THE USE OF RED TROPICAL SOILS AS POZZOLANAS:

REACTIONS, PRODUCTSAND PROPERTIES

by

Charles Amobi Nwakanma, B. Eng (Civil)

A thesis submitted in fulfilment of the requirements

for the degree of Doctor of Philosophy

Department of Civil Engineering

The University of Leeds

April 1979
 1.

ACIWOWLIDGMMS

The author wishes to acknowledge the very valuable help

rendered by the following towards the production of this, thesis:

Mr. J. G. Cabrera, Dipl. Ing., Dip. Dev.Adm., C. Eng., M. I. M.M,

F. R. M. S. not only for his supervision of the work and encourage-

ment throughout the period of the research, but for providing the

samples and part of the data.

Professor A. M. Neville, the former Head of Civil Engineering

Department and Professor A. R. Cusens, B. Sc., Ph. D., C. Fng.,

F. I. C. E., F. I. Struct. E., F. R. S. E. the present Head of Department

who allowed the use of facilities in the Department.

The academic staff and technicians of the Civil Engineering

Department. The following deserve special mention:

Messrs. R. Wood for helping to provide for the chemical analysis,

R. Duxbury who prepared the scanning electron microphotograph,

G. Cardwell for preparing the figures and Mrs. J. Alwnanah for

typing the manuscript.

The Department of Earth Sciences for providing the standard

samples used in the quantitative determination of the reaction

products and facilities for x-ray fluorescence.

The Scientific Division of the Central Electricity Genera-

ting Board, Harrogate, for generously permitting the use of their

atomic absorption spectrophotometer and Joel scanning electron

i
microscope.
The Association of Commonwealth
Universities for awarding
 ii.

the scholarship under which the research was undertaken and the
British Council for their efficient administration of this award.
Research colleagues who directly or, indirectly assisted,

especially Mr. K. Marquaye who helped with the computer search

program.

Finally his family who from thousands of miles away, in

Nigeria, provided that which none of you was capable. of.

. .,

t
 111.

ABSTRACT

Nineteen red tropical soils from the States of Paraiba and

Pernambuco which are located in the North East of Brazil were

treated with high calcium lime, they reacted showing a wide range

of variation of maximum unconfined compressive strengths with -

4.5% or 6.0% lime generally required to produce maximum strength.

The reaction between the soil and lime is mainly a diffusion

controlled process. Most of the reaction takes place within

five to seven days and the reaction within this period obeys
Jander's diffusion equation for solid state reactions. Based on

the constant rate of reaction obtained from this equation and on

the quantity of the reactive components of the soil, an activity'

index, Ai, is proposed and its validity tested against the uncon-

fined compressive strength of the soil-lime system. A valid

statistical correlation is presented linking Ai value with the

empirical reactivity parameter of Thompson.

Identification of the types of the reaction products by

x-ray diffraction which is greatly aided by the use of a semi-

automated computer search program reveals that the soils form

cementitious reaction products varying from dense and close

textured calcium silicates of the tobermorite-like type through

alumino-silicates to calcium alumino-ferri-silicates with

relatively larger voids when the soils are treated with lime and

cured for two years at 22°C. The increase in strength of the

soil-lime system not only depends on the volume of the reaction

products but also on their types and purity.

 iv,

Other constituents of the soil-lime system affect the

strength to varying degrees. While the increase in the amount of

unreacted material at the expense of cementitious reaction

products is not desirable, an amount of unreacted material is

essential for other engineering requirements other than strength

to be satisfied. Unlike other cemented composites, the strength

of the soil-lime system is not related to the overall porosity

of the system.
 v'.

'TABLE'0F CONTENTS

PAGE NO.

CHAPTER1 INTRODUCTION 1

CHAPTER2 LITERATUREREVIEW
GENERAL 4

2.1 Genesis and Composition 4

2.2 Classification 7

2.3 Physical and engineering

characteristics 9

2.3.1 Texture 9

2.3.2 Atterberg Limits 10

2.3.3 Density and compaction
characteristics 11

2.3.4 Strength characteristics 13

2.3.5 Swelling, permeability and

consolidation 14

2.4 Fozzolanic properties of red tropical

soils 15

2.4.1 Red tropical soils as pozzolanas 15

2.4.2 Soil-lime reactions 15

2.4.2.1 'Manifestations of 'the

reactions 15

2.4.2.2 Ameliöration in clays 16

2.4.2.3 Amelioration in red

tropical soils 20

2.4.2.4 The reaction process 23

2.4.2.5 The reaction products 25

2.4.3 Engineering characteristics of

lime treated red tropical soils 27

2.5 Chapter Summary 28

 vi.

PAGE NO.

CHAPTIIZ3 SCOPEOF THE INVESTIGATION 30

CHAPTER4 PROPERTIESOF THE MATERIALS USED IN THE
INVESTIGATION 32

4.1 Introduction 32

4.2 Location and environmental characteristics

of the area 32
4.3 Sampling 32

4.4 Chemical and mineralogical composition 37

4.5 Morphology and distribution of minerals 42

4.5.1 Kaolinite 42

4.5.2 Silica 45

4.5.3 Alumina 45

4.5.4 Iron minerals 46

4.6 Physical and engineering properties 46

4.6.1 Strength determination 46

4.6.2 Lime reactivity 49

4.7 Lime used 49

CHAPTER5 POZZOLANICACTIVITY AND MECHANISMOF REACTION

OF RED TROPICAL SOIL-LIME SYSTEMS 51

5.1 Introduction 51

5.2 Review of pertinent literature and

methods 51.

5.2.1 Ability of-soils to react with

lime 51

5.2.2 Mechanisms of reaction - the

reaction models 54

5.3 Experimental method 57

5.4 Measurement by atomic absorption

spectrophotometer 58

5.4.1 Basic principles 58

 VM,

PAGE ID.

5.4.2 Features of an atomic absorption

spectrophotometer 61

5.4.3 Operation of a spectrophotometer 62

5.4.4 The atomic absorption spectrophoto-

meter used 62

5.4.5 Calcium determination 63

5.5 Results and discussion 64

5.5.1 Lime consumption 64

5.5.2 Mechanism of reaction 71

5.5.3 Pbzzolanic activity 83

5.5.4 Activity index and unconfined

compressive strength 84

5.5.5 Activity index vs Lime reactivity 89

5.5.6 Activity index and classification 90

5.6 Uses of and using the activity index 92

5.7 Chapter Summary 95

CHAPTER6 RED TROPICAL SOIL-LIME REACTION PRODUCTS 97

6.1 Introduction 97

6.2 Pertinent literature and methods 98

6.2.1 Identification of reaction products 98

6.2.2 Quantification of the-reaction

products 100

6.2.3 Reaction products and strength 104

6.3 Experimental methods 107

6.3.1 Chemical analysis 107

6.3.2 Qualitative x-ray diffraction 108
6.3.3 Quantitative x-ray diffraction 115

6.3.3.1 The internal standard

used 115
 viii.

PAGE NO.

6.3.3.2 Standard minerals used 116

6.3.3.3 d-spacing used for peak
measurement in the test
samples - 116

6.3.3.4 Procedure 116

6.4 Results and discussions 121

6.4.1 Identification of reaction

products 121

6.4.2 The identified products 122

6.4.3 Consecutive and multiple reactions 124

6.4.4 Volume of reaction products vs

strength 134
.
6.4.5 Overall C/S, A/S ratios vs
strength 138

6.4.6 The relation between the activity

index and volume of reaction
products 139

6.4.7 Activity index vs C/S ratio 147

6.5 Limitations of the methods used 147

6.6 Chapter Summary 147

CHAPTER7 SEMIAU1tvIATED IDENTIFICATION OF CRYSTALLINE

REACTION PRODUCTS 1S0

7.1 Introduction 150

7.2 Review of literature and methods 150

7.2.1 The d-spacings and intensities 151

7.2.2 Error width or window 151

7.2.3 Data base 152

7.2.4 The "elements present" option 153

7.2.5 The program procedure 153

7.3 Features of the semi-automated search

used - 154
 ixt

PAGE NO.

7.3.1 The automated search 154

7.3.1.1 The data base 155

7.3.1.2 The input data 156

7.3.1.3 Design of the search

program 157

7.3.1.4 The output 157

7.3.2 Manual interpretation 163

7.4 Running the package 163

7.5 Test search identification 166

7.6 Chapter 9iwmary 166

CHAPTER8 SCANNINGELECTRONMICROSCOPYOF THE SOIL-LIME

SYSTIMS 168

8.1 Introduction 168

8.2 Basic principles and the design of the

SEN 168

8.3 The SFMs used in the investigation 169

8.4 Sample preparation 170

8.5 Results and discussion 170

8.6 Chapter' Summary 193

OF REDTROPICALSOIL-LIME
CHAPTER9 OTHERCONSTITUENTS
SYS IMS 195

9.1 Introduction 195

9.2 Literature review and methods 195

9.2.1 Calcium carbonate 195

9.2.2 Unreacted lime 196

9.2.3 'Porosity 200

9.3 Thermogravimetry 202

9.3.1 Introduction 202

9.3.2 Basic principles 203

9.3.3 Instrument design 204

 X.

PAGE NO.

9.3.4 Theni gravimetric-analysis of

soils and clays 205
9.4 Experimental. methods 206

9.4.1 Calcium carbonate 207

9.4.2 ' Unreacted lime 207

9.4.3 Porosity 208

9.4.4 Unreacted soil 208

9.4.5 Thermogravimetric analysis of

lime treated red tropical soils 208
9.5 Results and discussion 211

9.5.1 Calcium carbonate by gas evolution 211

9.5.2 Unreacted lime-, by extraction 214

9.5.3 Unreacted soil 223

9.5.4 Porosity vs strength 225

9.5.5 lhermogravimetry and Derivative
Thermogravimetry 225

",' 9.6 Chapter Summary 233

CHAPTER10 CONCLUSIONS 234

CHAPTER11 SUGGESTIONSFOR FUTURERESEARCH 238

REFERENCES 240

APPENDIXA. 1 X-ray diffraction traces 250

A. 2 Powder x-ray, data for calcium salts formed 257

APPENDIXB. 1 Rapid determination of carbon dioxide 263

B. 2 Calibration of carbon dioxide tube 266

B. 3 ASIM method for free lime determination 269

B. 4 Modified Franke method for free lime

determination 271

B. 5 Solvent Variation and Time Variation Methods

for free lime determination 274

B. 6 TG and DTG curves of the soils 276

 xt.

LIST OF PLATES PAGE D.

8.1 Electron micrograph of soil TH showing dense

microstructure 173

8.2 Electron micrograph of soil TH showing inter-

locking rods and plates of C-S-H 175

8.3 Electron micrograph of soil TH showing dense

and hexagonal gehlenite hydrate 177

8.4 Electron micrograph of soil TH revealing

unreacted material embedded in cementitious
reaction products 179

8.5 Electron micrograph for soil All showing massive

plates of C-S-H 181

8.6- Electron micrograph of calcium-alumino-ferri-

silicates formed by soil JII 183

8.7 How the reaction products of plate 8.6 bonds

particles together 185

8.8 Poorly crystallised reaction products formed by

JII 187
8.9 Energy dispensive elemental analysis for calcium,
silicon and aluminium 189

8.10 Energy dispensive elemental analysis for calcium,

silicon, aluminium and iron 191

9.1 Multiple extraction of unreacted lime 210

 X11.

LISTS OFFIGURES PAGEW.

4.1 Area of research in. relation to the world 34

4.2 Details of area of research 36

5.1 Relations between consumption of lime and time 68

5.2 Relations between consumption of lime and time 70

5.3 Relations between first term of Jander's equation

F(x) and time of reaction up to 5 days 76

5.4 Relations between first term of Jander's equation

F(x) and time of reaction up to 5 days 78

5.5 Relations between the first term of Jander's

equation F(x) and time of reaction up to 28 days
for four of the soils 80

5.6 Relation between activity index and the unconfined

compressive strength 88

5.7 Relation between activity index and the lime

reactivity parameter of Thompson 88

6.1 X-ray diffraction traces for NF 112

6.2 X-ray diffraction traces for SIA 114

6.3 Variation of peak intensity with density or

packing of the soil 120

6.4 Calibration curve for standard minerals 120

6.5 Relation between volume of reaction products

and total strength 136

6.6 Relation between volume of reaction products and

lime reactivity parameter of Thompson 136

6.7 Relation between overall C/S ratio and total

strength of the soil-lime systems 141

6.8 Relation between overall C/S ratio and lime

reactivity 141

6.9 Relation between overall A/S ratio and total

strength of the soil-lime systems 143

6.10 Relation between overall A/S ratio and lime

reactivity 143
 xW.

PAGE NO.

6.11 Relation between volume of reaction products

and the activity indexes 145

6.12 Relation-between overall-C/S ratio and activity

indexes 14S

7.1 Algorithm for search program 160

9.1 Weight of lime vs volume of solvent used in

, 216
extraction

9.2 Weight of lime vs volume of solvent used in

extraction 218

9.3 Relation between strength and reacted line as %

of added lime 220

9.4 Relation between lime reactivity and reacted

lime as s of added lime 220

9.5 Relation between volume of reaction products -

and reacted lime as % of added lime 222

9.6 Relation between strength and volume of

unreacted soil 227

9.7 Relation between strength and overall porosity 227

9.8 TG and DTG curves of lime treated soils 229

9.9 TG and DTG curves of lime treated soils 229

 41T.

LISTS OF TABLES PAGE NO.

2.1 Removable protons as a function of pH 19

4.1 Nomenclature of the soils 39

4.2 X-ray fluorescence analysis of the soils 41

4.3 Clay size fraction extractable by O. SN NaOH 44

4.4 Physical and engineering properties of the soils 48

5.1 Fraction of lime consumed at varying times for

different fractions of two of the soils 60

5.2 Fraction of lime consumed at various times by

the soils 66
5.3 Numerical values of the first term of Jander's
equation F(x) for various times of reaction 74
5.4 Rate constants and activity indexes for the soils 86

5.5 Tentative classification of red tropical soil-

lime for use in road structures 94
6.1 Standard conditions used to determine iron,
calcium, aluminium and silicon by atomic
absorption 110
6.2 Chemical analysis by XRF of standard minerals 118

6.3 Summary of likely minerals formed 126

6.4 Chemical analysis of acid soluble reaction

products 128

6.5 Approximate overall chemical composition of

reaction products 130

6.6 Quantification of reaction products from XRD 133

7.1 Part of typical output of search program 162

7.2 Computer program in Fortran-language 165

9.1 Constituents of the soil-lime systems other than

cementitious reaction products 213
 CHAPTER1

INTRODUCTION

Approximately one-third of. the land surface of the earth is

contained within the huge belt limited by the tropics of Cancer

and Capricorn. Befitting attention is now being given to tropical

weathering and the resulting soil products which are the dominant
features of the landscape.

The nomenclature of the soils of the tropics has increased

not only in size but in complexity. The confusion and controversy

originated by the very nomenclature proposed by researchers in

their-desire to precisely define these materials have led to

underutilization of the data gathered on their hehaviour. Terms

like laterite, läteritic soil, latosol, ferricrete, ferruginous

soil, ferrisol, oxisol, ferrasol, etc. have been used for the same

type of material in one case and applied to various types of

materials in other cases.

Attempts have been made to find commongroimds for agree-

ment. However, most of the attempts have not drawn'a clear

distinction between classification and nomenclature. To the

engineer, it is enough that nomenclature be an ordered system of

naming which reflects peculiar physical properties of the soils

while classification be a tool for predicting their engineering

behaviour. Hence in this work, 'use is made of the term "red

tropical soils" and this is broadly a soil which fits the

definition of "laterite" given by Sivarajasingham et al (1). Thus

it includes 'highly weathered material (a) rich in secondary forms

of iron, aluminium or both; (b) poor in humus; (c) depleted of

bases and combined silica; (d) with or without non-diagnostic
 2.

substances such as quartz, limited amounts of weatherable primary

minerals or silicate clays " The term "laterite" is

....

restricted to the original definition given by Buchanan (2) who

wrote " it is full'of cavities and pores and contains a very

...
large quantity of iron in the form of red and'yellow ochres. In

the mass, while excluded from the air, it is so soft that any iron

instrument readily cuts it, and is dug up in square masses with a

pick axe, and is immediately cut into the shape wanted with a

trowel or large knife. It very soon after becomes as hard as a

brick.... As it is usually cut into the form of bricks for

building, in several of the native dialects it is called the

brickstone (Itica cullu) ... 'Me most proper English name would

be Laterite, from lateritis, the appelation that may be given to

it in Science. "

However, the word "laterite" will also appear where passages

are quoted from works in which the term is used.

Red tropical soils have been used to a large extent on a

"trial and error" basis. They have been used successfully as

subgrades and as sub-bases or bases of low quality; in some

instances the aggregates comprising red tropical soils have served

as suitable mineral components of concrete and asphalt.

Their availability and abundance make them very attractive

as sources of "stabilized" or improved engineering materials.. To

improve their engineering properties and expedite construction

they are usually treated with cement, lime and rarely bitumen.

Reports of successful treatments as well as failures are abundant.

There is no doubt that contradictory results have their origin in

the empirical nature of the techniques employed for their

 3.

improvement.

This is particularly true of the use of lime to upgrade the

engineering properties of red tropical soils.. Although the studies

of lime-soil systems are numerous for temperate soils, knowledge

of the response of tropical soils to lime treatment is scanty and

limited. Like any other branch of applied science, progress on

the efficient and economic use of red tropical soil-lime systems

will only be possible when the fundamental mechanisms of the

reactions and the nature of the reaction products are properly

elucidated. This will provide a basis for adequate technology for

maximising the successful utilization of the lime treated soils

under varying conditions and permit more refined engineering
decisions to be made.
 4.

CHAPTER2

GENERALLITERATURE REVIEW

Each chapter of this thesis is commencedwith a review of

pertinent literature and methods. This chapter therefore

summarises literature on some aspects of pedology and engineering

behaviour of the soils.

2.1 Genesis and Composition

Red tropical soils can be formed in-situ or transported.

When formed in-situ, parent rocks like basalt, sandstone and schist

are decomposed under a wann and humid climate of the tropics

(3,4,5,6). Expansion and contraction associated with high and low

temperatures which occur during the dry and wet seasons respec-

tively, leave cracks in huge masses of rocks through which warm

surface water rich in carbon dioxide and bacteria filter. This

chemically active water decomposes the rocks (aluminous silicates).

An upward suction of moisture to the surface-in the dry season and

drainage of rainwater over the gentle and moderate relief of the

tropics remove the dissolved minerals of the parent rock in

solution. This process of relative accumulation is called

leaching. The residuum, comprises minerals like aluminium, iron,

titanium, manganese, chromium and vanadium oxides, quartz and clay

minerals. The types of minerals depend on the parent rock (7 - 10)

while their nature depends on other soil forming factors like

topography, drainage and climatic - vegetational conditions.

Aluminium exists in the form of gibbsite (aAl(OH)3),

boehmite (yA10(OH)) and in amorphous forms. The chemical composi-

tion of amorphousalumina and other amorphousmineral constituents

 5.

of red tropical soils are-not Imown.

Iron occurs corrunonly as goethite (aFe20(OH)) haematite

,
(aFe203), lepidocrocite (yFe00H). Maghemite (yFe203) and

magnetite (yFe304). have also been reported. (9,11).

Titanium minerals include rutile, anatase - and ilmenite.

Chromium and vanadium are resistant primary minerals and their
.
oxides can be present in small amounts depending on: the. type of

the parent rock unlike manganese oxide which readily goes into

solution in acidic conditions and isLsubject. to removal by

leaching. However, concentrations of manganese in forms like

lithiophorite and birnessite are present or-occur-where drainage

is poor or it is absent. Silica occurs as quartz, which can-be

residual or alluvial, and in the amorphous form. The clay mineral

that occurs most couuiionly in the soils is the 1: 1 type, i. e. ý,

kaolinite and/or halloysite. Kitagawa et al (12): and-Yoshinaga

and Aomine (13) have reported allophane (poorly-defined combina-

tions of silica and. sesquioxides usually regarded as hydrates of

alumino silicates (14,15)) as the dominant clay mineral of young

volcanogenous soils (soils derived mainly from pyroclastic

materials) occuring in the tropics. Solid state reaction within

an amorphous gel has been postulated to : be. the process of forma-

tion of kholinite (16) with a slow rate of crystallisation.,. On

the other hand, Kitagawa (15) postulated, from calculations of the

chemical composition of allophanes, "that halloysite could-be"the

end member of-the amorphous - allophane series. This and the fact

that metahalloysite is difficult to distinguish from kaolinite

suggests that the clay reported as kaolinite by"many could be

metahalloysite especially where such clay occurs alongside

 6.

amorphous minerals. Illite and montmorillonite are rare in red

tropical soils.

Further complications arise when the distribution and nature

of the minerals within the profile are'considered. - There is

little agreement by pedologists on the mechanism responsible for

the distribution'of the minerals. The concept of'the soil as a

residuum with'relative immobile constituents was modified by a

second phase of formation which involves the resolution of the

residuum by'groinmdwaters. The-recognition of pallid zones under=

lying some of the soils led to the concept of red-tropical soils

as precipitates. The enriched zone was believed to be'caused by

the precipitation of iron and alumina in'solution, with'capilla-'

rity and the seasonal fluctuations of the water table being the

mechanisms responsible for this (7, pp. '91- 108). Thomas (17)

postulated the removal of "rock alteration products" from deep

within by ionic-diffusion as a possible mechanism. Recent mmodels

proposed include the detritus models and those which' compromise"

between the residuum and' precipitation concepts.

It is difficult, if not impossible to discount groundwater

as an important factor which determines the nature and type of

some of the constituents of red tropical soils (18,19).

Harder (19) found among other things, for synthetic clay mineral

formation that:

a) amorphous hydroxides of Al, Fe, 'Mn Mg, Zn, Co, Ni, etc., are

capable of coprecipitating Si02 even'from very dilute (weathe-king)

solutions. '

b) the most important factor for silica mineral formation is the

silica concentration in the solution which should not be too high

t
 7.

or too low. Silica minerals form only in those precipitates from

solutions undersaturated with respect to amorphous silica (100 ppn

Si02 at 20°C); lower silica content allows formation of quartz or

feldspar while higher contents makes the precipitate stay

amorphous even after long ageing times.

c) a sufficiently slow rate of°precipitation is necessary for the

synthesis of clay minerals., The precipitation can be as a result

of changing pH. These findings suggest that in tropical conditions

clay minerals cannot be formed in the wet season since the silica

content of the groundwater is generally very low whereas their

formation is possible during the dry season below the water table

level. The presence of-allophanes and/or amorphous minerals is an

indication that during their formation the silica content of the
groundwater was too high for quartz or clay mineral formation.
2.2 Classification

Geologists, pedologists, agriculturalists and engineers are

interested in different aspects of behaviour of the soils and con-

sequently have classified them, from their professional view points.

