Table 2.3 Performance Graded Asphalt Cement Grades ..................................................14
Table 2.4 Summary of AASHTO MP1 Requirements .....................................................14
Table 2.5 Target Shear Stress and Strain Values ..............................................................31
Table 3.1 Testing and Conditioning Sequence .................................................................59
Table 3.2 Sample Set Conditioning Sequence ..................................................................65
Table 3.3 Testing Sequence ..............................................................................................66
Table 4.1 RTFO Mass Loss Results .................................................................................67
Table 4.2 RV Temperature Sweep Test 17 Results ..........................................................70
Table 4.3 RV Temperature Sweep Test 18 Results ..........................................................71
Table 4.4 RV Single Temperature Tests 19 through 32 Results........................................72
Table 4.5 RV Temperature Sweep Tests Precision Results..............................................77

vi
Table 4.6 RV Single Temperature Tests Precision Results ..............................................77
Table 4.7 DSR Test Results ..............................................................................................88
Table 4.8 DSR Tests Precision Results.............................................................................89
Table 4.9 BBR Test Results ..............................................................................................93
Table 4.10 BBR Tests Precision Results ..........................................................................94

1
CHAPTER 1
INTRODUCTION
Hot mix asphalt (HMA) pavements are an extensive part of the roadway system in
the United States. Building and maintaining HMA highways is an expensive task,
therefore, it is vital that a quality product be consistently produced. Efforts are
continually being made to increase the quality and performance of the asphalt concrete
pavements. Asphalt concrete consists of aggregate and asphalt cement binder. The
overall performance of the mixture is largely dependent on the type of asphalt cement
binder and its quality. The testing and grading of asphalt cement binders has
progressively changed to meet the demanding needs for quality HMA pavements.
1.1 HISTORY OF ASPHALT BINDER GRADING METHODS
Roberts, et al. (1996) present a review of asphalt binder grading methods used in
the HMA pavement industry, which served as the basis for the following discussion.
Prior to 1987, asphalt binders were tested and graded by two primary methods:
penetration grading and viscosity grading. The American Association of State Highway
Officials (AASHO) published the standard specifications for penetration graded asphalt
cements in 1931. The penetration grading system was created to establish different
asphalt grades for varying climates and applications. The primary test for penetration
grading is the penetration test. The penetration test is an empirical consistency test that
measures the penetration of a standard needle into a sample of original asphalt cement at
25 o C under a standard load and duration. A retained penetration test is also performed on
the asphalt binder following short-term aging (hardening) in a thin- film oven. Additional
tests are conducted to evaluate the asphalt binder’s flash point, purity, and ductility.
The penetration test is purely empirical. It fails to measure the consistency of the
binder in fundamental scientific units. In addition, the performance of the binder during
testing at 25 o C, which is close to the average pavement service temperature, may not be
applicable to the performance of the binder at lower or higher service temperatures.
The American Association of State and Highway Transportation Officials

(AASHTO), American Society for Testing and Materials (ASTM), Federal Highway
2
Association (FHWA), asphalt industry, and several state highway departments sought to
replace the empirical tests of the penetration grading system with scientifically-based
viscosity tests. The new specification system would measure the consistency of the
asphalt at temperatures approximating the maximum pavement surface temperature. The
resulting asphalt cement viscosity grading system was initiated in the early 1960’s and
soon became the most widely used grading system in the United States. The asphalt
cement viscosity grading system, also referred to as AC viscosity grading system,
characterizes asphalt consistency based on viscosity at 60 o C. Asphalt consistency at near
mixing/compacting temperatures is evaluated by conducting viscosity tests at 135 o C.
Penetration tests at 25 o C are performed to evaluate the asphalt’s consistency at average
service temperatures. Viscosity and ductility tests are performed on thin- film oven aged
asphalt samples as well. Additional tests are also conducted to evaluate the asphalt
binder’s flash point, ductility, and solubility.
A variation of the AC viscosity grading system is the aged residue viscosity

grading system, or AR viscosity grading system. The AR viscosity grading system
characterizes asphalt using aged residue from the rolling thin film oven, RTFO. Like the
AC viscosity grading system, the consistency of the RTFO-aged asphalt is characterized
based on viscosity at 60 o C. Additional specifications include a minimum penetration
and minimum viscosity at 25 o C and 135 o C, respectively. The AR viscosity grading
system has been used primarily by western states.
While the viscosity grading system was an improvement to the penetration

grading system, there were still shortcomings. The system still failed to provide a means
for evaluating the low temperature performance of asphalt binders. In addition, the
system neither evaluated modified asphalt binders nor considered long-term aging of
asphalt binders in service.
From October 1987 through March 1993, the Strategic Highway Research
Program (SHRP) conducted a $50 million research effort to develop performance-based
tests and specifications for both asphalt binders and HMA mixtures. The research led to
the development of the Superpave mixture design system, which stands for Superior
3
Performing Asphalt Pave ments. The Superpave system provides an improved means of
asphalt mixture design and analysis.
In addition to the Superpave system, Performance Graded (PG) binder tests and
specifications for evaluating asphalt binder properties and performance were also
developed during the SHRP research. The new SHRP binder specifications were
developed to address the shortcomings of the previous asphalt grading systems. The PG
binder tests include four physical tests and two conditioning methods. The conditioning
methods are the Rolling Thin Film Oven (RTFO) and the Pressure Aging Vessel (PAV).
The physical tests are the Rotational Viscometer (RV), Dynamic Shear Rheometer
(DSR), Direct Tension Tester (DTT) and the Bending Beam Rheometer (BBR). The PG
binder tests are described in detail in Chapter 2.
1.2 PROBLEM STATEMENT
The Asphalt Technology Laboratory at West Virginia University purchased the

following SHRP performance graded testing units:
• Despatch Rolling Thin Film Oven

• Applied Test Systems, Inc. (ATS) Pressure Aging Vessel
• NAPCO Model 5831 Vacuum Oven
• Brookfield Model DV-III Rotational Viscometer (RV Series)
• Bohlin Instruments DSR II Mechanical Dynamic Shear Rheometer
• ATS Bending Beam Rheometer
The equipment was new to the Asphalt Technology Laboratory and untried;
therefore, an evaluation of the equipment was needed. As with any new testing method
and protocol, there was also a learning curve and an adjustment period and that needed to
be satisfied to develop confidence in the use of the equipment and application of the
testing protocol. A confidence in the testing protocol and equipment performance was
necessary before the equipment could be used for continued research and industry testing.
1.3 RESEARCH OBJECTIVES
The objective of this research was to critically evaluate the SHRP binder testing
equipment in the Asphalt Technology Laboratory at West Virginia University. The
research effort was intended to evaluate the equipment calibration, operation and testing
procedures. A testing program was established to evaluate the equipment precision and
4
repeatability, as per the applicable AASHTO standard specification precision and bias
statements. Equipment problems were to be identified and corrected as necessary.
Potential sources of equipment and operator error, if any, were to be identified and
documented.
One of the stated advantages of the Performance Graded system is that the tests
measure fundamental material characteristics. This implies the test methods are
controlled by mechanistic theories. However, the testing methodologies were developed
under several research contracts within the Strategic Highway Research Program.
Therefore, part of the objective of this research was to consolidate and document the
underlying theories of the rotational viscometer, the dynamic shear rheometer, and the
bending beam rheometer.
1.4 SCOPE AND LIMITATIONS
This research was limited to an evaluation of the asphalt binder conditioning and
testing equipment available in the Asphalt Technology Laboratory at West Virginia
University. The equipment required for the performance grading of asphalt binder is
complete except for a direct tension tester, DTT. The final specifications for the direct
tension tester were not yet available at the time the West Virginia University Asphalt
Technology Laboratory was purchasing the other SHRP testing units. It was not
considered prudent to purchase the direct tension tester when the laboratory was
established. Therefore, the DTT equipment was not evaluated as part of this research.
This research was limited to evaluating existing equipment and test methods.
There was no effort to develop new equipment or test procedures. The AASHTO
Standard Specifications for Transportation Materials and Methods of Sampling and
Testing were the sole testing specification used. The AASHTO MP1-97 standard
specification was used to evaluate the test results of each unit. The dated AASHTO
standard specifications used to conduct the conditioning and testing procedures are as
follows:
Rolling Thin Film Oven - AASHTO T240-97
Pressure Aging Vessel - AASHTO PP1-97

5
Rotational Viscometer - AASHTO TP48-97
Dynamic Shear Rheometer - AASHTO TP5-98
Bending Beam Rheometer - AASHTO TP1-98
The testing program was conducted using a single operator. All tests were
performed with a single grade of asphalt binder, PG 70-22. This is the asphalt grade
predominately used for Superpave mixes in West Virginia. The Marathon Ashland
Petroleum LLC of Findlay, Ohio was the sole source for the asphalt binder.
Silicone molds were used to prepare the asphalt test specimens for all DSR tests.
Aluminum molds were used to prepare the asphalt beam specimens for all the BBR tests.
Concern was expressed in reports from the asphalt industry that the BBR silicone molds
were producing irregular shaped asphalt beam specimens, thus resulting in the poor test
results. Therefore, the decision was made to use the BBR aluminum molds.
1.5 ORGANIZATION OF THE REPORT
This report is organized into five chapters. Following this introductory chapter is
a background literature review. The literature review of Chapter 2 provides background
information for each performance grading test and related specifications. The research
approach is presented in Chapter 3. The chapter includes a detailed outline of the testing
program and methods. The analysis of the test results and the evaluation of the
equipment performance and testing protocol are provided in Chapter 4. Concluding
remarks and recommendations on the research are provided in Chapter 5.
6
CHAPTER 2
LITERATURE REVIEW AND BACKGROUND
2.1 INTRODUCTION
Asphalt binders play a critical role in the quality and performance of HMA
pavements. Therefore, pavement engineers must have a thorough understanding of
asphalt binder behavior. Proven testing equipment and test specifications are necessary
to accurately evaluate and select the appropriate asphalt binder for a pavement structure.
This chapter provides a summary of key asphalt rheology concepts that

characterize the behavior of asphalt binders. A brief discussion of the penetration
grading system and the viscosity grading specifications is presented fo llowed by an
overview of the development of the performance grading binder tests and the notable
features of the test specifications. The main focus of this chapter is the detailed
description of each PG binder test. The description includes the purpose and scope of
each test, a summary of method, equipment hardware, test sample and equipment
preparation, test procedure summary, unit calibration, and theory of analysis discussion.
Some of background information documented in this chapter was developed while setting
up and using the equipment and test procedures at the West Virginia University Asphalt
Technology Laboratory.
2.2 ASPHALT RHEOLOGY
Asphalt binders deform when subjected to loads. The properties of asphalt also
change with varying temperatures. The deformation is a combination of elastic response
and viscous flow. The magnitude of deformation, or mechanical response, is dependent
on load magnitude, duration, and rate of application and the temperature state of the
material. Since asphalt binders display both elastic and viscous response properties, they
are classified as viscoelastic materials. Anderson, et al. (1994) present a review of
asphalt rheology, which served as the basis for the following discussion.
The typical elastic, viscous, and viscoelastic responses to an applied stress is

shown in Figure 2.1. An elastic material experiences recoverable deformation when
7
subjected to a constant (or creep) load, Figure 2.1a. An elastic material will immediately
deform and maintain a constant strain when loaded, Figure 2.1b. The material will
immediately return to its initial shape when the creep load is removed. A viscous
Newtonian material, when subjected to a constant load, will deform at a constant rate
until the load is removed, Figure 2.1c. The deformation of the viscous material, however,
will remain after the load is removed; hence, a viscous material experiences non-
recoverable deformation.
(a) Applied
Stress
Time
(b) Elastic
Response
Time
(c) Viscous
Response
Time
(d) Viscoelastic
Response
Time
Figure 2.1 Mechanical Response of Elastic, Viscous, and Viscoelastic Materials
A viscoelastic material, when subjected to a creep load, experiences an immediate

deformation followed by a continued time-dependent deformation, as shown in Figure
2.1d. The immediate deformation corresponds to the material’s elastic response and the
time-dependent deformation corresponds to the material’s viscous response. Once the
8
load is removed, the viscous deformation component immediately ceases, but this
deformation is not recovered. The delayed elastic deformation component is slowly
recovered at a decreasing rate. Thus, a viscoelastic material experiences only a partial
recovery of the deformation resulting from creep loading. The viscoelastic behavior of
asphalt can be characterized by its deformation resistance and the relative distribution of
that resistance between the elastic component and the viscous component within the
linear range. The relative distribution of the resistance between the elastic component
and the viscous component is dependent on the asphalt cement characteristics and
temperature and loading rate.
The previous loading-response descriptions are for responses within the linear
range, which is characterized by the deformation being directly proportional to the
applied load at any time and temperature. Nonlinear loading responses are difficult to
model for viscoelastic materials such as asphalt. Linear response models, however, are
sufficient for the engineering analysis of asphalt binder response to the loading conditions
and environmental stresses encountered in the field.
Roberts, et al. (1996) and Bahia and Anderson (1995c) describe the mechanisms
that change the rheological properties of asphalt over time. The most familiar of these is
age hardening, also referred to as oxidative aging. Asphalt binders consist of
hydrocarbons that tend to oxidize when exposed to oxygen. The binder also undergoes
volatilization when exposed to the environment. Oxidation and volatilization increases
the binder’s stiffness (hardness) and viscosity. The time of exposure to the environment
(i.e. aging), as well as the service temperatures, directly affect the rate and extent of age
hardening.
The greatest extent of age hardening occurs during the production of hot mix
asphalt. During the mixing process, the thin film of asphalt binder covering the aggregate
is subjected to air temperatures up to 163 o C. The extreme production environment
accelerates asphalt binder oxidation and volatilization. Age hardening continues, at a
slower rate, during the transportation to the project site and during construction.
Thereafter, the rate of age hardening significantly reduces such that extended periods are
required to identify changes in asphalt rheological properties.
9
2.3 PENETRATION GRADING SYSTEM SPECIFICATIONS
There are five standard penetration grades of asphalt cement used for paving
operations, as per ASTM D946. These grades include 40-50, 60-70, 85-100, 120-150,
and 200-300. The grades are based on penetration units, where one penetration unit
equals one tenth of a millimeter penetration. The lower the penetration grade, the
“harder” the asphalt. Therefore, 40-50 grade is the hardest asphalt and 200-300 is the
softest. The grades 60-70 and 85-100 are commonly used in the United States. The
penetration graded binder specifications are provided in Table 2.1.
10
Table 2.1 ASTM D946 Requirements for Penetration Graded Asphalt Cements
Penetration Grade
40 50 60 70 85 100 120 150 200 300
Mi Ma Mi Ma Mi Ma Mi Ma Min Ma
n x n x n x n x x
Penetration at 77°F 40 50 60 70 85 100 120 150 200 300
(25°C),
100g, 5s
Flash point, °F 450 - 450 - 450 - 425 - 350 -
Cleveland open
cup)
Ductility* at 77°F 100 - 100 - 100 - 100 - 100 -
(25°C), 5 cm/min,
cm
Solubility in 99. - 99. - 99. - 99. - 99. -
trichloroethylene, 0 0 0 0 0
%
Retrained 55 - 52 - 47 - 42 - 37+ -
penetration after + + + +
thin- film oven test,
%
Dectility at 77°F - - 50 - 75 - 100 - 100 -
(25°C) +
5 cm/min, after
thin- film
oven test, cm
* If ductility at 77F (25C) is less than 100 cm, material will be accepted if
ductility at 60F (15.5C)
is 100 cm minimum at a pull rate of 5 cm/min.
2.4 VISCOSITY GRADING SYSTEM SPECIFICATIONS
There are six AC viscosity grades for asphalt binders specified in ASTM D3381:
AC-2.5, AC-5, AC-10, AC-20, AC-30, and AC-40. Poise is the standard unit of viscosity
measurement used by the viscosity grading system. Asphalt graded as AC-2.5 has a
viscosity of 250 poises at 60 o C. The lower the asphalt grade, the “softer” the asphalt.
Thus, AC-20 is “softer” than AC-30. There are five AR viscosity grades for aspha lt
11
cement: AR-1000, AR-2000, AR-4000, AR-8000, and AR-16000. An AR-1000 grade

represents an RTFO-aged asphalt with a viscosity of 1000 poises at 60 o C.
The ASTM viscosity graded binder specifications are provided in Table 2.2. As
presented in ASTM D3381, Table 2.2 actually includes three tables of requirements:
Tables 1 and 2 for AC grades and Table 3 for AR grades. Agencies specifying
AC graded asphalt cements could implement either Table 1 or 2 into their specifications.
12
Table 2.2 ASTM D3381 Requirements for Viscosity Graded Asphalt Cements
o o
Table 1 Requirements for Asphalt Cement, Viscosity Graded at 140 F (60 C)
Note - Grading based on original asphalt
Viscosity Grade
Test
AC-2.5 AC-5 AC-10 AC-20 AC-40
o o
Viscosity, 140 F (60 C), P 250 ± 50 500 ± 100 1000 ± 200 2000 ± 400 4000 ± 800
o o
Viscosity, 275 F (135 C), min, cSt 80 110 150 210 300
o o
Penetration, 77 F (25 C), 100g, 5 s, min 200 120 70 40 20
o o
Flash point, Cleveland open cup, min, F ( C) 325 (163) 350 (177) 425 (219) 450 (232) 450 (232)
Solubility in trichloroethylene, min, % 99.0 99.0 99.0 99.0 99.0
Tests on residue from thin-film oven test:
o o
Viscosity, 140 F (60 C), max, P 1250 2500 5000 10 000 20 000
o o A
Ductility, 77 F (25 C), 5 cm/min, min, cm 100 100 50 20 10
A o o
If ductility is less than 100, material will be accepted if ductility at 60 F (15.5 C) is 100 minimum at a pull rate of 5 cm/min.
o o
Note - Grading based on original asphalt
Viscosity Grade
Test
AC-2.5 AC-5 AC-10 AC-20 AC-30 AC-40
o o
Viscosity, 140 F (60 C), P 250 ± 50 500 ± 100 1000 ± 200 2000 ± 400 3000 ± 600 4000 ± 800
o o
Viscosity, 275 F (135 C), min, cSt 125 175 250 300 350 400
o o
Penetration, 77 F (25 C), 100g, 5 s, min 220 140 80 60 50 40
o o
Flash point, Cleveland open cup, min, F ( C) 325 (163) 350 (177) 425 (219) 450 (232) 450 (232) 450 (232)
Solubility in trichloroethylene, min, % 99.0 99.0 99.0 99.0 99.0 99.0
Tests on residue from thin-film oven test:
o o
Viscosity, 140 F (60 C), max, P 1250 2500 5000 10 000 15 000 20 000
o o A
Ductility, 77 F (25 C), 5 cm/min, min, cm 100 100 75 50 40 25
A o o
o o
Note - Grading based on residue from rolling thin-film oven test.
A Viscosity Grade
Test on Residue from Rolling Thin-Film Oven Test:
AR-1000 AR-2000 AR-4000 AR-8000 AR-16000
o o
Viscosity, 140 F (60 C), P 1000 ± 250 2000 ± 500 4000 ± 1000 8000 ± 2000 16000 ± 4000
o o
Viscosity, 275 F (135 C), min, cSt 140 200 275 400 550
o o
Penetration, 77 F (25 C), 100g, 5 s, min 65 40 25 20 20
o o
% of original penetration, 77 F (25 C), min 40 45 50 52
o o B B
Ductility, 77 F (25 C), 5 cm/min, min, cm 100 100 75 75 75
Tests on original asphalt:
o o
Flash point, Cleveland open cup, min, F ( C) 400 (205) 425 (219) 440 (227) 450 (232) 460 (238)
Solubility in trichloroethylene, min, % 99.0 99.0 99.0 99.0 99.0
A
Thin-film oven may be used but the rolling thin-film oven test shall be the referee method.
B o o
13
2.5 PERFORMANCE GRADED BINDER SPECIFICATIONS
The Performance Grading specifications and associated laboratory procedures

were designed to relate the asphalt binder rheological characteristics to pavement
performance. The Performance Graded binder tests and specifications possess the
following notable features (Roberts, et al., 1996):
Engineering principles are used to directly relate measured physical properties to

field performance.
Asphalt is evaluated at three critical stages of asphalt life via special conditioning
procedures. The first stage represents the transporting and handling of original asphalt
binder before mixing. The second stage represents the short-term aging following HMA
production and construction. The third stage considers the continued long-term aging
during the pavement service life.
The physical properties remain constant for all PG grades, but the temperature at
which the properties must be achieved changes in consideration of the asphalt grade
required for the project site climate, traffic loading, and speeds.
The complete range of service temperatures at the project site is considered.
Tests and specifications are established to reduce high temperature rutting,

intermediate temperature fatigue cracking, and low temperature thermal cracking.
Tests and specifications are established for modified asphalt binders.
The specifications maintain the solubility criteria.
A summary of the Performance Graded asphalt binder grades and specifications

from AASHTO MP1 are provided in Table 2.3 and Table 2.4, respectively. Each grade is
represented by two numeric values representing the temperature rating of the binder in
degrees Celsius. For example, a PG 70-22 is rated for a climate having an average 7-day
maximum design temperature of 70 o C and a minimum pavement design temperature of
-22 oC. The upper temperature rating is based upon the highest 7-day running average
temperature of the pavement, 5 cm below the pavement surface. The lower temperature
14
Table 2.3 Performance Graded Asphalt Cement Grades
High Temp.1 °C Low Temp.2 °C