This is reflected in the great volume of work devoted to classifi-

cation which unfortunately is confusing and of limited use, to a

civil engineer. - ldeally, "a classification should. serve not only

as a tool to transmit information butýalso as a tool to predict

behaviour when experience is not available.

Early attempts to classify red tropical soils used chemical

content as a basis. Martin and Doyne'. (20) introduced the concept

of silica - alumina ratio as 'a measure of laterization; a concept

that has since been modified. named duricrusts on the

-Dury'(21)
basis of their chemical content, the name being compounded from

those of the minerals present, thus: -

 8.

Silitic -. Si02

Siallitic Si02, A1203

Fersilitic Fe203, Si02 I

Fersiallitic Fe203, Fe00H, Si02, A1203nff O,, A100H

Ferrallitic Fe203, FeOOH, A1203nH2(), A100H '

Ferritic Fe203, FeOOH

Fermagnitic FeZ03, MnO2

Tiallitic TiO2, A1203nH20,

Allitic A1203nHz0;. A100H

These are at the best systems of nomenclature as they lack

scientific precision. They have not taken into account the

following which influence any classification with engineering

validity.
a).. Pedological and geological factors such as parent rock,.

genesis and environment.

b) The type and, more important, nature of the mineralogical

constituents of the soil "for it is the form in which the chemical

contents occur which varies consistently in accordance with the

physical properties of the soils" (7, pp. 17).

D'Hoore (22) used the terms ferruginous, ferralitic and

ferrisol to describe these soils depending solely on climatic

conditions, using annual rainfall intensity as a measure. Lyon

Associates (9) noted that most of the samples they tested showed a

good correlation between the index properties and D'Hoore's

pedological classification. Their other report (10) related South

American soils to the FAO -- UNESCOclassification which-is a com-

promise retaining some traditional pedological terms like pedosols

and. recent popular terms like. vertisols and andosols. These

 9.

classifications can only be used as a very rough guide to-the

properties of the soils.

Attempts by engineers to use particle size and plasticity,

which have proved adequate classification parameters for temperate

zone soils, have been unsuccessful because the conventional

methods for determining these parameters do not yield reproducible

results (4,5,6,23 - 28). This has led to a search for other

consistent parameters for classification of red tropical soils.

De Graft-Johnson et al (18) and Gidigasu (29) have shown that

'lateritic gravels' and fine grained 'laterite' soils can be

classified on the basis of geology, amount of clay, linear shrin-

kage and climatic conditions while Wesley (28) showed that basic

engineering properties are related more closely to mineralogical

composition than to pedological classification. It thus seems

that for any single quantifiable parameter to be suitable for

classifying these. soils, it must incorporate the intrinsic and

apparently extrinsic properties of the soils. Such apparent

extrinsic properties are climatic condition or the behaviour of

the soils in an environment of particular conditions.

2.3 'Physical "and 'engineering 'characteristics

2.3.1 Texture

Variations in particle size distribution with methods of

pretreatment and testing which have been widely reported can be

accounted for by a number of reasons. Wet sieving, for instance,

has been shown to increase the silt and clay fraction from 7% to

20% (30) while the dispersant used affects the percentage of the

clay fraction. The problem does not seem'to be with dispersion

but preventing the dispersed particles from re-flocculating (28),

 lQ+.

hence Quinones (31) found that a halloysitic clay from Kenya

.
dispersed with sodium oxalate gave 20% to. 30% clay-fraction but

with sodium hexametaphosphate gave 40% to. 50%. The sedimentation

test is further affected by the method of drying. Moh and Mazhar

(23) reported less clay in oven dried samples than in air-dried

or as received samples with increased silt and sand fraction.

Most workers believe that the coarse fractions of. the soils. result
from micro-aggregations of the clayey constituents by sesquioxides

(4,25,27). Pretreatments like wet sieving, dispersion-and.

remoulding dismembers some of the aggregations into finer clayey

clusters due to the friable nature of the sesquioxidic cementing

agents while drying results in an irreversible process, of dehy-

dration of the amorphous constituents and hydrated halloysite with

the free iron and alumina cementing and coagulating the-clay

particles (24,25,28). The indurated soils are. less affected by

pretreatment than the fine grained soils although they (indurated

soils) have high content of fines and therefore do not fit into

the classification system for temperate zone coarse-grained soils.

2.3.2 ' 'Atterberg 'Limits

The nature of the predominant clay mineral influences the

levels of plasticity of the soils (9). For example, montmori-

llonitic tropical clays of Ghana were found to plot above-the

A-line while the halloysitic clays over volcanic ash plot below

the A-line although both are very plastic. However, the coarse

grained soils show a less erratic plasticity characteristics; most

plot above the A-line with liquid limits generally below 60%.

Apart from the nature of=the clay it is also observed that the

position of the sample within the soil profile influences the

 lit

plasticity (32). Due to differential weathering, movement and

deposition of dissolved materials, ýthe variation of plasticity

with depth is very difficult to predict even in two similar

profiles on different topographical sites.

Another problem associated with the determination of
Atterberg limits is the effect of pretreatment and'handling.
Sherwood (33) found that the, liquid limit increased with increased

working while the plastic limit was not significantly affected.

The usual explanation is that remoüiding-. or working breaks down the

aggregations of clay particles by sesquioxides. While remoulding

could break down friable and weak aggregations of clays, it does

not seem likely that the average energy associated with mixing can
break down the normal bond. between'the, clay and sesquioxides. It

has also. been suggested that mixing affects the allophanes and- -

amorphous constituents by releasing the water held in the structure

of these minerals-, (28,34).. This explanation also seems to fall

short of the energy input required to release water held in

minerals.

Drying is said to partially or completely convert the clay-

fraction to a non-reactive silt-like fraction with reduction or

total loss of plasticity (8). Gidigasu (27) noted that wet-, forest

zone materials-and deep layer dry savanah zone soils are more

sensitive to changes in plasticity than highly dessicated top

soils from the dry savanah zone when they are dried. Red tropical

soils generally have low colloidal activity.

2.3.3 Density and compaction' characteristics I

(then compacted from dry to wet, the dry density increases

with increasing moisture content up to a maximum beyond which the

dry density decreases. With increased compactive effort, the

 12+.

maximum dry density generally increases and the optimum moisture

content generally decreases. The moisture-density relationships

are influenced by factors which can be'divided into two broad

groups. The first group of factors- derive from genesis while the'

second derive from pretreatment prior to testing. The failure to

consider all the factors in any single'study has placed limitations

on the applicability of the-results available on their density and

compaction characteristics.

The most important genetic factor is texture or grading.

This is illustrated by the fact that the gravels have maximum dry

densities ranging between 1842 -2323 kg/m3 and optimum moisture

content between 6- 16%; while for sandy clays. the maximum dry

densities are 1602 - 1922 kg/m3 and optimum moisture content '

16 - 24%. The maximum dry densities for clays are generally below

1602'kg/m3 with the'optimun moisture content above 24%. Compaction

breaks down weak aggregations in the soils (4,5,24,35). Studies

on the effect of reusing materials, which results in further break-

down of weak aggregations, on the compaction characteristics of

some red tropical soils show that whereas the maximum dry density

is increased, the optimum moisture content-appears unaffected (35).

jVhether higher or lower densities are obtained after compaction

depends on whether the breakdown leads to an improvement or dete-

rioration of the grading.

Pretreatment such as drying (air, or oven) affects the densities

and optimum moisture contents of the soils irrespective of texture.

Generally, drying gives higher maximum dry density and lower

optimum moisture content than values obtained at their natural

moisture contents (9,25,27). As has been pointed out, drying

 13,

alters the state of some of the mineral constituents, irreversibly

and as such the density determined from a dried sample of such

soils is, in effect, for a different. material. Hirashima (36)

reported that for Hawaiian 'lateritic soil' formed under conditions

of continuous high moisture content in the profile there was no

peak when the material was compacted. wet to dry but when after

drying it was compacted dry to wet, a peak occurred. This

illustrates the need to test samples from areas with. high regular

rainfall in their undried state.

2.3.4 Strength characteristics

The strengths of red tropical soils are usually determined

by unconfined compression strength test (U. C. S), California, bearing

ratio test (C. B. R), triaxial and penetration tests; UCS and CBR

being more popular. In their untreated state, genesis and pedology

influence significantly the strength of the soils. Lyons

Associates (9) using D'Hoore's pedological, classification (22)

found that ferruginous soils have higher CBRtvalues than ferrisols

which in turn have higher values than ferrallitic soils. The most

important genetic factor is texture. Unusually high internal

friction has been reported for even the clayey soils (4,24,25,

30,35). To evaluate the hardness of laterite concretions, Nova-

Ferreira and Correia (37) proposed a hardness index which is a

modified form of the fineness modulus used in assessing mineral

aggregates; while de Graft-Johnson et al (18) converted grading

into a single parameter called the fineness index which is a

modification of the hardness index. They (18) showed a significant

correlation to exist between the CBR and this index for some

'lateritic' gravels. The hardness index apparently excludes 'soft'

 14,

concretions which contribute to strength despite their breakdown

during compaction while the fineness index does not take into

account the breakdown of the concretions.

Other genetic factors such as mineralogy and the profile

position of the sample can be tied into texture as far as their

effects on strength are concerned. For example, mineralogy will

influence the breakdown of the aggregations during compaction

manifesting in improvement or deterioration of the final texture

after compaction.

2.3.5 Swelling; 'permeability and'cöiisbliddtion characteristics

Since the occurrence of swelling minerals is rare in the

soils, they are not susceptible to excessive swelling. Swelling

is affected by their moisture content and apparently, drying. The

higher the moisture content, the less the swell while oven-dried

samples prior to compaction seem-to swell less than samples

compacted at their natural moisture content'(38).

While 3n-situ the porous and granular type of the soils are

highly. permeable but exhibit low permeabilities when compacted or

remoulded ' (6) . Typical in-situ

.permeability values range from

10-2 to 10-1 cm/sec'but decrease to 10-4 to 10-6 cm/sec after

compaction. Malomo (39) found that the strength of red tropical

soils depends on the level of the stress applied and that they

change from over-consolidated clay behaviour at low moisture

content to normally consolidated clay behaviour at higher moisture

contents. He attributed this to particle degradation. He

suggested that the phenomena of breakdown of aggregations under

stress occurs more readily with increase in moisture content.

Hence samples with-higher moisture contents will breakdown'or

 ýSý

disaggregate more than those compacted at lower

moisture content.
2.4 Pozzolanic'properties of red tropical soils
2.4.1 Red tropical soils as pozzolanas "'

BS 892 defines a pozzolana as "a naturally occuring or

artificial material which has in itself little or no cementing

property but which will combine with lime in the presence of

water to form a stable insoluble compound having cementitious

properties". Literature abound to show that most clayey soils of

the tropics as well as temperate zone clays satisfy this'defini-

tion. The study of the pozzolanic properties of clayey soils have

generally been termed "lime soil stabilization". This' term is

usually associated'with highway' engineering simply because' the

soils so treated have been used in sub-bases, bases and founda-

tions of highway strictures'. The additional use of red tropical

soil-lime systems in parts of 'superstructures as blocks, bricks

and in concrete (3) calls for the use of'a term broader than

"soil stabilization". In view' of'this and the fact that they

satisfy, the BS 892 definition given above, ' red tropical soils are

considered as pozzolanas'in this work.

2.4.2 'Söil-lime 'reactions

2.4.2.1 Mänifestätioris 'öf 'tfie 'reäctä öns P

The reactions which occur when lime is mixed with a clayey

soil in presence of water* account for the ' changes* ahn.i.ts

engineering properties. These changes are broadly recognised as

two phases: -

a) a phase resulting in improved workability and reduced

plasticity (amelioration);

b) in. strengthening (cementation). Among the

a phase resulting
 - 16.

many mechanisms proposed to account for amelioration are

(i) flocculation due to cation exchange, (ii)ý, carbonation, '
(iii) pH-dependent increase in cation exhange capacity (cec),

(iv) adsorption of calcium hydroxide molecules, (v) pozzolanic

reaction leading to diffuse cementation. It-is unanimously agreed

that "pozzolanic reaction" is responsible for the phase' resulting

in strengthening.
i
2.4.2.2 Amelioration '3n 'clays

Diamond and Kinter (40), -.'Stocker (41) and the Transportation

and Road Research Board, USA (42)-have very exhaustively. reviewed

and summarised -current opinion. Cation exchange and carbonation

cannot be discounted'-but have, been accepted-as auxiliary effects,

under appropriate 'circumstances, inadequate for the explanation of

the changes associated with amelioration. They are unavoidable

lime consuming processes'.

Hilt and Davidson ý(43) in proposing the pH-dependent cec

theory, 'whIch involves "a crowding of additional cation to the

surfaces of the clay" after'the normal cation'exchange, noted that

significant increase in, the unsoaked strength of the soil-lime,

system occurs only after a certain threshold lime content is'''

exceeded. - The threshold they termed the "lime retention"point".

This implies that "pozzolanic reaction" (responsible` for strength-

increase) can only take place beyond the lime retention point

which is usually measured after two hours. ` Further clarification

of the lime retention point was given by Ho and Handy-(44) who

observed that "lime retention requires an alkaline pH, 'little or

no retention occurs below pH'11.0 and retention is maximum at

about pH 11.75. Calcium hydroxide provides a near optimum ratio

 17.

of Ca++ and OH- ions, Mg++ and Na+ ions are relatively

ineffective". They further advanced the following fundamental

explanation of the lime retention point phenomena. "The cec of

clays is ordinarily measured at pH 7.0. At high pH's, more

cations are adsorbed, perhaps because of increasing dissociation

of weakly acidic SiOH-groups exposed on clay crystal edges and is

usually attributed to isomorphous substitutions in the clay

mineral structure". They went on to say "The increased adsorption

of Ca++ to clay surfaces at high pH may be the cause of better

bonding between particles, evidenced by bigger flocs, - higher

suspension viscosity and high plastic limit". They concluded that

"the lime-montmorrillonite reactions are divided into three

levels. In the first, the total. cec is ordinarily measured at

pH 7.0. In the second, the additional lime retention is. dependent

on an alkaline pH. In the third, after Ca++ holding capacities

are fulfilled, pozzolanic reaction with additonal lime can

proceed".

While admitting that pH-dependent cation exhange can and

does occur, Stocker (45, pp. 35) questioned the validity of the

claim that this phenomenon sufficiently accounts for amolioration.

He argued that pH-dependent cation exchange consumes less, much

less than I$ lime if the soil pH is initially 7, less still if the

pH is even greater initially. Based on Jackson's (46) identifica-

tion of five ranges associated with the removal of protons from

different sites (see Table 2.1) he postulated pozzolanic reaction

leading to diffuse cementation as the mechanism responsible for

amelioration. He stated that when pH-dependent cation exchange is

maximised pH is around 10 or 11 and at this value rapid dissolution

Table 2.1 Removable protons as a function of p11
(After Jackson (46)).
 19.

. t'

ýt

`
t

Range pH Removable protons

Range I Approx., 'pH < 4.0 : Exchangeable H+ or H30+

.
Range II Approx. pH 4.0 - 5.6 Exchangeable Al+++
:
Range III Approx. pH 5.6 - 7.6 Positive - hydroxyl, alunina
; : -
,
polymers (containing A1-OH2
terminal groups).

Range IV Approx. pH 7.6 -z 10: Residual, weaker, Al-OHZ

groups in hydroxyl alumina
and. possibly layer silicate
edges.

Range V. Approx. pH > 10 Alumina silicate dissolution

reaction.

- ý,

;,ý
 20.

and reaction (Range V, Table 2.1) is initiated leading to genera-

tion of tiny cementitious reaction products, at, or near the edges,

responsible for amelioration. He rather favours diffusion as a

mode of attack than dissolution and believes lime retention point

to be the quantity of lime used up in reaction in two hours by

which time many of the physical changes are maximised.

The adsorption theory of Diamond and Kinter (47)-recognises

the reaction of lime with clay-but suggests that this is preceeded

by physical adsorption of calcium hydroxide molecules on to the

clay surfaces, a largely reversible process-, ' This theory is in

closer agreement with that of Stocker, the difference being that

while Diarand and Kinter see amelioration and. strengthening as

being accounted for by two'distinct mechanisms, Stocker believes

one mechanism is responsible for both. Current' opinion therefore

seems to recognise cec, carbonation and'pH-dependent cec as

processes that take place but each is insufficient to explain

amelioration. Amelioration in clays therefore seems to-. be a

result of one or a combination of these processes supplemented by

formation of tiny reaction products.

2.4.2.3 Ameliorätloii ' zui 'red 'trdpicäl 'söils

If amelioration in red tropical soils is not explained by

the theories enumerated for clays, the deviation should, -be as 'a

result of the presence of the following: -

a) the colloidal fraction (fraction less or equal to clay, size

consisting of clay minerals, amorphous oxides and/or allophanes).

b) the very poor crystallinity of the clays.

c) the crystalline iron compounds.

The effects of these on cec, physical adsorption and pozzolanic

 21.

reaction leading to diffuse cementation theories are examined.

The carbonation theory is omitted because it is independent of the

three factors enumerated.

Unusually high cec have been reported for red tropical soils

(48,49,50) which do not contain high cec crystalline clay

minerals such as montmorillonite,. vermiculite, and illite.

Crystalline oxides of aluminium and silicon have low capacities

and cannot account for this apparent abnormality nor could poor

crystallinity of the clay present. The 'culprit' therefore-seems

to be the colloidal fraction. Fieldes et al (51) showed that the

capacities of the colloidal hydrous oxides of, aluninium and

titanium were so high that if they are present in small quantities

they could cause the high cec observed in red tropical soils.

Because the differential thermal analysis curves of the laboratory

precipitated' colloidal hydrous oxide of aluminium contained

various endothermic peaks similar to. the unusual peaks of the

soils,. he concluded that the principal exchange material in the

soils was amorphous hydrous aluminium oxide. Capacities for the

colloidal hydrous oxides of, iron and silica were lower but

considerable. These findings imply that where allophanes are

present (they have. low Si/Al ratio) in the., soils, high cec may be

expected. Birrell et al (52) not only confirmed this expectation

but further concluded that "physical adsorption of cations by

soils containing allophane, palagonite. and certain amorphous

oxides is responsible to a greater or less extent for the apparent

high cec shown by these soils... ". He was led to suspect-"some

kind of adsorption mechanism" because he observed that the cec

depended on the strength of the leaching-solution and the cation

 22.

used, and that the exchange complex was extremely easily

hydrolysed.

It is therefore reasonable to say that cec in collodial

fraction takes place and is pH-dependent because the allophanes

and amorphous minerals are likely to have surfaces characterised

by broken bonds (53). The exchange capacity will depend on the

composition of the equilibrium liquid since the surface potential

will be determined by the pH of the equilibrium liquid. This has

been demonstrated by Aomine et al (54) who showed that the cec of

allophane separated In an alkaline medium (pH 10.5) was about

three times that of the same fraction separated in acid medium

(pH 3.5) . o1lowing a ,method of approach similar to that 'of'Birrell

et al, Diamond and Kanter (47) concluded among other things that

calcium hydroxide is physically adsorbed from solution at a very

rapid rate by calcium-saturated clay and that this process which

is largely reversible at very early stages is soon followed by

reaction to produce calcium silicate hydrates. They measured

adsorption isotherms at 5 mies. for bentonite and 10 rains. for

kaolinite suspensions. It thus appears that pH-dependent cec

through physical adsorption of cations takes, place in the collo-

dial fraction of red tropical soils. The question then is, is

this the process responsible for amelioration or is it the begin-

ing of the process?

The answer can be'found by considering the time required for

amelioration in practice and that required for physical adsorption.

The low concentration region of the isotherms of Diamond and

Kinter fitted the B. E. T. adsorption equation theory. If 5 rains.

or 10 rains. is suitable for this process in crystalline clays and

 23.

15 rains. is suitable for the determination of cec (an indirect

measure of adsorption) of allophanes as shown by Aomine et al (54)

then it is likely that pH-dependent cec through adsorption seems

to be the initiation of the process actually responsible for

amelioration. Diamond (55) demonstrated that'lime reacts almost

instantaneously with hydrous alumina of high specific area to

generate tetracalcium aluminate hydrate. Recently, De Carvalho

(56) found that while amorphous silica and alumina contents were

significant, pH-dependent cec is an excellent indicator of a

soil's lime reactivity. Lime reactivity is as is, measured by

Thompson (57) i. e. the difference between the strengths of the

lime treated soil and the untreated soil. Since strengthening-is

a result of the formation of some cementitious reaction products,

pH-dependent cec through adsorption appears therefore to be the

initiation of the actual, probably one, process-responsible for

both amelioration and strengthening. Amelioration in red soils

containing amorphous minerals appears to be a result of the rapid

reaction of the collodial fraction to produce cementitious

reaction products with the initiation of the process of reaction

manifesting (at very early stages) in pH-dependent cec-through

adsorption of lime.

2.4.2.4 The *reaction 'pröcess

The reaction rate between, soil and lime is influenced by a

number of factors among which are: -.

(i) the fineness or specific surface area of the minerals
(ii) -solubilities of the minerals
(iii) pH of the medium
(iv) intimacy of the mix (mineral + lime)
 24.

(v) temperature.

In a practical red tropical soil-lime system, factors iii, iv, v,

are controllable and since the pH depends on the percentage of

lime used, factors i and ii are considered at the optimum lime %
for maximumstrength (pH is nearly that of saturated lime water,

12.4).

As-has been suggested in the foregoing section 2.4.2.3, very

rapid early reaction takes place between lime and the colloidal

fraction.: The mechanism is not yet clear. While investigators

like Eades et al (58,59) postulate a 'through-solution' reaction

mechanism, others like Stocker (45) favour reaction through solid

state transformation. The probable mechanism or combination of

mechanisms responsible for strengthening in red tropical soils is

a subject of Chapter 5. However, the order of reaction of the

minerals can be postulated. ' If two minerals react, the finer one
is or both are the faster is. the reaction. The collodial fractions

of the soils are very fine as measured by their high specific

surface areas (50,53) and hence will be more reactive than the

crystalline constituents of the soil. They are also more soluble.

Using a monomer model, Marion. et al (60) showed that at different

pH, there is considerable difference in the total soluble alumina

that can exist in equilibrium with the different aluminium com-

pounds found in soils. For instance, at a pH of 10 the total

soluble alumina increases in the order of the aluminium compounds

Kaolinite-> Gibbsite > Halloysite > Amorphous A1(OH)3. This order

is significant if solubility is a measure of reactivity. From the

view point of chemical composition only Kitagawa (15) concluded

that halloysite may be the end member of the allophane series.

 25.