PG 46 -34, -40, -46
PG 52 -10, -16, -22, -28, -34, -40, -46
PG 58 -16, -22, -28, -34, -40,
PG 64 -10, -16, -22, -28, -34, -40,
PG 70 -10, -16, -22, -28, -34, -40,
PG 76 -10, -16, -22, -28, -34, -40,
PG 82 -10, -16, -22, -28, -34, -40,
Notes:
1
Average 7-day maximum pavement design temperature
2
Minimum pavement design temperature
Table 2.4 Summary of AASHTO MP1 Requirements
Test Performance Asphalt Binder Specification Specification Test Temp.1

Parameter Condition State Limit °C
RV Flow ability Neat Viscosity 3 Pa.s 135
DSR Rutting Neat G*/sin δ @ 10 1.0 kPa (Min) High
resistance rad/sec
DSR Rutting RTFO-aged G*/sin δ @ 10 2.2 kPa (Min) High
resistance rad/sec
DSR Fatigue cracking PAV aged2 G*sin δ @ 10 5000 kPa Intermediat
resistance rad/sec (Max) e
BBR Thermal PAV aged2 Creep Stiffness, S 300 Mpa Low+10°C
cracking @ 60 sec.
resistance m-value @ 60 0.300 Low+10°C
secv.
DTT Thermal PAV aged2 Failure Strain @ 1.0% Low+10°C
cracking 1.0mm/min
resistance
Additional Specifications
Flash point Temperature = 230°C (Min)
RTFO Mass Loss = 1.0% (Max)
Notes:
1
High temperature - see Table 2.3
Low temperature - see Table 2.3
Intermediate temperature = (High temp. + Low temp.)/2 + 4°C
2
Pressure Aging Vessel Conditioning Temperatures per PG Grade:
PG46 90°C PG70 - 100°C (110°C for desert climates)
PG52 90°C PG76 - 100°C (110°C for desert climates)
PG58 100°C PG82- 100°C (110°C for desert climates)
PG64 100°C
15
rating is based on the coldest estimated pavement temperature 5 cm below the pavement
surface.
The rotational viscometer and dynamic shear rheometer have been used for
rheological evaluation in many industries for years (Petersen, et al., 1994a). The rolling
thin film oven, which is an improved variation of the thin film oven used in the viscosity
grading method, was developed prior to the SHRP study. A comprehensive study of the
existing rolling thin film oven was considered as part of the SHRP project. However,
since the evaluation of the rolling thin film oven would have consumed too much of the
project resources and no long-term field aging specifications existed, the SHRP
researchers never initiated the study.
Instead of evaluating the rolling thin film oven, focus was placed on developing a
long-term aging procedure. The pressure aging methods used for years in the asphalt
research field and the rubber products industries were evaluated for use with the PG
binder specifications. The study led to the adoption of the pressure aging vessel test, a
modified form of the pressure aging methods from other industries.
Several attempts have been made in the past, with very little success, to develop
an inexpensive device for evaluating the low temperature performance of asphalt binders
(Bahia and Anderson, 1995b). Prior to the SHRP research, Pennsylvania State University
introduced a bending beam rheometer for evaluating low temperature performance. The
bending beam rheometer was refined during the SHRP research project and was
incorporated into the PG binder specification.
No test methods existed for characterizing the fatigue or the fracture properties of
asphalt binders prior to the SHRP research (Petersen, et al., 1994b). The direct tension
test was therefore developed to evaluate the tensile failure properties of asphalt binders at
low temperatures. The final test specifications for the direct tension tester were not yet
available at the time the West Virginia University Asphalt Technology Laboratory was
purchasing the other SHRP performance graded testing units.
16
2.6 ROLLING THIN FILM OVEN
Introduction
Purpose and Scope
Significant age hardening occurs during the production and construction of hot
mix asphalt (Roberts, et al., 1996). The rolling thin film oven test (RTFO) is a
conditioning procedure that simulates the age hardening asphalt undergoes during the
production and construction of HMA. The conditioning procedure is used to determine
the effect of he at and air on a moving film of asphalt and to evaluate the resistance to
aging during the production and construction of hot mix asphalt structures.
Summary of Method
The AASHTO T240 standard test method requires that the RTFO continuously
expose aspha lt specimens to both heat (163 o C) and airflow (4000 ml/min) for 85 minutes
to achieve the accelerated age hardening of the asphalt. Eight asphalt specimens (35
grams each) can be conditioned during each RTFO procedure. The RTFO conditions the
asphalt specimen for further conditioning and physical testing. The PG specifications of
AASHTO MP1 limit the mass loss due to RTFO conditioning to one percent or less.
Equipment
Hardware
The Despatch RTFO evaluated during this research is shown in Figure 2.2. A
RTFO consists of an oven chamber that houses a vertical circular carriage. The carriage,
which holds eight RTFO specimen bottles, rotates about its center. A single air jet is
located in the oven. Hot air is blown into the center of each RTFO bottle as it passes in
front of the jet. A fan continually circulates the air within the oven chamber. The system
requires a clean, dry, and pressurized air supply.
17
Figure 2.2 Despatch Rolling Thin Film Oven
The AASHTO T240 specifications require that the RTFO specimen bottles are
139.7 mm long and 64.0 mm in outside diameter. The opening at the end of each bottle
is 31.8 mm in diameter. The bottles are made of heat resistant glass with a 2.4 mm wall
thickness.
Data Collection
No data is collected during the RTFO conditioning procedure. The average
asphalt mass loss, however, is recorded at the conclusion of the procedure. The average
asphalt mass loss is determined by weighing the asphalt samples from two RTFO bottles
before and after the conditioning procedure and computing the difference as follows:
Initial Mass − Final Mass

Percent change in mass = 100 (2-1)
Initial Mass
Testing Protocol
Test Specification
AASHTO T240-98 Effect of Heat and Air on Rolling Film of Asphalt defines the
procedures followed during RTFO conditioning.
18
Test Sample and Equipment Preparation Summary
The RTFO oven is preheated to the 163 o C test temperature 16 hours prior to the
test. The carriage rotation is set at 15 rpm and the air jet flowrate is set at 4000 ml/min.
At least 350 grams of neat asphalt is heated until fluid to pour. Upon reaching the desired
consistency, 35 grams of the heated asphalt is poured into each RTFO specimen bottle
and then allowed to cool to room temperature. If change in mass determination is
desired, two RTFO bottles first must be weighed empty. The bottles are then filled with a
35 grams asphalt sample and placed in a desiccator while cooling to room temperature.
The filled RTFO bottles are weighed again upon cooling.
Test Procedure Summary
Once the 16- hour preheating requirement is complete and the asphalt samples
have cooled to room temperature, the RTFO bottles are placed in the RTFO carriage and
the test is initiated. The sample bottles remain in the oven for 85 minutes. The rotating
carriage continuously exposes fresh asphalt to the heat and circulating air. The hot air is
blown into the bottle with each pass in front of the air jet. The constant rolling action of
the binder in the sample bottles prohibits the formation of an age inhibiting skin, unlike
the previous thin film oven tests (Roberts, et al., 1996). The rolling action also benefits
in the conditioning of modified asphalt since the modifiers remain dispersed in the
binder.
At the conclusion of the 85- minute conditioning procedure, the two specimen
bottles designated for mass loss determination are removed from the RTFO and
immediately placed in a desiccator while they cool to room temperature. The six
remaining specimen bottles are removed from the RTFO and their asphalt residue is
collected in a container. The AASHTO procedure requires that the final RTFO bottle be
removed from the RTFO within five minutes of the removal of the first bottle. Once the
asphalt sample from the final bottle is collected in the container, the container is covered
and stored for further conditioning and/or testing.
The two bottles designated for mass loss determination are weighed once they
have cooled to room temperature and the average mass loss from the two bottles is
calculated. The asphalt from these bottles is then discarded.
19
Unit Calibration
The rolling thin film oven requires a periodic verification of the RTFO flow
meter. The required 4000 ml/min airflow is verified by using wet-test meter or other
displacement methods. The oven temperature controller must be periodically verified.
The oven temperature RTFO is measured by a proportional control thermostat-sensing
element. An ASTM Loss on Heat Thermometer 13C is located within the oven as well.
The thermostat-sensing element is initially calibrated via a temperature offset to the
ASTM thermometer. The temperature offset is programmed into the temperature
controller. Agreement between the ASTM thermometer temperature and the temperature
controller display should be verified prior to each RTFO conditioning process.
2.7 PRESSURE AGING VESSEL
Introduction
Purpose and Scope
The pressure aging vessel (PAV) simulates the age hardening of asphalt during
the first 5-10 years of pavement service life (Roberts, et al., 1996). The pressure aging
vessel conditioning allows further testing by the dynamic shear rheometer and bending
beam rheometer to evaluate the binder’s performance following aging.
Summary of Method
While the Pressure Aging Vessel procedure is often referred to as a test, it is only
a conditioning procedure. The AASHTO PP1 standard test method requires that the PAV
expose RTFO-aged residue to high pressure, 2.1 MPa, and high temperature over a 20-
hour period to achieve accelerated hardening of the asphalt. The conditioning
temperature, which is dependant upon the asphalt grade being conditioned, ranges from
90 o C to 110 o C. Ten asphalt specimens, 50 grams each, can be conditioned with each
PAV run. The residue from the PAV is degassed in a vacuum oven immediately
following conditioning.
20
Equipment
Hardware
The primary components of a pressure aging vessel include a pressure vessel and
a heating unit. Two different types of pressure aging vessel units are specified in the
AASHTO test method. One type of pressure aging vessel unit is an integral system in
which the pressure vessel is permanently enclosed in an oven. The other available type is
set up such that the pressure vessel is a separate unit placed in a forced draft oven. Both
types of units are equipped with a pressure regulator and transducer, temperature
controller, thermocouple or resistance thermal detector (RTD) and a data acquisition
system. The ATS pressure aging vessel unit evaluated by this research is shown in
Figure 2.3. The pressure vessel houses a sample rack that holds 10 asphalt specimen
pans. The pressure aging vessel unit requires a separate compressed air cylinder to
supply clean, dry air to the pressure vessel. A vacuum oven is also required for degassing
the PAV-aged residue following the pressure aging vessel conditioning procedure.
Figure 2.3 ATS Pressure Aging Vessel

21
Data Collection
The pressure aging vessel conditioning procedure requires no data collection.

AASHTO PP1 does require the monitoring of the pressure and temperature, however, at
regular intervals throughout the duration of the conditioning procedure. The aging
temperature must be reached within 2 hours of the initiation of the procedure, or the
asphalt samples must be discarded. The pressure must remain at 2.1 ±0.1 MPa and the
temperature within ±0.5 o C of the test temperature during the conditioning procedure.
The pressure must be reduced to atmospheric pressure within 9±1 minutes at the
comp letion of the 20-hour conditioning procedure. The data acquisition system
automatically monitors the pressure and temperature. The pressure and temperature data
log can be reviewed on the ATS 504D pressure aging vessel; however, a hardcopy of the
data cannot be printed.
Testing Protocol
Test Specification
AASHTO PP1-98 Standard Practice for Accelerated Aging of Asphalt Binder

Using a Pressurized Aging Vessel (PAV) defines the protocol required for PAV
conditioning of asphalt cement.
The pressure aging vessel unit is preheated to the required conditioning

temperature. The test temperatures specified in AASHTO MP1 include 90 o C, 100 o C, or
110 o C, depending on the asphalt grade. Asphalt grades PG 46 and PG 52 are
conditioned at 90 o C. Asphalt grades PG 58, PG 64, PG 70, PG 76, and PG 82 are
conditioned at 100 o C. Asphalt grades PG 70, PG 76, and PG 82 are conditioned at
110 o C for desert climates (see Table 2.4). The pressure controller is preset to the
2.1 MPa test pressure. A covered container of RTFO-aged residue is heated in an oven
until fluid to pour. A 50-g sample of the heated RTFO residue is poured into each
specimen pan.
22
The PAV specimen pans are loaded in the sample rack and placed into the
pressure vessel/oven unit. The conditioning test is initiated and the pressure vessel
automatically pressurizes. The 20-hour conditioning period begins once the temperature
stabilizes following the vessel pressurization. The test temp erature and pressure are
maintained throughout the 20-hour conditioning process. The pressure vessel
automatically depressurizes at the conclusion of the procedure.
The specimen pans are removed from the pressure vessel and placed in a
preheated to 163 o C oven for 15 minutes. The specimen pans are removed from the oven
and the PAV-aged residue is collected in a container. The container of PAV-aged asphalt
is immediately placed in a vacuum oven, preheated to 170 o C, without the vacuum
applied. Following a 10- minute equilibration period, the vacuum valve is opened to
reduce the oven pressure to 15 kPa absolute. The vacuum is maintained for 30 minutes.
At the end of 30 minutes, the vacuum is released and the container of PAV-aged asphalt
is removed. The container is inspected for air bubbles on the surface of the asphalt
residue. Any air bubbles present are removed with a hot knife or flashed with a torch.
The container is then covered and the sample is stored for subsequent physical testing.
Unit Calibration
The pressure aging vessel must have the resistance thermal detector (RTD) within
the pressure vessel and the pressure gauge calibrated every six months. These calibration
procedures are typically performed by a commercial calibration service.
2.8 ROTATIONAL VISCOMETER
Introduction
Purpose and Scope
Asphalt binders must remain sufficiently fluid, or workable, at the high

temperatures necessary during the plant mixing, field placement and compaction of hot
mix asphalt (Roberts, et al., 1996). The binder reaches temperatures exceeding 135 o C
during these procedures. The rotational viscometer (RV) is used to evaluate binder in an
unconditioned or neat state. This represents the state of material in the tank at the asphalt
23
plant. The rotational viscometer measures the rheological properties of asphalt binders to
evaluate their pumpability during delivery and plant operations.
The rotational viscometer is also used to establish equiviscous temperature ranges

for selecting HMA mixing and compaction temperatures (Asphalt Institute, 2001).
Mixing and compaction temperatures are determined from viscosity-temperature graphs.
A typical log-log viscosity vs. log temperature graph is shown in Figure 2.4 (Asphalt
Institute, 2003). Viscosities at two temperatures are plotted on the log- log viscosity vs.
log temperature graph. The equiviscous temperature range for mixing and compaction
correspond to where the plotted line passes through the respective viscosity range.
Summary of Method
The rotational viscometer determines the asphalt viscosity by measuring the

torque necessary to maintain a constant rotational speed of a cylindrical spindle
submerged in an asphalt specimen held at a constant temperature, as per the AASHTO
TP48 standard test method. Unlike the capillary viscometers used with the viscosity-
graded method, the rotational viscometer can evaluate modified asphalt binders
(Anderson, et al., 1994). The viscosity of asphalt binders can be measured within the
range of 0.01 Pa·s (0.1 poise) to 200 Pa·s (2000 poise) (Petersen, et al., 1994b).
Temperature,
Figure 2.4 Typical Equiviscosity Curve for Asphalt Binder

24
Equipment
Hardware
The major components of a rotational viscometer system consist of a rotational

viscometer, temperature controller, thermo-chamber and a personal computer (not
required with some rotational viscometers). The rotational viscometer is a rotating
spindle-type viscometer. It is comprised of a rheometer head and base unit. The
rheometer head unit consists of a motor, rotational transducer, spindle, digital readout
display and a control panel. The base unit acts as interface between the rheometer head
and the personal computer. The thermo-chamber is a heating chamber that houses the
asphalt sample holder during the test. The temperature within the thermo-chamber is
monitored by a resistance thermal detector (RTD) and controlled by the temperature
controller unit. The Brookfield DV-III rotational viscometer evaluated during this
research is shown in Figure 2.5.
Data Collection
A rotational transducer within the rheometer head measures the torque required to
turn the spindle. The electronic circuitry within the base unit automatically converts the
measured torque to viscosity and sends the test results to the computer. Viscosity
readings are taken at one- minute intervals for three readings.
Testing Protocol
Test Specification
AASHTO TP48-97 Standard Test Method for Viscosity Determination of Asphalt

Binder Using Rotational Viscometer defines the protocol for measuring the viscosity of
asphalt binders at high temperatures. This test method does not cover procedures for
measuring the temperature-viscosity relationship for asphalt binders.
The test operating speed of the rotational viscometer is set. The temperature
controller of the thermo-chamber is set at the test temperature and allowed to preheat.
The sample holder and spindle are preheated in an oven heated to the test temperature as
well. A test temperature of 135 o C is specified for evaluating pumpability in AASHTO
25
Figure 2.5 Brookfield DV-III Rotational Viscometer
MP1. Rotational viscosity tests at two temperatures are needed when creating
viscosity-temperature graphs for establishing equiviscous temperature ranges. The
Asphalt Institute recommends taking the first viscosity measurement at 135 o C, and the
second at 165 o C (Asphalt Institute, 2001). The lower temperature measurement is taken
first as it is much quicker to raise the temperature in the thermo-chamber than to cool the
sample for the second viscosity measurement. Thus, prior to conducting tests for
equiviscous temperature ranges, the thermo-chamber should be preheated at the lower
temperature. Likewise, the sample holder and spindle are preheated in an oven heated to
the lower temperature.
A container of neat asphalt is placed in the oven and heated until fluid to pour.
The heated asphalt is poured into the sample holder to a level that will just cover the
upper conical portion of the spindle when it is lowered into the sample holder.
26
The sample holder is placed in the preheated thermo-chamber. The spindle is

removed from the oven, attached to the viscometer, and lowered into the asphalt sample.
The system is allowed to equilibrate for 30 minutes and the test is initiated.
Three viscosity measurements are taken at each test temperature. Thus, when
evaluating asphalt pumpability, three viscosity measurements are taken at 135 o C. When
establishing equiviscous temperature ranges, three measurements are taken at the lower
temperature and at the higher temperature.
If the displayed torque is out of the allowable viscosity range for the selected
spindle and operating speed, the spindle or speed is changed, as per the manufacturer’s
recommendations, and the test is restarted with a new asphalt specimen. The test is
terminated at the conclusion of the viscosity readings. The sample holder is removed
from the thermo-chamber and the asphalt specimen is discarded.
Unit Calibration
The test method requires the verification of the RV temperature controller. An

asphalt sample is placed in the testing chamber and heated to the test temperature. The
indicated temperature of the sample is verified via an ASTM thermometer. In addition to
the temperature controller verification, the accuracy of the rotational transducer is
checked by performing a viscosity test on a standard viscosity reference fluid.
Theory of Analysis
Most asphalt binders behave like Newtonian fluids when heated to temperatures
exceeding 100 o C (Bahia and Anderson, 1995a). A Newtonian fluid is characterized by a
totally viscous response, in which the viscosity is independent of the rate of shear strain.
Since asphalt has a totally viscous response at high temperatures, it is sufficient to
characterize the workability of the asphalt during mixing and construction with a measure
of its viscosity.
The working equations relating shear stress to torque and shear rate to angular
velocity are as follows (Petersen, et al., 1994b and Whorlow, 1992):
Shear Rate (s-1 )