Fieldes(34) distinguished between the various forms of allophanes

and indicated that they age-to clay in the sequence: -

allophane B -} allophane AB -} allophane A -> metahalloysite -}

kaolinite. This order holds for 'the increasing specific surface

areas of the minerals. Using solubility and specific surface

areas as factors for assessing reactivity the following order of

decreasing reactivity of some minerals found in red tropical soils

follows: - Amorphous Si, Al > Allophanes > Metahalloysite >

Halloysite > Gibbsite > Kaolinite. The role and position of

crystalline iron minerals is more difficult to assess since in an

alkaline medium such as that provided by red tropical soil-lime

system, iron oxides are insoluble. Their roles will be considered

in latter chapters.

2.4.2.5 The reaction' products

Chemically, the reactions 'that take place when lime reacts

with soil minerals can be represented by the following over,

simplified equations.

Ca (0H) 2 -> Ca" "+2 (OH) (2.1)

Ca++ + 2(OH) -+ Si02 (Soil silica) -} CSH

...
(2.2)

Ca++ + 2(OH)- + A1203(Soil alumina) -> CAH (2.3)

...
Kaolinite + lime -} CSH and/or CAH and/or CASH (2.4)
...
Montmorillonite + lime -+ CSH and/or CASH (2.5)
...

where C= CaO; A= A1203; S= Si02; H= H20; F= Fe203.

Equation (2.1) represents the dissociation of hydrated lime in

water. The solubility of lime at normal temperature is low;

approximately 0.20 gm/100 ml. This means that most of the lime

added remain as solid particles. The dissociated and undisso-

ciated lime react with the soil minerals to produce calcium

 26.

silicate hydrates, calcium aluminate hydrates and calcium alumino-

silicate hydrates as indicated by equations 2.2 - 2.5. The main

factors controlling the development and type of reaction products

are the inherent properties and characteristics of the soil.

Reaction products identified generally as calcium silicate

hydrates have been reported (58,59,61,62,63). Others have

reported calcium silicate hydrates as belonging to one of the

three sub-groups:

a) tobermorite gel which has a high C/S ratio and is attributed

with strength gains that occur in hydrating cements.

b) CSH (1) has a low C/S ratio than tobemorite and is generally
found in poor crystalline or gel form.

c) CSH (2) has a high C/S ratio but with doubtful properties (40).

Handy and Glenn (64), Ormsby (65) and Diamond et al (66) reported

the formation of CSH (1) and CSH (gel). Ruff and Ho (67) went

further to state that strength gain-of lime-clay mixtures is due

to the presence of the different phases of the complex reaction:

lime + clay -> CSH (gel) - CSH (2) + CSH (1) -ý tobermorite; the

further the reaction went, the higher the strength. The formation

of calcium aluminate hydrates alongside calcium silicate hydrates

was also reported by some of the authors quoted. Some authors

(40) have suggested that the aluminium present in some products is

as a result of isomorphous substitution for silicon within the

calcium silicate hydrate lattice. The identification of quater-

nary compounds (compounds in the lime-silica-alumina-water system)

have been limited. Ariizumi et al (68) reported gehlenite,

similar to that reported by Benton (69), to form at normal tempe-

rature when lime was used to treat loam soils consisting mostly
 2.7t

of amorphous clay mineral and allophane. It thus appears that

quaternary phases in which silicon and aluminium atoms occur in

distinquishable lattice'position. are produced when lime reacts

clay or with amorphous clay minerals

with calcined and allophanes.

It may therefore be speculated that in red 'tropical soils treated

with lime, quaternary compounds of lime-silica-alumina and'

possibly lime-silica-alumina/-iron may be formed.

The reaction conditions chosen by different workers can

account for the apparent differences in the reaction products

reported for the same clays treated with lime. Many used pastes

or systems of high liquid content, high lime/clay ratio and high

curing temperatures.

2.4.3 Engineering- characteristics'of lime'treated red tropical soils

This has been aptly summarised by de Graft-Johnson (8).

The ameliorative effects have been discussed in sub-sections

2.4.2.2 and 2.4.2.3. There is a fair agreement that lime

decreases the plasticity index by considerably increasing the

plastic limit of the soil. The compaction characteristics of

treated soils show that increasing the lime content decreases the

maximum dry density and increases the optimum moisture content.

The effect of lime on strength varies according to the type

and quantity of the reaction products formed. While some soils

form products that increase their strengths significantly, others

form products associated with small strength increase (70).

Curing temperature and time affect the amount of reaction product

and hence the strength. Sherwood (33) reported that red tropical

soils containing high proportions of aluniniwn oxide may show a

 28.

decrease in strength with time due to the formation of CAHio

and C2AH8 which later change to a more stable and weaker form

C3AH6. Unless this is the only reaction product, which is

unlikely, the reduction in strength may be masked by a general

increase in strength. With lime treatment most'red'tropical

soils lose the little swell they had and permeability is greatly

reduced.

2.5 Chapter Summary

a) The formation of clay minerals under tropical weathering

conditions is discussed and it is noted that because of the low

silica content of. groundwaters in the wet season, clay minerals

cannot form but their formation is possible during the dry season

below the water table level.

b) The reasoning that mixing releases the water held in the

amorphous constituents or breaks aggregations of clay size

particles to explain the changes that occur in the index pro-

perties as a result of increased mixing is questioned.

c) Red tropical soils-satisfy the definition of pozzolanas and

their versatile use after treatment with lime calls for the use

of a term broader than "soil-stabilization" to describe the lime

treatment of the soils.

d) Following the theories advanced for amelioration in clays and

the explanations given for the high cation exchange capacity in

red tropical soils, it is suggested that amelioration in those

soils which contain amorphous minerals is due to the rapid

reaction of the colloidal fraction with lime. This reaction

produces cementitious products with the initiation of the process

of reaction manifesting (at very early stages) in pH-dependent

 29.

cation exchange possibly through adsorption.

e) Finally the order of reaction of some of the minerals in the

soils is postulated viz: - amorphous silica, alumina >

allophanes > metahalloysite > halloysite > gibbsite > kaolinite.

These reactions could lead to the formation of quaternary

compounds. - .,

'v
 30.

CHAPTER3

SCOPE'OF THE *INVESTIGATION

The preceeding chapters show the lack'of information on the

fundamental properties of red tropical soils and red tropical

soil-lime systems despite the actual and potential wide utiliza-

tion'of the soils in both the natural and lime'treated states.

The response of red tropical soils to lime treatment is'assessed

only by empirical methods. Methods with theoretical basis,

tested by laboratory results are desirable; furthermore where

experience is not available, as in most countries where red

tropical soils occur, a simple classification system with

engineering validity will be beneficial.

An-understanding and eventual solution to these problems

will be facilitated by the following objectives set out in this

work:

a) to examine and if possible determine the mechanism of

reaction between the soil and the lime,

b) to study the nature and types of the reaction products and

their relative effects on the strength of soil-lime systems and,

c) to investigate the contributions, if any, of the other

constituents of the red tropical soil-lime systems to strength.

It was decided to use the unconfined compressive strength (UCS)

of the soil-lime systems as a measure of the effects of the

reactions and reaction products since strength is one of the most

important engineering properties considered in a soil-lime

mixture.
 31.

The mineralogical and chemical compositions of the soils

and soil-lime systems are studied using x-ray diffraction, x-ray

fluorescence and thermal analysis while the micromorphology of

'the reaction products is assessed by scanning electron microscopy.

The mechanisms"of the soil-lime reactions are studied-in detail,

and the role of the type and volume of the reaction products and

quantity of other constituents of the system are assessed in

relation to the changes of unconfined compressive strength.

 32.

rue vrc n

PROPERTIESOF THE MATERIALS USED IN THE INVESTIGATION

4.1 Introduction

The purpose of this chapter is to give a brief description

of the location and environmental characteristics of the area,

sampling procedure, chemical and mineralogical composition, and

some of the engineering properties of the soils used. The origin

and chemical composition of the lime used to treat the soils is

also given.

4.2 Location 'and 'environmental 'characteristics of the area

The nineteen soils were collected from the states of Paraiba

and Pernambuco which are located in the North East of Brazil.

Figure 4.1 shows the area in relation to the world. Three

distinct climatic conditions prevail over the area. A section

(A in Fig. 4.2) enjoys a hot and humid climate with mean annual

temperature ranging between 22°C to 26°C. Annual precipitation in

this section range between 700 mm - 1500 mm with about 80%

humidity. While another section (B in Fig. 4.2) is hot and humid

as well, irregular rainfall (annual precipitation around 800 mm)

gives it an arid characteristics. The third section (C in Fig.

4.2) is dominated by very hot climate with temperatures never

below 24°C. It is semi-arid with low annual precipitation of

around 400 mm.

4.3 Sampling

The localities in the states of Paraiba and Pernambuco from

which the samples were collected are shown in Fig. 4.2. The
Fig. 41 Area of research in relation to the world
 34.

"Y ell
;f3

1-
%N Y.
` ol r. .

i]"j .
ý

rM1^ ýý 1y

LL V
7
r, r, a
aw
wa
z" ww

ov"

LIN

2
4i
O

cl,
A4 c a
J o CL
ýQ v
LL zd
w
121
0
or -.o ce- t
Fig. 4.2 Details of area of research showing also the

climatic zones A, B, C referred to in the text

 -I

3i
cN
03

c
4n
"1
co

NV3OO (L
w n=. 311NVI1V
10

0
0
I

V W
vý H
1a.
" U
W
N ý
m
v
N
Q r1i
v m?

W Q co
Z ý 4i LC

ä
" ýN
L " "
z " Y
I- " Q
w W
Ö "g
y
y
=- o
Ali
O C3
(a
C11 cr- U- %tu `: Z
O "i äý
ö" äa W ca
Z x
z" <g cn
0 I
o - o
0 W
cn

W Um a
z _ a
Z °c
ac
~ n.
o 0

I- cc Fý xn l`

y
pW
W,..
4 " ý~
0C

h v =gä
U W

W
- ý oCd

0w Q
cri c
0 Z a

O 'Q
O cC

" W

CL

o
`° ybHJ °
Co
 37.

choice of samples was based on such factors as experience, local

information and the performance of those materials that have been

used in roads and dams. The samples, normally taken from the B
horizon, were placed in polythene bags and sealed before being

taken away for storage.

Samples were named after the towns or the nearest town from

where they were taken. Throughout this study, samples are identi-

fied by the abbreviations for the names of the towns followed by

roman numeral to indicate where more than one sample is collected

from a town (see Table 4.1). The samples from the city of Joao

Pessoa (JP) are identified according to their positions in the

profile, i. e., A (acima = top), M (meio = middle), and B (abaixo =

base); hence JPA, JIM, JPB. All the samples used for the study

were less than 2 mm with the sieving done according to BS 1377

(71). Where the clay size fraction (< 0.002 nun) was needed it was

obtained by sedimentation under gravity using Stoke's law and

dried at 600C to avoid possible changes in the material

characteristics.

4.4 Chemical and *minerälogicäl 'composition

The total silica, alumina, iron and other oxides in the clay

size fraction were determined by x-ray fluorescence and the

results are shown in Table 4.2. The results show that total

silica and alumina each constitutes about a third of the total

minerals (including water) in the clay fraction of the soils

except JPA, SIA, and JII in which iron is substantial (about one

third). Total iron in most of the rest of the soils account for

about one-tenth of the total mineral constituents. The silica/

alumina and silica/sesquioxide ratios range between 1.05 to 2.67.

LEEDS
LIBRARY
UNIVERSITY
Table 4.1 Nomenclature of the soils
 39.

Soil Name Soil Symbol

....

Joao Pessoa Acima JPA

Joäo Pessoa Meio JPM

Joffo Pessoa Abaixo JPB

Cuite Cr

AreisI Al

Areia II A II

SolaneaIA SIA

Solanea IB S IB

Solanea II S II

Nova Floresta NF

Junco I JI

Junco II J II

Teixeira I TI

Teixeira II T II

Recife RC
Usina S. Maria USM

S. Bananeiras SB

Penha PH

Sage Mari SH....

. ..... .......
(

Table 4.2 X-ray Fluorescence Analysis of the Soil-

Fraction less than 0.002 mm (% by weight

based on samples dried at 60°C)

 41.

N
N p Ln'. O -t r -I M 01 r-1 r""1 N 01 M 8 M Qn co L1) 00 H cf"
2: M MOO ¢t '. 0 M N %0 N '.O ` '.O t. N
.. . .
Cn } e--1rý O r-I r-1 r-1 r-1 -I r-4 e4 r-1 r1 r -I ri i- 1 ri r1 rI r4

N c» M m' NO r--1 00 N M N. LA C) N LI) O '0 Ql N LE N 00

20 Ln Ln CO '.D Ln Lf '.o M N. Ln -r p N. M N. 00 00 ýr In
N .. . .. .. . . .. . . . . . .
(n ß; r4 rl O r- I r-1 r1 r-4 r1r1 ri r-1 r-i r1r1 ri ri ri i4 r-i

c+)
pN
N0 I O8N
N. N
O
C
%DN
00N
NC-N
rn1orn
M
rn
MOO
C) r-4
I%D
C:) r
R:
C)
' CO
rn
Ln
rn
O
O
O
N.
00
03
ý 4 4 4 4
V) Q ri N ri r- 1 rl N 1, 11 r1N Nr1 e , N r-1 rl

M CO M co r--1 LP) CIS N) M N L/) N. N al d cP '.O 0 r1

'
r. { 00 C71 O Ict \.D 00 \O [- C31 M r-- N L- 00 d t-- %O ºn N
Cd . . " ". .. . . "- " . .
0 8,7, ý8ý 8 88 ýS 8 0) 8 8
1 r-I
V
Cd
. C'S Q) \D d rl ý! N tf) N. CO O O N. N N. ýo 00
NO
N
rnC
t1) [M N
.o \0 r1M
.
4
't
000
N. N) ýt
O
L)
'o
NN
.
00
NM
P-4,O CC
r-1
rn
U ")
N.
M
N
4
v
4

aeÖ r-1 i--1 r1 r-4 r-1 r-I ririr1 r1 r-1 e- 1 rý r-1 ri ri ri r4 " r°I
.9

p d C) N ct Np 0) d cY 00 tt' '.0 0 1Y' 00 N. 0 N O

Ol N. Ql M C) NO N. C) N 0 M '.D 1:31
N r4 r-4 r-1
ono z0 4;
004 O ßr46 ýO -; .t cv O O .4 M

uf ýO N CO Llý p Ln -p M N. C) - '0 M r-i CO \0 . ý'

O ei vn un 1"-4 O \p ' rn 0M co r- 03 - r--4 Ln n co p M
N p NN O ri r-i M fV p 11) If)
rl ri r4 e4 r-1 r4 e-4

u) C) Ln V) oo d- '.O 'D d r-i N. 0 0) 01 M

C) 00 OO M r-4 e-I r"-1 vM N. 00 . -I . -i r-1 p r1
°ýý. ýtN do
00 O 00 000 O OO O O O O O

Md r-1 t- Np 0 00 00 M wt 00 u'1 M O N 0 m t.n

p ýnýn ýn ýn ýn ýn ýn 19 v Ln Ln Li
000 O co odo 0 co 00 O 0 0 0 0

00 N 't V) 0 t! ) 00 fV LPG -t "0 CM M et N. 0 R

O CO
-q
0 CV M N rl fV -- (V r-1 D %9 M t"K) rl fV N
_M'.
-"oo 0,0
_ 000 0 00 -. 000 O 00 00 O O O O O
.
r-A MN r-I C14r-4 - r-I d fV r-1 (V
r-1
8 C>" 000
r-1
0 .. ' 0 0 Oo 8 8 0 8
o OOo io --,
OOO O'_'+OO<-000 o coo _'OO O O O O O

N cN 0 CO un pct C'4 Ln
rn 00 CO co O r-1 \0
OM
Ln 'o Ln O O
zY OI M %D N. N. r1 Q) 00 M N. M Q1 O M
4
M C. (:; H N C4 rl r-i ri r-1 M9 ri ri fi O ct N

tM Ln %D Ln fV 00 N. (V '.O O co N r-1 Lf) 00 N. 0 [-

r-i
0 ci o0 Ol CO M o0 0) M IJ) 11) rl CO CO M It" r-4 CO to M
N """ " .. . " . " 4
ono to M 01 Q', CO N. O ýO M te M d 0) CO
r-4 e--1 r {

p M om
M
0 o0 V o0 c-' -. 4 0 NN co t /) ý7 C71 On 00 lý
M MM M ON M (Stn
MMN tom) N C%4 M (V M M M M

et N0 0O 0 O O. d COcN %D %0 lzt 'IT 00 0 00 N

0 N vt
0
pM et et Ln O)
. -i 00 p v) '.O rl t"M N. N 0)
cv
COD M N
M N) M
M M
M MMM M MN M qzt

rl O
U3
h I-A v)U3 hti E
U Q4 E~-ý UPI)
 42.

Table 4.3 shows the constituents, in the clay size fractions

of the soils, which are extractable by O. SN sodium hydroxide.

The extracts represent the different minerals in the amorphous

states (56). An examination of the percentages of the amorphous

silica shows that there is a wide range of values from 2% for SIA

to 11% for SII whereas the range of values is narrower for

amorphous alumina (from 3% to 8%). Amorphous iron is nearly non-

existent in RC but has the highest percentage value of 2.5 in JPA.

The mineralogical composition was determined by x-ray

diffraction; it will be discussed in detail in Chapter 6. These

results are shown*in appendix A. 1; they reveal the presence of the

clay mineral kaolinite with quartz, haematite and goethite

present.

4.5 Morphology'and distribuitiöri 'of 'minerals

A discussion of the morphology of the minerals identified in

the soils is carried out in this section with a view to examining

their possible roles in the soil-lime reactions.

4.5.1 Kaölznite

The kaolinite as shown by x-ray diffraction is disordered,

giving a slight shift in the (001) reflection (d-spacing 7.211

d-spacing 0
7.14A for kaolinite).
as against of well crystallised

The disordered clay mineral has been reported as kaolinite rather

than metahalloysite purely from x-ray diffraction peaks.

Metahalloysite is difficult to distinguish from disordered

kaolinite. It has been suggested that the presence of partially

hydrated halloysite can be detected by heating the soil to about

400°C and observing the 7.20A d-spacing (72). An enhancement of

the peak indicates the presence of partially hydrated halloysite

 Table 4.3 Constituents in'the clay size fractions

extractable by 0.5 N NaOH (as determined

by De Carvahlo (56)).

4ý
ýý
 44.

-,
Constituents extracted by
0.5N NaOH
Soil Name Soil Symbol
%SiO2 %A1203 %Fe203
Joäfo P. Acima. JPA 2.81 3.10 0.46

Joäo P. Meio JPM 10.91 "7.97 0.68

Joao P. Abaixo' JPB 6.71 6.11 2.51

Cuite CT 7.84 4.12 0.80

Areia I AI 8.11 6.95 1.07

Areia II All 9.69 5.89 0.62

Solanea IA SIA 2.08 3.28 1.01

Solanea IB SIB 4.13 3.97 0.79

Solanea II STI 11.29 7.56 0.21

Nova Floresta NF 8.31 7.02 0.43

Junco I Jr 5.02 6.11 0.78

Junco II JII 6.69 4.07 2.05

Teixeira I TI 8.12 6.98 0.30

Teixeira II TH 4.68 4.02 0.87

Recife RC 9.21 6.16 0.10

Usina S. Maria USM 3.71 4.41 0.19

S. Banareiras SB 8.19 6.84 0.28

Penha PH 2.91 3.80 0.29

Sape Mari SCI 7.98 6.48 0.78

 45.

in the sample. Thermally, goethite reacts at this temperature and

since the x-ray diffraction peaks are affected by the presence and

state of the iron compounds (39), such a method of differentiation

may not apply to the soils. Methods for the determination of the

degree of crystallinity and distinction between kaolinite and

metahalloysite include x-ray"diffraction, infra-red and thermal

analysis. The crystallinity of the soils have been studied using

infra-red spectroscopy (73) and by the use of kinetic thermogravi-

metric'methods. Both methods show that the kaolinites present in

the red tropical soils studied are poorly crystallised. These

studies did confirm that clay mineral halloysite is-absent.

4.5.2 Silica

X-ray diffraction shows'silica to occur in crystalline form

as quartz and in kaolinite. In the clay size fraction the range

of total silica is narrow (Table 4.2). Thus it will appear that

if the soils react with lime according to the total silica not

much difference in the strengths of the lime treated soils will be

observed. Relatively, the range of values of the amorphous silica

is wider and if the soils should react with lime in proportion to

the amorphous silica present, the range of values of their

strengths will be wide. It is important to point out that high

values of amorphous silica are not necessarily associated with

high values' of total silica.

4.5.3 Alumina

No other'form of crystalline alumina e. g. gibbsite, boehmite

was detected. Crystalline alumina occurs only in kaolinite. The

range of values of total alumina is narrow whereas it is fairly

wide in the amorphous state. As in the case of silica, amorphous

 46.

alumina does not vary as the total alumina.

4.5.4 Iron minerals

Iron exists in the soils as haematite, goethite and in the

amorphous state. The range of values of total iron is very wide

whereas the range is narrow for the amorphous state. A relation-

ship seems to exist between the values of the amorphous and total

iron in most of the soils. It would appear that the higher the

value of the total iron, the greater is the value for the

amorphous iron. It thus seems that iron as measured in terms of

total or amorphous will have the same effect (if any) on the lime

reactivity of the soils.

4.6 Physical 'arid 'engineering 'properties

Table 4.4 shows the textural composition, compaction and

strength characteristics of the soils and soil-lime mixtures.

The conditions under which the maximum strengths of the soils were

determined are given since strength is the engineering property

used in assessing the lime activity of the soils in this study.

4.6.1 'Strength 'determiriätion

Each soil was treated with 0,1.5,3.0,4.5 and 6.0% lime

(by weight of soil). The strength of the soil is measured for

each lime percentage at different moisture contents. For any

particular lime content, the soil was mixed with water, statically

compacted into a 5.08 cm diameter x 5.08 cm high split mould using

energy equivalent to the BS intermediate compactive effort

(approx. 1090 KJ/m3). The density and moisture contents were

determined. The moisture content was varied and the procedure

repeated. The compacted specimens were cured at 22°C for 28 days

when their strengths were determined in unconfined compression.

Table 4.4 Physical and engineering properties of the

soils used for the investigation

 48.
N Q1 M 0
M .4 p1 M %0 Q1 N co '0 t- 0 VI Q1 in
01 N e! 1ý v1 O Vf Q1 ý7 Iý M Q1 O f
u ý""1 . "4 ýD
9Qý O M O p p O O p
4 . ""1 p p ."1 p n"1 Q p
r"1 r"1
GAG"N

N vº O in in in in LA 0 O O O O
in in in u1 of of O
MI- g ä
ý
Ö
e
t
w
v v
ýD
V v
ý
ö
ý
Ö
ý
Ö
ý
C
v of v. ýÖ

10 >

'^ W 0
-4 M
M .-4
in
vA in
0 al
in .4 .o 0% v in
F-
N in oh c) .M"i v00 rn
ý N %0 r4 O. M -4 N F- in ri Q
j "
{r 'A O r"; N O O N O O M N N M p
r"1 ri .- "i
4-A Z,
H

4yý 4"+ YC >. O O O n

u. ºn in O in in in 0 O O in O O in O O
r, 41 Vl N ri M in N F- F- M N r-1 . 00 st of of in
" r""4
O O t4 v ýD U M N of 'D at M t-4
Gl H a+
ý
. rte
r"1 . "-1 LA
1-4 Cý
'. -"1 ri r1
. --
r-4 . -4"1 r-i . "i ri r-1 1-4 -H . -"1 '"-"1 r1

m
-! 0 ý? ' N Co Co "0 %0 %0 Co .0 N Q1 O 0 F- in
"^"1 .--4
E_ Q Q in O co "-4 in t') Co 8 in - 4 co of
y ý . -1 '0 1- 1 t- '0
41 E CQý (71ý c1 co O. in n C. c, rn C% in n n in n co n co
i-1 -4 1""1 r-4 r1 -4 ri r1 u-I ri u-I - N rl """1 "-1 rl r"1 r4 rl
'O 1G
.

i-"
S-1-4 5
b J, 0 O. 0 0 0
^y y in In GO N %0 00 M N
C" " m
eN 00 r4
ý""1
<Y V1
ri
Lfl 4 O 00 Ln V1 N N N 00 in
00 01
r-1 Ot
C71
fd y r-4
$' 0140 O O O O O O O O O O O p
(D .i .i .r .1 .1
4+ 7V to

t0
V O 0 0N Q O 0 O o O
M in n
u.
M in Cý g
" co M V 00 N u- I 01 '.O in i rý
yC
1r
r
O cli C-
M t- M n -: r r1 a O M O N M %0 N n N ß
Q QÜ 1-4 "'4 14 ri " -4 r-1 1-4 ""1 r-1 r/ r-4 rl r-4 r-4 """1 14 H
F. 40
v J

ý "
CA N) Ql N 0 M N Q1 M
, tn E -4 r-1
in N1
rl C1 "cý 00 N 'O co OO
in ' ""1 N '.O co in t- in ýp M !- M in r"4 N
° ß 0 1"1 . -1
y ß ß C71 co 1 ß 1- Ch u
co ý'VS4 N Ch
tO r -1 1-4 e-4 '"4 . .u 4 - 4 "-1 " -1 N r "4 u-I r -1 -4 . -1 rß-1 i
rco '--i

`
^ 9's " t- %0 r- -tr W) - m °
V 1') r4 r-4 Go ýq t-n
.), r4 1 -4 14 N r / N M M M M N N
ý
t0 O
V
V.

N
8
N O Q O co 00 F- 0 C1 N N O O p M
C co in at
"-4 1-4 e1 N N r-1 Pl Pl M M N N
"""1 '-1 u-I 1-4 r1

01 O Pl co 0 in %D .0 NG co in 'D OD O Q1
in ""- %D tY in
F- %D CO %0 %D .0 in in %D in C- ¢Y' ": r '.D V st NO in
E
N

ý"- Q ý H H H H H
. U
N

s
in. a4 c: ° 8 ` LA
1-0
s
 49.

For a lime content, the optimum moisture content was determined

and the strength corresponding to this moisture content on a

strength vs moisture content plot was taken as the unconfined

compressive strength for that lime content. ' The UCS vs lime

content was plotted and the optimum lime content and maximum UCS

determined. The values are shown in Table 4.4; it can be seen

that there is a wide range of values of maximum strength produced

by 4.5% or 6.0%-lime.

4.6.2 'Lime' reactivity

Lime reactivity is used as defined by Thompson (57). It is

the difference between'the maximum unconfined compressive strength

of the lime treated soil and the maximum UCS of the natural soil

cured for 28 days-at 220C. - The values are shown in Table 4.4.

4.7 Lime used

A comercial hydrated calcitic lime produced under the name

of Brancal, meaning white lime, by the company Argilas e Minerios

Nordestinos S. A. of Joäo. Pessoa, Paraiba, Brazil was used in

treating each soil. The chemical composition of lime is given as

(after De Carvalho (56)): -

CaO 68.1%

Ca(0F02 89.9% (calculated by molar ratio from

CaO present)

MgO 1.68%

Si02 1.19%

A1203 + Fe203 1.06%

Loss on ignition 28.9%

 50.

4.8 Chapter Summary

a) The nineteen soils investigated were sampled from an area of

North-East Brazil with the climate varying from hot and humid to

very hot and semi-arid.

b) The soils contain kaolinite as the clay mineral with quartz,

haematite, goethite and amorphous silica, alumina and iron

present.

c) While amorphous silica and alumina do not vary as the total

silica and alumina respectively, amorphous iron generally varies

as the total iron.

d) When treated with hydrated calcitic lime, the soils show a

wide range of variation of maximum strengths with 4.5% or 6.0%

generally required to produce maximum strength.

 51.

CHAPTER 5

POZZOLANICACTIVITY'AND MECHANISMOF REACTIONOF

RED TROPICAL SOIL-LIME SYSTPMS

5.1 Introduction

This chapter presents a study of the mechanisms of the

reaction that-takes place when red tropical soils are mixed with

lime. Because these soils contain many minerals in various forms,

such mechanisms and indeed-the reaction compounds are very

complex. Nevertheless, their understanding is essential for the

development of improved methods of red tropical soils utilization.

5.2 Review of' pertinent'literätüre'and methods

Two principal methods have-been used to study soil-lime

reactions. One method is based on the ability of the soil to

react with lime and the other on the. mechanism responsible for

such reactions. The review is therefore treated under two

sub-sections.

5.2.1 'Ability 'öf 'the 'soils 'tö 'react with 'lime

The classical method of assessing pozzolanic materials for

use with.. cement is that proposed by Fratini (74) and modified by

Rio (75). This method involves the determination of the amount of

lime left in solution after a mixture of the pozzolana and

Portland cement is kept in water at äpproximately 28°C for eight

days. Qualitatively, a reactive material"is, that which leaves

a low calcium ion concentration at the end of this period.

Soil engineers have assessed reactions between soil and

lime by measuring a change in some engineering property of the

 52.

soil. Hilt and Davidson (43) utilised the changes in plasticity

as a measure of the initial early reaction, while Thompson (57)

used changes in strength as a measure of a soil's lime reactivity.

He defined lime reactivity as the difference between the

unconfined compressive strength of a soil treated with lime for

maximumstrength, cured for 28 days at 22.8°C and the strength

of the untreated soil. This method has been used by De Carvalho

(56), and Harty and Thompson (70), among others, to assess 'the

lime reactivity of red tropical soils. While the use of such an

indirect method to study soil-lime reaction is acceptable and

useful, there is need to explore the use of the fundamental

reaction properties such as reaction rate, volume of reaction

products etc. Such reaction property as rate (which is measurable

at ordinary temperatures) can provide full or partial information

on the mechanism of reaction.

The methods in actual use for'studying the reactions of other

pozzolanas fall into' two''broad groups; a method based on solid "'

state transformation and a. method using a suspension of the

pozzolanic material in a solvent. The former employs high tem-

perature to effect the reaction between the material and lime.

Because of this, the method is useful in studying the kinetics of

pdrtland cement clinker phases as recently demonstrated by

Mackenzie and Banerjee (76).

Various materials, conditions of experiment and methods of

analysis have been used in the second group by researchers who

employ monodisperse suspensions. The materials include silica and

alumina in various states, pulverised fuel ash (PFA), and

naturally occuring soils in suspension of either lime solution or

 53.

hydrofluoric acid (or a mixture of hydrofluoric acid and nitric

acid). Barret et al (62) and Taylor (77) used different types of

amorphous silica in lime solutions while Greenberg. (78) used both

amorphous silica and quartz in lime solution for his investiga-

tion. Diamond and Kinter (47) used clays (kaolinite and

nxntmorillonite) in calcium hydroxide solution. The use of

compounds other than lime has so far been confined to workers

studying the reactions, of PFA. Hydrofluoric acid was used by

Raask and Bhaskar (79)ißn place of a mixture of the acid and

nitric acid which they considered too aggressive on PFA. The use

of such selective solvent as hydrofluoric acid which dissolves

only silica is not. desirable inýthe study of-red tropical soil-

lime reactions because the soils contain other reactive minerals

apart from silica. Thus the use'of the soil in lime solution is

preferred.
Lime solution is generally used because it is easier to
follow the removal of lime from solution not only at early, periods

of reaction but throughout the duration of the reaction. The

progress of the reaction has been followed in a number of ways.

The conventional volumetric analysis as used by Diamond and Kinter

(47) is time consum3,ng, cumbersome and hazardous. Greenberg (78)

followed the depletion of lime by a method based on the reduction

in electrical conductivity of the solution as reaction proceeds.

He assumed, in this method, that conductance is directly pro-

portional to the concentration of calcium hydroxide; the reaction

products and silica were said not to contribute to the con-

ductivity of the solution. In red tropical soils this assumption

needs verification because of the presence of iron compounds.

 54.

Recently, Barret et al (62) followed the progress. of the reaction

using atomic absorption spectrophotometry to determine silica and

lime. This method is neat and fast and is therefore favoured for

use in red tropical soil-lime solution system.

5.2.2 Mechanisms of reaction- ' the 'reaction 'models

Strengthening in soil-lime system is a result of cementi-

tious reaction products generated by the reaction between the lime

and the soil minerals (pozzolanic, reaction). The mechanism res-

ponsible for pozzolanic reaction is not clear. Two views emerge

from literature. While some postulate a dissolution of the soil

minerals to precede reaction, others favour a reaction through a

diffusion controlled process. So far, a qualitative approach has

been employed to study the mechanism. Scanning electron micros-

copy and sometimes transmission electron microscopy have been used

by most workers. Useful as these techniques are, the interpre-

tation of the microphotographs is difficult and calls for much

experience. Even where experience exist, the interpretation of

the same microphotograph may differ very widely indeed. Hence it

is felt that mechanisms based on mathematical models should be

considered. This if used'in conjunction with photographic

evidence should give a better resolution of the problem. The

mathematical theories of the two mechanisms are thus considered.

The derivation of the equations governing dissolution was

presented by Greenberg (78) and recently more lucidly by Raask and

Bhaskar (79). The rate of dissolution of a given mass is assumed

proportional to its surface area, thus: -

 55.

dw CA (5.1)
=- ... ... ... ... ... ...
dt

where = dissolution rate at time t

) C= rate constant

A= total surface area.

For uniform sized spherical particles equation (5.1) becomes

dw C1W2/3 (5.2)
=- ... ... ... ... ...
dt

where W= weight of undissolved material at time t

i. e. C2t = Wo1/3 - W1/3 (5.3),
... ... ... ...

i. e. C3t =1- Wmi/3 (5.4)

... ... ... ...

where Win -W
- mo
,

and Wo = weight of undissolved material at time zero.

Equation (5.4) is valid for a monodisperse suspension.

For materials of particle size normally distributed, equation

(5.4) becomes
I
Cot =1- Wma (5.5)
... ... ... ...

For materials of particle size log normally distributed,

equation (5.4) becomes

Wm1/5 (5.6)
CCt=1-
... ... ......

Hence if the dissolution theory is valid, experimental results on

lime consumption with time should obey equation (5.5) or (5.6).

Kinetic models for solid state reactions based on diffusion

processes have been considered by Jander (80),

controlled

Ginstling and Brounshtein (81) and Carter (82). Jander's model

is the simplest but does not consider the effect of the varying

volume of reaction products on the rate of formation of new

reaction products as does Carter's. Jander based his model on

 56.

the reaction which occurs between two spherical particles and on

the assumption that the rate of thickening of the reaction

products is inversely proportional to its thickness, thus:

.
(5.7)
...... ..,: .........

dt y

where dy/dt = rate of thickening of the reaction products

k= rate constant

y= thickness of the reaction products.

Integrating equation (5.7):

y2 = 2kt (5.8)
... ... ... ... ...
The volume of unreacted material V, at a time t is given by:

V= (r-y)3 (5.9)
... ... ... ... ...

4
or V=3r 3( 1-x) (5.10)
... .... ... ... ...

where r= radius of reacting sphere

x= fraction of sphere which has reacted.

From equations (5.9) and (5.10):

y= rEl - (1-x)1/3J (5.11)

... ... ...
Substituting y from equation (5.8) in (5.11) gives

2kt
[1 - (1-x)h/3]2 _ =, Kt (5.12)
... ...
r2

Equation (5.12) is the well known Jander equation from which the

rate constant of reaction can be determined. The equation can be

written as:
F(x) = [1 - (1-x)1/3]2 = Kt (5.13)
... ...
Hence a plot of F(x) against time should give a straight line.

It may be worthwhile to point out that neither of the two

models considered truly represents the actual practical red

 57.

tropical soil-lime system, since the lime in the system is-,

partly dissociated and partly solid.

5.3 Experimental method

The reaction between soil and lime was studied by a method

similar to that used by Barret et al (62) which consists basically

in measuring the amount of lime left in solution at predetermined,

time intervals using atomic absorption spectrophotometry. The

detailed procedure is as follows: -

a) Preparation of lime solutions by mixing 0.15 gm or

0.10 gm of hydrated lime with C02 free distilled water;,

b) Mixing 2.5 gm of soil with appropriate concentration of

lime solution so as to obtain either 6% or 4.5% lime

concentration-in the soil. These percentages correspond

to the optimum percentages of lime required for maximus

strength (see Table 4.4). Thus a soil requiring 6% for

maximum strength is mixed with lime concentration of

1.5 gm/1 while 1.0-g/l is used for that requiring 4.5%

lime for maximum strength;

c) Placing the mixtures of soil and lime solution in

plastic containers tightly sealed to prevent the carbo-

nation of the lime;

d) Storing the sealed containers in an oven at a constant

temperature of 2S°C;

e) Shaking the containers periodically and measuring the

lime concentrations at predetermined times. To measure

lime concentration by atomic absorption it is necessary

to have very clear solutions. To achieve this, an

aliquot of the soil-lime mixture is taken after the

 58,

containers have been shaken. The aliquot is centifuged

at 2,500 rpm for 5- 10 rains. and the lime concentration

is determined from 1 ml of the supernatant.

A preliminary study was conducted with two soils using three

different size fractions of each, i. e. 2 mm - 0.074 um, 0.074 iron -

0.002, mm and < 0.002 mm. This was done to determine the reactive

fractions and to assess the magnitude of contribution to lime

consumption by the coarser fractions of the soil. It is a known

fact that many of the coarser particles in red tropical soils are

only aggregations of the clay size fractions (3,4) and thus they

may influence the overall soil-lime reactions.

The results of this preliminary study are presented in

Table 5.1. The clay fraction is 3 to 5 times as reactive as the

sand fraction and 2 to 3 times as reactive as the silt fraction.

The two' soils used' in the preliminary study represent the most

reactive and one of the least reactive of the nineteen soils.

The ratios quoted above will increase further, in favour of the

clay size fraction, when all the fractions are mixed together in

presence of lime since they will be exposed to a "competition"

for lime which was not necessary when each fraction was separately

mixed with lime. Therefore it is apparent that by using only the

clay size fraction, a representative result can be obtained for

any of the soils. Consequently, the study was conducted with the

clay size fraction of each of the nineteen soils.

5.4 Measurement by atomic'absorption'spectrophotometer

5.4.1 Basic principles

If a solution containing a metallic ion is aspirated into a

flame, such as air-acetylene flame, an atomic vapour of the metal

Table 5.1 Fraction of lime consumed at varying times for

different fractions of two of the soils investigated

 60.

tn 00 (0) r- \O N
q i4 00 ý r4 M
rß-1 Ln

tý O O O O O

C)
4.4 00 t-
cd 0) N. ýi
A Ln O M

C) Ln O r-I
O O O' O O

U
Cd
H
G)
O cq %0 C)
00 r-I
co ý °°
a Ä \0
O
0 rl N
c) rl fit'
0
44 .
O O O O O O O
ýr

..... .. ..
ßl1 r-
V1 Ln
cn M Ö
%D
Ö Mt
N O O O O

00 N
cq -t
'O (V Ql O 0
A O N 0
.ý
O O O O O O
r-1

8
ice ö o
O O
"' N N
ý

o ö ö O
N O v N O V

O
ý
Q)
S
z H
"o
cn a, ý
r+ ý.
rl rl O "-' "r I r-ý
v
"i 2R3
9 h Q
6
 61.

will usually be formed. Someof the atoms of the metal maybe

raised to an energy level sufficiently high to emit the characte-

ristic radiation of that. metal. This phenomenon is utilised in

the emission flame photometry technique. However, an over-

whelmingly larger percentage of the atoms of the metal will remain

in the non-emitting, "ground" state. These ground-state atoms can

receive light radiation of'their specific resonance wavelength

(generally the wavelength being the same wavelength they would

emit if excited). Thus, if light of this wavelength is passed

through a flame containing atoms of the metal, part of that light

will be absorbed, and the absorption will be proportional to the'

density of the atoms in the flame. This phenomenon is exploited

in atomic absorption spectroscopy.

5.4.2 Features of an atomic absorption spectrophotometer

An atomic absorption spectrophotometer consists essentially

of the following components:

a) a stable light source, emitting the sharp resonance line of

the element to be determined;

b) a flame system into which the test solution may be aspirated

at a steady rate. The temperature of this flame should be suffi-

cient to produce an atomic vapour of the required metal'from the

compounds present in the test solution;

c) a monochromator lens to isolate the resonance line and focus

it upon a photomultiplier;

d) a photomultiplier that detects the intensity of the light

energy falling upon it, and which is followed by facilities for

amplification and readout.

 62.

The light source is usually a lamp having a hollow cathode made

.
of the element'to be'determined. The emission from this lamp is

modulated so that its radiation only, and not that emitted from

the flame by some of the atoms of the element in the test

solution, will be recorded in the galvanometer signal of the

readout.
5.4.3 Operation 'of 'a 'spectrophötometer

In using an atomic absorption spectrophotometer to determine

the elements in a test solution, a calibration curve for that

element has to be constructed first. A meter or the needle

reading is adjusted to read zero when 'a blank solution (distilled

water) is sprayed on to the flame. In this condition, the

"unobstructed" light of the hollow-cathode lamp passes on to the

photomultiplier. When'standard solutions of the element to be

determined'(i. e. solutions in which the concentrations of the

element are known) are sprayed'on to. the flame, part of the light

from the lamp is absorbed resulting in a diminution of light

intensity falling upon the photomultiplier. The readings of the

meter are noted and a graph of concentration of the element

against meter reading is plotted.

The test solution is sprayed on to the flame and the reading

of the meter is noted. From the calibration curve, the concentra-

tion of the element in the test solution can be determined.

5.4.4 The atomic 'absorption 'spectröphötometer 'tised

A Perkin-Elmer (603 series) atomic absorption spectrophoto-

meter owned by the Scientific Division of the Central Electricity

Generating Board at Harrogate was used for the elemental analysis

in this study. The instrument is capable of being calibrated

 63.

internally so that the concentration of the element being

measured in parts per million (ppm) is read directly on a digital

display. The reading displayed is an average of the readings over

a time interval which is chosen by the operator. One minute time

interval was used.

5.4.5 Calcium determinätion

When an air-acetylene flame system is employed, depression

of calcium response occurs in the presence of aluminium, magnesium

and phosphate ions. Such depression can be due to the formation

of refractory compounds that are not dissociated in the flame.

This interference is overcome by the addition of lanthanum

chloride.

To determine calcium, 1 m1 of the supernatant solution of

the centrifuged aliquot was made up to 50 ml and the ppm of

calcium determined after 1,2,4,5,7,14,21,28 days. at the

following standard conditions.

Wavelength: 211. uv

Slit setting: 4(1.4 nm)

Light source: ' hollow-cathode lamp

Flame type: air-acetylene, oxidising (lean blue)

Because of the possible presence of aluminium lanthanum

chloride was added each time a measurement was made.

The fraction of lime reacted after time t is calculated as
follows:

Extractable calcium at time ppra X Vol. ' Used x, d. f

= Ca(t)
sample wt. X 103

where d. f = dilution factor

final 'diluted 'volume ' iii ml. ''
_
volume of aliquot taken for dilution in ml.
 44.

74.08
Calcium as Ca(OH)2 = x Ca(t) =y... (5.14)
40.08

(5.15)
. ': Fraction of lime reacted =1- _Y_ """
0.15
or
=1- (5.16)
___
0.10 ...

The values of these fractions as a function of time are given in

Table 5.2.

5.5 Results and *discussion

The results of the soil-lime reaction measurements are

discussed below.

5.5.1 Lime 'consumption

The relations between lime reacted or consumed and time,

obtained in the study, are presented in Figs. 5.1 and 5.2; for

clarity of presentation, the soils have been divided in four

groups as shown in the figures. The shapes of the curves are

similar to those obtained by Diamond and Kinter (47) for

adsorption of calcium hydroxide by kaolinite vs time in mins., by

Barret et al (62) for the depletion of lime by amorphous silica

vs time in days and Greenberg (78) for consumption of lime by

amorphous silica (measured by specific conductance of the

solution) vs time (mins).

The relations between lime consumption and time in Figs.

5.1 and 5.2 show that lime is consumed at a "fast" rate in the

initial stages of the reaction, that is, up to five to seven days.

After this period the consumption continues but at a much reduced

rate. This continuing consumption is in line with the general

opinion that pozzolanic reaction is time dependent and therefore

strength development is gradual but continuous for long periods

Table 5.2 Fraction of lime consumed at various times by

the soils
 66
.

SOIL FRACTIONOF LIME REACTED

AFTER
SYMBOL
1 DAY 2 DAYS 4 DAYS 5DAYS -7-DAYS 14 DAYS 21 DAYS 28 DAYS

JPA 0.218 0.28 0.404 0.449 0.49 0.510 0.502 0.53

JIM 0.324 0.378 0.458 0.538 0.60 0.623 0.640 0.651

JPB 0.146 0.206 0.298 0.324 0.420 0.434 0.451 0.467

CT 0.20 0.227 0.289 0.333 0.362 0.359 0.371 0.380

Al 0.111 0.164 0.278 0.298 0.320 0.333 0.356 0.372

All 0.20 0.253 0.324 0.38 0.394 0.410 0.425 0.437

SIA 0.125 0.167 0.25 0.28 0.291 0.310 0.308 0.326

SIB 0.191 0.233 0.35 0.40 0.421 0.437 0.445 0.462

SII 0.34 0.413 0.52 0.56 0.580 0.587 0.598. 0.589

NF 0.262 0.324 0.404 0.422 0.417' 0.434 0.447 0.458

JI 0.167 0.27 0.43 0.48 0.52 0.527 0.532 0.547

JII 0.19 0.217 0.50 0.52 0.533 0.547 0.558 0.562

TI 0.247 0.32 0.41 0.45 0.437 0.462 0.473 0.482

TII 0.33 0.417 0.46 0.48 0.52 0.533 0.538 0.546

RC 0.20 0.20 0.253 0.262 0.289 0.292 0.298 0.316

USM 0.156 0.173 0.224 0.28 0.301 0.318 0.333 0.345

SB 0.191 0.227 0.317 0.346 0.396 0.407 0.416 0.427

PH 0.20 0.431 0.44 0.493 0.498 0.510 0.519 0.523

'IN 0.20. 0.271 0.316.. 0.369.. 0.381.. 0.389 0.398.. 0.407

. . .
Fig. 5.1 Relations between consumption of lime and time

for the nineteen soils investigated.