27
 2 Rc2 
γ& = ω 
( 
 2Rc − R s 
2 2
) (2-2)
Shear Stress (dynes/cm2 )
M
τ = (2-3)
2πRs2 L
where:
ω = angular velocity of spindle (rad/s)
Rc = radius of sample holder (cm)
Rs = radius of spindle (cm)
M = applied torque (dyne-cm)
L = effective length of spindle (cm)
The viscosity is calculated by the following relationship:
τ
µ= (2-4)
γ&
The unit of viscosity may be converted to pascal-seconds (Pa·s) using the

conversion 1 centipoise = 0.001 Pa·s.
Parameter Specifications
AASHTO MP1 requires a maximum viscosity limit of 3 Pa·s when tested at

135 o C. The maximum limit ensures that the asphalt binder is sufficiently fluid for
pumping during delivery and plant operations.
New mixing and compaction viscosity specifications were not developed during
the SHRP research (Asphalt Institute, 2001). The industry standard for the past twenty
years has been the viscosity ranges of 170±20 centistokes for mixing temperatures and
280±30 centistokes for compaction temperatures. The same viscosity ranges are still
recommended for Superpave; however, the units have been converted the Pascal-second
metric equivalent; 0.17 ± 0.02 Pa·s for mixing and 0.28 ± 0.02 Pa·s for compaction.
28
2.9 DYNAMIC SHEAR RHEOMETER
Introduction
Purpose and Scope
Roberts, et al. (1996) and Bahia and Anderson (1995a) present a description of
the purpose and scope of the dynamic shear rheometer test, which served as the basis for
the following section. Asphalt concrete pavements can be prone to wheel path rutting in
the early stages of their life. Repeated traffic loads increase the consolidation of the
pavement structure after construction, thus causing depressions in the traffic wheel paths.
Traffic induced rutting can also result from the lateral flow of the asphalt concrete
materials in the wheel paths. Although the rutting potential of asphalt concrete
pavements is influenced primarily by the quality of construction (compaction), mix
design, and aggregate angularity and texture, the consistency of the binder must still be
considered. Rutting is more prevalent at high service temperatures due to the increased
fluidity of the asphalt. The rutting potential of HMA pavements decreases with time
since the asphalt binder stiffens with age hardening.
Asphalt concrete pavements also develop the potential for fatigue cracking later in
their service life. Repeated load-associated stresses are the primary cause of fatigue
cracking. Asphalt binder properties play a major role in the fatigue life of thin HMA
pavements. Excessive asphalt binder hardening can significantly decrease the fatigue life
of thin HMA pavements.
The dynamic shear rheometer (DSR) was adopted to characterize the viscoelastic
behavior of asphalt binders at intermediate and high service temperatures. The DSR
provides an indication of the rutting resistance of asphalt cement immediately following
construction. Resistance to rutting at high service temperatures in the early stages of
pavement life is also evaluated. The DSR also provides an indication of the resistance to
fatigue cracking at immediate service temperatures in the later stages of service life.
Summary of Method
The viscoelastic properties of asphalt are determined by evaluating the behavior

of an asphalt specimen when subjected to oscillatory (sinusoidal) stresses. The AASHTO
29
TP5 standard test method requires that a thin asphalt specimen be sandwiched between
two parallel metal plates held in a constant temperature medium. One plate remains fixed
while the other oscillates, at an angular frequency of 10 radians per second for 10 cycles,
with respect to the other. The dynamic shear rheometer evaluates the specimen’s
response to the sinusoidal stresses and calculates the asphalt’s complex shear modulus
and phase angle. The complex shear modulus and phase angle of a binder, which are
indicators of an asphalt’s resistance to shear deformation in the viscoelastic region, help
predict the rutting potential and fatigue life of hot mix asphalt pavements (Bahia and
Anderson, 1995a).
Equipment
Hardware
A dynamic shear rheometer system includes the rheometer, temperature

controller, data acquisition unit and a personal computer. The rheometer itself consists of
a loading assembly, motor, load and strain transducer, parallel plates (an upper oscillatory
spindle plate and a lower fixed plate) and an environmental control chamber with a
resistance thermal detector (RTD) mounted inside. The Bohlin Instruments DSR II
dynamic shear rheometer evaluated during this research is shown in Figure 2.6.
The There are two types of dynamic shear rheometers: controlled stress and
controlled strain. Controlled stress rheometers apply a sinusoidally varying stress and
determine the resulting strain. Controlled strain rheometers apply a sinusoidally varying
strain and determine the resulting stress.
DSR uses two different sized parallel plate sets for evaluating asphalt binder: 8-
mm diameter plates and 25- mm diameter plates. The 25- mm plates are used with RTFO-
aged asphalt specimens when evaluating the rutting resistance parameter. The 8- mm
plates are used with PAV-aged asphalt specimens when evaluating the fatigue cracking
resistance parameter. Silicone molds are available for making asphalt test specimens for
use with both the 25- mm plates and the 8- mm plates.
30
Figure 2.6 Bohlin Instruments DSR II Mechanical Dynamic Shear Rheometer
The environmental chamber houses the parallel plates, a platinum resistance

thermometer, the test specimen and the medium used for heating or cooling the asphalt
specimen to a constant test temperature. Two types of environmental chambers are
available for controlling the asphalt specimen temperature. One type uses a circulating
gas medium (nitrogen or dried air) to surround the test specimen. The other type of
environmental chamber uses a water bath to surround the test specimen. A circulating
bath unit is required if a fluid medium is used.
Data Collection
The data acquisition unit records the test temperature, applied torque, loading
frequency and deflection angle during the 10 cycles of the test and sends the test data to
the personal computer. The computer software reduces the data and calculates the shear
stress, shear strain, complex modulus and phase angle.
31
Testing Protocol
Test Specification
AASHTO TP5-98 Standard Test Method for Determining the Rheological

Properties of Asphalt Binder Using a Dynamic Shear Rheometer (DSR) defines the
protocol for conducting the dynamic shear rheometer test.
Test Sample and Equipment Preparation
The system software is initialized and the test parameters entered. The test
temperature is set, as per AASHTO MP1 criteria. The test temperature selected depends
on the asphalt type and the performance parameter. For rutting resistance evaluation,
neat and RTFO-aged asphalt binders are tested at high service temperatures. Test
temperatures range, depending on the grade of asphalt, from 46 o C to 82 o C. PAV-aged
asphalt is tested at intermediate service temperatures for fatigue life evaluation. Test
temperatures range from 4 o C to 40 o C.
The loading angular frequency is set to 10 radians per second. The shear stress
amplitude or shear strain amplitude, for controlled stress mode testing or controlled strain
mode testing, respectively, is set. The input target values for the shear stress amplitude
and the shear strain amplitude depend on the type of asphalt being tested, as shown in
Table 2.5.
Table 2.5 Target Shear Stress and Strain Values
Asphalt Tested Target Shear Target Shear

Strain, % Stress, kPa
Neat 12 0.12
RTFO-Aged 10 0.22
PAV-Aged 1 50
The appropriate plate size is selected, 25- mm diameter plates for neat and RTFO-
aged asphalt and 8-mm diameter plates for PAV-aged asphalt, and mounted on the
rheometer. The gap between the upper and lower plates is then set. The gap between the
plates is a critical test parameter. The accuracy of the gap measurement is directly related
to the accuracy of the asphalt specimen evaluation. A micrometer wheel is used to
32
measure the gap between plates (i.e. when the micrometer wheel is set on 1 mm and the
upper plate is fully lowered, a 1- mm gap will be maintained between the plates). A gap
measurement verification procedure, called setting the zero gap, must be conducted to
ensure that the micrometer reading and the actual gap between the plates is the same.
The zero gap is set by lowering the upper plate in small increments until the upper and
lower plates just touch, or reach zero gap. The micrometer wheel is then set to zero when
zero gap between the plates have been achieved. Before setting the zero gap, the
temperature controller is turned on and the environmental chamber is preheated, or
cooled, to the test temperature. The zero gap is set after the medium surrounding the
plates stabilizes at the test temperature.
A covered container of the asphalt is placed in an oven and heated until fluid to
pour. Two methods are accepted for fabricating an asphalt test specimen. One technique
employs the use of the silicone mold. The heated asphalt is poured into the mold and
allowed to cool. The asphalt specimen is then demolded and placed on the bottom
loading plate. The other technique is to pour the heated asphalt directly onto one of the
loading plates.
With the asphalt specimen properly placed on one of the loading plates, the upper
plate is lowered to squeeze the asphalt specimen between the plates. The upper plate is
lowered such tha t the gap between the plates is 0.05 mm greater than the test gap. The
test gap for the 25-mm plates and the 8- mm plates is 1 mm and 2 mm, respectively. The
excess asphalt that is squeezed from between the plates is removed by trimming around
the periphery of the plates. The upper plate is then lowered to the test gap. The asphalt
specimen should slightly bulge around the periphery of the plates.
The trimmed specimen is brought to the test temperature, via the fluid or air
medium, and allowed to equilibrate for 10 minutes. At the conclusion of the 10- minute
equilibrium period, the specimen is conditioned by applying the required stress or strain
at the 10 rad/sec frequency for 10 cycles. The test is then run by applying the same stress
or strain for an additional 10 cycles. The data from the second set of cycles are reduced
and used to calculate the complex shear modulus and phase angle.
33
Unit Calibration
The calibration procedures for the dynamic shear rheometer include a resistance
thermal detector (RTD) calibration/verification and overall verification calibration. The
fluid bath RTD is initially calibrated by using a reference thermister that is inserted into a
silicone wafer the size of an asphalt testing specimen. The thermister/wafer apparatus is
inserted between the DSR testing plates in the fluid bath. The resistance created by the
thermister is read by an ohms resistance meter and then converted to a temperature
reading. An appropriate temperature correction to the DSR RTD temperature
measurement is applied if the measurement is not within ±0.1 o C of the reference
thermister measurement. The RTD measurement must be periodically verified using the
reference thermister.
The overall calibration of the DSR is checked by performing a test on a specimen

made from a standard viscosity fluid. The complex modulus measurement and the test
frequency are used to determine the viscosity of the reference fluid. The resulting
viscosity value must be within an allowable range or recalibration of the DS R by the
manufacturer is required.
Theory of Analysis
Since the dynamic shear rheometer evaluates asphalt binder properties at

intermediate to high service temperatures, the binder responses are in the
viscoelasticrange (Bahia and Anderson, 1995a). Therefore, viscosity measurements
alone are not sufficient to characterize the asphalt behavior. Properties need to be
established which characterize the resistance to deformation and the relative distribution
of the resistance between the elastic component and the viscous component. Dynamic, or
oscillatory, testing is a common technique for evaluating viscoelastic behavior.
The DSR evaluates the behavior of an asphalt specimen by subjecting it to

oscillatory (sinusoidal) stresses. A thin asphalt specimen is sandwiched between two
parallel metal plates held in a constant temperature medium. One plate remains fixed
while the other oscillates, at an angular frequency (ω) of 10 radians per second for 10
cycles, with respect to the other. A complete DSR loading cycle is shown in Figure 2.7
(Roberts, et al., 1996). When torque from the DSR motor is applied, the oscillating plate
34
Applied shear
stress or strain
Rotating plate
Asphalt
Fixed plate
Figure 2.7 Configuration and Load Cycle of Dynamic Shear Rheometer
moves from point A to point B. The plate then passes back through point A to point
C. The cycle of oscillation is completed as the plate passes back through point A again.
The 10 rad/sec angular frequency is equivalent to a frequency (ƒ) of 1.59 cycles

per second (1.59 Hz), as per the relationship ω=2πƒ (Petersen, et al., 1994a). The
10 rad/sec angular frequency corresponds, with sinusoidal loading, to a 0.1 second
loading time, where loading time (t) is determined from the relationship t=(2πƒ)-1 . The
0.1-second loading time represents the loading time within a pavement structure resulting
from the pass of a truck tire traveling at 50 mph.
The basis for the 0.1-second loading time was not found in the literature.
However, the relationship between tire contact area and load duration presented in Huang
(1993) may serve as the basis for this loading time. For example, the load per tire of a
dual tire, 18-kip axle, is 4500 pounds. A tire pressure of 70 psi yields a tire contact area,
Ac, of 64.29 in2 .
35
Tire Load
Ac = (2-5)
Tire Pressure
For a combined contact area of the dual tires, the combined contact radius, Rc, of
dual tires is 0.53 ft.
 Ac 
Rc = 2  (2-6)
 p 
For a speed, S, of 50 mph, the load duration, D, is 0.087 seconds.
12 Rc
D= (2-7)
S
By comparison, tire pressures of 75 psi and 110 psi yield load durations of 0.084
seconds and 0.070 seconds, respectively. These times represent the duration of the tire
patch contact. The actual load duration is greater than these values, however, due to the
“stress wave” surrounding the tire. Thus, the above relationships yield loading times that
are reasonably close to the 0.1-second loading times documented in the literature.
Dynamic testing provides an indication of a binder’s resistance to deformation

and the elastic/viscous component distribution by determining the binder’s complex
modulus and phase angle (Roberts, et al., 1996). The complex modulus, G*, commonly
referred to as G star, represents the total deformation resistance when loaded or sheared.
The complex modulus is defined as the ratio of the absolute value of the peak-to-
peak shear stress to the absolute value of the peak-to-peak shear strain:
Complex Modulus (Pa)
τ max − τ min
G* = (2-8)
γ max − γ min
The following relationships are used calculate τmax and γmax :
2Tmax
τ max =
πr 3 (2-9)
36
θ max r
γ max = (2-10)
h
where:
τmax = absolute value of the peak-to-peak shear stress (Pa)

γmax = absolute value of the peak-to-peak shear strain (%)
Tmax = maximum applied torque (N- m)
r = radius of specimen plate (mm)
θmax = maximum deflection angle (rad)
h = specimen height (mm)
The phase angle, δ, represents the relative distribution between the elastic
response and the viscous response to loading (Roberts, et al., 1996). The phase angle
indicates the delayed strain response, or lag, of the binder to the applied shear stress,
during steady state conditions.
A graphical description of the phase angle with respect to time, applied shear
stress, and shear strain is shown in Figure 2.8. For the elastic response, Figure 2.8 (a),
the shear strain sinusoidal curve mirrors the applied shear stress sinusoidal curve at all
points from (O) to (B) (Roberts, et al., 1996). For the viscous response, Figure 2.8 (b),
there is a gradual strain response to the applied shear strain from point (O) to point (A).
The strain response has not yet reached steady state conditions. However, at all points
from (B) to (D) the strain response has a consistent time lag of equal duration between
the applied shear stress and the strain response. Thus, the strain response has reached
steady state conditions and is “out-of-phase” with the applied shear stress.
The time lag at steady state conditions can be represented graphically as a 90o
shift between the maximum applied shear stress and the maximum shear strain. Thus, the
strain response of a viscous material is 90o “out-of-phase” with the applied shear stress,
Figure 2.8 (b); therefore, the phase angle is 90 degrees. Conversely, there is no time lag
between the applied shear stress and resultant strain response of an elastic material,
Figure 2.8 (a); therefore, the strain response is “in-phase” with the applied shear stress
and the phase angle is zero degrees.
37
Shear stress
Elastic response
Shear strain
Time
a. Elastic response, δ = 0°
Shear stress
Time lag
Shear strain
Steady state response

Time
a. Viscous response, δ = 90°
Figure 2.8 Phase Angle s for Elastic and Viscous Materials

38
The phase angle is the product of the time lag and the frequency, as shown in
Equation 2-11.
δ = ∆t × ω (2-11)
where:
∆t = time lag (sec)
ω = angular frequency (rad/sec)
The time lag between the maximum applied shear stress and the maximum shear
strain response can be determined from Equation 2-11. For example, the phase angle for
a totally viscous response at the test angular frequency of 10 rad/sec is 90 degrees.
Substituting the phase angle and angular frequency into Equation 2-11 and solving for
time lag:
90o = 0.5π = (∆t)(10 rad/sec) ⇒ ∆t = 0.157 sec. (2-12)
Thus, a 0.157 second lag exists between the maximum applied shear stress and
the maximum shear strain response of a material having a totally viscous response. The
time lag can be verified using the relationship between frequency (ƒ) and period, or cycle
time (T), T=ƒ-1 . An angular frequency of 10 rad/sec is equivalent to a frequency (ƒ) of
1.59 cycles per second (1.59 Hz). A 1.59 Hz angula r frequency yields a period, or cycle
time (T), of 0.629 seconds as per the relationship T=ƒ-1 . One fourth of the 0.629-second
cycle time, which corresponds to 0.5π, is 0.157 seconds, which is consistent with
Equation 2-11 and Equation 2-12.
A graphical description of the phase angle with respect to the complex modulus is
shown in Figure 2.9 (Roberts, et al., 1996). The complex modulus consists of the storage
modulus, G’ and the loss modulus, G”. The storage modulus, which is the elastic
(recoverable) component, represents the amount of energy stored in the sample during
each loading cycle. The loss modulus, which is the viscous (non-recoverable)
component, represents the amount of energy lost during each loading cycle. When the
phase angle is zero degrees, elastic behavior, the complex modulus consists solely of the
storage modulus. Likewise, when the phase angle is 90 degrees, viscous behavior, the
39
Viscous
Part
*
Viscous
Part
G"
G" *
δ δ
G' G'
Elastic Part
Elastic Part Asphalt B
Asphalt A
Figure 2.9 Relationship Between Complex Modulus and Phase Angle
complex modulus consists solely of the loss modulus. It is therefore necessary to

determine both the complex modulus and the phase angle within the viscoelastic range of
response to adequately characterize asphalt binders.
The complex modulus must be measured within the linear viscoelastic range,
which is the region of behavior in which the shear modulus is independent of shear stress
or strain, to ensure test repeatability (Petersen, et al., 1994b). The limit of linear
viscoelastic behavior is the point beyond which the complex modulus decreases to 95%
of the measured value at zero-strain, as shown in Figure 2.10 (Petersen, et al., 1994b).
The relationship derived during the SHRP research between the complex modulus and
shear stress and strain within the linear viscoelastic range is shown in Equation 2-13 and
Equation 2-14:
Shear Stress (kPa) τ = 0.12 (G*)

0.71
(2-13)
Shear Strain (%) γ = 12.0 / (G*)

0.29
(2-14)
The target shear stress and shear strain parameters from Table 2.5 were
established from Equation 2-13 and Equation 2-14, respectively. The AASHTO TP5 test
40
method requires that the shear stress and shear strain be controlled to ±20 percent of the
target values to ensure test repeatability.
Shear Modulus, kPa

10
8 Go *
0.95Go *
4 Linear
Region
2
0
0 20
Shear Strain, Percent
Figure 2.10 Graphical Description of Linear Viscoelastic Region
Rutting Parameter
The work of Bahia and Anderson (1995a) and Roberts, et al. (1996) served as the
source of the following discussion on the rutting parameter development. The magnitude
of the complex modulus and the degree of phase angle are required to determine the
relationship between asphalt stiffness and the type of deformation: recoverable and non-
recoverable. This is especially true when considering rutting resistance at high service
temperatures. A higher G* and a lower δ are desired for rutting resistance. An asphalt
with a high G* is stiffer and provides increased resistance to deformation. An asphalt
exhibiting a lower δ has a greater elastic component, thus allowing more of the total
deformation to be recovered.
41
Rutting is assumed to be the primary result of deformations within the surface

layer. Rutting is considered a stress-controlled, cyclic loading phenomenon. Work is
being done to deform the surface layer with each loading cycle. A portion of the work is
recovered in the elastic rebound of the surface layer, while the remaining work is
dissipated through permanent deformation and heat. For a viscoelastic material, the work
dissipated per loading cycle, Wc, is a function of stress and strain:
Wc = (π)(σ) (ε)(sin δ) (2-15)
Since rutting is assumed to be a stress-controlled (σo ), cyclic phenomenon, the

work dissipated per loading cycle can be written as:
Wc = (π)(σo )(ε)(sin δ) (2-16)
By convention in the literature, the strain component in Equation 2-16 is

computed as:
σo
ε= (2-17)
G*
Note that Equation 2-17 does not follow conventional engineering notations as G*
is a shear modulus while e and s are traditionally used to indicate normal strains and
normal stresses, respectively. Substituting Equation 2-17 into Equation 2-16, Wc takes
the final form:
 
( )
Wc = π σ o2  *
G
1 

(2-18)
 sin δ 
The relationship G*/sin δ was chosen as the parameter for SHRP specifications
with respect to rutting. It can be seen in the preceding equation that an increase in G*
42
and decrease in sin δ will both decrease the amount of work dissipated per loading cycle
within a pavement’s surface layer. This relationship follows the rationale that a binder
with a high G* value is stiffer, which increases its resistance to deformation, and a binder
with a low sin δ value is more elastic, whereby its ability to recover part of the
deformation is increased.
Fatigue Cracking Parameter
Fatigue cracking is the primary pavement distress at intermediate service

temperatures (Roberts, et al., 1996). Pavement fatigue cracking is considered a strain-
controlled distress in thin pavement layers, less than 2 inches, because deformations in
the asphalt layers are typically the result of poor subsurface layer support and not so
much the effect of decreases in pavement stiffness (Huang, 1993). Pavement fatigue
cracking is considered a stress-controlled distress in thick pavement layers, greater than
6 inches, as the pavement is the main load-carrying constituent. A combination of both
stress-controlled and strain-controlled distresses exists with intermediate thickness HMA
pavements.
Fatigue cracking occurs primarily in thin pavement layers; therefore, the distress
is modeled as a strain-controlled phenomenon (Bahia and Anderson, 1995a and Roberts,
et al., 1996). The work per cycle equation can be written for strain-controlled cyclic
loading as follows:
Wc = π (σ )(ε o )(sin δ ) (2-19)
where ε o is the applied strain amplitude.
By conve ntion in the literature, the stress component in Equation 2-19 is

computed as:
( )
σ = (ε o ) G * (2-20)
43
As with Equation 2-17, the relationship of Equation 2-20 does not follow
conventional engineering notation. Substituting the above relationship into Equation 2-
19 yields:
( )[
Wc = π ε o2 G* sin δ ] (2-21)
The work done during a loading cycle can be dissipated by cracking, crack
propagation, heat, and plastic flow. All of these dissipation mechanisms are damaging to
pavement structures; therefore, it is necessary to limit the amount of energy dissipated. It
can be seen in Equation 2-21 that strain-controlled work per cycle is directly proportional
to G*sin δ. The G*sin δ relationship was therefore selected as a parameter for the SHRP
specification. By limiting the G*sin δ parameter, decreasing G* and/or sin δ, the energy
dissipated per cycle is limited as well. This limiting parameter follows the rationale that
a binder with a low G* is softer, which allows it to deform without developing high
stresses, and a binder with a low sin δ will be more elastic, which enables the pavement
structure to return to its original condition without dissipating energy.
The parameter specification review by Roberts, et al. (1996) served as the basis
for the following discussion. The Federal Highway Administration Asphalt Binder
Expert Task Group (ETG) established the SHRP binder specifications for the DSR test
parameters. The ETG is comprised of professionals from the asphalt industry, academia,
and specification agencies. The ETG established three separate specification
requirements for the DSR tests: minimum limits of G*/sin δ for both unaged and RTFO-
aged binder tests, and a maximum limit of G*sin δ for PAV-aged binder tests.
The minimum limit of G*/sin δ for unaged asphalt testing is 1.0 kPa. The ETG
selected 1.0 kPa as the minimum limit based upon the results of performing DSR tests on
AC-10 viscosity graded asphalts. Unaged AC-10 asphalts yielded G*/sin δ values of
approximately 1.0 kPa when tested at the required 10 radians per second. Since AC-10
viscosity graded asphalts were providing reasonable service in moderate climates, the
44
ETG concluded that the 1.0 kPa value of G*/sin δ for unaged asphalt was a reasonable
specification limit.
The minimum limit of G*/sin δ for RTFO residue is 2.2 kPa. This limit was
derived from the 1.0 kPa minimum limit for unaged binders. Following the review of
RTFO test data, the average binder aging index was found to range from 2 to 2.5, where
the average binder aging index is defined as the binder viscosity after RTFO conditioning
divided by the binder viscosity prior to RTFO conditioning. Thus, the binders became 2
to 2.5 times stiffer following RTFO conditioning, with 2.2 being the average of the range.
The ETG, therefore, used the RTFO test data to establish the minimum 2.2 kPa limit for
RTFO-aged binders.
A maximum limit of 5000 kPa was established for the G*sin δ parameter. The
ETG originally called for a maximum 3000 kPa; however, when this limit was
investigated by testing 42 asphalt binders, only half would meet the initial specification.
The ETG considered the initial specification too restrictive and chose to investigate a
higher limit. After a review of the test data from 42 binders analyzed, it was noted that
approximately 15% failed to achieve a 5000 kPa value. The ETG considered the 5000
kPa value reasonable limit and established it as the maximum limit until field validation
could be compiled and analyzed.
2.10 BENDING BEAM RHEOMETER
Introduction
Purpose and Scope
The following section is based on the work of Bahia and Anderson (1995a) and
Roberts, et al. (1996). Asphalt concrete pavements are susceptible to non- load associated
distresses. Low-temperature cracking, commonly termed thermal cracking, is the most
recognized non- load associated distress. Thermal cracking is caused by thermal
shrinkage induced stresses resulting from environmental cooling. The asphalt binder in
HMA pavements stiffens and shrinks during environmental cooling. The HMA layer,
however, is restrained by the friction created with the underlying layers, which
experience less shrinkage either because they remain warmer or because they have a
45
smaller coefficient of thermal contraction. The friction- induced restraint creates tensile
stresses within the asphalt layer that, if not relaxed by the flow of the asphalt binder, have
the potential to exceed the tensile strength of the HMA layer and thus result in thermal
cracking. The magnitude of stresses the HMA pavement layer experiences is dependent
upon the stiffness of the asphalt binder, or its resistance to deformation, and its ability to
relax those stresses by dissipating energy via permanent flow.
Thermal cracking can result from two different environmental conditions. The
thermal distress can result from a single thermal cycle alone, in which a critically low
temperature is encountered. This critical temperature is often referred to as the limiting
stiffness temperature. Or, thermal cracking can result from thermal cycling where the
temperature cycles up and down, but always remains above the limiting stiffness
temperature.
The asphalt binder plays a critical role in the thermal cracking potential of HMA
pavement. Hot mix asphalt pavements that have a high stiffness modulus at low
temperatures are susceptible to thermal cracking. Asphalt binders become harder, or
stiffer, as the ambient temperature decreases. Given that the stiffness of a HMA
pavement is directly proportional to the asphalt binder stiffness, high binder stiffness at
low temperatures increase the potential of thermal cracking. The bending rheometer was
developed to determine the stiffness of asphalt binders at low service temperatures and to
evaluate the binder’s potential for thermal cracking.
Summary of Method
A PAV-aged asphalt beam specimen is subjected to a 980 mN load for 240

seconds, as per the AASHTO TP1 standard test method. The beam is simply supported
with the load applied at mid-span. The asphalt beam dimensions are 6.35 mm thick by
12.70 mm wide by 127 mm long. The asphalt beam is held at a constant low temperature
throughout the test procedure. The load and deflection of the asphalt beam are measured
and used to calculate the creep stiffness and the creep rate (m- value). The creep stiffness
is an indicator of the specimen’s ability to resist the constant creep load and the creep rate
is the rate at which the creep stiffness changes with loading time (Roberts, et al., 1996).
46
Equipment
Hardware
The primary parts of the bending beam rheometer system consist of a loading
frame, base unit, mechanical refrigeration unit, specimen molds and a personal computer.
The specimen supports that the sample beam rests upon are mounted to the loading
frame. A loading shaft assembly for applies the creep load. An air bearing and
pneumatic piston controls the loading shaft’s vertical movement. A linear variable
differential transformer (LVDT) is attached to the loading shaft to measure the sample
beam’s deflection. A load cell is mounted with the loading shaft to measure the applied
creep load. The ATS Bending Beam Rheometer evaluated during this research is shown
in Figure 2.11.
The base unit houses a cooling fluid bath, system control and data acquisition
electronics and air regulators that control the loading shaft. The cooling fluid bath houses
the lower half of the loading frame, which includes the sample beam. The bath maintains
the sample beam at a constant specified temperature during the test. The fluid bath must
remain fluid and optically clear throughout the test temperature ranges. The mass density
of the fluid bath is not to exceed 1.05 kg/m3 at test temperatures, as per AASHTO TP1
specifications, so that the sample beam remains neutrally buoyant during the test. A fluid
mixture of 60 percent glycol, 15 percent methanol, and 25 percent water is recommended
in AASHTO TP1 to achieve the required mass density. A bath agitator maintains the
homogeneity of the fluid mixture and the homogeneity of the fluid temperature. The
mechanical refrigeration unit cools the fluid bath. A resistance thermal detector is
located in the fluid bath to monitor the temperature.
The specimen molds form a 6.35 mm thick by 12.70 mm wide by 127 mm long
asphalt beam. Aluminum specimen molds and silicone specimen molds are available for
preparing the asphalt specimen beams. The specimen beam’s geometry closely follows
the guidelines for determining the flexural properties of elastomers established in ASTM
D790.
47
Figure 2.11 ATS Bending Beam Rheometer
Data Collection
The beam deflection, applied load and loading time is acquired by the data
acquisition unit and sent to the personal computer during the test. Immediately following
the test, the computer software automatically reduces the test data and computes the creep
stiffness and creep rate. The creep stiffness and creep rates are reported at six loading
times: 8, 15, 30, 60, 120, and 240 seconds.
Testing Protocol
Test Specification
AASHTO TP1-98 Standard Test Method for Determining the Flexural Creep
Stiffness of Asphalt Binder Using the Bending Beam Rheometer (BBR) defines the
protocol used for conducting the bending beam rheometer test.
Test Sample and Equipment Preparation
The loading frame is placed in the fluid bath. The temperature controller is set to
the test temperature and the fluid bath is subsequently cooled. The fluid bath is allowed
to stabilize at the test temperature for at least 30 minutes prior to testing. A covered
container of PAV-aged asphalt binder is placed in an oven and heated until fluid to pour.
48
The heated binder is poured into the specimen mold and allowed to cool to room
temperature.
Once the sample has cooled, the excess binder is trimmed from the top face of the
mold with a hot spatula, such that the asphalt sample is flush with the top of the mold.
The mold is placed in a freezer or ice bath for approximately 60 seconds so the sample
beam can be easily demolded. Once the sample beam is demolded, it is immediately
placed in the BBR fluid bath, which has been stabilized at the test temperature. The
sample beam remains in the fluid bath for 60 minutes.
At the conclusion of the 60- minute conditioning period, the sample beam is
placed on the loading frame specimen supports and the test is initiated. The BBR
electronics and air regulators automatically control the load application and test duration.
The sample beam is discarded at the conclusion of the 240-second test.
Unit Calibration
The bending beam rheometer has many devices that need to be verified and
calibrated. The load cell, linear variable differential transformer (LVDT), and resistance
thermal detector (RTD) all require an initial calibration. The load cell is calibrated using
a set of calibrated weights. The linear variable differential transformer (LVDT) is
calibrated with a calibrated step-block. The BBR manufacturer provides both the
calibrated weights and step-block. The fluid bath RTD must be initially verified via an
ASTM thermometer. If the difference in temperature between the BBR display and the
thermometer exceeds a specified value, the RTD requires recalibration. The recalibration
is accomplished by establishing a temperature correction offset within the BBR system.
The overall calibration of the BBR is accomplished by performing a confidence check.
The confidence check is conducted by performing a BBR test on a calibrated beam of
known modulus. The resulting stiffness values must be within an allowable range or
additional calibration steps are required.
After the initial calibration, a verification of calibration procedure is conducted

prior to each day of testing. The verification of calibration procedures is an abbreviated
49
variation of the initial load cell, LVDT, and RTD calibration procedures. A confidence
check is to be conducted at the beginning of each testing day as well.
Theory of Analysis
The asphalt beam dimensions were selected such that the elementary Bernoulli-
Euler theory of bending of prismatic beams would be applicable to the analysis of the test
results (Anderson, et al., 1994). From the elementary bending theory, the maximum
deflection of an elastic pris matic beam in three-point loading occurs at the beam midspan.
The deflection of the beam at midspan, δ, is calculated as:
PL3
δ = (2-22)
48 EI
where:
P = applied load (N)
L = span length (mm)
E = modulus of elasticity (Pa)
I = moment of inertia of section (mm4 )
The modulus of elasticity, E, can also be defined by the following relationship,
where D(t) is the extensional creep compliance:
E = 1 D(t ) (2-23)
In addition, the inverse of the extensional creep compliance is equivalent to the

time dependent creep stiffness, S(t), of a viscoelastic material. Based upon the elastic-
viscoelastic correspondence theory, the assumption can be made that the stress
distribution in a viscoelastic beam is the same as that in an elastic beam under the same
applied load. Thus, the time-dependant strains and displacements of an asphalt specimen
can be derived from the elementary bending theory for elastic materials by substituting
S(t) for E (Anderson, et al., 1994):
PL3
δ (t ) = (2-24)
48 S (t )I
50
Considering that the moment of inertia, I, is defined as bh3 /12, where (b) and (h)
are the width and height of a prismatic beam, respectively, and rearranging to solve for
S(t), the elementary bending theory can take the following final form for evaluating an
asphalt specimen:
PL3
S (t ) = (2-25)
48bh δ (t )
3
Thermal Cracking Parameters
The magnitude of stresses the HMA pavement layer experiences is dependent

upon the stiffness of the asphalt binder and its ability to relax those stresses by dissipating
energy via permanent flow (Bahia and Anderson, 1995a and Roberts, et al., 1996). The
asphalt properties should be such that it flows under stress and have a lower elastic
component of response. Thus, parameters where selected that evaluate both the stiffness
and its stress relaxation potential.
The creep stiffness parameter, S(t), was selected to evaluate the extent of thermal
stresses that develop in the HMA pavement during thermal contraction. The creep
stiffness is characterized by measuring the creep response of asphalts with the BBR at
critical temperatures. An HMA pavement composed of a binder with a high creep
stiffness has the potential to develop high tensile stresses during thermal contraction.
The logarithmic creep rate, m(t), is an ind icator of the asphalt’s ability to relax
stresses (Roberts, et al., 1996). The logarithmic creep rate, commonly called the m-
value, is a measure of the rate at which creep stiffness changes with loading time. The
m- value is graphically defined as the slope of the log creep stiffness versus log loading
time on a master curve, as shown in Figure 2.12 (Roberts, et al., 1996). As the m- value
decreases, the rate of relaxation decreases, thus, stresses build as the binder loses its
ability to relieve those the rmal- induced stresses by permanent flow.
The work of Bahia and Anderson (1995b) and Roberts, et al. (1996) served as the
sources for the following discussion on the BBR test parameter specifications. Both S(t)
51
Log Creep
Stiffness, S(t)
Slope - m value
Log Loading Time, sec
Figure 2.12 Graphical Definition of Creep Rate, m-value
and m(t) are functions of loading time; therefore, a loading time which correlates with
thermal cracking was specified. A review of asphalt literature during the SHRP research
showed that loading times ranging between 3,600 and 20,000 seconds had been
correlated to thermal cracking. Loading times of that duration were not practical for
laboratory tests. Therefore, the SHRP researchers used the time-temperature
superposition principle to shorten the loading times required to reflect thermal cracking
conditions. Using the time-temperature superposition principle allows the response of
long loading times to be estimated by testing at higher temperatures for a shorter
duration.
To apply the time-temperature superposition principle, creep stiffness versus

loading time curves for different temperatures are generated on a log- log plot. A
reference temperature is then selected and the curves from the other temperatures are
shifted relative to the reference temperature curve until the curves merge into a single
function, the master curve, as shown in Figure 2.13a (Bahia and Anderson, 1995b). The
log of the shift factors (x1 , x2 , etc.) needed to overlap the logarithmic curves in
Figure 2.13a are also plotted versus temperature, as shown in Figure 2.13b (Bahia and
Anderson, 1995b). A temperature shift function is established from the log shift factors
versus temperature plot of Figure 2.13b.
52
Log Flexural Creep
Log Shift Factor, x

x1 Temp Curve A
x2
Stiffness (Pa)
Temp Curve B x1
x2
Temp Curve D Temp Curve C x3
x4
x3 Reference
Temp Curve E
x4
Log Loading Time, sec Temperature, °C
Figure 2.13 Development of Time-Temperature Shift Factors
It was observed by the SHRP researchers that the shift functions generated from
their research were linear with a slope that ranged from only 0.18 to 0.2 log seconds per
degree Celsius. The shift function similarities ultimately led to the agreement that a
single offset temperature could be used for estimating the binder response at longer
loading times from shorter duration tests conducted at higher temperatures. Upon further
examination of the shift factors, it was concluded that an offset of 10 o C above the lowest
specification temperature could be used to equate the BBR stiffness at 60 seconds loading
time to the asphalt binder stiffness at 7,200 second loading time in the field.
During the BBR test, the creep stiffness and creep rates are determined at six
loading times: 8, 15, 30, 60, 120, and 240 seconds. The shape of the log creep stiffness
versus log loading time master curve, Figure 2.12 and Figure 2.13a, for asphalt binders is
determined from the data of the six loading times and the following relationship:
Log S(t) = A + B log(t) + C[log(t)]2 (2-26)
where:
S(t) = asphalt binder stiffness, Pa
t = time, sec.
A, B, and C = constants
53
Equation 2-26 and the resultant master creep stiffness curve are used to determine
the specification parameters. A maximum limit specification is required for the creep
stiffness parameter and a minimum limit specification is required for the m- value
parameter. Placing a maximum limit on the creep stiffness restricts the level of stresses
that will develop in the HMA pavement. The specification criterion established for creep
stiffness at 60 seconds, S(60), is 300 MPa maximum. The stiffness limit was established
based on previous studies that correlated thermal cracking with creep stiffness, as well as
additional evaluation tests conducted with the bending beam rheometer by the SHRP
researchers.
Placing a minimum limit on the m- value maintains the rate of relaxation such that
stresses can be relieved through permanent flow. The specification criterion for the
logarithmic rate of relaxation at 60 seconds, m(60), is 0.300 minimum. The specification
was established based on the experience of the Expert Task Group and on data collected
from a large number of asphalts tested during SHRP research.
During the SHRP testing, a new hardening behavior was observed in asphalts
tested at low temperatures. This phenomenon, termed physical hardening, is a time-
dependant increase in S(t) and the decrease in m(t) which is caused by time-delayed
asphalt volume shrinkage. The asphalt hardening was found to increase at a rapid rate
within the initial period of isothermal conditioning. Asphalt hardening was observed to
increase up to 50 to 100 percent within 24 hours. The physical hardening phenomenon
was found to be highly asphalt specific and its consequences on asphalt mixtures were
observed to be difficult to evaluate. A standard isothermal conditioning time of
60 minutes was observed to be outside the range of rapid hardening, while still being
practical for laboratory testing. Therefore, the SHRP researchers established the
specification requiring the measurement and reporting of S(60) and m(60) following
60 minutes of isothermal conditioning in the BBR fluid bath. It was also decided to
include the requirement of testing after 24 hours of isothermal conditioning to provide an
indication of the potential of the binder for physical hardening; however, the report of
S(60) and m(60) following 24 hours of isothermal conditioning is fo r information
purposes only.
54
2.11 SUMMARY
The performance grading system is a definite improvement over the previous

asphalt binder grading systems. Engineering principles are now used to relate measured
physical properties to field performance. A complete range of temperatures is now
considered. A constant criterion is used for all PG grades, with the temperature at which
the asphalt properties changing in consideration of the asphalt grade for the climate,
traffic loadings and speed.
New and revised conditioning procedures enable the performance of asphalt

binders to be evaluated all three critical stages: the transporting and handling of original
asphalt binder before mixing, the short-term aging of binders following HMA production
and construction, and the continued long-term aging during the pavement service life.
Performance parameters are in place to evaluate a binder’s resistance to aging

during construction and during its service life. The parameters ensure that the binder is
workable during the mixing and placement of HMA pavements. Performance parameters
also control rutting, fatigue cracking and thermal cracking pavement distresses.
55
CHAPTER 3
RESEARCH APPROACH
3.1 INTRODUCTION
The objective of this research was a straightforward evaluation of the

performance grading equipment at the Asphalt Technology Laboratory and the related
testing protocol. However, due to the newness of both the test protocol and equipment,
accomplishing this objective was fraught with difficulties and complication. The test
protocols and specifications are under constant scrutiny, review, and occasional revision.
To make this problem manageable, it was decided to use the test procedures in force
when the research was initiated. Training in the proper methods required a considerable
effort. Calibration and verification of the equipment was time consuming. In some
cases, supplemental equipment and tools were developed to assist with test performance.
Once confidence was developed in conducting the tests and in the equipment
performance, a test program was developed. Since the emphasis of the research was on
the equipment and testing protocol, a single grade and source of asphalt was used for all
tests.
3.2 TESTING PROGRAM PREPARATION
Operator Training
A thorough review of the equipment manufacturer’s documentation, related