 68.

Reaction
time, days
0 10 20 30
0

0.2

0.4

0.6

0B
d
C. ý

Eo

4.
O. A

0"b

0.6

0-8
 t

Fig. 5.2 Relations between consumption of lime and

time for the nineteen soils investigated

ý-

-ý
 70.

time, days
Reaction
1n 90 30
0

02

0.4

0.6

0.8
-o
a,
v
d
d
I-
0)
E

0
-T;
v
d
U-

oýý

0.1

0.8
 71,

of time. The ratios of the unconfined compressive strengths of

lime treated soils after curing for 28 days to their strengths

after curing for. about 7 days for moderately reactive soils have

been reported to vary between 2 and 3 (70). If this substantial

increase in strength is totally a result of-the formation of more

cementitious reaction products, this should result in substantial

lime consumption between 7 and 28 days because the amount of lime

which disappears from the solution equals that which reacts with

the soil to produce cementitious reaction products. However, the

figures show that further lime consumption after the initial

rapid period-of about 7 days is. limited. It is interesting to

point out that the total lime consumption between 7 and 28 days

corresponds only to eight percent of the lime consumed during the

initial, seven days. Therefore, it appears that the increases in

-,
strength of a soil-lime system beyond the. seven day, period cannot

be totally explained in terms of generation of pozzolanic reaction

Thus it is suggested that the increase in strength

products.

shown by, these soils beyond seven days is mainly due to the

changes that take place in the structure of the cementitious

formed during pozzolanic reaction, i. e., hydration and

products

increases in crystallinity of the reaction products.

5.5.2 Mechanism 'of 'reaction

The numerical values of the lime consumption vs time rela-

Lions were used to test the reaction mechanisms based on the

mathematical models discussed previously. It was found that the

experimental results do not follow the relations based on the

equations which represent the dissolution theory,. despite the fact

that the experimental conditions were such as to give monodisperse

 72.

suspensions for which the equations are intended to be valid.

This finding coincides with the results of Raask and Bhasker (79)

who reported that the dissolution model was not applicable to

pulverised fuel ashes.

Since the volumes of the reaction products formed with time

were not measured, the experimental data were tested with Jander's

diffusion controlled model. The experimental results were found

to approximate very closely to the values obtained with the

function F(x) according to equation (5.13). It may be pertinent

to point out that Mackenzie and Banerjee (76) used Ginstling and

Brounshtein's equation (81) also based on a diffusion model to

analyse the formation kinetics of tricalcium aluminate (C3A).

They found that the fraction of lime reacted vs time obeyed the

equation and they deduced the rate constant of reaction from the

relationship. The experimental points and the theoretical lines

are shown in Figs. 5.3 and 5.4. They were plotted from values of

F(x) in Table 5.3. The point at which the experimental results

deviate from the mathematical model coincides with the time at

which the rate of lime consumed changes i. e., 5-7 days. The

linearity of the plots in Figs. 5.3 and 5.4 up to 5 or 7 days is

taken as an indication of a degree of satisfactory fit. Fig. 5.5

shows the relation between F(x) and time for the total period of

time investigated for some of the soils. The deviation of the

plots after 5-7 days may be explained by the fact that one of

the weaknesses of the Jander model is its inability to incorporate

the effect of the varying volume of reaction products being formed

and therefore their influence on the reacting system. It is

apparent then, that the increased volume of reaction products

 4

14

Table 5.3 Numerical values of the first term of Jander's

equation F(x) for various times of reaction

'3

1I

ýi

ýi

lk
 74.

SOIL THE FUNCTIONF(x) ACCORDING

TO JANDER'SEQUATION
SYMBOL 1 2 DAYS 4 DAYS 5 DAYS 7 DAYS 14 DAYS 21 DAYS 28 DAYS
DAY

JPA 0.00619 0.0108 0.0251 0.0325 0.0404 0.0448 0.0432 0.0495

JPM 0.015 0.0214' 0.0341 0.052 0.0693 0.0785 0.0833 0.0876

JPB 0.00263 0.00548 0.0124 0.015 0.0276 0.0302 0.0328 0.0358

CT 0.00514 0.00676 0.0115 0.016 0.0194 0.019 0.0205 0.0217

AI 0.0015 0.0034 0.0108 0.0124 0.0146 0.016 0.0186 0.0206

All 0.00514 0.00858 0.0150 0.0217 0.0236 0.0260 0.0284 0.0304

SIA 0.0019: 0.00349 0.00836 0.0108 0.0117 0.0135 0.0133 0.0152

SIB 0.00465 0.00716 0.0179: 0.0245 0.0277 0.0304 0.0318 0.0348

SII : 0.0167: 0.0265: 0.0471. 0.0573 0.0631 0.0640 0.0658 0.0686

0.0093: 0.015 0.0251: 0.0279 0.0271 0.0299 0.0321 0.0341

JI 0.00349 0.0099' 0.0292: 0.0384 0.0471 0.0488 0.050 0.0538

JII 0.0046: 0.0161. 0.0426" 0.0471 0.0502 0.0538 0.0568 0.0579

TI 0.00814 0.0146' 0.0265 0.0321 0.0304 0.0348 0.0370 0.0388

TH 0.0156 0.0171: 0.0345" 0.0384 0.0471 0.0502 0.0515 0.0536

RC 0.00514 0.00514 0.0084' 0.0093 0.0115 0.0118 0.0124 0.0141

USvI: 0.00302 0.00376 0.0066 0.0108 0.0127 0.0141 0.016 0.0173

SB 0.00465 0.0068 0.0142 0.0174 0.0239 0.0256 0.0269 0.0287

PSI 0.00514 0.0164 0.0309 0.0411 0.0421 0.0448 0.0469 0.0478

S4 0.00514 0.010 0.0141.. 0.0202 0.0218 0.02 29 0.0242 0.0256

. .". . . . .
Fig. 5.3 Relations between the first term of Jander's

equation F(x) and time of reaction up to 5 days

 76.
006

0.05

x
0-04
L 0.03

0.02

001

Time,days

0.05

a04

X0.03
LA-

0.02

001

Time,days
Fig. 5.4 Relations between the first term of Jander's

equation F(x) and time of reaction up to 5 days

 78.
0.06

0.05

004
LA-

0.03

0.02

0.01

Time,days

0.0

0.01

O.0'
rx

0.O

0.0

Time,days
Fig. 5.5 Relations between the first term of Jander's

equation F(x) and time of reaction up to 28 days

for four of the soils
 80.

'M

0.08

0.07

006

0.05
PA

x
0
L.,..."0L,

PB
0.03

0.02

001

0 268 10 12 14 16 18 20 22 24 26 28
Time,
days
 81.

reduces the rate of diffusion as shown in Fig. 5.5.

While caution must be exercised when deducing mechanistic

information from mechanistic reasoning, it seems most likely that

the rate of reaction in red tropical soils is controlled by a

diffusional process. While this may partially support Stocker's

(45) deductions, dissolution is difficult to discount as a

possible mechanism through which amorphous minerals present in the

soils react with lime (62). In reviewing the reactions of

aggregates in Portland cement concrete, Hansen (83) discussed the

dissolution of quartz and siliceous minerals in general and

concluded that the dissolution of silica involves:

.
a) Consecutive-penetration of the electron cloud of an 02-
ion by two protons to form OH ions; penetration of two
OH- ions-into. the electrical field of the Si4+ ion to

produce the unstable Si(OH)4 group (monosilicic acid), .

b) two more OH ions enter the Si(OH)4 electrical field

to, form an Si(OH)6 ion,

c) the Si(OH)6 ion being 6 coordinated-can no longer form

bonds with, the other silicon atoms in the silica lattice

which are 4 coordinated,

d) for proper screening (electrically) this ion combines

with two protons to give the molecule H2Si(OH)6 which is

free to migrate i. e., to dissolve.

For dissolution to occur, therefore, the material should be able

to provide surfaces consisting of oxygen and unbonded OH- ions as

required by step (a). After cation exchange is satisfied, the

available off ions from the dissociation of the very small per-

centage of the lime added to the soil is limited. Fripiat (84)

 82.

concluded that the 'hydroxyl' surfaces of amorphous silica and

alumina hydrates were unlike true hydroxyl surfaces; they could

more correctly be described as oxygen surfaces permeated by a

cloud of delocalised protons. In characterising some soil

allophanes, De Mumbrum and Chester (85) noted that soil allophanes

have chemical and structural characteristics similar to the

mettalic silicates of aluminium and manganese and appear to be

highly hydrated, to have many exposed, unbonded OH groups and that

their amorphous nature suggests "solid-solutions" of alumina and

silica. Therefore, it will appear that these accounts of the

"abnormal" state of existence of the surfaces of amorphous

minerals and the OH ions will satisfy the pre-requisite for

dissolution (step a) as given by Hansen (83). As a consequence,

the amorphous silica and alumina of the red tropical soils used

in this investigation might initially and for a short period

react through a dissolution mechanism. Because there is no sharp

transition from a non-crystalline to crystalline state in these

soils, it is difficult to draw a line beyond which dissolution

is unlikely to occur.

The rate of reaction of the crystalline minerals is lower,

the mechanism not withstanding. Hence their reaction with lime

will follow in the presence of reaction products formed by a more

rapid reaction outlined above. The situation of the clays more

or less resembles the state of the clay minerals when cement is

Ito treat soils containing only crystalline clay minerals.
used

Taking the red tropical soil-lime system as a whole, it appears

that apart from a short rapid reaction of the amorphous minerals

possibly through dissolution, the reaction is mainly a

 83.

diffusion controlled process.

5.5.3 Pozzolanic activity

The pozzolanic activity of a soil is defined as the ability

of some or all of its constituents to react with lime to produce

cementitious products. This ability is directly related to the

type and quantity of the reactive constituents and the reaction

rate. The latter depends on such factors as degree of crysta-

llinity, surface area or fineness of the reactive constituents,

pH of the reaction medium, temperature, intimacy of mixing etc.

Thus a determination of the rate constant of reaction under

particular experimental conditions is an indirect measure of

these factors.
From equation (5.13) in sub-section 5.1.2, the slopes of the
in Figs. 5.3 and 5.4 give 2.
lines presented the quantity

Assuming the same mean radius for the clay fraction of all the

soils investigated, the slopes of the lines in Figs. 5.3 and 5.4

are measures of the rate constants of reaction. Since the clay

size fraction is the most reactive, the quantity of the reactive

constituents is expressed by the fraction of material smaller than

0.002 mm although it is recognised that most of the active

components in the clay are really the amorphous and semi-

crystalline materials.

It is proposed here that the pozzolanic activity of a red

tropical soil should be expressed by an activity index which is

related to the rate constant and the percentage of the reacting

compounds by the following equation:

 84.

Ai = Rix ... (5: 17)

... ... ... ...
where Ai = activity index

K= rate constant

n= fraction less than the clay size

x=a constant depending on the type of reacting

compound in the clay fraction.

Although the introduction of the constant x may seem arbitrary, it

is based on the fact that different soils may have different types

of reacting compounds within their clay size fractions. For red

tropical soils these compounds are silica, alumina and possibly

iron; thus x is defined as a positive number whose function is to

convert the reactive silica, alumina and iron to equivalent

silica. It is proposed that. x takes a value of 2.

Hence Ai = Kn2 (5.18).

... ... ... ...
The values of the rate constants and the activity indexes

for the soils using equation (5.18) are presented in Table 5.4.

The Ai values range from 0.12 x 10-4 per day to 4.08 x 10-4 per

day, giving a wide range for classification purposes.

The usefulness of the activity index should be proved

against other accepted parameters. This was done by correlating

the activity index with the unconfined compressive strength of the

soil-lime systems and also by investigating its relation with the

well-known lime reactivity parameter of Thompson (57).

5.5.4 Activity index and unconfined compressive'strength

The relationship between the activity index Ai, and the

unconfined compressive strength (UCS), is shown in Fig. S. 6 with

the UCS as the dependent variable and Ai as the independent

variable. A stepwise multiple regression analysis was used to

relate Ai to UCS. The aim was to produce a linear combination of

Table 5.4 Rate constants and activity indexes for the soils

investigated
 86.

Pozzolanic
"Rate Constant" Activity
Soil Symbol K Index
per day Ai x 10-4
per day

JPA 0.0035 0.172

JPM 0.0052 4.08

JPB 0.0015 0.12

CT 0.0016 0.291

AI 0.00124 1.116

All 0.00234 0.791

SIA : 0.00108 0.461

SIB 0.00245 0.120

SH 0.00543 3.13

N 0.00279 0.472

JI 0.00384 0.384

JII 0.00471 0.567

TI 0.00340 1.646

TH 0.00384 3.689

RC 0.00118 1.062

USM 0.00108 1.248

SB 0.00174 2.512

PH 0.00411 1.812

94 0.00202. 1.07...
. .. ..... .
 &

Fig. 5.6 Relation between activity index and the unconfined

compressive strength of the red tropical soils

1ý

Fig. 5.7 Relation between the activity index and the

reactivity parameter of Thompson

e
 88t

60

50

6.0
clý

E
3.0

2.0

1.0

0 0.5 1.0 1.5 1U1.5 su il5 4-u

Activity index, Ai x 10 per day

4.0
C4
E

ý: 3.0
u
d
0.)

0,2.0
E_
J

1.0

0
Activity index Ai x 10-4 per day
 89,

Ai and/or powers of Ai which correlates as highly as possible with

the'UCS. It was assumed that UCS is either a linear function of
Ai or a power function which can be expressed by polynomials. The

Statistical Package for the Social Sciences (SPSS 5), a ready

computer programme available in. Leedý University. computer,

was used for this purpose. The equation'of best fit which

expresses the relation between, UCS and Ai was found to be:

UCS = 1.639 Ai + 0.0135 Al

- 0.0276 Ai + 0.33 (5.19)
...
where UCS = unconfined compression strength in MN/m2

Ai = activity index x 10-4" per day.

This equation is statistically highly significant with a multiple

correlation coefficient of 0.973 at 95% confidence. The standard

error 15 0.359.
.
Seventeen out of the nineteen soils investigated seem to

behave in the pattern as shown'by equation (5.19). The deviations

of the two soils (Teixeira II and. Penha) may be due to error in

their strength determinations or due to the type and cementing

quality of the reaction products formed (a subject of chapter

six).

5.5.5 Activity ' index "ys Lime 'reactivity

Figure 5.7 shows the variation between activity index and

lime reactivity as defined by Thompson (57). A stepwise multiple

regression was used to find the lime of best fit with lime

reactivity as the dependent variable and the activity index as the

independent variable. Unlike equation (5.19), the best fit is

given -by
LR = 0.72Ai + 0.278 ... ... (5.20)
... ... ...
 90.

where LR = lime reactivity in MN/m2

Ai = activity index x 10-4 per day.

The standard error of prediction is 0.468.

This equation is highly significant with a correlation coefficient

equal to 0.88 at 77% confidence. Three soils (Teixeira II, Penha,

Areia I) out of the nineteen investigated, probably due to the

reasons given in sub-section 5.4.4, do not obey this relation.

The linear relationship of equation (5.20) suggests that

the activity index and lime reactivity have a common basis. Lime

reactivity is a measurement (strengthwise) of the effect of the

reaction products formed by reactions the characteristics of which

determine the activity index. On the other hand, the polynomial

nature of equation (5.19) suggests that the total strength of the

soil-lime system is not determined solely by the reactions between

soil and lime i. e., the reaction products do not exclusively

determine the strengths of red tropical soils when they are

treated with lime for optimum strength. The activity index thus

seems to be a valid representation of the pozzolanic activity of

red tropical soils investigated in this study.

5.5.6 Activity ' index 'and 'cIassificdtion

As pointed out in Chapter 2; section 2.2, the parent

material from which the soils are derived and climatic conditions

which are significant for classification of red tropical soils

influence or determine "the type and'nature of their mineralogical

constituents". De Graft-Johnson et al (18) and Gidigasu (29)

showed that tlateri, tict gravels and fine grained 'laterite' soils

can be classified on basis of geology, amount of clay content,

linear shrinkage and climatic conditions. ' In proposing the
 91.

activity index the clay content (n) and the type of reacting

compounds (the factor x) have been considered and used. Further-

more the reaction rate which may indirectly measure the nature of

the reacting compounds was used in defining the activity index.

Thus the activity index which incorporates' the significant

parameters for classification purposes, may be a valid parameter

for the classification of the soils.

The engineering characteristics used for classification will

vary depending on the objectives and field service environment.

The criteria for classification can be*broadly divided into two:

a) that based on situations where immediate improvement

in plasticity, improved workability, innnediate strength

increase and reduced'swell potential are required;

.
b) that based on a 'long time' strength improvement

associated with cementation by. reaction products.

Classification based on criterion (a) is not considered here as

V

the engineering properties used in (a) were not determined for

this work.
Classification based on strength usually divides the soils

into groups based on ranges of strengths. The range varies

according to the use to which the material will be put, for

example, the range will be higher for base materials than for sub-

base materials since stresses and durability conditions differ for

various depths in the road pavement. The strength ranges so

chosen can be validated only on basis of experience and actual

field. performance. Literature is virtually non-existent on the

actual field performance of red tropical soils treated with lime.

The result-has been that workers who attempt to classify the soils
 92.

have chosen arbitrary ranges. Harty and Thompson (70) divided

the twenty-six soils they investigated into five arbitray groups

on basis of their 28-day cured strengths after lime treatment for

maximum strength. Minimum values of UCS suggested by other

workers as an indication of lime treated soil's suitability vary

very widely. For instance, for use as road bases the Central

Road Research Institute of India suggests a minimum UCS value of

150 p. s. i (ti 1.03 MN/m2) while Clare and Mohan (86) used 300 psi

(x 2.06 MN/m2), for assessing the suitability of "stabilized

laterites" from Nigeria.

It is not the intention here to add a few more arbitrary

ranges without field validity. However, because the activity

index is obtained much faster than strength at 28 days, it may be

worthwhile to illustrate its potential use as a classifying

parameter. In doing this, use is made of the American Association

of States Highway Officials (AASI-D) requirements for assessing

suitability of lime treated temperate soils for use in road

pavement sections. For use as a sub-base (minimum strength of

0.70 MN/m2) an activity index of approximately 0.4 x 10-4 per day

is required while for suitability as base (minimum strength of

1.03 MN/m2) an activity index of approximately 0.75 x 10-4 per

day is required. Based on this, a tentative classification of

the lime treated soils investigated, for use as road materials is

proposed in Table 5.6.

5.6 Uses of 'and 'using 'the activity 'index

The activity index is much faster to determine than the

lime reactivity. It also requires less labour and materials.

These facts cannot be overlooked where skilled labour and/or

 0

Table 5.5 Tentative classification of red tropical

soil-lime for use in road structures

 ö

cj
w
0

tf)

C)
Ü
P.
.H

"Ü
U ýO
cd

ba

z x

p O O O
U
45 ý N

i-1 tai Q
v
41
r-1
U

C)

8 8
ý-o 9 n

CD o

ý--ý N
 95.

material is scarce. This index is not only useful in estimating

the strength and lime reactivity of soil-lime mixtures but also

potentially useful for classification purposes.

However, it should be exphasized that estimates made using

the Activity Index is subject to error, the error increases with

decreasing Ai value. It is therefore suggested that forecasts

based on the index should serve as initial estimates and as

useful aid in selecting materials, followed by actual laboratory

and/or field strength testing. Much care and thought should

accompany the use of the regression equations for soils other than

red tropical soils as they may not be applicable because of the

role played by the amorphous constituents of the red tropical

soils.

5.7 Chapter Summary

a) Formation of reaction products as represented by lime

consumption takes place during the initial five to seven days;

therefore the gains in strength beyond this period appear to be

due mostly to the hydration and increased crystallinity of the

reaction products.

b) Under the experimental conditions, the bulk of the reaction

between soil and lime is a diffusion controlled process apart

from the initial rapid reaction of the amorphous minerals

possibly through dissolution.

c) The activity index based on the diffusion model and the

quantity and type of the reacting compounds of the soil is

statistically related to the unconfined compressive strength of

the soil-lime system.

 96.

d) A highly significant statistical linear correlation links

the empirical lime reactivity as defined by Thompson and the

proposed Activity Index. This suggests a common basis for the

two methods and therefore allows'the use of either of the two for

assessing the pozzolanic properties of the soils.

e) The activity index is faster to obtain than lime. reactivity

and is potentially very useful tool for classifying red tropical

soils if field performance of the soil-lime system. is not

available. It does not replace laboratory testing but. serves

mainly. to help in recognising the active soils.

 97.

ri-TAPTFR 1

RED TROPICAL SOIL-LIME REACTION PRODUCTS

6.1 Introduction
The mineralogical composition of the soils and their

reactions with lime have been considered in the preceeding

chapters. The reactions lead to the formation of cementitious

products which may be partly responsible for amelioration, but

definitely responsible for strengthening in those soil-lime

systems that show increase in strength.

In this chapter, the nature, type and characteristics of

the reaction products in such a complex system as red tropical

soil-lime are considered. The mineralogical composition of the

soils points out to the possible formation of CAHip and/or C2AH8.

These compounds are known to form during the hydration of some

high alumina cement and later convert to C3AH6 (a more stable

form) resulting*in decreased strength. The role of iron in the

reactions needs to be assessed. Lea (87) observed that "some

ferric iron compounds in aggregate, for example, appear to react

with lime without deleterious effect, but there is evidence that

oxidation of ferrous iron is one factor contributing to the

deterioration of some dolerite concretes". Could this be true of

red tropical soil-lime systems?

Thus the objectives of this chapter are:

a) to identify and possibly quantify the cementitious reaction

products.

b) to examine the influence of the reaction products on strength.

 98.

Only the main cementitious reaction products are considered;

other products of reactions such as calcium carbonate are

considered in latter chapters.

6.2 Pertinent literature and'methods

The subject of this sub-chapter is treated under. three

sections.
6.2.1 Identification of reaction' products

The products formed by reactions between lime and clays have

been extensively studied. In Chapter 2, sub-section 2.4.2.5, the

formation of hydrates of calcium aluminates and silicates (of the

tobermorite-like type) when clays are treated with lime has been

noted. On the contrary, there is virtually no literature on the

reaction products of red tropical soil-lime systems.