SHRP documentation and the AASHTO standard specifications test methods was first
conducted to establish a full understanding of the equipment’s scope of use, limitations,
and testing protocol and procedures. A trial period of testing was conducted to learn the
testing protocol and procedures and to become proficient in asphalt sample preparation
and handling, unit calibration and operation, and testing technique via hands-on
experience. The trial testing period was also intended to provide the opportunity to
identify and document potential sources of equipment and user error.
A one-day training workshop provided by Applied Test Systems, Inc. (ATS) for
the pressure aging vessel and bending beam rheometer was attended. The workshop
56
provided instruction on equipment calibration, operation, maintenance, and sample

specimen preparation. As a result of difficulties experienced with the DSR during the
trial testing period, Dennis Burke, a senior technician for the Citgo Asphalt Refining
Company and an experienced operator of the Bohlin DSR and Brookfield RV, graciously
visited the WVU Asphalt Technology Laboratory to evaluate the performance of the
DSR. Mr. Burke provided instruction on sample specimen preparation and equipment
calibration/operation of the RV and DSR.
Equipment Preparation
The majority of the hardware and accessories required for the operation of each
unit (i.e. sample specimen molds, standard viscosity calibration fluid, etc.) was available
from the equipment manufacturer or by another asphalt industry supplier. There were
some accessories that were not available from either of the above and therefore had to be
fabricated.
A tool for scraping RTFO-aged asphalt residue from the RTFO specimen bottles
was fabricated. Two different styles of tools were developed in an effort to achieve the
AASHTO T240 requirement of 90 percent residue removal from the specimen bottles.
One tool was designed such that the residue was removed by scraping down the
longitudinal length of the specimen bottle. The other style of tool removed the residue by
scraping around the periphery of the bottle. The later tool design was more efficient than
the former; however, the required 90 percent removal crit erion could never be achieved
with either style of scraping tool.
A DSR specimen trimming tool was not available from a commercial supplier;
therefore, a trimming tool was fabricated. The tool was designed such that the asphalt
specimen was trimmed by dragging the edge of the tool around the periphery of the
parallel plates. The trimming tool design was evaluated via multiple trial tests and
worked well. Therefore, the tool was used for the testing program.
The calibration of each unit was verified and recalibrated as necessary. The flow
rate of the RTFO was verified with a flow meter and the oven temperature was verified
with a NIST traceable ASTM Loss of Heat Thermometer 13 o C. The pressure and
temperature calibration of the pressure aging vessel is to be conducted by a commercial
57
calibration service provider. The PAV calibration was not verified prior to the testing
program. Mr. Burke conducted the RV temperature controller and rotary transducer
calibration verification. The temperature controller was verified using an ohmmeter and
a NIST traceable thermister. The rotary transducer was verified using Brookfield
Viscosity Standard Fluid, No. HT100000. Mr. Burke also verified the calibration of the
dynamic shear rheometer. The temperature controller was verified using a Fluke 77III
multimeter and a Cannon wafer thermister, No. 9728-V95. The overall calibration of the
unit was verified using Cannon Viscosity Standard Fluid, No. N2700000SP. The
bending beam rheometer LVDT and load cell was calibrated using the standard weights,
thickness gauge, and confidence beam supplied by ATS. The ATS technical
representative stated that the resistance thermal detector on their units do not require
calibration; therefore, it was not calibrated.
The equipment calibration of each unit was checked multiple times at random
during the trial period in order to maintain a level of confidence in the stability of the
equipment. Multiple trial tests were conducted until consistency in results was
established.
Testing Protocol
The AASHTO standard specifications were the testing protocol source for each
unit. The protocol for the selection of the rotational viscometer spindle type and spindle
speed, however, was not straightforward. The AASHTO TP48 test method requires that
a test first be conducted at 20 RPM. If the resultant torque is out of range for the selected
spindle and speed, then the operator is to change the spindle or speed as per the
manufacturer’s recommendations. The specifications, however, are silent regarding the
spindle types. In addition, the precision and bias criterion of the AASHTO test method is
established only for 135 o C test temperatures; no criterion is stated for higher test
temperatures that are encountered when performing equiviscosity tests.
The AASHTO TP48 protocol was compared to AASHTO MP1, and SHRP and
Asphalt Institute literature for additional direction with regard to spindle type selection.
The ASTM D4402 protocol was referenced in AASHTO MP1. The ASTM D4402 states
that testing with more than one spindle might be required to properly select the
58
appropriate spindle type. The ASTM standard also states that testing should start at 20
RPM for the Brookfield RV, HA, and HB series and 12 RPM for the LV series. The
standard states that changes to spindle speed or spindle type may be required if viscosity
readings are not within a specified range. No recommendations are presented in the
specifications for a particular spindle type. A SC4-27 spindle is recommended in the
SHRP literature (Petersen, et al., 1994b). A 20-RPM operating speed is also
recommended by the SHRP researchers in a different volume of the final report
(Anderson, et al., 1994). However, neither volume of the SHRP final report contains
documentation to which Brookfield series (LV, RV, HA, or HB) their spindle and speed
recommendations apply.
The Asphalt Institute recommends conducting viscosity tests at 135 o C and 165 o C
when establishing temperature-viscosity curves (Asphalt Institute, 2001). However, the
Asphalt Institute literature provided no direction on what spindle types or speeds to use
when conducting the tests.
The rotational viscometer testing protocol seemed ambiguous when comparing

the direction provided by AASHTO, ASTM, SHRP, and the Asphalt Institute. Therefore,
the decision was made to follow the AASHTO TP48 protocol, with the exception that the
SC4-27 spindle was used with an operating speed of 20 RPM, as per the SHRP literature.
Additionally, the testing program would include tests to temperatures of 165 o C, per
Asphalt Institute methodology. The AASHTO standard specifications were, however,
strictly followed for every other test.
3.1 TESTING PROGRAM
A graduate student was employed for conducting the testing program after the
testing preparation procedures described in the previous section were completed. The
student was trained on the equipment operation and sample preparation. The student also
performed multiple trial tests with each unit until a consistency in test results was
established.
Once the objectives of the trial testing period were accomplished, the actual
testing program was initiated to evaluate the equipment precision and repeatability.
59
Rotational viscometer, dynamic shear rheometer, and bending beam rheometer tests were
conducted on PG 70-22 asphalt. The tests and conditioning procedures are summarized
in Table 3.1.
The PG 70-22 neat asphalt underwent rolling thin film oven and pressure aging
vessel conditioning procedures, where required. Asphalt sample sets were created. The
sample sets consisted of neat, RTFO-aged, and PAV-aged asphalt stored in three-ounce
tins. Each three-ounce tin from a sample set was used to make a test specimen for a
specific test. One test specimen, either a rotational viscometer, dynamic shear rheometer,
or bending beam rheometer test specimen, was made from each three-ounce tin. A
summary of the testing program is provided in Figure 3.1 and Figure 3.2.
Table 3.1 Testing and Conditioning Sequence
Test Temperature1 Conditioning Test Number

DSR High RTFO (Drain Residue) 1,2,3,4
DSR High RTFO (Scrape Residue) 5,6,7,8
DSR Mid RTFO (Scrape); PAV 9,10,13,14,
BBR Low RTFO (Scrape); PAV 11,12,15,16
RV Sweep None 17,18
RV 135 o C None 19,20
RV 140 o C None 21,22
RV 145 o C None 23,24
RV 150 o C None 25,26
RV 155 o C None 27,28
RV 160 o C None 29,30
RV 165 o C None 31,32
DSR High None 33,34,35,36
1
High temp = 70 o C
Mid Temp = 28 o C
Low Temp = -12 o C
Sweep = 135 o C to 165 o C
63 60
2 Gallons of Neat Asphalt
Split Neat Asphalt into (12) 1-Pint Cans Split Neat Asphalt into (30) 3-oz tins
Sample Set 7
See Figure 3.2
Pint 10 Pint 11 Pint 12 Pint 2 Pint 3 Pint 4 Pint 6 Pint 8
RTFO RTFO RTFO RTFO RTFO RTFO
Same procedure as Pint 10
Same procedure as Pint 12

Conditioning Conditioning Conditioning Conditioning Conditioning Conditioning
w/ Mass Loss w/ Mass Loss
Determination Determination
Scrape Residue Scrape Residue
Drain Residue Scrape Residue into 1-Qt Can into 1-Qt Can
into 1-Qt Can into 1-Qt Can
PAV Conditioning PAV Conditioning
Split Residue Split Residue
into (4) 3-oz tins into (4) 3-oz tins Collect Residue Collect Residue
in 1-Gal Can in 1-Gal Can
Sample Set 1 Sample Set 3
Vacuum De-gas Vacuum De-gas
Tin Tin Tin Tin PAV-residue PAV-residue
4 2 4 2
Split Residue Split Residue
Test Test Test Test into (6) 3-oz tins into (6) 3-oz tins
1 2 5 6
(DSR) (DSR) (DSR) (DSR) Sample Set 5 Sample Set 6
Sample Set 2 Sample Set 4

Tin Tin Tin Tin Tin Tin Tin Tin
Tin Tin Tin Tin 4 5 2 6 4 5 2 6
4 2 4 2
Test Test Test Test Test Test Test Test
Test Test Test DSR 9 10 11 12 13 14 15 16
3 4 7 8 (DSR) (DSR) (BBR) (BBR) (DSR) (DSR) (BBR) (BBR)
(DSR) (DSR) (DSR) (DSR)
Figure 3.1 Testing Program Summary

64
See Figure 3.1

Split Neat Asphalt into 1-Pint cans
Test 17 (RV Temperature Sweep) Tin 12
Figure 3.2 Testing Program Summary – continued

Test 18 (RV Temperature Sweep) Tin 15
Test 19 (RV Test @ 135 o C) Tin 11

2-Gallon Neat Asphalt
o
Test 25 (RV Test @ 150 C) Tin 29

Sample Set 7

Split Neat Asphalt into (30) 3-oz tins
o
Test 32 (RV Test @ 165 C) Tin 8
Test 33 (DSR Test) Tin 22

61

62
The testing program required seven sample sets:
Sample Set Test Conditioning

1 DSR RTFO (Drain)
2 Replicate of Sample Set 1
3 DSR RTFO (Scrape)
5 DSR; BBR RTFO (Scrape); PAV
7 RV; DSR None
Sample Sets 1 and 2 are replicates. Each set required one-pint of binder that was
conditioned in the RTFO and then split into sample tins for testing with the DSR. The
RTFO conditioning bottles were drained, rather than scraping, to remove the binder
following conditioning.
Sample Sets 3 and 4 are replicates. The laboratory protocol was similar to Sets 1
and 2, with the exception of scraping the RTFO-aged material following conditioning
rather than draining.
Sample Sets 5 and 6 are replicates. Each set required two pints of binder. The
samples were conditioned in the RTFO. The RTFO-aged material was scraped from the
bottles following conditioning. The RTFO-aged material was then conditioned in the
PAV and vacuum degassed. The PAV-aged material was split into sample tins for testing
with the DSR and BBR.
Sample Set 7 was used for RV testing and DSR tests on neat asphalt; therefore, no
conditioning procedures were conducted prior to testing. Two replicate tests were
performed using a RV temperature sweep procedure. Seven pairs of test tins were used
for replicate single temperature tests. The remaining tins were used for DSR tests.
Two gallons of PG 70-22 neat asphalt binder was required for the testing
program. The gallon containers of neat asphalt were heated in a 135 o C oven once and
immediately split into 6 one-pint cans and 15 three-ounce tins. Approximately 350 grams
63
of the neat asphalt was poured into each one-pint can and approximately 40 grams of the
neat asphalt was poured into each three-ounce tin. The filled cans and tins were covered
and stored at room temperature. Thus, 12 one-pint cans and 30 three-ounce tins were
available for further conditioning and testing.
As shown in Figures 3.1 and 3.2, only 8 one-pint cans and 20 three-ounce tins of
neat asphalt were required for the testing program; the additional 4 one-pint cans and 10
three-ounce tins served as a reserve supply in case problems were encountered during the
conditioning and testing procedures. The pints and tins used for testing program and for
sample reserves were randomly selected.
The neat asphalt from each of the 8 one-pint cans was conditioned with the
RTFO, as per AASHTO T240. Each of the one-pint cans was heated in a 135 o C oven
until sufficiently fluid to pour into the RTFO bottles. A mass loss test was conducted
with the RTFO conditioning procedures for Sample Sets 1 through 4. The asphalt residue
from each RTFO conditioning procedure was collected in a one-quart can. The residue
from the RTFO bottles was collected by two different methods. One method was to
allow the residue to drain from the RTFO bottles. The residue from the RTFO
conditioning procedures for Sample Sets 1 and 2 was collected in this manner. The
residue from the remaining RTFO conditioning procedures was collected by scraping the
residue from the RTFO bottles. The one-pint cans containing the RTFO residue was
covered and stored at room temperature for further conditioning and testing. DSR tests
were later conducted on both the drained RTFO residue and scraped RTFO residue to
determine if scraping yields significant differences in binder properties.
The residue collected from the RTFO conditioning procedures for Sample Sets 5
and 6 underwent PAV conditioning, as per AASHTO PP1. The PAV residue for each
sample set was collected in a one- gallon can and then degassed in a vacuum oven as
specified in AASHTO PP1. Each can of PAV-aged residue was covered and stored at
room temperature for future testing.
Each of Sample Sets 1 through 4 consisted of 4 three-ounce tins containing

approximately 40 grams of the RTFO aged asphalt residue. The one-quart cans of RTFO
residue were heated in a 135 o C oven until sufficiently fluid to pour into the tins. The
64
three-ounce tins were covered and stored at room temperature. Only 2 of the 4 tins in
each sample set were used for testing. The remaining 2 tins served as a spare tins in case
problems were encountered during testing. As shown in Figure 3.1, the residue stored in
the three-ounce tins was used for DSR testing, as per AASHTO TP5. Only one DSR test
specimen was made from each three-ounce tin. Each DSR specimen was made using a
silicone specimen mold. The RTFO residue from each three-ounce tin was heated in a
135 o C oven until sufficiently fluid to pour into the silicone specimen mold.
Each of Sample Sets 5 and 6 consisted of 6 three-ounce tins containing

approximately 40 grams of the PAV-aged asphalt residue. The one-gallon cans of PAV
residue were heated in a 163 o C oven until sufficiently fluid to pour into the tins. The
three-ounce tins were covered and stored at room temperature. Only 4 of the 6 tins in
each sample set were used for testing. The remaining 2 tins served as a spare samples.
The PAV residue stored in the three-ounce tins were used for DSR testing, as per
AASHTO TP5, and BBR testing, as per AASHTO TP1. Only one DSR or BBR test
specimen was made from each three-ounce tin. The DSR specimens were made using a
silicone mold and the BBR specimens were made using aluminum bar molds.
Sample Set 7 was made during the initial splitting of the 2 gallons of asphalt.
Approximately 40 grams of neat asphalt was poured into 30 three-ounce tins. Only 20 of
the 30 tins were used for testing; the remaining 10 tins served as spares.
As shown in Figure 3.2, the neat asphalt in each three-ounce tin was used for RV
and DSR testing. RV tests were conducted with 16 of the specimen tins in Sample Set 7
while DSR tests were conducted on the remaining tins. Only one RV specimen or DSR
specimen was made from each three-ounce tin. Each tin was heated in a 135 o C oven
until sufficiently fluid to pour into the RV sample holder or the DSR silicone specimen
mold. The DSR and RV tests were conducted as per AASHTO TP5 and AASHTO TP48,
respectively.
As shown in Figure 3.2, two duplicate RV temperature sweep tests were

conducted. Each temperature sweep test was performed on a single test specimen.
Viscosity measurements were taken in 5 o C increments beginning at 135 o C up to 165 o C.
A 10-minute equilibrium period at each test temperature was allowed before the viscosity
65
measurements were taken. Three viscosity measurements were taken at 1- minute

intervals at each 5 o C incremental temperature. Duplicate single temperature RV tests
were also performed at temperatures ranging from 135 o C to 165 o C, at 5 o C intervals, as
shown in Figure 3.2.
The sample set conditioning sequence is shown is Table 3.2. The sample set
conditioning sequence, the pints conditioned for each sample set, the selection order of
replacement tins, the selection order of tins for testing, the selection order of replacement
tins, and the tests conducted from each sample set were randomly selected. The testing
sequence is shown in Table 3.3. The testing order was randomly selected as well.
Table 3.2 Sample Set Conditioning Sequence
Sample Pint No.'s Selection No. of Selection Selection Tests

Set1 Conditioned Order of Tins in Order of Order of Conducted
for Sample Replacement Sample Tins for Replacement from
Set Pints Set Testing Tins Sample Set
4 2 9-5-7-1 4 4-2 1-3 7,8
5 3,4 9-5-7-1 4 4-5-2-6 1-3 9,10,11,12
1 10 9-5-7-1 4 4-2 1-3 1,2
2 11 9-5-7-1 4 4-2 1-3 3,4
6 6,8 9-5-7-1 6 4-5-2-6 1-3 13,14,15,16
3 12 9-5-7-1 6 4-2 1-3 5,6
1
Listed in order of conditioning sequence
66
Table 3.3 Testing Sequence
Test No.1 Test Description Sample Set Tin No. Selection Order of
33 DSR (Neat) 7 22 10-18-9-5-30-14-20-28-17-4
34 DSR (Neat) 7 2 10-18-9-5-30-14-20-28-17-4
o
27 RV - 155 C 7 7 10-18-9-5-30-14-20-28-17-4
12 BBR 5 6 1-3
o
29 RV - 160 C 7 19 10-18-9-5-30-14-20-28-17-4
o
28 RV - 155 C 7 1 10-18-9-5-30-14-20-28-17-4
8 DSR (RTFO) 4 2 1-3
7 DSR (RTFO) 4 4 1-3
o
32 RV - 165 C 7 8 10-18-9-5-30-14-20-28-17-4
o
23 RV - 145 C 7 21 10-18-9-5-30-14-20-28-17-4
o
24 RV - 145 C 7 6 10-18-9-5-30-14-20-28-17-4
19 RV - 135o C 7 11 10-18-9-5-30-14-20-28-17-4
18 RV - Temp. 7 15 10-18-9-5-30-14-20-28-17-4
o
26 RV - 150 C 7 26 10-18-9-5-30-14-20-28-17-4
o
20 RV - 135 C 7 3 10-18-9-5-30-14-20-28-17-4
36 DSR (Neat) 7 25 10-18-9-5-30-14-20-28-17-4
10 DSR (PAV) 5 5 1-3
6 DSR (RTFO) 3 2 1-3
14 DSR (PAV) 6 5 1-3
35 DSR (Neat) 7 16 10-18-9-5-30-14-20-28-17-4
15 BBR 6 2 1-3
2 DSR (RTFO) 1 2 1-3
3 DSR (RTFO) 2 4 1-3
5 DSR (RTFO) 3 4 1-3
o
21 RV - 140 C 7 27 10-18-9-5-30-14-20-28-17-4
11 BBR 5 2 1-3
o
30 RV - 160 C 7 24 10-18-9-5-30-14-20-28-17-4
31 RV - 165o C 7 13 10-18-9-5-30-14-20-28-17-4
1 DSR (RTFO) 1 4 1-3
16 BBR 6 6 1-3
4 DSR (RTFO) 2 2 1-3
22 RV - 140o C 7 23 10-18-9-5-30-14-20-28-17-4
9 DSR (PAV) 5 4 1-3
17 RV - Temp. 7 12 10-18-9-5-30-14-20-28-17-4
o
25 RV - 150 C 7 29 10-18-9-5-30-14-20-28-17-4
13 DSR (PAV) 6 4 1-3
1
Listed in order of testing sequence. The test number is defined in Figures 3.1 and 3.2.
67
CHAPTER 4
TEST RESULTS ANALYSIS AND EQUIPMENT EVALUATION
4.1 INTRODUCTION
The purpose of this chapter is to provide an analysis of the test results. The test
results are used to draw conclusions on the performance of the SHRP testing equipment.
The test data were compiled and summarized to characterize the equipment’s ability to
meet the SHRP performance test specifications and the AASHTO precision criteria. A
discussion of potential causes for test errors was prepared. This chapter also includes a
discussion on the time and effort that was required to set up and calibrate the equipment
and to perform trial tests such that a confidence in the equipment operation was achieved.
4.2 ROLLING THIN FILM OVEN
Testing Results Analysis
The average mass loss for each RTFO test is presented in Table 4.1. Precision of
mass loss test criteria was not available in AASHTO T240 at the time of this research;
therefore, no analysis of the results was conducted.
Table 4.1 RTFO Mass Loss Results
Sample Pint RTFO Pre-Conditioned Post- Mass Loss Average

Set No. Bottle No. Mass (grams) Conditioned (grams) Mass Loss
Mass (%)
(grams)
1 10 Bottle 1 34.992 34.821 0.00489 0.3069
Bottle 2 35.056 35.012 0.00126
2 11 Bottle 1 35.399 35.387 0.00034 0.0411
Bottle 2 35.211 35.194 0.00048
3 12 Bottle 1 34.826 34.796 0.00086 0.1064
Bottle 2 34.713 34.669 0.00127
4 9 Bottle 1 34.852 34.800 0.00149 0.0933
Bottle 2 34.785 34.772 0.00037
68
Equipment Evaluation
The Despatch RTFO was simple in operation and required minimal time to
become proficient in its use. The procedures outlined in the AASHTO T240 test method
for the neat asphalt heating and handling during the preparation of the sample specimens
are very basic. There is little potential for user error when conducting the conditioning
procedure as per the test method.
The only notable difficulty experienced during the testing was collecting the
RTFO residue from the bottles. The AASHTO specifications require that if the RTFO
residue is removed from the bottles via scraping, rather than allowing the residue to drain,
at least 90% of the residue must be collected from each bottle. Each of the eight RTFO
bottles are filled with 35 grams of neat asphalt prior to conditioning; therefore, 31.5
grams of the RTFO residue must be collected from each bottle when scraping, as per
AASHTO specifications. No minimum amount to be collected is specified when
allowing the RTFO-aged asphalt to free drain.
Collecting 90% of the RTFO residue by scraping proved to be very difficult.