The methods used for the identification of reaction products

can be broadly grouped as: -

a) methods based on radiation, (b) thermal analysis,

c) chemical methods and (d) optical methods. The method of

electron optics is considered in Chapter 8. X-ray diffraction

(XRD) is a typical example of radiation methods while differential

thermal analysis and thermogravimetry usually are typical thermal

Volumetric chemical analysis of extracts is used to

methods.

chemically determine the constituents and overall composition of

the reaction products.

Thermal methods are based on weight loss/weight gain or

reactions of compounds heated in a known

exothermic/endothermic

atmosphere. The temperatures at which the reactions occur charac-

terize the compounds and are therefore used to identify them

 99.

qualitatively. These methods are therefore limited to those

compounds that react when heated in a known environment. Like

many other methods they do not resolve most overlaps i. e. do not

identify minerals that react to heat at the same temperature and

same environment when they are present in a mixture.

Chemical methods determine the overall chemical composition

of the products but do not give information on the composition of

the individual minerals constituting a material.

The most widely used method of identification is XRD. - The

principle of the use of XRD in soil mineral analysis is based on

the fact that the orderly atomic structure of the crystals present

in a soil causes them to behave to x-rays as a diffraction grating

would to visible light. The atoms near the beam of single wave-
length vibrate and re-emit x-rays as if their source were the atom
itself. Similarly the atoms in the deeper layers are penetrated
by the x-rays and behave as the surface atoms. At some angle of

incidence of the x-rays the reflections from the successive layers

reinforce one another to give a diffracted beam which obeys

Bragg's equation:

nX = 2d sin e (6.1)
... ... ... ...
where A= wavelength of x-rays
d= spacing (basal) between each of the diffraction

" planes
e= angle of incidence of x-rays

n= order of reflection, usually 1 for the diffracted

beams to reinforce one another.

If the crystals of a mineral are rotated in an x-ray beam, the

possible sets of the diffracting planes of its atoms can be

 100.

obtained as each set of planes momentarily satisfies Bragg's

equation and gives a diffracted beam. In a practical x-ray

diffractometer, this is achieved by grinding the crystals so that

their random orientation ensures that a few will be in the correct

position for any set of planes. The angle 0 is usually increased

from low to high values while the diffracted beam is'monitored.

Typical diffraction traces are shown in Fig. 6.1 and Fig. 6.2.

The limitation of this method is that it can only be used

for crystalline minerals; it cannot identify amorphous reaction

products. XRD like any other method, seldom achieves unambiguous

identification of the reaction products due to the pattern simi-

larities of many closely related hydrated silicates, alumino-

silicates and alumino-ferri-silicates. However such ambiguities

are resolved with the aid of some of the-methods mentioned earlier

as supplements. Hence in this work it was decided that identifi-

cation will be done by XRD and supplemented by thermogravimetry

(TG), scanning electron microscopy SFM (Chapter 8), and chemical

extraction analysis. Because the results of TG not only serve to

identify but may be used'to quantify products of the soil-lime

system, like lime and calcium carbonate, (not considered here)

they will be discussed in Chapter 9.

6.2.2 Quatitificatiöri 'of the 'reaction 'pröducts

The quantitative determination of reaction products is

difficult, their precise composition is seldom:. known and they

occur in a mixture of many other minerals. In this investigation,

three methods of approach based on different principles are

considered i. e., thermal, chemical and radiation.

 1oi .

Thermal methods, employ the selective reaction of the,

cementitious products in an atmosphere, monitoring the weight

changes that occur during the reactions and quantifying the

reactive product from the stoichiometry of the transformation.

Principal thermoponderable reactions met in cement chemistry, like

the equilibria solid-gas thermal reactions of the hydration-

dehydration (see eqn. 6.2), carbonation-decarbonation (eqn. 6.3)

types, have been used to measure some constituents of cement (88).

Ca(OH)2 CaO + H2O (6.2)

... ... ...
solid solid gas

CaCO3 CaO + C02 (6.3)

... ... ...
solid solid gas

By adapting the equations, 6.3 for instance, tricalcium silicate

(C3S), a cement clinker component, can be quantified if ground

cement is heated in carbon dioxide. atmosphere. Theoretically, it

will react with C02, after the possible elimination of free lime,

according to the following equation:

C3S + CO2 -1, C2S + CaCO2 (6.4)

... ...
This equation is a simplified representation of the actual complex

reaction involved. By measuring the weight changes C3S can be

quantified using equation 6.4. Similarly by first reducing the

alumino-ferrite phase of a cement clinker in a hydrogen atmosphere

according to equation 6.5 and oxidising the reduced

C4AF + 3H2 -ý C4A + He + 3H20 (6.5)

...
4Fe 302 -ý 2Fe2O3 (6.6)
-+ ... ...
(oxygen results from the dissociation 2CO2 0 2C0 + 02)

iron in a C02 atmosphere (eqn. 6.6) at 1000oC, the iron content of

the ferrite phase can be estimated. While TG holds out a

potential method of quantifying reaction products of soil-lime

 102.

systems, it can be seen from the examples given that the use of

TG requires a knowledge of the actual chemical composition of the

reaction product which is not certain.

Chemical methods of quantification depend on the ability to

extract the cementitious products using a selective solvent. The

quantification consists of the determination of the different

minerals in the extracts. The methods used for separating

reaction products in pozzolanas from the unreacted material rely

on a method proposed by Steopoe (89)-and described in old German

works. The method is based on the fact that whereas the silicates

and aluminates in a pozzolana are relatively insoluble in dilute

hydrochloric acid under certain conditions, the silicates and

aluminates in the reaction products are extremely acid soluble

under, the same conditions and thus may be extracted. It is only

the optimum conditions for maximum extraction that different users

of this method have changed. Ruff and Hö (6? ) used 50 mis of

0.2N Hcl in extracting 0.5 gm of lime treated soil dried in

vacuum. They employed an extraction time of 30 minutes. The acid

concentration is such as to give a final supernatant pH of between

1 and 2 because not only is aluminium ion stable at this pH range

(it slowly polymerizes and precipitates as Al(OH)3 at pH > 4.5)

but also the rate of polymerization of the silicic acid dissolved

is minimum in this range (90). Gray (90) used the same conditions

as Ruff and Ho (67). The following Steopoe's method as

modified

communicated to the author by Jambor (91) was tried out and found

to give a final supernatant pH of about 1.45 for red tropical

soils.
 103.

The hardened lime-pozzolana paste is ground to a fineness

approximately < 9Opm. One gramme Of the ground paste is gradually

poured, under constant stirring,. into a beaker containing 60 ml

of HC1 (density 1.12) and cooled by water at 12 - 13°C. The

solution is left to clear up for 10 rains. and is filtered and the

extract analysed.

Quantitative 'determination 'by 'XRD, makes use of a standard

mineral whose composition is the same as that of the mineral to be

quantified in the soil (72). Using varying quantities of the

standard mineral (by weight), the diffraction intensities (above

the background) at a selected d-spacing are measured preferably

using the areas under the peaks as measures of intensities. A

calibration graph is drawn by plotting the intensities as

ordinates against the weights of the standard minerals as

The test sample (soil-lime mixture) is x-rayed and the

abscissa.

intensity of the peak at corresponding d-spacing above the back-

is measured. Using this value of intensity for the test

ground

sample and the calibration graph, the weight of the mineral can

be determined. Because the x-ray absorption rate varies with the

constituents present, the intensity of a mineral present in the

test sample depends on other minerals present (diluents) in the

soil.
The use of an internal standard in addition to the standard

mineral almost completely eliminates the error due to the x-ray

absorption by the sample. The peak intensities of the standard

mineral and the test samples are measured relative to the corres-

peak intensity of the internal standard. Any mineral can

ponding

be used as an internal standard provided that:

 104.

a) it is not present in the test sample,

b) it has a peak intensity as close as possible to the d-spacing

at which the diffraction intensities are measured for the standard

mineral and the test sample.

Calcium fluoride, ammonium chloride, quartz have been used as

internal stardards.. (72). Recently, Tuncer et al (92) modified this

method. Their modification comprises the comparison of the XRD

peak intensities of the mineral in the sample with the intensities

of the same peak in specimens that are a mixture of the original

soil sample and known additional amounts of the mineral to be

quantified, i. e. the component being determined is used as an

internal standard. This modification relies on precisely knowing

the composition of the mineral to be quantified so that the peak

of the mineral in the-test sample and the internal. standard are

coincident. This is very difficult, if not impossible to achieve

in red tropical soil-lime systems. Hence this modification was

not adopted.

6.2.3 Reaction 'prödücts 'and 'strength

The study of the effects of the amount of silicates,

aluminates, alumino-silicates and alumino-ferri-silicates on

strength has been extensively carried out on concretes and

mortars. Their effects on the strengths of clay-lime systems have

not been studied as extensively while such studies are non-

existent for red tropical soil-lime systems.

The strength of a cementitious paste depends on the type and

quantity of binders present. Kalousek and Prebus (93) associated

plate-like tobermorite with high strength and low shrinkage and

CSH I as well as fibrous CSH II with low-strength and high

 105.

shrinkage. Powers (94,95) attributed strength to cement gel and

proposed that it is proportional to the third power of the volume

of gel present. He empirically determined the relationship to be:

S= a0V3 ... (6.7)

... ... ... ...

where S= strength of cement paste

v9 = strength of gel of maximumdensity without

capillary voids

V= proportion of volume occupied by gel in a unit

volume of cement paste.

Some other investigators have found the strength of concrete to be

a power function of the quantity of hydration products. Among

them is Schiller (96) whose formula

P=K log ek (6.8)

e ... ... ...

where ek = critical porosity for which the strength, P=0,

K= coefficient independent of over-all porosity but

dependent on pore shape and nature of crystal

contacts,

can be modified to apply to the calculation of strengths of

cement pastes, mortars and concretes.

Ruff and Ho (67) found that the strength gain of lime-clay

mixtures cured at different temperatures is due to the presence

of the different phases of the complex reaction,

lime + clay -1-CSH(gel) -º CSH(II) -ý CSEH(I)-º tobermorite.

The further the reaction proceeded, the higher the strength. They

found no consistent relationship between time, temperature,

strength and the silica/alumina ratios of the reaction products

but in order to achieve high strengths they noted that the

apparent calcium oxide/silica ratio had to be less than 2.

 106.

Jambor (63,97) working on lime-pozzblana pastes found the follow-

ing relationship to exist between strength and volume of new bind-

ing materials formed depending on type of the binding material.

S= av3 + bv2 + cv +. ... (6.9)
... ...

where S= strength

v= volume of hydration products of types I-V

a, b, c = constants,
Type I tobennorite-like phases

Type II CSH(l)

Type III 70 to 80% gehienite hydrate (C2ASHn) and

20 to 30% of CSH(l)

Type 'IV '70 to 80% hydrogarnet (C3ASnH6_2n) and

20 to 30% of CSH(l)

Type V C3AH6and/or hydrogarnet phase.

The effects of the different types'of cementitious materials is

further illustrated'by Alexander et al (98) who found that the

strength of mortars is most sensitive to C3A content of the cement.

Most work done on the influence of the quantity and type of

cementitious products on the strength of pastes has been on

mortars and concretes. Very little has been done in this direc-
"iion for practical soil-lime systems. In his investigations

Jambor employed the following conditions not normally used in a

practical soil-lime system: -

a) He used among other pozzolanas, activated kaolin made by

burning washed kaolin for 2 hours at 650°C.

b) All his pozzolanas were ground to less than 90 microns.

c) He used high percentages of lime viz. 34%, 50%, 66%.

These conditions led to almost complete reaction between the lime

 1oß.

and the pozzolanas for his XRD traces detected only weak lines of

unreacted pozzolanas. Hence the strengths of the lime-pozzolana

systems he measured were attributed to the new binding materials

formed.

6.3 Experimental methods

The lime treated soils used were the broken pieces of the

mixture that gave maximum strength as outlined in Chapter 4. They

were tested after curing for about 2. years. Details of the

methods used for x-ray (qualitative and quantitative) and chemical

analyses are given*in sections under this sub-chapter.

6.3.1 Chemical analysis

Selective chemical extraction of the reaction products was

performed using the modified Steopoe's method communicated to the

author by Jambor (91); the details of which were given in section

6.2.2. The lime treated soil was ground to less than 75 microns.

No selective method of extraction leaves other constituents per-

fectly intact especially in cementitious pastes made from sili-

ceous admixtures with high alumina and iron oxide (R203) contents

and forming low contents of the reaction products (97). In view

of this, similar chemical treatment was applied to the natural

soil. The extracts were analysed using atomic absorption spectro-

photometry as described in Chapter 5. The acid soluble calcium,

silicon, aluminium and iron in both the lime treated and the

untreated soils were measured under the standard conditions shown

in Table 6.1 The elements were calculated thus:

pp 'x 'volume used 'x ' d. f=

Extractable element = Ca or Al or Si
sample weight x 103 or Fe

where d. f = dilution factor

 108.

finäl
diluted völtune ' iri ml .
volume of aliquot taken for dilution in ml.

56.08 Ca 1.40 Ca
Calcium as CaO = =
40.08

60: 09 2.14 Si
Silicon as Si02 = Si =
28.09

101.96 3.78 Al
Aluminium as A1203 = Al =
26.98

159.70
Iron as Fe203 = Fe = 2.86 Fe
55.85

The values of the elements as oxides in the reaction

products were obtained by subtracting the results for the

untreated soils from those for the treated soils.

6.3.2 qualitative XRD

Identification of the reaction products was carried out by

XRD using molybdenum Ka radiation obtained from a Phillips x-ray

diffractometer. Molybdenum Ka radiation was used in preference to

a copper Ka radiation because it gives better diffraction peaks

for iron-rich materials such as red tropical soils (39). The

operating conditions are 42 KV, 14 mA, scanning rate of 10 per

minute, rate meter of 4 and time constant of 4. Among the factors

that affect the intensity of a diffraction peak are particle size,

chemical composition, crystal imperfection, sample packing, x-ray

absorption, crystal orientation and amorphous substances (72).

Methods adopted for the controlable. factors are as follows:

Particle size was controlled by gently grinding the soil-

lime mixture with a rubber pestle to pass a 75 micron sieve.

Control on sample packing was effected by investigating the varia-

tion of the diffraction peak at 2e = 100 with different weights of

-" the soil contained in a sample holder of a constant volume. The area
i'ý i

V4

IL

Table 6.1 Standard conditions used to determine iron,

calcium, aluminium and silicon in the extracts

by atomic absorption spectrophotometry

tA
 110.

Element Iron Calcium Aluminium. Silicon

.I

Wavelength 248 uv 211 uv 309 uv 1 252 uv

Slit setting .3 (0.2 nm) 4(1.4 nm). 4 (0.7 nm) :j3 (0.2 run)

Light source Hollow Ca Lamp

Flame type : Air-acetylene flame Nitrous oxide-acetylene

oxidising (lean, blue). reducing (rich, red)
 x

Fig. 6.1 X-ray diffraction traces for NF showing the 4

lime treated and the natural soils. Shaded

areas represent positions of new peaks.

IL

Nk

t
 112.

N
C,
YN

Code
G- Goethite =
He Hematite Y
Ur'
K- Kaolirwte
M
S Silica

AA Y

iIIIII1
36 34 32 30 28 26 24 22 20 is 16 14 12 10 a6

-0---- 2 90

Me, Ku radLstion
Fig. 6.2 X-ray diffraction traces for SIA showing the

lime treated and natural soils. Shaded areas

represent positions of new peaks.

 114.

JIII 32 30
II
28
IIII
26 24 22 20 t6 16 14 12 10 86
36 34

-0- 2 9°
Mo, Kot radiation
 115.

of the peak was measured in each case with a planimeter. The'

results are presented in Fig. 6.3 and show that the maximum peak

occurs when about 0.5 grams of the soil is used in that particular

sample holder. This weight of soil was therefore used throughout

the investigation.

The XRDof the lime treated and natural soil of each sample

was run. Typical results are shown'in Figs. 6.1 and 6.2 for soils

NF and SIA. The diffraction traces for, some of the rest of the

soils are shown in appendix Al. By comparing the XRDs of the lime

treated and natural soils the 29 values at which new peaks occur

were abstracted (shown shaded in the figures). These 20 values

were fed into a computer search program similar to that used by

Willoughby et'a1 (61). Details of the search program are given

in Chapter 7. `ý The search was for the strong peaks of ASZM

standard hydrates of calcium silicates, aluminates, alumino-

silicates and alumino-ferri-silicates whose composition-include

elements present in the initial reactants. The possibility of

peaks overlapping those of the unreacted soil minerals was taken

into consideration.

6.3.3 Quantitative XRD

The method of internal standard was used. `

6.3.3.1 The internal standard used

Anunoniumchloride was used as the internal standard because
it has a diffraction peak at 20 150 - 160 which does not overlap

the peaks of the minerals in the soil. The position of this peak
is near to the d-spacings at 20 = 130 - 140 used to`quantify the

test sample. The XRD tracing of the ammonium chloride is shown

in appendix-Al.
 116.

6.3.3.2 Standard minerals used

Naturally occuring silicates (okenite), alumino-silicate

(prehnite) and alumino-ferri-silicate (epidote), obtained from the

Department of Earth Sciences of Leeds University, were used as

standard minerals for the lime treated soils forming silicates,

alumino-silicates and alumino-ferri-silicates respectively. The

peaks of the standard minerals at 29 z 130 - 140 were used.

Naturally occuring minerals were preferred to their chemically

synthesized counterparts because they are more likely to be

impure as discussed later. Their chemical compositions were

studied by x-ray fluorescence and are shown in Table 6.2 while

their XRDs are shown in Appendix Al..

6.3.3.3 d-spacing 'used ' för 'peäk 'measürenlerit ' in 'the 'test 'samples

Most calcium silicates, aluminates, alumina-silicates and

alumino-ferri-silicates have diffraction peaks at 2e equal to 13°

approximately. The intensity of this peak was used as a measure

of the reaction products in each test sample. Lime and calcium

carbonate have peaks in this region as well. Chemical and thermo-

gravimetric analyses showed the lime treated soils to contain

appreciable quantities of unreacted and carbonated lime (see

Chapter 9). The reaction products obtained as a percentage of the

lime treated soil were corrected for the presence of unreacted and

carbonated lime.

6.3.3.4 Procedure

A constant quantity (0.15 gin) of the internal standard (gi)

was added to 0.05,0.075,0.10,0.125,0.15 and 0.20 gms of the

standard mineral (gm). Untreated soil (diluent) was added to make

up each mixture to weigh 0.5 gms. All three materials were

Table 6.2 Chemical analysis by XRF of minerals used as

standards in XRD quantification by the method

of internal standard
 118.

OKENITE PREHNITE EPIDOTE

:

Si02 56.61 41.40 39.65

.
A1203 0.638 22.40 23.18
Fe203 0.042 0.99 11.48
CaO 23.28 24.29 20.07
H2O 18.28 3.68 10.20
 Fig. 6.3 Variation of peak intensity with density or

packing of soil sample

Fig. 6.4 Calibration curves for the standard minerals

used in quantifying the reaction products

1, s

f
 4 Mý.
120.

E
u
r

oý 01 0.1 0-3 0.4 0.5 0.6 01 r

0
Sampleweight,g
.ý
ýi

e 't

i' !tý

0.9 (C2 AS3H1

Prehnite

0.8

0.7 0

0.6 /Epidote(C2RMS3F0.51
,
ý/ (CS)
Wollastonite
.,
0.5
0 .o: Ckenite
jý °''ý [CS2Hzl
0.6

0.3
iý
0.2 13

0.1

0 0.1 0.4 0.6 0.8 1.0 1.1 1.4

gmm/gmi
 121.

thoroughly mixed and x-rayed. The intensities of the peaks of the

standard mineral (Im) occuring between 20 = 13° - 14° and of the

internal standard (Ii) occuring between 2e = 15° - 16° were

measured with a planimeter. The ratios Im/Ii and their corres-

ponding gm/gi values were calculated. A calibration curve of

In/Ii as ordinate vs. gm/gj as abscissa was plotted for each

standard mineral (see Fig. 6.4).

To quantify the reaction products in each soil, 0.15 gm of

the internal' standard was mixed with 0.35'gm of the lime treated

soil and'x-rayed under the same conditions as for the mixtures

used for the calibration curve. The intensities of the peaks at

20-z 13° for the reaction products and at 20 = 15° - 160 for the.
internal standards were measured and the ratio Im/Ti calculated
for each test sample. Using these values and the calibration

curves the weights of the reaction products in the soils are

determined. The calibration curve used for each soil depends on

whether that soil forms silicates, alumuno-silicates or alumino-

ferri-silicates as determined from their qualitative XRDs. The

volumes occupied by the reaction products per unit volume of

compacted lime treated soil were calculated using the following

relative densities for the powdered' standard minerals:

0.43 for okenite, 0.32 for prehmite and 0.29 for epidote.

6.4 Results and discussions

The results are given and discussed under appropriate

sections.
6.4.1 Identification of reaction products
The computer search program greatly aided the identification

of the reaction products. Tables A-E in Appendix A2 show the

 122.

ASTM x-ray data for calcium salts concordant with the new lines

observed in the lime treated soils. It can be seen that it is

difficult to distinguish between. the many closely related

silicates, alumino-silicates and/or alumino-ferri-silicates.

6.4.2 The identified products

Table 6.3 shows a summary of the likely minerals formed by

each soil reacting with lime. The Table is prepared from Tables

A-E in Appendix A2. The mineral, plomberite gel, (AS1M ref

10 - 416) seems to be formed in most of the soils. The ASIM index

characterises the mineral as "a natural gel, poorly crystallised

and probably of mixed hydrates". Its chemical composition is

given as (%): - Si02 = 40.4, CaO = 32.6, A1203- 2.4., Fe203 = 0.8,

MgO = 0.3, H2O = 23.3. The next most common minerals in the soils

is tobermorite of the type xCaO. Si02. yH20"(ASTM ref 6- 0010)

with x-1, y not known. This is also characterised as a "poorly

crystallised, mixed hydrates; probably contains interstratified

mixed hydrates, x-ray pattern is very variable especially

d= 12.5A".

From the characteristics of the two silicates that occured

in most of the soils it appears that after about two years curing,

most of the reaction products are still in poorly crystallised

states but cystalline enough to be detectable by x-rays. It would

also appear that most of the silicates do not occur in pure forms;

they seem to be mixed hydrates. It was for this reason that

naturally occuring minerals were preferred to their chemically

synthesized counterparts as standards for quantitative XRD. The

mixed nature of the hydrates would suggest ionic substitution in

the lattice of the tobermorite formed. Kalousek (99) reported

 123,

that "up to 5- 6% altmiinium was substituted beforea hydrogarnet

C3ASH4 was detected by XRD as a second phase"; below this percen-

tage he observed that "the tobermorite with substituted aluminiun

exhibited the same XRD pattern as the aluminium-free tobermorite

except for a small increase in values of basal spacings with

increasing amount of aluminium". Diamond (100) substituted up to

10% aluminium before detecting a hydrogarnet by XRD. Using

goethite as a source of iron, Diamond et al (101) were able to

substitute iron into the lattice of tobermorite. Copeland et al

(102) reacted high-lime tobermorite gel with aluminates, ferrites

and sulphates at room temperature and found that iron substituted

in amount of 7.1%. They pointed out that the results of iron

substitution into tobermorite were not very reproducible. The

mineralogical composition of red tropical soils strongly favours

ionic substitution by aluminium or iron or both in the lattice of

the calcium silicates formed thus bringing about the inpure nature

of the reaction products. Only soil SI seems to have formed a

hydrogarnet and it thus appears that aluminium does not substitute

in large enough amount as to make a second phase x-ray detectable

in most of the soils.

Some soils like JI, RC, ULI, Cr have formed mostly

aluminates and alumino-silicates. In few other soils alumino-

ferri-silicates have been identified. None of the soils formed

iron compounds only as a distinct phase in contrast to the forma-

tion of calcium silicates and calcium aluminates. It would there-

fore appear that the role of iron seems to be to substitute

into the silicates, aluminates and/or alumino-silicates formed a5

distinct phases. Indeed it is also noteworthy that the formation

 124.

of quarternary compounds of lime-silica-alumina-iron hydrates in

which silicon, aluminium and iron occur in distinguishable lattice

positions is not detected in most of the soils.

Table 6.4 shows the results of the chemical analysis of the

reaction products from which their approximate overall composi-

tions are calculated as shown in Table 6.5. Their approximate

overall compositions show clearly that iron is substituted in

different quantities in the silicates and alumino-silicates.

Because only the overall composition of the reaction products can

be calculated from the chemical extractions rather than the

chemical composition of the different phases that make up the

overall products, the amount of iron required to substitute to

enable the detection of a quartenary ferrite phase by XRD cannot

be determined. No CAHio or C2AH8 was detected to form in any of

the soils.

6.4.3 Consecutive and multiple' reactions

Table 6.3 shows that nearly all the soils form more than one

product'. Thus most of the reactions are of the second order

generally represented as: -

A+B; C+D... (6.10).

... ... ...
in his study of lime-quartz pastes, Taylor (103) gave a general

picture of the reaction that takes place. He was of the opinion

that surface reactions on quartz gave a lime rich substance simi-

lar to CSH (II). If the overall C/S ratio is low, the reaction

proceeds until all lime is depleted. The CSH (II) then reacts

with more quartz to give CSH (I) with low C/S ratio which

recrystallises to tobermorite depending on time and temperature.

 Table 6.3 SLUrunaryof likely minerals formed

(with % of strong peaks present in brackets)

I"'k
 126,

EL3

I-i " i-1 ý"

Fi
! >

W cn UH .-1
-
N pm ,1

ca

O
W NO l r. . p-s

r -f 4-J Cl
C/) U Uf 'ý V) Uri . - - . .
cc
v, xcc r,

10
C,
X
rýN l J

m 00
l ll
u
týM rý ^
Q N
Ný 00
d UrI - .

mM N

U r-1
N Q1 rý

00
UM

N
UM

U
H LA
a
ý+ >4
Zý ö p Ö Ö Ö Ö Ö
9 O O O O
00 Nil
I (M ý-I Q1 C( v v 110 . v OO
OOý
> > ! ! !

0
Lr)

"N

x O
c
00
ýýa ýH
Q S" ý H
HC ý ýL Lý( Zýi H H U
ý
""' H
d
H H H H
ýn W
aº E-
-ýv ý ti º-ý cn cn
Table 6.4 Chemical analysis of the acid soluble reaction

products
 128.

Chemical'Composition (%I) IN PRODUCT(mols) AS

"
CaO in
SOIL
SYM-
hydra-
ýL
BOL Fe203 tion CaO Si02 A1203 Fe203
. Ca(OH)2 Ca003
pro-
ducts

PH 4.24 8.78 3.41 0.30 1.65 22.62 0.40 0.071 0.086 0.021

JPM 11.56 17.35 1.05 0.26 0.67 15.70 0.28 0.192 0.170 0.0066

TH. 9.37 13.27 3.68 0.42 1.20 38.45 0.69 0.156- 0.130 0.023

SB 14.25 14.90 1.58 0.45 1.40 25.74 0.46 0.237 0.146 0.0099

NF 5.26 11.63 13.48 0.34 1.68 22.55 0.40 0.088 0.114 0.084

JPB 7.06 6.94 1.05 0.76: 0.40 24.69 0.21 0.118 0.068 0.0066

JPA: 8.99 15.51" 5.78 0.43 0.70 14.14 0.25 0.15: 0.152 0.036

SIB 7.32 10.61 15.75 1.29: 2.16: 35.33 0.63 0.122 0.104 0.098

JI . 10.79 15.92 11.81 0.61: 1.53 21.67 0.39 0.18. 0.156 0.074

SII' 16.69 15.10 3.68 0.72 1.69 39.47 0.70 0.278 0.148 0.023

RC 10.66 16.53 0.26 0.34: 1.50 25.0: 0.45 0.177 0.167 0.0016

USM 14.0 23.68 2.28 0.53: 1.52 33.86 0.60 0.233 0.232 0.014

CT 9.63 13.68 5.51 0.79; 1.61. 27.99 0.50 0.16: 0.134 0.034

AI 16.56 27.35 4.90 0.68: 1.52- 30.31 0.54 0.276' 0.268 0.031
AII: 8.47 21.02 3.24 0.66: 1.40 33.47 0.60 0.141 0.206 0.0203
SM 14.38 23.88 7.09 0.45. 1.55 31.64 0.56 0.24 0.234 0.044
TI 24.52 31.23 10.33 0.89* 1.16 37.41 0.67 0.408 0.306 0.065

SIA 5.13 6.94 3.15 1.40. 1.01 25.22 0.81 0.085 0.068 0.020

JII 12.58 11.84 5.78 0.83 1.10 23.40 0.42 0.209 0.116 0.036
Table 6.5 Approximate overall chemical composition of

reaction products
 130.

SOIL CaO/Si02 Approx. overall

A1203/SiO2 Fe203/Si02
SYMBOL composition

PH 5.67 1.22 0.30 C5.67 SA1.22 FO"30 H??

JFM 1.46 0.89 0.03 C1.46 SA0.89 Foos H?

TH 4.42 0.83 0.15 04.42 SA0183 Fo"15 H? :

SB 1.94 0.62 0.04 C1.94 SA0.62 FO. 04 H?

NF 4.57 1.30 0.06' C4.57 SA1"30 F0"06 H?

JPB 5.78 0.58 0.06 C5.78 SA0"s8 F0.06 H?

JPA 1.67 1.01 0.24 01.67 SA1.01 F0024 H?

SIB 5.16 0.85 0.81 C5"16 SA0.85 F0"81 H?

JI 2.17 0.87 0.41 C2.17 S40187 F0.41 H?

SH 2.52 0.53 0.08 C2.52 SAo"53 Foos H?

RC 2.54 0.92 0.009 C2.54 50.92 F0"009H'

U34 2.58 1.00 0.06 C2"58 X0.06 H?

CT 3.13 0.84 0.22 03.13 SAo"84 FO-22 H'

AI 1.96 0.97 0.11 C1.96 SA0.97 F0.11 H?

All 4.26 1.46 0.14 04.26 SA1.46 F0"14 H?

SM 2.34 0.98 0.19 C2034 SA0"98 F0.19 H?

TI 1.64 0.75 0.16 C1.64 SA0.25 F0.16 H?

SIA 5.48 0.80 0.23 C5"48 SA0"80 F0023 H?

111 2.01 0.56 0.17 C2.01 SA0"56 F0.17 H?

 131.

This sequential reaction suggested by Taylor implies that

for some reactions of red tropical soils, equation 6.10 can be

expressed as a consecutive reaction of the form,

A+ BýCID... (6.11)
... ... ...
where k is the rate constant for the first step and 1 is the rate

constant for the second step C67). If 1>k, C is not formed and

if 1<<ka negligible amount of D will form. If k is slightly

higher than 1, both C and D will be present. These inequalities

will, possibly hold only for those red tropical soils that form

one type of product, say silicates. Where more than one type of

product is formed, say silicates and alumzno'-silicates, the con-

sequences of the inequalities between 1 and k as applied to

equation 6.11 are not valid. They would rather apply to the

multiple reactions represented by equation 6.12.

A+B ýC
(6.12)
A+BID3... ... ... ...

The implication of equation 6.12 is that the reactions can be

simultaneous irrespective of the values of k and 1. The behaviour

of the mixture from the point of view of strength will depend on

whether C or D is in greater quantity at any particular time.

These facts, in addition to the effect of the volume of reaction

products formed with time, may explain the shapes of the kinetic

function-time curves shown in Fig. 5.5. The figures show two

distinct reaction rates. One reaction rate occurs between 0-5

or 7 days and the other thereafter to 28 days. It would appear

that one rate is k and the other 1. Where consecutive reaction

occurs then k corresponds to reaction rate measure between 0-5

or 7 days whereas where multiple reactions occur the greater of

Table 6.6 Quantification of reaction products of the
lime-soil systems from XRD
 133.

Vol. per
OIL Ifli Vol. per
Im gm 35 gm
S Ii
gm lime cc o f
(cm2) Ii gl
BOL (CM2) treated compacted
soil soil (40)

PH '0.50 7.1 0.07 0.205 0.031 0.072 3.36

JPM 0.80 5.6 0.12 0.36 0.054 0.126 6.96

TH 1.00 12.50 0.08 0.24 0.036 0.084 3.97

SB 1.10 10.00 0.11 0.31 0.047 0.108 5.65

NF 0.70 10.0 0.07 0.20 0.030 0.070 3.69

JPB 0.50 11.10 0.045 0.13 0.02 0.045 2.57

JPA 0.40 6.20 0.065 0.19 0.029 0.066 3.54

SIB 0.80 6.20 0.13 0.15 0.023 0.07 3.64

JI 0.80 5.20 0.096 0.20 0.030, 0.075 3.82

SII 0.50 5.0 0.10 0.30 0.045 0.105 5.30

RC 0.60 4.30 0.14 0.225 0.034 0.105 5.59

USM 0.50 8.0 0.063 0.13 0.020 0.05 2.33

CT 0.50 5.6 0.089 0.10 0.015 0.035 1.81

AI 0.40 4.40 0.091 0.27 0.041 0.094 4.60

All 0.40 4.70 0.085 0.18 0.027 0.093 4.60

SM 0.70 7.80 0.09 0.27 0.041 0.094 5.01

TI 0.70 6.40 0.11 0.23 0.035 0.12 5.73

SIA 0.60 15.0 0.04 0.04 0.006 0.019 1.02

JII 0.50 6.60 0.076 0.15 0.023 0.052 3.04

 134.

the two rates is represented by the 0-5 or 7 days rate in

Fig. 5.5.

6.4.4 Volume of reaction'product, vs strength

Table 6.6 shows the volumes of the crystalline reaction

products as measured by the internal standard method. The volumes

were plotted against the total strength of the soil-lime mixtures,

the resultant relation is shown in Fig. 6.5. Its shape is

similar to that obtained by Jambor (63,97) who also showed that

the strength of lime-pozzolana pastes depended on the type of

reaction product formed. In any of the systems under study, a

amount of the soil is unreacted, it may contribute to

substantial

the overall strength of the lime treated soils. It. will therefore

be logical to ascribe increase in strength to the type and

volume of the reaction products. Hence Fig. 6.6 shows a plot of

lime reactivity as measured by increase in strength against volume

of the reaction products. It can be seen that with the reaction

products broadly divided as:

type I- silicates

type II - silicates and alumino-silicates

type III - alumino-silicates and/or alumino-ferri-silicates,

a family of curves can be drawn.

The total strength of a red tropical soil treated with lime

has two components viz.

a) Some fraction of the strength of the natural soil (So) at the

content as'the treated soil. This component depends

same moisture
and grain to grain contact (4,24,25). The reaction
on cohesion

products will reduce the effects of cohesion and friction; the

extent depending on the volume of the reaction products. This.

 Fig. 6. S Relation between volume of reaction products

and total strength of the soil-lime systems

Fig. 6.6 Relation between volume of reaction products

and lime reactivity parameter of Thompson

 136.

50

4.0
N
E

3.0

20

1.0

05 1234567
Volumeof reactionproducts per vol of compactedsoil V °/o

4.0
)e 1

3.0
E

c,
2-0
0
ai
o,

E_
J

1.0

0-
0"5 1234567
Volume per vol of compactedsoil, V °/o
of reactionproducts
 137.

component of total strength is represented by k(v)S0 where k is a

reduction factor.

b) The strength of the reaction. products which depends on the

volume and type of reaction product present. This component is a

power function of the volume which can be represented by a

polynomial of third degree as suggested and demonstrated by

Jambor (0) .

Hence total strength (St) can be written as

St = k(v)S0 + av3 + bv2 + cv (6.13)
... ... ...
where v is the percentage volume of the reaction products in a

unit volume of lime treated, compacted soil and a, b, c are

constants. The following boundary conditions apply to k.

(i) when v=0, k=1; (Ii) when v= 100, k=0. Equation 6.13

assumes that St is an arithmetic sum of the two components.

Assuming that k~1 for small volumes, a stepwise multiple

regression was carried out between lime reactivity as the dependent

variable and v, v2, v3 as the independent variables on one hand

and between total strength. as the dependent variable and v, v2, v3

as the independent variables on the other. Multiple regression has

been similarly used by Alexander et al (98)-to study how strength

and hydration are related to the composition of Portland cement.

The regression equations obtained for the soil-lime mixtures are:

a) For total strength:

St 0.0136v3 + 0.0552v2 - 0.4338v + 1.368 (6.14)

..
For increase in strength:

b) type I: St - So = 0.0183v3 - 0.0791v2 + 0.184v + 0.0156 (6.15)

c) type II: St - S6 = 0.0482v3 - 0.3374v2 + 0.7361v + 0.0248(6.16)

d) type III: St - So = 0.0799v2 - 0.1477v + 0.06263 (6.17)

...
 138.

Equation 6.14 is highly significant (multiple R =-0.929) at 86%

confidence with standard error of 0.5340. While equations

6.15,6.16,6.17'are significant. (multiple Rs = 0.996,0.905,

0.930 respectively) at 99%, 82%, 86.5% confidence and with

standard errors of 0.1974,0.4116,0.3359 respectively, they can

not be used for prediction purposes because the number'of samples

in each group do not ensure a statistically significant

population.
From Fig. 6.6 it can be seen that for small volumes of

reaction products, within the experimental range, the order of

strength increase is type III '> type II > type I whereas at-

larger volumes the order is reversed. This contrasts Jambor's

findings. A possible explanation that can be offered is Ithat

ionic substitution in the'lattices of the silicates'`seem to

reduce their effective binding strength and this reduction is more

pronounced in small volumes'than in larger volumes.

6.4.5 Overall C/S; A/S 'ratios 'vs 'strength

The overall C/S ratio (Table 6.5) was plotted against the

total unconfined compressive strength and lime reactivity

(Thompson parameter) as shown'in Figs. 6.7 and 6.8 respectively.

From these figures, it can be seen that a relationship appears to

exist between the overall C/S ratio and the total strength on one

hand and lime reactivity on the other for type I reaction products.
The lower the C/S ratio the higher the strength. Ruff and Ho (67)

found a similar pattern for lime-bentonite-water mixtures and

concluded that high strength is achieved if the apparent C/S

ratio is less than 2. ' A similar conclusion cannot be drawn here.

No consistent relationship appears to exist for types II and III

 139,

reaction products which'are results of multiple reactions.

Figures 6.9 and 6.10 show the plot of total strength, lime

reactivity vs A/S ratio. No consistent relationship can be found

between this ratio and strength or lime reactivity for any of the

types of the reaction products.

6.4.6 The relation between ' the ' activity y*iiidex 'and 'volwne of reaction

products

The relation between the activity index and the volume of

reaction products is shown graphically in Fig. 6.11. It can be

seen that for a relationship to exist between them, the type of

reaction product formed has to be taken into account. This can

be explained by the fact that the äct ity index depends on the

fraction of lime reacted, the time and the quantity and type

of the reactive clay size fraction of the soil. On the other

hand, the volume of the reaction products depends only on the

fraction of lime reacted and time. Hence any relationship

between the activity index and volume of the reaction products

must take into account the effect of the quantity and type of the

reactive clay size fraction. The effect of this parameter in a

soil-lime reaction is not only the volume but also the type of the

reaction products.
The lime reactivity parameter of Thompson was fotmd to be

similarly dependent on the volume and type of the reaction

products. In Chapter 5, it was found from statistical analysis

that lime reactivity is a linear function of the activity index,

Ai. The similarities of Figs. 6.6 and 6.11 not only supports the

variation between the activity index and lime reactivity but also

confirms the conclusion that the former is the theoretical

Fig. 6.7 Relation between overall C/S ratio and total

strength of the soil-lime systems

Fig. 6.8 Relation between overall C/S ratio and lime

reactivity parameter of Thompson
 141.

5.00

a.00
N
E

3.00
U

2-00

1.00

0113456
Overall C/S ratio

N
E

u
d
oý
ý,

aº
E_
J

p123456
Overall C/S ratio
\

Fig. 6.9 Relation between overall A/S ratio and total

strength of the soil-lime systems

Fig. 6.10 Relation between overall A/S ratio and lime

reactivity parameter of Thompson

-.1

143.

5.00

4.00

3.00

CV
E

2'00

1.00

0ý-
0.2 04 0.5 0.8 1.0 1.2 14 1.5
OverallA/S ratio

T.,.... I
4.00

N
E
" 3-DD

8 2.00
a,
E
J

1.00

0I
p-2 04 01 o-B 1.0 1.2 1.4 1.5
Overall A/S ratio
Fig. 6.11 Relation between volume of reaction products

and the activity indexes of the soils

Fig. 6.12 Relation between overall C/S ratio and activity

indexes of the soils

 145.
I

5.0

Type1
4.0 -

CU /
I/

° 3.0 /

ai
c
1.0
Type111

/A Type11
1'0
A

05 1134.567
Volumeof reaction productsper volumeof compactedsoil, V'/.

5-0

4.0

a.

° 3.0
x
1

ac
a,
c 2"Q

u
Q

1.0

0
23656
OverallC/S ratio
 146,

counterpart of the latter. The only difference that seems to

exist between Figs. 6.6 and 6.11 is that relative to the position

of the curves that form type I reaction products, the curve for

the soils that form type II reaction products in Fig. 6.11 has

been displaced downwards. At the present state of knowledge of

the properties of the activity index, no explanation can be

offered for 'this.

Using a stepwise multiple regression analysis with the

activity index as the dependent variable and the volume as the

independent variable, it is found that the activity index (Ai) is

a power function of the* volume of reaction products (v) which can

be expressed' as:

a) For type I products:

Ai =-0.003v3 + 0.163v2 - 0.395v + 0.0105 ...

'(6.18)

b) For type II product's:

Ai = 0.0306v3 - 0.216v2 + 0.440v + 0.0028 (6.19)

...

c) For type III products:

Ai = 0.0295v3 - 0.226v2 + 0.6122v + 0.0115 (6.20)'

...

where Ai = activity index x 10-4 per day

v= percentage volume of reaction products.

Equations 6.18,6.19'and 6.20 are statistically highly significant

(multiple Rs = 0.9984,0.9775 and'0.9816 respectively) at 99.7%,

95.5%, 96.4% confidence and with standard error of estimate of

0.201,0.126,0.144 respectively. These statistics show that the

activity index correlates better with volume of reactiön'products

than does the lime reactivity parameter of Thompson. The ''

equations cannot be used for prediction purposes because the

number of samples in each group do not ensure a statistically

 147,

significant population. Soils TII, USM, SH and PR do not comply

with the results in this section.

6.4.7 Activity index vs'C/S ratio

Figure 6.12 shows the variation between the activity index

and the overall C/S ratio. No consistent relationship appears to

exist between them.

6.5 Limitations of the 'methods 'tised

The user of the information presented in this chapter

should be aware of the following:

a) XRDmeasures only minerals with specific order; short range

order or amorphous reaction products cannot be quantified whereas

if the reaction products are in this state they will contribute
to the strength of the soil-lime systems.
b) The minerals used as standards are not exactly the same as

the reaction products whose exact chemical compositions are not

even known.

c) The reaction products were classified very broadly into three

categories. Since their strengths depend on their specific types

(97,98) a better" classification would be to further divide the

broad groups into. groups comprising more specific minerals. The

method. of identification used cannot resolve this problem. The

possible effect of ionic substitution as pointed out in the

silicates makes the resolution of this problem worthwhile.

6.6 Chapter Summary

While recognising the limitations of. the test methods and

materials used, this chapter shows that for the soils studied: -

a) Cementitious reaction products varying from silicates of the

tobermorite-like type through alumino-silicates to alumino-ferri-

qp-
 148.

silicates are formed by consecutive or multiple reactions.

b), Where silicates of the tobermorite-like type are formed ionic

substitution in the lattice of such silicates owing to the minera-

logical composition of the soils is pointed out:

c) The role of iron seems to be to substitute into silicates and

aluminates formed as reaction products.

d) With the reaction products broadly'calssifled as

type I- 'silicates'

type II" - silicates and alumino-silicates

type III - alumino silicates and/or alumino-ferri-silicates,

the functional relationship between increase in strength and

volume of reaction products is:

.
St - So = av3 + bv2 + cv for types I and II

St - Sý = bv2 + cv for type III

where St = total strength of lime treated soil in MN/m2

So = strength of the natural soil in MN/m2

a, b, c= constants

v= percentage volume of reaction product in a unit

volume of compacted soil.

When the total strengths of the soil-lime systems are considered,

the functional relationship between total strength and volume of

reaction products is:

St = av3 + bv2 + cv +k

where St = total strength of lime treated soil in MN/m2

v= percentage volume of reaction products in a

unit volume of compacted soil

a, b, c, k= constants.
 149.

e) Ionic substitution in the lattice of silicates seem to have

a more pronounced effect on strength for small values of v than

larger values.

f) No consistent relationship exists between overall C/S, A/S

ratios and strength for types II and III reaction products whereas

there appears to be one between C/S ratio and strength for type I

reaction products.

g) The activity index is a power function of the volume of

reaction products and the functional relationship between them can

be expressed as:

Ai = av3 + bv2 + cv for types I, II and III reaction

products,

where Ai = activity index' x 10-4 per day

.
v= percentage volume of reaction products in a unit

volume of compacted soil

a, b, c= constants.

A better correlation exists between the activity index and

volume of reaction products than between the lime reactivity

parameter of Thompson and the latter.

h) No consistent relationship appears to exist between the

activity index and the overall C/S ratio.