Several different scraping tools and techniques were tried prior to the start of the testing
program; however, none of the attempts were successful in achieving the required
amount of residue from each bottle. The average amount of RTFO residue that was able
to be collected from each bottle by draining was approximately 24 grams, or 69 percent.
The average amount of RTFO residue collected from each bottle by scraping was
approximately 28 grams, or 80 percent.
The decision was made to proceed with the testing program and to note any
differences observed between the DSR tests conducted from drained RTFO residue and
from scraped RTFO residue. No significant differences were observed in the DSR test
results on residue collected by the different methods.
4.3 PRESSURE AGING VESSEL
The pressure aging vessel conditioning procedure yields no data to be evaluated.

69
The ATS pressure aging vessel and the NAPCO vacuum oven operation requires
minimal user input. Once the asphalt specimens are placed in the PAV, the unit software
controls the conditioning procedure. The operator must follow the AASHTO PP1
procedures for the RTFO-aged asphalt sample preparation as well as the vacuum oven
degassing procedures following the PAV conditioning. When these procedures are
followed, the potential for user error is minimal.
The ATS PAV consistently maintained the temperature and pressure range
requirements during conditioning, as per AASHTO criteria. Likewise, the NAPCO
vacuum oven functioned properly during the testing program. The vacuum oven
maintained the temperature and pressure during degassing, as per AASHTO precision
criteria.
4.4 ROTATIONAL VISCOMETER
Summaries of the test results for duplicate rotational viscometer temperature

sweep Tests 17 and 18 are shown in Table 4.2 and Table 4.3. The test results for single
temperature rotational viscometer Tests 19 through 25 are summarized in Table 4.4.
The test results in these tables are shown in SI units, with viscosity and shear
stress expressed as milliPascal-seconds (mPa⋅ s) and Newtons per square meter (N/m2 ),
respectively.
The equivalent centimeter ⋅ gram ⋅ second (CGS) units are as follows:
SI CGS
Viscosity 1 mPa ⋅s = 1 cP (centiPoise)
Shear Stress 1 N/m2 = 10 dyne/cm2
10 dyne•sec 
where: 100 cP =   = 10 P = 1 Pa • s
 cm 2 
70
Table 4.2 RV Temperature Sweep Test 17 Results
Test Actual Torque Shear Stress Shear Rate Viscosity Average

Temperature Temperature (%) (N/m2 ) (sec-1 ) (mPa·s) Viscosity
o o
( C) ( C) (mPa·s)
135 135.10 3.86 33.15 6.80 487.50
135 135.10 3.83 32.30 6.80 475.00 479.2
135 134.90 3.83 32.30 6.80 475.00
140 140.10 2.68 22.95 6.80 337.50
140 140.10 2.66 22.95 6.80 337.50 337.5
140 139.90 2.69 22.95 6.80 337.50
145 145.20 1.92 16.15 6.80 237.50
145 145.10 1.92 16.15 6.80 237.50 237.5
145 144.80 1.91 16.15 6.80 237.50
150 150.20 1.37 11.90 6.80 175.00
150 149.90 1.34 11.05 6.80 162.50 166.7
150 149.80 1.32 11.05 6.80 162.50
155 155.20 0.92 7.65 6.80 112.50
155 155.20 0.90 7.65 6.80 112.50 112.5
155 155.10 0.89 7.65 6.80 112.50
160 160.10 0.58 5.10 6.80 75.00
160 159.90 0.55 5.10 6.80 75.00 75.00
160 159.80 0.55 5.10 6.80 75.00
165 165.00 0.31 2.55 6.80 37.50
165 164.80 0.33 2.55 6.80 37.50 37.50
165 165.10 0.31 2.55 6.80 37.50
71
Table 4.3 RV Temperature Sweep Test 18 Results
Test Actual Torque Shear Stress Shear Rate Viscosity Average

Temperature Temperature (%) (N/m2 ) (sec-1 ) (mPa·s) Viscosity
(o C) (o C) (mPa·s)
135 135.10 4.16 35.70 6.80 525.00
135 135.10 4.09 34.85 6.80 512.50 512.5
135 135.10 3.98 34.00 6.80 500.00
140 140.10 2.76 23.80 6.80 350.00
140 140.20 2.73 22.95 6.80 337.50 341.7
140 140.10 2.74 22.95 6.80 337.50
145 144.90 2.00 17.00 6.80 250.00
145 145.20 1.99 17.00 6.80 250.00 250.0
145 145.10 1.99 17.00 6.80 250.00
150 150.00 1.45 12.75 6.80 187.50
150 150.10 1.44 11.90 6.80 175.00 179.2
150 150.20 1.41 11.90 6.80 175.00
155 155.20 1.01 8.50 6.80 125.00
155 154.80 1.00 8.50 6.80 125.00 125.0
155 154.80 0.98 8.50 6.80 125.00
160 160.00 0.69 5.95 6.80 87.50
160 159.90 0.69 5.95 6.80 87.50 87.50
160 159.90 0.67 5.95 6.80 87.50
165 165.10 0.43 3.40 6.80 50.00
165 164.80 0.41 3.40 6.80 50.00 50.00
165 164.80 0.42 3.40 6.80 50.00
72
Table 4.4 RV Single Temperature Tests 19 through 32 Results
Test Test Temp. Actual Temp. Torque (%) Shear Stress Shear Rate Viscosity Average
Number ( o C) ( o C) (N/m2) (sec-1) (mPa·s) Viscosity
(mPa·s)
19 135 135.00 4.01 34.00 6.80 500.00
135 135.10 3.94 33.15 6.80 487.50 491.67
135 135.10 3.92 33.15 6.80 487.50
20 135 135.10 4.06 34.85 6.80 512.50
135 135.10 4.02 34.00 6.80 500.00 504.17
135 135.00 3.98 34.00 6.80 500.00
21 140 140.10 2.83 23.80 6.80 350.00
140 139.80 2.81 23.80 6.80 350.00 350.00
140 139.90 2.78 23.80 6.80 350.00
22 140 139.90 2.71 22.95 6.80 337.50
140 139.80 2.70 22.95 6.80 337.50 337.50
140 140.00 2.67 22.95 6.80 337.50
23 145 145.00 2.03 17.00 6.80 250.00
145 145.30 1.98 17.00 6.80 250.00 250.00
145 145.10 1.97 17.00 6.80 250.00
24 145 145.10 2.10 17.85 6.80 262.50
145 145.00 2.06 17.85 6.80 262.50 262.50
145 144.80 2.06 17.85 6.80 262.50
25 150 149.80 1.95 17.00 6.80 250.00
150 149.90 1.96 17.00 6.80 250.00 245.83
150 150.10 1.94 16.15 6.80 237.50
26 150 150.30 1.45 12.75 6.80 187.50
150 150.20 1.43 11.90 6.80 175.00 179.17
150 150.00 1.40 11.90 6.80 175.00
27 155 155.20 1.63 13.60 6.80 200.00
155 154.80 1.61 13.60 6.80 200.00 200.00
155 154.70 1.62 13.60 6.80 200.00
28 155 154.90 1.76 15.30 6.80 225.00
155 155.30 1.74 14.45 6.80 212.50 216.67
155 155.30 1.71 14.45 6.80 212.50
29 160 160.10 1.37 11.90 6.80 175.00
160 159.80 1.33 11.05 6.80 162.50 166.67
160 160.10 1.33 11.05 6.80 162.50
30 160 160.00 0.57 5.10 6.80 75.00
160 159.90 0.57 5.10 6.80 75.00 75.00
160 159.90 0.55 5.10 6.80 75.00
31 165 165.00 0.27 2.55 6.80 37.50
165 164.80 0.26 2.55 6.80 37.50 37.50
165 164.90 0.26 2.55 6.80 37.50
32 165 165.10 0.42 3.40 6.80 50.00
165 164.80 0.45 3.40 6.80 50.00 50.00
165 164.80 0.42 3.40 6.80 50.00
73
The SHRP binder specifications for pumpability require a maximum binder

viscosity limit of 3 Pa⋅s when tested at 135 o C. As shown in Tables 4.2 through
Table 4.4, each viscosity reading at 135 o C is well below the 3 Pa ⋅s limit, thus meeting
the SHRP binder performance specifications.
There are no specification criteria for tests at temperatures greater than 135 o C.
However, higher temperature tests are required to establish equiviscosity curves. As
shown in Table 3.1, higher temperature testing was performed using both a single
temperature per sample and temperature sweep methodologies.
Review of the data in Tables 4.2 through Table 4.4 demonstrates the sweep
technique produces a decreasing viscosity with increasing temperature, as expected.
However, some of the results for the tests conducted at single temperatures do not display
expected behavior.
Test 25, which was conducted at 150 oC, appears to be erroneous since its
average viscosity is significantly higher than the average viscosity of the duplicate Test
26. The Test 25 average viscosity is also much greater than the average viscosity at
150 o C of duplicate temperature sweep Tests 17 and 18.
The average viscosities of duplicate Tests 27 and 28 are expectedly high as well.
The average viscosities of Tests 27 and 28, which were conducted at 155 o C, are
significantly higher than those from duplicate temperature sweep Tests 17 and 18 at
155 o C. Furthermore, the viscosity of asphalt at 155 o C should be less than the viscosity
at 150 o C; however, the average viscosity results of duplicate Tests 27 and 28 are greater
than the average viscosity of duplicate temperature sweep Tests 17 and 18 at 150 o C. The
average viscosity results of duplicate Tests 27 and 28 are also greater than the average
viscosity of Test 26, which was conducted at 150 o C.
Test 29 appears to be invalid as well. Its average viscosity is much greater than
the average viscosity of duplicate Test 30. In addition, the average viscosity of Test 29,
is significantly greater than the average viscosity of duplicate temperature sweep Tests 17
and 18 conducted at 160 o C.
74
A graphical display of the above observations is provided in Figure 4.1. The

average viscosities of temperature sweep Test 17 are plotted and connected to create a
viscosity-temperature curve. A viscosity-temperature curve for temperature sweep Test
18 is plotted on Figure 4.1 as well. As shown, both viscosity-temperature curves closely
follow each other. The average viscosity of single temperature Tests 19 through 32 are
plotted as data points only. Notice that Tests 19 through 24 and Test 31 and 32 produced
average viscosities which are consistent with the average viscosities produced by
temperature sweep Tests 17 and 18. Tests 25, 27, 28, and 29, however, produced
unexpectedly high average viscosities, as previously discussed.
Tests 27, 29, and 28, were the first three RV tests conducted, as shown in
Table 3.3. Therefore, it would be reasonable to assume that user error was a contributing
factor to the erroreous results. However, Test 25 was the last test conducted. Therefore,
user error with Test 25 is questionable.
A linear relationship between viscosity and temperature is desired for

equiviscosity curves. To achieve this linear relationship, viscosity is plotted, by
convention, on a log-log scale and temperature on a log scale. A log-log viscosity vs. log
temperature plot of the Test 17 results is shown in Figure 4.2. The results of Test 17,
however, do not exhibit the expected linear relationship. Thus, the results are suspect.
The precision criteria of AASHTO TP48 requires that the difference in duplicate RV test
results for a single operator in a single lab, expressed as a percent of their mean, shall not
exceed 7.3 percent. A precision comparison of the average viscosity results at each test
temperature for the duplicate RV temperature sweep Tests 17 and 18 is shown in Table
4.5. The average viscosity results from tests taken at and above 155 o C failed the
AASHTO precision criteria. As shown in Table 4.6, duplicate single temperature RV
Tests 25 and 26, Tests 27 and 28, Tests 29 and 30, and Tests 31 and 32, of which test
temperatures range from 150 o C and up, failed to meet the AASHTO precision criteria.
75
600
Test 20
500
Test 19
Test 18 Temp. Sweep
400
Average Viscosity, (mPa*s)
Test 21
Test 22
300
Test 24
Test 25
Test 23 Test 28
200
Test 26 Test 27
Test 17 Temp. Sweep Test 29
Curve
100
Test 32
Test 30
Test 31
0
135 140 145 150 155 160 165
o
Temperature, ( C)
Figure 4.1 RV Viscosity vs. Temperature Plot

77
76
10.0
5.0
3.0
2.0
Viscosity (Pa-s, log-log scale)
1.0
0.7
0.5
0.4
0.3
0.2
0.1
100 110 120 130 140 150 160 170 180 190 200
O
Temperature ( C, log scale)
Figure 4.2 RV Test 17 Log-Log Viscosity vs. Log Temperature Plot

77
Table 4.5 RV Temperature Sweep Tests Precision Results
Test 17 Test 18 Maximum

Average
Test Temp. Average Average Allowable
Viscosity % Results
(o C) Viscosity Viscosity %
Difference
(mPa·s) (mPa·s) Difference1
135 479.17 512.50 6.7 7.3 PASSED
140 337.50 341.67 1.2 7.3 PASSED
145 237.50 250.00 5.1 7.3 PASSED
150 166.67 179.17 7.2 7.3 PASSED
155 112.50 125.00 10.5 7.3 FAILED
160 75.00 87.50 15.4 7.3 FAILED
1 Per AASHTO TP 48 precision criteria
Table 4.6 RV Single Temperature Tests Precision Results
Test Test Temp. Average Average Maximum Results

Number (o C) Viscosity Viscosity Allowable
(mPa·s) % Difference % Difference1
19 135 491.67
2.5 7.3 PASSED
20 135 504.17
21 140 350.00
3.6 7.3 PASSED
22 140 337.50
23 145 250.00
4.9 7.3 PASSED
24 145 262.50
25 150 245.83
31.4 7.3 FAILED
26 150 179.17
27 155 200.00
8.0 7.3 FAILED
28 155 216.67
29 160 166.67
75.9 7.3 FAILED
30 160 75.00
31 165 37.50
28.6 7.3 FAILED
32 165 50.00
1
Per AASHTO TP48 Precision Criteria
78
The observation that the average viscosity results failed the AASHTO precision
criteria from test temperatures greater than 145 o C and 150 o C, for the individual tests and
temperature sweep tests, respectively, raised questions regarding the precision of the
Brookfield DV-III rotational viscometer. Therefore, the Brookfield Operating
Instructions manual was reviewed to determine the manner in which the DV-III
determines viscosity. The Brookfield DV-III viscometer determines the viscosity by first
measuring the percent torque required to rotate the spindle at a constant speed. The
torque is applied through a calibrated spring. The degree to which the spring is wound is
used to calculate the applied torque. The percent torque value is rounded to one tenth of
a percent and used in Equation 4-1 to calculate the viscosity.
 10000 
ν = ∗TK ∗ SMC  (TP ) (4-1)
 RPM 
where:
υ = viscosity, mPa•sec
RPM = Spindle speed
TK = Model spring constant
SMC = Spindle multiplier constant
Tp = Percent torque expressed in decimal form
The spindle multiplier constant, SMC, of Equation 4-1 is based on the spindle
type. The model spring constant, TK, of Equation 4-1 is based on the calibrated spring of
the unit. There are four different model series of the Brookfield DV-III: LV, RV, HA,
and HB model series. Each model series uses a different calibrated spring. An RV
model series Brookfield DV-III, RVDV-III, was used during this research.
The product of 
10000 
∗ TK ∗ SMC  in Equation 4-1 is referred to as the full-scale
 RPM 
viscosity range. The full-scale viscosity range for a RVDV-III equipped with a SC4-27
spindle and a 20-RPM operating speed is 12,500 mPa⋅s.
The shear rate is calculated as:

79
•
γ = RPM × SRC (4-2)
where:
•
γ = shear rate, sec-1
SRC = Spindle shear rate constant
Shear stress, τ, is calculated as:
τ = 100 (TK × SMC × SRC × Pt ) (4-3)
The Brookfield DV-III viscometer determines the percent torque and then uses
that value, rounded to the nearest tenth of a percent, to calculate the viscosity with
Equation 4-1. Based on Equation 4-1 and the measuring capabilities of the Brookfield
DV-III, a tenth of a percent change in torque yields a 12.5 mPa⋅s change in viscosity
when using an RV series viscometer with an SC4-27 spindle at 20 RPM. Thus, the
smallest change in torque, a tenth of a percent, will result in 12.5 mPa ⋅s incremental
changes in viscosity. AASHTO TP48 requires that the difference in duplicate test results
for a single operator, expressed, as a percent of their mean, shall not exceed 7.3 percent.
Since viscosity changes in increments of 12.5 mPa⋅s and the maximum difference
between test results shall not exceed 7.3%, the minimum viscosity difference between
two duplicate tests can be approximated as follows:
 Viscosity Difference 
Max. Allowable Percent Difference = 100   (4-4)
 Average Viscosity 
 12.5 mPa • sec 

7.3% =  
 Average viscosity 
It follows from Equation 4-4 that duplicate tests, conducted with the Brookfield
RVDV-III using a SC4-27 spindle at a 20-RPM operating speed, that produce average
viscosities of less than 171 mPa⋅ s will fail the precision and bias criteria of AASHTO
TP48. As shown in Tables 4.5 and 4.6, the duplicate tests on the neat PG 70-22 binder at
80
temperatures above 150 o C yielded average viscosities less than 171 mPa⋅s. As a result,
those tests failed the precision and bias criteria of AASHTO TP48.
Tests yielding viscosities lower than 171 mPa ⋅s, which meet the precision and
bias requirement of AASHTO TP48, can be conducted using the SHRP recommended
SC4-27 spindle by reducing the full- scale viscosity range indicated by Equation 4-1.
Reducing the full- scale viscosity range results in a smaller incremental change in
viscosity per each tenth of a percent change in torque, which, in turn, yields a lower
average viscosity that can be tested while meeting the precision and bias criteria of
AASHTO TP48, as per Equation 4-4.
The full- scale viscosity range can be reduced via two different means. Increasing
the spindle speed is one method for reducing the full-scale viscosity range. For example,
by increasing the spindle speed to 30 RPM, the full- scale viscosity range of an RV series
viscometer with a SC4-27 spindle is reduced to 8,333. Thus, a tenth of a percent change
in torque will result in an 8.333 incremental change in viscosity. The minimum average
viscosity that would meet the AASHTO precision and bias criteria for two duplicate test
results using a SC4-27 spindle at 30 RPM would be approximately 114 mPa ⋅s, which in
turn would allow tests to be conducted at higher temperatures.
Different spindles can also be used to reduce the full-scale viscosity range. An
RV series viscometer with SC4-21 spindle operating at 20 RPM, for example, has a full-
scale viscosity range of 2,500. The viscosity will change by 2.5 mPa ⋅s with each tenth of
a percent change in torque. As a result, the minimum average viscosity, which would
meet AASHTO precision and bias criteria, between two duplicate test results for a SC4-
21 spindle at 20 RPM would be approximately 68.5 mPa⋅s.
A different model series DV-III viscometer must be used if the required full-scale
viscosity range cannot be achieved by changing the spindle type and/or operating speed.
The viscometer must be selected with the appropriate spring constant to work in
conjunction with the spindle type and operating speed combination to achieve the desired
full-scale viscosity range.
As a result of the viscosity measuring method and capabilities of the Brookfield

DV-III, operators must be aware of the minimum viscosity limits, which in turn limit the
81
maximum test temperatures, in order to achieve the precision and bias criteria of
AASHTO TP48. These limits are dependant upon the asphalt binder grade, viscometer
spring cons tant, spindle type, and operating speed.
In addition to the DV-III viscosity measuring method and capabilities observed

during the review of the Brookfield Operating Instructions manual, it was also noted that
the range of viscosity measurements that sho uld be taken from the Brookfield DV-III
viscometer extends from a minimum viscosity of 10% of the full-scale range to a
maximum viscosity of 100% of the full-scale range. The accuracy of tests producing
viscosities outside of the 10% to 100% full-scale range cannot be guaranteed. Thus, the
minimum viscosity that can be reliably measured when using an RVDV-III with a SC4-
27 spindle at 20 RPM is 1250 mPa⋅s. The asphalt viscosities measured during the testing
at 135 o C ranged from 475 mPa⋅s to 525 mPa ⋅s. The viscosities yielded by these tests are
well below the 10% of full- scale range minimum allowable viscosity criterion of the
Brookfield Operating Instructions manual. Thus, the accuracy of all the rotational tests is
suspect.
An asphalt supplier was contacted to determine typical viscosity ranges of

PG 70-22 asphalt and with what type of spindle and speed their tests were conducted.
Typical viscosities of PG 70-22 at 135 o C range from 555 mPa ⋅s to 600 mPa⋅ s, as per the
supplier’s quality control data. The supplier uses a Brookfield RVDV-II viscometer; the
typical viscosities provided by the supplier were obtained using a SC4-21 spindle and a
20-RPM operating speed. A Brookfield technician was contacted to determine if the
Brookfield DV-II models used by the supplier have the same measuring and precision
capabilities as the DV-III model used for this research. The Brookfield technician
verified that the DV-II and DV-III viscometer models share the same measuring and
precision capabilities. In addition, the spring constant for the RV model series is the
same for both the DV-II and DV-III models (Thibodeau, 2003).
The allowable viscosity range using a SC4-21 spindle operating at 20 RPM is 250
to 2,500 mPa⋅s (i.e. 10% to 100% of the 2,500 full-scale viscosity range); therefore, the
viscosities at 135 o C obtained by the supplier are within the allowable limits of the
Brookfield RVDV-II. Although the results obtained during this research are close to the
82
typical viscosity range provided by the asphalt supplier, they are out of the accuracy
range of the RVDV-III with a SC4-27 spindle at the 20-RPM operating speed. Therefore,
using an SC4-27 spindle at 20 RPM with RV model series Brookfield viscometers should
be avoided when testing PG binders at temperatures that produce viscosities less than
1250 mPa ⋅s. When using an RVDV-II or RVDV-III to perform tests at 135 o C to verify
the pumping ability of PG 70-22 asphalt binder, spindle speeds of at least 45 RPM would
be needed in combination with the SC4-27 spindle in order to achieve viscosities above
the 10% full- scale range limit, as per the manufacturer’s criterion.
Based upon the above information, the combination of a SC4-27 spindle at 20

RPM is not applicable to all tests. The selection of spindle type and speed is dependant
upon the viscometer spring constant, asphalt grade and test temperatures. Unfortunately,
ambiguity on this detail exists in asphalt industry protocol, as discussed in Chapter 3. A
SC4-27 spindle is recommended in the SHRP literature (Petersen, et al., 1994b). A 20-
RPM operating speed is also recommended by the SHRP researchers in a different
volume of the final report (Anderson, et al., 1994). No recommendation was found that
united the spindle selection and operating speed. The SHRP researchers used a
Brookfield DV-II viscometer for their testing; however, the viscometer series (LV, RV,
HA, or HB) was not documented (Anderson, et al., 1994). A plot of viscosity-
temperature curves generated using capillary and Brookfield DV-II viscometers during
the SHRP research is shown in Figure 4.3. The Brookfield tests used to construct the
viscosity-temperature curve of Figure 4.3 were conducted using an SC4-27 spindle at 20
RPM. The test results are less than the minimum viscosity that can be reliably measured
using that spindle-operating speed combination with an RVDV-II viscometer, as shown
graphically in Figure 4.3 with the addition of the reliability limit. If the SHRP
researchers used an RVDV-II viscometer, then SHRP test results such as those in Figure
4.3 are suspect.
The use of incorrect spindle and operating speed combinations can lead to
erroneous equiviscosity curves. A linear relationship is expected creating log-log
viscosity vs. log temperature plots. This linear relationship allows for the development of
equiviscosity curves by conducting tests at only two temperatures, as per both the SHRP
and Asphalt Institute methods. When developing equiviscosity curves at only two
83
temperatures; however, this linear relationship is assumed. For example, consider

Figure 4.4, the log- log viscosity vs. log temperature plot of the temperature sweep
Test 17 data, which is considered suspect.since the tests were conducted outside the
measuring capabilities of the Brookfield RVDV-III due to the spindle and operating
speed combinatio ns used. If an operator was unaware of the importance of proper
spindle-operating speed selection and only plotted the Test 17 data points at 135 o C and
165 o C, as per the Asphalt Institute recommendations, the result would be an
equiviscosity curve that does not reflect the actual test data curve. In this case, the
assumption that the a linear relationship exists would be incorrect, thus leading to an
erroneous equiviscosity curve. Therefore, it is critical that the tests are conducted within
the measuring capabilities of the rotational viscometer to produce accurate equiviscosity
curves when testing at only two temperatures. Furthermore, testing at three or more
temperatures would also help verify the accuracy of the equiviscosity curves.
100
Brookfield
10
Viscosity, Pa's
Kinematic
+
1
+
0.1
Reliability limit = 1.25 Pa.s (min.) for Brookfield
RVDR-III models with SC4-27 spindle, 20 RPM
0.01
90 100 110 120 130 140 150 160 170 180 190 200
Temperature, °C
Figure 4.3 Viscosity Measured by SHRP Researchers with Capillary and Brookfield
Viscometers, SC4-27 Spindle at 20 RPM
84
10.0
5.0
3.0 Test 17 Data Curve

2.0
Viscosity (Pa-s, log-log scale)
Equiviscosity Curve as per

Asphalt Institute Protocol
1.0
0.7
0.5
0.4
0.3
0.2
0.1
100 110 120 130 140 150 160 170 180 190 200
O
Temperature ( C, log scale)
Figure 4.4 RV Temperature Sweep Test 17 Equiviscosity Curve

85
The concern that other research facilities and the asphalt industry could
misinterpret the existing testing protocol and other reference sources (i.e. AASHTO
standard specifications, Asphalt Institute publications, etc.) was justified following the
inquiry with the aforementioned asphalt supplier for their typical viscosities of PG 70-22
asphalt at 165 o C. The range of viscosities for PG 70-22 asphalt at 165 o C varied from
150 mPa ⋅s to 160 mPa⋅ s, as per the supplier’s quality control data. These results were
from tests conducted with an RVDV-II using a SC4-21 spindle operating at 20 RPM. As
described above, the minimum viscosity that can be reliably measured with an RVDV-II
viscometer using a SC4-21 spindle at 20 RPM is 250 mPa⋅s. Thus, the viscosity range
cited by the asphalt supplier is suspect since it is below the allowable limits of the
Brookfield RVDV-II viscometer.
The rotational viscometer test required minimal user input. The unit was simple
to calibrate and operate. Once the basic sample handling and preparation steps where
completed, the rotational viscometer software controlled the remainder of the test.
Therefore, user proficiency was quickly developed.
Problems with test temperature recording were experienced early in the trial
testing period. The test temperature recorded on the Thermosel unit did not match the
temperature displayed on the computer monitor, as determined by the Brookfield
software. The AASHTO specifications require that the temperature of the test specimen
be maintained to within ± 0.1 o C of the test temperature. Temperatures differences
between the Thermosel and the Brookfield software, however, were observed to be
greater than the ± 0.1 o C allowable difference. A Brookfield technician was contacted to
obtain guidance on the observed temperature discrepancy. The technician advised that
the computer display was correct; therefore, the Brookfield software temperature display
was used as the standard calibration and testing throughout the testing program.
As previously described, 1250 mPa ⋅s is the minimum viscosity that can be

reliably measured by the Brookfield RVDV-III when using an SC4-27 spindle at 20
RPM. The viscosity results obtained from the testing program; however, were much
lower than 1250 mPa ⋅s, ranging from 525 mPa⋅s to 37.50 mPa⋅ s, as shown in Tables 4.2
86
through Table 4.4. Therefore, the erroneous results of Tests 25, 27, 28, and 29 are
expected since their viscosities could not be reliably measured. Furthermore, none of the
tests results can be considered reliable.
Test Protocol and Specifications
The performance grading specification for pumpability of AASHTO MP1

requires that the viscosity at 135 o C must not exceed 3 Pa·s. It is noteworthy that a
material specification for viscosity-grading system exists for the viscosity at 135 o C as
well. The viscosity- grading system criterion, which is provided in ASTM 3381,
established minimum viscosities ranging from 0.08 Pa·s to 0.3 Pa·s at 135 o C, depending
on the asphalt cement grade. Presumably, this specification was established to control
asphalt drain-down during storage and transportation. A minimum specification is not
provided in the PG system; it only requires a maximum viscosity for pumpability. No
documentation was found in the literature to justify this fundamental change in
specifications.
Equiviscous temperature ranges have been used for the past 20 years for
determining laboratory mixing and compaction temperatures (Asphalt Institute, 2001).
The capillary tube viscometer was used with the Marshall mix design methodology for
determining equiviscous temperatures. The capillary tube viscometers tests were
conducted at 60 o C and 135 o C, which, in turn, required extrapolation to estimate mixing
and compaction temperatures that can exceed 150 o C. The AASHTO TP48 test
procedure for rotational viscometer testing recognizes the need for determining asphalt
equiviscous temperatures for mixing and compaction. However, the AASHTO test
method does not provide test temperature specifications for creating viscosity-
temperature profiles. The Asphalt Institute (2001) recommends determining asphalt
viscosities at 135 o C and 165 o C test temperatures, which allows for the determination of
equiviscous temperatures via interpolation. The Asphalt Institute recommendation is
appealing since it avoids extrapolation issues. However, the very low viscosity of asphalt
at 165 o C requires appropriate selection of the rotational viscometer spindle and test
speed.
87
4.5 DYNAMIC SHEAR RHEOMETER
A summary of the dynamic shear rheometer test results are shown in Table 4.7.
The results of the test are checked against SHRP performance specifications G*/sin δ and
G*sin δ, which were discussed in Chapter 2. Test 6 was the only test that failed to meet
the SHRP specifications; it did not achieve the minimum G*/sin δ requirement of 2.2 kPa
for RTFO-aged binder. A second test, Test 6b, was conducted from a reserve tin in
Sample Set 3. It failed to achieve the minimum G*/sin δ requirement as well. Test 6 was
the 7th of 16 DSR tests conducted, as shown in Table 3.3. As shown in Table 3.2, the
sample set from which Test 6 was conducted, Sample Set 3, was the last sample set to
undergo RTFO conditioning. Thus, user error with the RTFO as a contributing factor to
the failed test result is minimal; however, user error with the DSR is possible.
A comparative precision analysis between the duplicate test results is provided in

Table 4.8. The single-operator precision criteria of AASHTO TP5 requires that the
difference in duplicate DSR test results for a single operator in the same lab, expressed as
a percent of their mean, shall not exceed 9.5 percent when testing neat asphalt, 11.0
percent when testing RTFO-aged asphalt, and 22.4 percent when testing PAV-aged
asphalt. A comparison between Tests 5 and 6 was not applicable since Test 6 was not
valid. Duplicate Tests 35 and 36 were the only tests to fail the precision requirements of
AASHTO TP5. Test 36 is an outlier when compared to the results of Tests 33, 34, and
35; of which all were conducted on neat asphalt. Test 36 was the 5th DSR test conducted;
therefore, user error is a possible factor in the unexpected results.
88
Table 4.7 DSR Test Results
HIGH TEMPERATURE RANGE TEST RESULTS: RTFO-AGED BINDER

Test Frequency Final Strain Modulus, G* Phase Angle, δ G*/sin Minimum Test
Number (Rad/sec) Temp. Amplitude (kPa) (Degrees) δ (kPa) Allowable Status
(o C) (Percent) G*/sin δ
(kPa)
1 10.08 70 10.01 2.546 83.2 2.564 2.2 Passed
2 10.08 70 9.89 2.564 83.6 2.580 2.2 Passed
3 10.08 70 10.01 2.636 83.1 2.656 2.2 Passed
4 10.08 70 10.02 2.714 82.7 2.737 2.2 Passed
5 10.08 70 10.29 2.463 83.1 2.481 2.2 Passed
6 10.08 70 9.93 2.120 83.4 2.134 2.2 Failed
6b 10.08 70 9.98 2.019 84.3 2.029 2.2 Failed
7 10.08 70 10.06 2.410 83.3 2.427 2.2 Passed
8 10.08 70 10.06 2.548 83.0 2.567 2.2 Passed
INTERMEDIATE TEMPERATURE RANGE TEST RESULTS: PAV-A GED BINDER

Test Frequency Final Strain Modulus, G* Phase Angle, G*sin Maximum Test
Number (Rad/sec) Temp. Amplitude (kPa) δ (Degrees) δ (kPa) Allowable Status
(o C) (Percent) G*sin δ
(kPa)
9 10.08 28 1.03 3879 50.6 2999 5000 Passed
10 10.08 28 1.01 3722 50.5 2873 5000 Passed
13 10.08 28 1.01 3641 50.4 2804 5000 Passed
14 10.08 28 1.00 3226 50.4 2486 5000 Passed
HIGH TEMPERATURE RANGE TEST RESULTS: UNAGED BINDER

Test Frequency Final Strain Modulus, G* Phase Angle, G*/sin Minimum Test
Number (Rad/sec) Temp. Amplitude (kPa) δ (Degrees) δ (kPa) Allowable Status
(o C) (Percent) G*/sin δ
(kPa)
33 10.08 70 12.06 1.207 84.4 1.213 1.0 Passed
34 10.08 70 12.07 1.200 84.6 1.205 1.0 Passed
35 10.08 70 11.99 1.224 85.1 1.229 1.0 Passed
36 10.08 70 11.83 1.074 85.1 1.078 1.0 Passed
89
Table 4.8 DSR Tests Precision Results
HIGH TEMPERATURE RANGE TEST RESULTS ANALYSIS:

RTFO-AGED BINDER
Test G*/sin δ Average Maximum Results
Number (kPa) G*/sinδ Allowable
% Difference % Difference1
1 2.564
0.62 11 Passed
2 2.580
3 2.656
3.00 11 Passed
4 2.737
5 2.481
N/A
6 *Failed the minimum G*/sinδ criteria*
7 2.427
5.61 11 Passed
8 2.567
INTERMEDIATE TEMPERATURE RANGE TEST RESULTS

ANALYSIS: PAV-AGED BINDER
Test G*sin δ Average Maximum Results
Number (kPa) G*sin Z Allowable
9 2999
4.29 22.4 Passed
10 2873
13 2804
12.02 22.4 Passed
14 2486
1
HIGH TEMPERATURE RANGE TEST RESULTS ANALYSIS:
UNAGED BINDER
Test G*/sin δ Average Maximum Results
Number (kPa) G*/sin δ Allowable
33 1.213
0.66 9.5 Passed
34 1.205
35 1.229
13.09 9.5 Failed
36 1.078
1
90
The Bohlin Dynamic Shear Rheometer required significant training time to

become proficient in its use. The sample specimen preparation and operation of the
rheometer allowed for potential sources of user error. Trimming the asphalt specimen
required great care. The sample must consistently be trimmed such that the vertical sides
of the asphalt specimen are flush with edges of the parallel plates. The edges of the
asphalt specimen should slightly bulge out when the plates are set to the test gap position
if the specimen was trimmed correctly. Poor trimming technique can produce test
specimens with either excessive material or insufficient material resulting in erroneous
tests. Heating the trimming tool is a common technique to aid in the trimming of the
asphalt specimen. Heating the tool, however, requires careful attention. If the trimming
tool is heated too much, the asphalt specimen will bleed and flow from between the plates
while trimming, thus leaving an inadequate sized specimen for testing. Conversely,
heating the trimming tool too little could cause the asphalt specimen to stick to the
trimming tool, resulting in the asphalt specimen being pulled from between the plates,
reducing the asphalt specimen size as well. Both scenarios would produce erroneous test
results.
Setting the zero gap is a very critical procedure as well. As described in Chapter
2, the zero gap is achieved when the gap between the upper and lower plates is reduced to
the point when the plates just touch. The actual test gap obtained, 1 mm or 2 mm,
depending upon the type of asphalt tested, is relative to the zero gap. For example, if the
zero gap is incorrectly set, such that a 20 µm gap actually exists between the plates,
thenthe actual test gap will likewise be 20 µm greater than the target test gap. Given that
the target gap is a first-order term in the calculation of the complex modulus, a gap error
can produce erroneous test results. For example, if the target test gap is 2.00 mm and a
20 µm gap error is experienced, then a 1% error in the complex modulus will result.
There was initial concern with the craftsmanship of the Bohlin Mechanical
DSR II. The upper and lower plates could never be aligned perfectly such that the
periphery of both the plates were flush. The upper and lower plates always maintained a
fraction of a millimeter offset even after repeated attempts to align the plates. There was
91
also concern regarding what appeared to be significant amount of play observed in the
upper spindle. The upper spindle can be slightly moved horizontally when light pressure
is applied by the operator’s fingers. Horizontal movement of the spindle, which was
exaggerated by the misalignment of the upper and lower plates, was also observed when
the trimming tool was dragged along the parallel plate’s periphery. There was concern
that this undesired movement would adversely affect the precision of the trimming
procedure, which in turn could lead to erroneous tests.
Consistent test repeatability could not be achieved during the trial testing period.
Therefore, Dennis Burke of Citgo visited the lab to evaluate the equipment performance
and user technique. Mr. Burke did not express concern over the spindle movement and
plate misalignment. Mr. Burke recalibrated the temperature controller and overall
calibration with the standard viscosity fluid.
The repeatability problems persisted during Mr. Burke’s evaluation. Therefore, a

service technician was contacted to check the DSR. The technician evaluated the DSR
and reported that the play in the upper spindle was typical of the mechanical bearing type
Bohlin dynamic shear rheometers. The technician did state that the play in the upper
spindle was eliminated with the newer air bearing type rheometers. The technician
concluded that the play in the upper spindle and the misalignment of the plates were
minimal and would not affect the operation of the DSR. The technician checked the
electronic circuitry unit and made adjustments. The DSR performance greatly improved
following the Bohlin technician’s service. Consistent test repeatability was finally
achieved. Based on the results of the testing program, the DSR appears to be operating
correctly.
Effect of RTFO Sample Retrieval
RTFO samples may be collected from the conditioning bottles by draining or

scraping. Based on the data in Table 4.8, the average G*/sin δ value of the drained
samples (Tests 1-4) and scraped samples (Tests 7 and 8) were 2.634 kPa and 2.497 kPa,
respectively. These results are within the precision requirements for the tests. Hence, it
appears from these data that the method used to retrieve the samples from the RTFO
bottles does not impact subsequent test results.
92
Precision and repeatability of the DSR tests are dependent on proper trimming of
the asphalt specimen. The heating of the trimming tool is critical to a properly trimmed
asphalt specimen. Therefore, the addition of a standard protocol for trimming tool
heating to the test method would improve test consistency. The development of a
standard trimming tool would be beneficial as well.
The AASHTO TP5 test standards provide single-operator precision statements for
the DSR tests on original binder, RTFO-aged binder, and RTFO/PAV-aged binder. The
precision statements are established for “duplicate” tests. With regard to the tests on
conditioned asphalt, whether RTFO-aged or PAV-aged, the standard is silent on when the
split of the asphalt material for the creating the test specimens is to be performed. The
material for the test specimens could be split from the neat tank material, conditioned
separately, and then tested. Or, the material for the test specimens could be split from the
same container of conditioned material. The addition of a standard sequence to the test
method would eliminate the existing ambiguity.
4.6 BENDING BEAM RHEOMETER
A summary of the bending beam rheometer test results is presented in Table 4.9.
The results of each test are checked against SHRP performance specifications for creep
stiffness, S(t), and the m- value, m(t). All of the BBR tests passed the SHRP specification
for the m- value. Tests 11, 15, and 16 each failed the SHRP specification for creep
stiffness by exceeding the maximum allowable 300 MPa limit. Test 12 was first and only
test that met the SHRP specification for creep stiffness. However, the test temperature at
60 seconds was -11.2 o C, which differed from –12 o C the target temperature by more than
the allowable ± 0.1 o C difference, as per AASHTO TP1. The creep stiffness produced by
Test 12 was only 246.4 MPa while Tests 11, 15, and 16 produced creep stiffness results
ranging from 301.2 MPa to 322.3 MPa. It is doubtful that a temperature difference of
+1.0 o C would result in a stiffness reduction of over 54 MPa, which is the case when
93
Table 4.9 BBR Test Results
Test Temp. Load, Deflectio Creep Max. S(60) m- Min. m(60)

No. 1 (o C) P n, Stiffness Allow. Test value, Allow. Test
(mN) δ , S(60)2 Results m(60) m(60)1 Results
(mm) S(60) (MPa)
(MPa)
11 -12.4 10.01 2.546 322.257 300 Failed 0.311 0.300 Passed
2
12 -11.2 9.89 2.564 246.392 300 Passed 0.329 0.300 Passed
4
15 -12.2 10.01 2.636 301.202 300 Failed 0.308 0.300 Passed
6
16 -12.5 10.02 2.714 302.951 300 Failed 0.301 0.300 Passed
1
Fluid bath temperature at 60 seconds; Target test temperature = -12
o
C
2
Per AASHTO MP1
Specifications
comparing Test 12 results to Tests 11, 15 and 16. Therefore, the BBR results for creep
stiffness are suspect.
A comparative precision analysis between the duplicate test results is provided in

Table 4.10. The standard specification AASHTO TP5 for single-operator precision
requires that the difference in S(60) and m(60) for duplicate BBR test results for a single
operator in the same lab, expressed as a percent of their mean, shall not exceed 9.1
percent and 4.0 percent, respectively. The S(60) and m(60)results for duplicate Tests 11
and 12 each failed the AASHTO precision criteria. The comparisons, however, are not
valid due to the test temperature problems encountered with Test 12. Duplicate Tests 15
and 16, however, passed the AASHTO precision criteria for both S(60) and m(60).
It is not certain why the BBR tests failed to meet the SHRP performance
specifications; several possibilities do exist however. One issue to consider is the pre-
conditioning procedures the asphalt specimens must undergo before testing. It is possible
that errors in the RTFO or PAV conditioning would be reflected in the BBR test results.
As discussed previously, the RTFO and PAV conditioning procedures are relatively easy
to conduct with minimal sources of error. Furthermore, the DSR tests on PAV-aged
94
Table 4.10 BBR Tests Precision Results
Test Creep Average S(60) Maximum Test Results

Number Stiffness,S(60) % Difference Allowable
(MPa) % Difference1
11 322.2572 26.68 9.1 Failed

12 246.3924
15 301.2026 0.58 9.1 Passed
16 302.951
Test m- value,m(60) Average m(60) Maximum Test Results

Number % Difference Allowable
% Difference1
11 0.311 5.63 4 Failed
12 0.329
15 0.308 2.30 4 Passed
16 0.301
1
specimens all passed the AASHTO criteria. Still, although unlikely, the possibility for
user error with the conditioning procedures cannot be discounted. There are definitely
potential sources of user error and equipment error associated with the BBR unit to
consider. These sources are discussed in detail in the following section.
The ATS bending beam rheometer operation and sample specimen preparation
and handling required a significant amount of time to develop user technique. The nature
of specimen beam preparation and handling provides several opportunities for user error.
The calculated stiffness of the beam is proportional to the thickness, raised to the third
power (Petersen, et al., 1994b). Therefore, it is critical that the prepared specimen beams
be to the exact dimension required and that the shape of the beam is not distorted during
demolding. Care must be taken to ensure that air pockets do not form while pouring the
95
PAV-aged residue into the specimen molds or erroneous tests may result. The operator
must ensure that the fluid bath remain clean. The specimen beams become very brittle
when submerged in the BBR fluid bath, which often results in fragments being collected
in the fluid bath. These fragments can collect on the beam supports and loading head and
lead to erroneous deflection results.
Significant difficulty was experienced when trying to calibrate linear variable

differential transformer (LVDT) of the BBR. The LVDT is calibrated via a factory-
calibrated stepped thickness gage. The stepped thickness gage is a disk with 4 slots; each
slot has a different calibrated thickness. Prior to BBR testing, a calibration verification
procedure is conducted by placing the stepped thickness gage in the BBR loading frame
and a thickness reading is taken at each of the 4 slots. The deviation between the
measured thickness and the actual gage thickness at each slot is recorded. The total
deviation is calculated and checked against the manufacturer’s LVDT precision
specifications. If the total thickness deviation is greater than the precision specification, a
calibration procedure, that is very similar to the calibration verification, is conducted
using the same stepped thickness gage.
The slots in the disk are approximately 12 mm wide. It was observed that the
thickness varied across the each slot. It is recommended by the ATS to always take
thickness readings at the same location on the disk: either at the center or the same edge
of the slot. The manufacturer’s direction was followed; however, the calibration of the
LVDT could never be achieved. Therefore, the BBR loading unit and the stepped
thickness gage were sent back to the ATS for investigation. No defect was found with
the loading unit and LVDT. The stepped thickness gage, however, was observed to be
out of specification and required additional milling.
The steps taken by ATS did not eliminate the problem; the LVDT calibration still
could not be verified consistently. Therefore, the location of thickness readings in each
slot was selected as necessary to “force” a successful LVDT calibration during testing.
While this procedure is not the correct calibration methodology, it was necessary to
proceed with the testing program. It is possible that the problems encountered with the
BBR test results are a consequence of the LVDT calibration difficulties.
96
As with the AASHTO test method for the DSR tests, the single-operator precision
statement for the BBR tests is silent on when the split of the asphalt material for the
creating the test specimens is to be performed. A standard sequence should be provided
in the test method as well. The AASHTO test standard specification for a suitable fluid
bath should be more specific. A standard fluid bath specification would be beneficial.
97
CHAPTER 5
CONCLUSION AND RECOMMENDATIONS
5.1 CONCLUSIONS
The objective of this research, to critically evaluate the SHRP binder testing
equipment in the Asphalt Technology Laboratory at WVU and the corresponding
standard test methods, was accomplished. The equipment calibration, operation, and
testing procedures were evaluated and the results documented. An evaluation of the
equipment precision and repeatability, as per the applicable AASHTO precision and bias
statements, was conducted. Equipment problems were identified and corrected when
possible. Potential sources of equipment and operator error were also identified and
documented.
The Despatch rolling thin film oven, the Applied Test Systems, Inc. pressure
aging vessel, and the NAPCO vacuum oven performed satisfactorily during this research.
However, their performance is more difficult to assess since they only condition asphalt
cements; no direct test data is produced to evaluate. The dynamic shear rheometer tests
provided good results when testing both RTFO-aged and PAV-aged asphalt specimens.
The bending beam rheometer test results on PAV-aged asphalt specimens were poor.
However, the performance of the BBR is suspect; therefore, it is doubtful that the poor
results are due to erroneous RTFO or PAV conditioning procedures.
No significant sources of equipment or operator error were identified with the

pressure aging vessel or the vacuum oven. The scraping procedure for RTFO residue
collection, however, proved to be problematic. As discussed in Section 4.2, the
AASHTO T 240 requirement of obtaining at least 90% of the RTFO residue when
scraping the RTFO bottles could never be achieved. No literature was found to document
the need for the scraping requirement or the consequences for not obtaining this result. In
addition, the AASHTO standards do not specify a minimum amount of RTFO residue
that must be collected when draining the bottles in lieu of the scraping procedure. The
DSR test results from this research showed no difference between the draining and
scraping RTFO-residue collection techniques. If the 90% retrieval specification for
98
scraping is maintained, a standard tool should be developed and specified in the

AASHTO standards.
The Brookfield RVDV-III rotational viscometer was easy to operate with simple
test sample preparation procedures. No significant sources of operator error were
identified. The Brookfield RVDV-III provided good repeatability, with the exception of a
few erroneous tests.
Problems were encountered with the test temperature display on the Brookfield.
The test temperature displayed on the Thermosel temperature controller consistently
differed from the temperature displayed on the computer monitor, as determined by the
Brookfield software, by more than the allowable variance in the test method. As per the
manufacturer’s verbal instruction, this problem was “resolved” by ignoring the
temperature displayed by the temperature controller.
The AASHTO precision and bias criteria were never achieved on several tests due
to spindle selection and test speed. Furthermore, all the tests were conducted outside the
measuring capabilities of the unit, again, due to the spindle selection and test speed used
throughout the testing program. The error in spindle/speed selection was due, in large
part, to the ambiguity that exists on the topic between the testing protocol and related
reference sources (i.e. AASHTO standard specifications, Asphalt Institute publications,
SHRP publications). The testing protocol should be altered with respect to spindle and
test speed selection to ensure the equipment is operated in a manner consistent with the
measurement capability of the unit and the precision required from the test.
Another problem noted with the AASHTO test method is the absence of test
temperature criterion for conducting tests for equiviscous temperatures. Inclusion of
standard test temperatures into the test method would be beneficial.
The dynamic shear rheometer and bending beam rheometer tests were the most
difficult to perform. Both the DSR and BBR tests required significant training time to
develop the skills necessary for proper specimen preparation and unit operation. Proper
sample preparation is critical due to the nature of the tests; therefore, user skill is a key
component to the precision and repeatability of both tests.
99
Although there was an initial difficulty in producing repeatable test results, the
Bohlin Instruments DSR II mechanical dynamic shear rheometer achieved satisfactory
results following service from the manufacturer and continued to perform well
throughout the testing program. The performance of the ATS bending beam rheometer,
on the other hand, was poor. The BBR failed to produce acceptable results. The poor
performance of the ATS BBR could be a result of the calibration difficulties experienced
during the research. The linear variable differential transformer could never be properly
calibrated even following service from the manufacturer. The calibration problems need
to be resolved before a complete evaluation of the unit can be obtained.
The standard specifications and test protocol for the DSR and BBR were
satisfactory for the most part; however, they do not provide a standard sequence for the
splitting of conditioned asphalt for creating DSR and BBR test specimens. The addition
of a standard sequence would be advantageous. The addition of standard trimming tool
and a standard method for heating the trimming tool to the DSR testing protocol would
be beneficial. The AASHTO test standard for a suitable fluid bath for the BBR should be
more specific. A standard fluid bath composition should be specified.
5.2 RECOMMENDATIONS
A level of confidence in the performance of the Despatch rolling thin film oven,
the Applied Test Systems, Inc., pressure aging vessel, the NAPCO vacuum oven, and the
Bohlin Instruments DSR II mechanical dynamic shear rheometer was achieved by the
conclusion of the research such that the units should be considered reliable and ready for
continued research and industry testing.
The evaluation of the Brookfield RVDV-III rotational viscometer should

continue. Additional tests should be conducted on PG 70-22 asphalt binder, as well as
other PG grades, using the appropriate spindle type and test speed such that the tests are
done within the measuring capabilities of the unit and the precision requirements of the
test method.
Exhaustive efforts were made during the research to bring the calibration of the
Applied Test Systems, Inc. bending beam rheometer into compliance. However,
100
calibration of the unit could never be achieved. The complete unit should be sent back to
the manufacturer for their assessment of the calibration problems. A further evaluation of
the Applied Test Systems, Inc. bending beam rheometer is not recommended until the
calibration problems are resolved.
The shortcomings of the testing protocol, as summarized in the Conclusions

section of this research, should be addressed. The suggested improvements to the test
method would provide a more consistent and comprehensive protocol and eliminate
ambiguity.
101
REFERENCES
Anderson, D. A., Christensen, D.W., Bahia, H.U., Dongre, R., Sharma,

M.G., Antle, C.E., Button, J., “Binder Characterization and Evaluation;
Volume 3: Physical Characterization,” Strategic Highway Research
Program Report No. SHRP-A-369, National Research Council, Washington,
D.C., 1994.
Bahia, H.U., Anderson, D.A., “Strategic Highway Research Program Binder
Rheological Parameters: Background and Comparison with Conventional
Properties,” Transportation Research Record 1488, TRB, National Research
Council, Washington, DC, pp. 32-39, 1995a.
Bahia, H.U., and Anderson, D.A., “The Development of the Bending Beam
Rheometer; Basics and Critical Evaluation of the Rheometer,” Physical
Properties of Asphalt Cement Binders: ASTM STP 1241, John C. Hardin,
Ed., American Society for Testing and Materials, Philadelphia, PA., pp. 28-
50, 1995b.
Bahia, H.U., and Anderson, D.A., “The Pressure Aging Vessel (PAV): A
Test to Simulate Rheological Changes Due to Field Aging,” Physical
Properties of Asphalt Cement Binders: ASTM STP 1241, John C. Hardin,
Ed., American Society for Testing and Materials, Philadelphia, PA., pp. 67-
88, 1995c.
http:// www.asphaltinstitute.org/superpav/chart.jpg, accessed 9/23/2003.
Huang, Y.H., Pavement Analysis and Design, Prentice Hall, Englewood
Cliffs, N.J., 1993.
“Laboratory Mixing and Compaction Temperatures,” Asphalt Institute
Technical Bulletin,
http://www.asphaltinstitute.org/superpav/mixcompt.html, accessed 8/19/01.
Petersen, J.C., Robertson, R.E., Branthaver, J.F., Harnsberger, P.M., Duvall,
J.J., Kim, S.S., Anderson, D. A., Christensen, D.W., Bahia, H.U., “Binder
Characterization and Evaluation; Volume 1,” Strategic Highway Research
Program Report No. SHRP-A-367, National Research Council, Washington,
D.C., 1994a.
Petersen, J.C., Robertson, R.E., Branthaver, J.F., Harnsberger, P.M., Duvall,

J.J., Kim, S.S., Anderson, D. A., Christensen, D.W., Bahia, H.U., Dongre,
R., Antle, C.E., Sharma, M.G., “Binder Characterization and Evaluation;
102
Volume 4: Test Methods,” Strategic Highway Research Program Report No.

SHRP-A-370, National Research Council, Washington, D.C., 1994b.
Roberts, F.L., Kandhal, P.S., Brown, E. Ray, Lee, D., Kennedy, T.W., Hot
Mix Asphalt Materials, Mixture Design and Construction, NAPA Research
and Education Foundation, Lanham, MD, 1996.
Thibodeau, L., Brookfield Engineering Laboratories, Inc., Verbal
communication via telephone, 11/18/2003
Whorlow, R.W., Rheological Techniques, Ellis Horwood Limited, West
Sussex, England 1992.

Aashto Codes2

Uploaded by

Copyright:

Available Formats

Aashto Codes2

Uploaded by

Document Information

Original Title

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

Aashto Codes2

Uploaded by

Copyright:

Available Formats

Evaluation of Performance Graded Asphalt

Binder Equipment and Testing Protocol

Asphalt Technology Program

Department of Civil and Environmental Engineering

Morgantown, West Virginia

Technical Report Documentation Page

1. Report No. 2. Government 3. Recipient's catalog No.

7. Author(s) 8. Performing Organization Report No.

1.1 History of Asphalt Binder Grading Methods 1

1.2 Problem Statement 3

1.3 Research Objectives 3

1.4 Scope and Limitations 4

1.5 Organization of the Report 5

Chapter 2 Literature Review and Background 6

2.2 Asphalt Rheology 6

2.3 Penetration Grading System Specifications 9

2.4 Viscosity Grading System Specifications 10

2.5 Performance Graded Binder Specifications 13

2.6 Rolling Thin Film Oven 16

2.7 Pressure Aging Vessel 19

2.8 Rotational Viscometer 22

2.9 Dynamic Shear Rheometer 28

2.10 Bending Beam Rheometer 44

Chapter 3 Research Approach 55

3.2 Testing Program Preparation 55

3.1 Testing Program 58

Chapter 4 Test Results Analysis and Equipment Evaluation 67

4.2 Rolling Thin Film Oven 67

4.3 Pressure Aging Vessel 68

4.4 Rotational Viscometer 69

4.5 Dynamic Shear Rheometer 87

4.6 Bending Beam Rheometer 92

Chapter 5 Conclusion and Recommendations 97

Figure 2.1 Mechanical Response of Elastic, Viscous, and Viscoelastic Materials.............7

Figure 2.2 Despatch Rolling Thin Film Oven ..................................................................17

Data Collection ..................................................................................................................17

Figure 2.3 ATS Pressure Aging Vessel ............................................................................20

Figure 2.4 Typical Equiviscosity Curve for Asphalt Binder ............................................23

Figure 2.5 Brookfield DV-III Rotational Viscometer.......................................................25

Figure 2.6 Bohlin Instruments DSR II Mechanical Dynamic Shear Rheometer..............30

Figure 2.7 Configuration and Load Cycle of Dynamic Shear Rheometer........................34

Figure 2.10 Graphical Description of Linear Viscoelastic Region...................................40

Figure 2.11 ATS Bending Beam Rheometer ....................................................................47

Figure 2.12 Graphical Definition of Creep Rate, m-value ................................................51

Figure 2.13 Development of Time-Temperature Shift Factors ........................................52

Figure 3.1 Testing Program Summary..............................................................................60

Figure 3.2 Testing Program Summary – continued ..........................................................61

Figure 4.4 RV Temperature Sweep Test 17 Equiviscosity Curve ....................................84

Table 2.3 Performance Graded Asphalt Cement Grades ..................................................14

Table 2.4 Summary of AASHTO MP1 Requirements .....................................................14

Table 2.5 Target Shear Stress and Strain Values ..............................................................31

Table 3.1 Testing and Conditioning Sequence .................................................................59

Table 3.2 Sample Set Conditioning Sequence ..................................................................65

Table 3.3 Testing Sequence ..............................................................................................66

Table 4.1 RTFO Mass Loss Results .................................................................................67

Table 4.2 RV Temperature Sweep Test 17 Results ..........................................................70

Table 4.3 RV Temperature Sweep Test 18 Results ..........................................................71

Table 4.4 RV Single Temperature Tests 19 through 32 Results........................................72

Table 4.5 RV Temperature Sweep Tests Precision Results..............................................77