 150.
.

7'"

CHAPTER 7

SF11I-AUTOMATED'
IDENfIFICATION'0F'CRYSTALLINE REACTION PRODUCTS

7.1 Introduction

Typical XRD.of some of the lime treated and natural soils

are shown in Figs. 6.1 and 6.2 and Appendix A. Identification of

the reaction products involves:

a) abstracting the 20 values of the Bragg angle at which new

peaks occurred by comparing the two traces;

b) calculating the d-spacings associated with each value of 8

using equation 6.1;

c) estimating the intensity ranks or calculating the relative

intensity value of each peak;

d) checking the d-spacings. and their intensities against those

of standard minerals.

A mineral is inferred if the measured d-spacings and their inten-

sities correlate totally or very strongly with its diffraction

patterns usually available in powder diffraction files (pdf).

This chapter therefore reviews methods available for

searching the pdf. The methods are adapted for use in matching

new peaks that occur in a soil-lime system with the pdf data base.

A pdf data base is a set of minerals in a powder diffraction file

whose diffraction patterns are stored in the computer and used as

a basis to evaluate. the measured diffraction patterns.

7.2 Review of literature and methods

Of the various automated systems available for identifying

 151.

correlations between measured diffraction patterns and pdf data

base, the Johnson-Vand system (104) has been promoted by the

Joint Committee on Powder Diffraction Standards and is now

widely used in many countries. The principal features of this

program are critically outlined.

7.2.1 The 'd-spacings 'and'üiterisifies

The input data are of the d-spacing --'relative intensity

type. ' Hence` the operator has to compute 'the d spacings from the

e- values and quantitatively measure their relative intensities

before he can operate the program. It would be much better if

the measured diffraction patterns can be input as 0,2e or

functions of 0. Ina mixture of minerals where peaks overlap,

quantitative measurement of intensities'is difficult because the

area under each of the overlapping peaks is difficult to define.

In such systems therefore, ' there is a case for the replacement of

the relative intensity values by intensity ranks using the

strong to very weak scale.

7.22 Error 'width. 'or 'window

An allowance for an error made for every observation in the

measured data depends on the apparatus used to obtain the data.

Thus in the program is specified an error'window (IW) given by

1=( 1000 (7.1)
x
. .... .... 0 ..

y+nd)

where d= d-spacing and Ad is the tolerable variation on

either side of d.
Hughes (105) confirmed Johnson's findings that for Nonius focusing

camera IW = 2; for Debye Scherra camera IW'= 3; for a 1o per

IW = 4; while 10
minute continous scan on a diffractometer for a
100
 152.

step scan IW = 2. The operator of the program therefore has to

input a value for 1Wwhich is used to calculate the error window,

od, for each d-spacing. If no value of IW is specified, the

program takes it to be 2.

It appears more rationale to assign error in measurement to

each Bragg angle of the measured data. The value of the error

window should depend on the variations in the-measured diffraction

patterns rather-than on the position of the peaks of the measured

diffraction pattern Cd-spacing) and the instrument used. For

instance, where a broad peak occurs, the error window should be

greater than if a very sharp peak had occurred at the same value

of d-spacing. Broadening of peaks is more, a result of the mineral

structure of the x-rayed material than the result of the

instrument used (7; ) .

7.2.3 Data base

The program holds about 20,000 compounds in the data base.

Thus when employed to identify crystalline reaction products of

soil-lime systems (broadly known' to be silicates, alumino-

silicates and alumino-ferri-silicates of calcium), it searches

through a data base consisting of irrelevant compounds. This

results in the fase of unnecessary computer time. Futhermore, when

such large data base of irrelevant compounds are used, huge

magnetic file tapes are required and large computers which can

handle such large data base are expensive to run. As far as time
is concerned, the program is not suitable for the identification

of a single phase compound. Johnson and Vand (104) concede that

for identification of a single phase, as far as time is concerned,

an experienced person will win easily while the chances are even
 153.

even for a two phase mixture., For a three or more phase mixture,

the program will win. -"

7.2.4 The "elements present"option

Chemical elements or radicals known to be present or absent

in the material x-rayed, can be inserted in the input and the

output will be biased accordingly. Such elements or radicals are

punched in the "elements present" card. Any compound that

contains at least one of the "elements present" and gives a

reasonable line match is considered by the program as, a chemical

fit. Thus if oxygenwere included in the "elements present" card,

every oxide, hydroxide, hydrate; ' carbonate, bi-carbonate etc. that

gave reasonable line matches will be considered possibilities.

7.2.5 The' ptOgräm'pröcedure

In the program search, the computer takes the d-, spacing

values with an error bar and their relative intensities and tries
to fit the most likely compound to them'. It then subtracts the

/S'IM d-spacings and their relative intensities of the first r

compound considered most likely from the measured d-Spacings and

their relative intensities. Using the residue, it tries to fit

the next most likely mineral. The process is repeated until most

of the measured' diffraction patterns have been accounted for. A

list of the ten most likely mineral phases present in the descend-

ing order of reliability is printed out.

Thus the program nearly eliminates completely the human
involvement in such identification. In a good review of the

merits and demerits of automated searches, Hughes (105) observed

that no automated search can be as reliable as an experienced

manual searcher. Automated systems can at their best serve as

 154.

assistants. The problem with manual searching is the tedium

associated with it. A way out of the problem is to reduce the

number of likely minerals to be considered to a minimum. In

achieving this reduction, it will be desirable to leave some of

the parameters for identification for use by the manual inter-

preter of-the results. Hence a blend of the automated and manual

searches is desirable (106), backed up by other methods such as

elemental analysis.

A semi-automated" method of identification was therefore

preferred for use in this investigation. This method is

described in the following sub-chapter.

7.3 Features 'of the 'semi=automated 'search 'tised

The system proposed'and used for the research comprises a

first stage of an automated' search followed'by a second of a

careful manual interpretation and assessment of the results of

the automated search. It is orientated to aid identification of

reaction products In a practical soil-lime system but can be

adapted for the identification of other crystalline minerals.

7.3.1 The automated' search

The basic function of the automated search is to eliminate

those compounds in the data base which do not match the measured

diffraction data on the basis of d-spacings only. Usually two

diffraction peaks match if their d-spacings and intensities are

coincident or differ within tolerable limits. Vhere the minerals

in a mixture have their peaks overlapping, it is very difficult

to quantitatively measure their intensities. In lime treated

soils many minerals are. present and peaks overlap (see Figs. 6.1,

6.2). Hence it was decided that intensity ranks rather than

 155.

intensity values should be used as measure of intensity. It was

also decided that this parameter should be employed when manual
interpretation is carried out; thus the use of d-spacings only in

the automated search. To make the search adaptable for identifi-

cation under other or simpler circumstances a provision was made

for the use of intensity as well as d-spacing for the search.
The automated package consists of a search program; 'a, data
base and an input data as different and nearly independent sub-

routines. The idea'of their "independence" is to permit the user

freedom to replace the data base with the relevant minerals

without virtually altering the program. Furthermore, it was

hoped that this will facilitate provisions for the use of any

suitable source of radiation.

7.3.1.1 'The 'dätä 'base

The data base consists of all the available calcium

silicates, aluminates, alumino-silicates, alumino-ferri-silicates,
ferrites, i. e. all the calcium salts that contain some or all of
the initial elements present in the reactants. Their diffraction

patterns were derived from the strong peaks of the minerals

occuring in the ASTM powder' diffraction file. held in The

Houldsworth School Applied Sciences'(Leeds University). Strong

peaks were selected for use in identification in order to filter

out matching weak reflections otherwise an alarming number of

ASTMminerals will be printed which have some lines in commonwith

the measured pattern. The number of strong peaks entered or

stored for each mineral in the data base varied according to how

many d-spacings of the-mineral have high relative intensity

values.
 56.
x,
.

The d-spacings of the strong peaks and their relative

intensity values were punched on cards and the peaks of each

mineral arranged in the descending order of their d-spacings.

Each mineral was identified by its ASIM file number punched on a

card which precedes*the card/s containing its d-spacings and their

intensities. One hundred and sixty minerals were used as data

base. Provided the number of minerals in any proposed data base

is less than two hundred', the new set can overwrite the old-with-

out alterations in the program. If it is greater than two

hundred, its dimension in the program should be increased

accordingly. The database was coded XRAYDATA2.

7.3.1.2 The input data

The first card of the input'data field contains the username

of the operator and the code of the input, XRAYDATA. On the next

three cards are entered a 'description of the test material (e. g.

JPM, 6% lime treated', cured 28 days at 22°C) and the detailed

conditions of the XRD such as chart speed, number of counts per

second, rate meter etc. The fifth card contains the value of the

wavelength of the radiation used. Thus it does not matter what

the'radiation source is provided its wavelength-Is known. If two

constraints (d-spacing and the intensity) are considered necessary

to identify a peak in the automated search, a constant error width

for the'measured intensity values is punched on the fifth card

after the value of the wavelength of the radiation. The error

width must be an integer.

The next set of cards contain the diffraction patterns of

the new peaks observed. The patterns are input as 26 values of

the Bragg angle. The value of 20 at which a new peak occurs is

 157.

referred to as 0 2. The - error window in measurement of e2 is

input as 20 values, referred to as e1, e3. e1 is the lower

limit (using the magnitudes of-the angles as a measure and

measuring from 00) and e3 the upper limit of e2., ° Thus a new

peak is described in the input by 92 61, e3 punched in this

,

order on a card. If the intensity of a peak- is. measured and is

to be considered in the automated search, its relative value ($)

is punched-immediately after 03 on the same card that contains

e2, Olt 03 values of the peak. The input is terminated by

punching three negative values of e2, e1, e3 on the last card.

7.3.1.3 Design 'Of the 'seärcfi 'program

The algorä. thm'for the search program is outlined in Fig. 7.1.

Measured values of 02, e1, @3, their intensity values if measured,

a constant error width for the intensity values and the wave-
length of the radiation used'are entered on cards. Using these,

the computer calculates the d-spacings di, d2, d3 corresponding

to 01-1 02-1 03 respectively. A standard mineral is read and its

d-spacings and intensities (if applicable) checked against the

measured values. A count of the matching. peaks Is kept. If no

value of intensity is given for a. peak, the. program ignores any

specified intensity error width. It finally expresses the number

of matching peaks as a percentage of the total strong peaks of

that mineral. - The search program is coded XRAYPROG

and is written

in Fortran language.

7.3.1.4 The output

The following phrases used in displaying the output are

defined or described:
 158.

a) "Calculated spacings" refers to the d-spacings calculated

using the values of e2.

b) "Intensity" refers to the measured intensity of a new peak.

c) "Standard spacing" refers to the d-spacings of the strong

peaks of a mineral in the data base.

d) "Relative intensity" refers to the intensity of a standard

spacing.
The output shown in Table 7.1 consists of the following
information:

a) the details of the sample and the conditions of x-ray are

listed;

b) 01, e2,03 and their associated d-spacings are listed in

that order;

c) the matching peaks of the measured diffraction pattern and

the diffraction pattern of the data base mineral are tabulated.

The details of the tabulation for each mineral are thus: - First

the data base mineral is identified by its ASDI reference number

referred to and written as "card number". The "standard spacings"

and their "relative intensities" are listed. The "calculated

spacing" and its "intensity" are written on the same line

(horizontally) as the "standard spacing" and its "relative

intensity" if both match. If no "intensity" values are measured

for the new peaks, the value 0 is printed under "intensity" when

a "calculated spacing" matches a "standard spacing". Finally the

total number of matching peaks is expressed as a percentage of the

total strong peaks of that mineral. This is printed as

"percentage of strong peaks present". The output is coded

XRAYOUT.
Fig. 7.1 Algorithm for seach program
 START 160.

Read lamda, Q

intensity limit
T-

1301 yes
L: 1,50 atc. spacing
80
STD; spacin

Read input data

92,01,03
1 no

Set IJ = J-1
92 = 50
-ve?

no
40
J: 1,3

IJ = 0? Set IJ =1

Compute spacings
D (L, J)

no

40

no

Write Os and
spacings
Spacing yes
limits Intensity
satisfied? given?

30

no yes

50
F Decrease L
U-L-1
ntensity,
SetIJ=J I tý _ ýý limits Increment
satisfied Kount by 1

Read MN
f
yes no
29 Write results
MN : 1, MN

Initalise kount
kount 0

ead and write card No

Read No of items on
card
Read and write STD
spacing and intensity 70

70 1, L
I:

DO
J: 1, N
 ý,
ý_
ý;,

''N

Table 7.1 Part of typical output of the search program

 -ý

162.

IUATOATA FOR ROMAN ROAD

"p YMITE

MOLYIeakUM TUor
---- -- -------"------------=--------------------"------------------------------------------------

TNETAI SPACINCI TNETA2 SPACIN62 TNETA3 SPACtNG3

5.67 16.58 5.76 15.34 5.90 14.97
6.30 14.02 6.40 13.80 6.62 13.34
9.50 9.30 9.55 9.26 9.65 9.16
10.60 8.34 10.70 8.26 10.80 8.19
12.00 7.37 12.30 7.19 12.60 7.02
13.60 6.51 13.60 6.41 14.00 6.32
16.55 5.35 16.65 5.32 16.75 5.29
17.8S 4.97 17.90 4.95 '18. OS 4.91
19.25 4.61 19.30 4.60 19.40 4.57
20.55 4.32 20.70 4.29 20.85 4.26
22.40 3.97 22.65 3.92 . 22.80 3.90
24.40 3". 65 24.55 3.62 24.80 3.59
26.50 3.36 26.70 3.34 26.80 3.32
26.10 3.17 28.25 3.16 28.30 3.15
29.00 3.08 29.10 3.07 29.20 3.06
29.80 3.00 29.90 2.99 30.05 2.97
31.90 2.80 32.00 2.80 32.10 2.79
34.30 2.61 34.40 2.61 34.65 2.59
34.85 2.57 34.90 2.57 34.95 2.57
------- ----------- ---------- ------------- --------- --------------------------------------------------

CARP MUMlER 9.469

STANDARD
SPACING 21.00 8.80 7.40 3.56 3.07 3.05 2.98 2.93 2.77 2.15 1.80

RELATIVE
INTENSITY 100 80 411 80 60 60 50 80 SO 40 80

CAICUTATED STANADARD RELATIVE

SPACING INTENSITY SPACING INTENSITY

3.07 0 3.07 60

2.99 0 2.98 SO

PERCENTAIE OF STRONG PEAKS PRtSENT 18.2

- _-- -------------------- __-___-_. --_-___--__--_------__-_. _-_---_---. --__---_-_-_--. -. -. _--Z...

CARD NUMBER 7.302

sTANDARD
$PACI%$ 19.00 9.40 7.65 4.65 4.13 3.80 3.47 3.14 2.83 2.63 1.84 1.75
,
RELATIVE
Iw1EwsITY 100 100 40 40 100 40 40 100 100 100 70 60

CALCULATED STANADARD RELATIVE

$FACING INTENSITY SPACING INTENSITY

ºEpCENTAll OF STRONG PEAKS PRESENT . 0.0 ''

....... ............................................................................................
 163.

7.3.2 Manual-*ixiterpretät ion

The first step taken in interpreting the output is to

consider that mineral with the highest "percentage of strong peaks

present". The "calculated spacings" used in its identification

are subtracted from the listed-"calculated spacings" at the top

of the output. The remainder of the calculated spacings are used

to check the next mineral with-the highest '.'percentage of strong

peaks present" and in so doing overlapping peaks are taken into

account. The procedure is repeated until all the listed

calculated spacings have been assigned.

Because of overlapping peaks, intensities of the new peaks

were measured by intensity ranks, in this study, using a very

strong to very weak scale. * The intensity ranks. were incorporated

during the above manual interpretation procedure. For example,

a "standard spacing" having 1.00 relative intensity should match a

"calculated spacing" having a very strong Intensity rank.

7.4 'RLmnIng *the 'packäge

To run the package, a data base consisting of the relevant

minerals of interest is read in. The input XRAYDATAis also read

in. The cards in the input are punched thus: -

Card number 1: IN : USERNAME,

XRAYDATA

2: XRAYDATA
FOR
..........
3: Details of the test sample and the
}
4: conditions of the x-ray
5: LAMDA = insert wavelength and next inten-

sity error width (integer)

6: 62 Al 03 (intensity if measured)

7: 02 01 03 (intensity if measured)
Table 7.2 The computer program in Fortran language
-060s MASTER XRAY
{ö006 REAL LANDA
0007 DIMENSION 12(200) 165,
DIMENSION
=_Q008
0009
D(200,3),
THETA(3),
DS(20). INT(20)
DIMENSION TITLE(1O)
=0010 DIMENSION THETB(3)
0011 DN(LAHDA, ARG) O. 5"LAMDA/SIN(O. 5. ARG)
=001? CONST O. 0174532925
0013 WRITE(5v700)
DO 20 I,. 1,3
-ää1s : READ(1010) TITLE
WRITE(3,600) TITLE
.
-0016
-00t7 20 CONTINUE
-0018 YRITE(S, A00)
0019 800 FOR1;AT(IH , 120(1H-), //)
WRITE(5,905)
_=0020
0021 905 FORMAT(1HOr6HTHETAI, 2X, 8HSPACINGI. 5X, 6KTHETA2.10H SPACING?,
^0022 fSX, 6HT. ETf3,1JH SPACING3)
ü023 10 FORMAT (I OA8)
: 0024 READ(1,15) LANDA. LIFi
0025 25 FORIIAT(3F0.0, I5)
-:0026 1S FOR,; AT(6X, F6.4, I2)
OOZ7 DO 3n L. 1,200
RFAD(1,25) THtTA(2), THETA(1). THETA(3), 11(L)
--0028 LT. 0.0) GO TO 50
0029 IF (THETA(2)
-8030 DO 40 J 1,3
0031 THETB(J)sCO;; ST*THETA(J)
-0032 D(L, J)ADN(LAMUA, THETB(J))
0033 40 CONTINUE
wRITE(5,000) (THETA(J). D(L. J). J 1.3)
-0031,
0035 900 FORf1AT(F6,2.5F1J. 2)
0036 30 CONTINUE
0037 SO L. L-1
0038 WRITE(5,800)
0039 RFAD(2,51lU) X4;1
-0040 00 400 1141,1uN
'0041 Kou;4T=u
RFAD(2r55) CARDI
-0642
0043 WQITE(5,56) CRDN
55 'FORMAT(A6)
--0044 RFAD(2.60)
0045 tl
0046 60 FORtIAT(IO),
0047 READ(2,65) (DS(I)'IHHT(I)"1*1, N)
0048 65 FORNAT(20(FO. 0, I0))
0049 WRITE(5,805) (DS(I), I 1, N)
0050 805 FORNAT(1H(#. 'STA; JDARD, /1H , 'SPACING9, T16,15F7.2)
0051 WRITF(5, A10) (INT(I), I 1, N)
0052 810 "FOR1tAT(1HOORELATIVE'/1N .' INTENSITY'.
T16,1517)
dRITE(5,460)
005 460 FORilAT(//, 1HO, (CALCULATED', 726, "STANADAkD#, 3X, IRELATIVEI/
0055 0N , 'SPACI! 1G'tT140'INTE; aSITY9.3X, 'SpACIyG", 4X,, INTENSITY')
-0056 110 FOR;! AT(1HO. F7,2,6X, I5.5X, F7.2,5X, I5)
0057 600 F0Kf1ATt140,1OAS)
0058 7e0 FORt1AT(1H1)
0059 00 70 la1, L
0060 00 dgl Ja1rN
0061 IF(0(I, 2). GE. DS(J)) Gu TO 302
0062 AO CObTI1UE
- 0063 302 IJaJ-1
0064 IF(IJ E., O) IJ'1
-Ws 75 IF<DS(! J). LT. D(3.1) AND. DS(IJ) GT. D(1.3)) GO TO 112
0066 IF(IJ. EV. J) GO TO 70 . 4

0067 IJ J
0068 GO TO 75
0069 112 IF(IT(I). EG. A) GO Ti) 115
-0070 IF(IAi$(II(I)"INT(IJ)). GT. LIM)GO TO 70
0071 115 KnU; 4T3KUUI1T+I
0072 I+RITE(5,110) D(I, 2). I1(I)'. DS(IJ), INT(IJ)
0073 70 COIfTIJ41JE
0074 56 FORIIAT(IHO, $CARD NUMBER ', T17, A6. /)
0075 5C0 FOR;4AT(10)
0076 COUNT KuUNT
0077 X. N
0078 PERC8100. U"000NT/X
0079 NRITE(5,920) PERC
0080 FORt1AT(//, lHO, 'PEHCEHTAGE OF STRONG PEAKS_PRESENT
920 a o,
0081 "F?. 1)
WRITE(5, A00)
-0082
0083 400 CONTINUE
0084 STOP
0085 END
 166.

Last card: -92 -ei -93.

The formats are shown in the program on Table 7.2. After the

input is read in, two job cards are required to operate the

program.
Card number 1: Job name,: useriiame, JD (specify job time)

2: contains the following below: -

PROGFORTRANXRAYPROG,FILE* CRO, = XRAYDATA, FILE*

cRJ = XRAYDATA2, FILE* LFO = XRAYOUf, TL 59

All cards are punched from line 1.

7.5 Test seärch'identification

The package successfully identified a mixture supplied by

National Bureau of Standards (America) to the Central Electricity

Generating Board (CEGB) in Harrogate for assessing the merits and

the demerits of automated" search against manual search. Quartz

silicate hydrate (CS2H2) were identified. Identifica-

and calcium

tion was enhanced by omitting the intensity values from the input

but considering them in the manual interpretation.

7.6 'Chapter Summary

a) A computer' search program i.s. developed to aid in eliminating

the many cisosely related silicates, altnninates', alunitho-ferri-

silicates and alummno-silicates' formed when red' tropical soils are

treated with lime.

b) A manual interpretation of the results of the automated

backed up by possible elemental analysis is necessary for

search

the final and conclusive identification of the reaction products.

c) The data base minerals against which the new peaks observed

in the XRD are checked. can be*replaced or edited to contain

minerals relevant in the search without altering the search

 167.

program.
d) The search program is capable of handling input data for

the peaks of°the new diffraction, patterns and their error windows

as 2e values, of Bragg's angle.

e) The program is flexible and can either use the d-spacings

alone or the d-spacings and their intensities to match the peaks.

.
 168.

CHAPTER8

SCAMVINGELECTRONMICROSCOPY
OF THE SOIL-LIME STYSTIMS

8.1 Introduction

Microscopic methods which are in use for studying the

atomic lattice structure or the fabric appearance of specimens

can be broadly divided into two i. e. optical microscopy and

electron microscopy. While optical microscopes use light energy

to form magnified images of specimens, electron microscopes use

high energy electrons for the same purpose. The electron image

formed by electron microscopes can be observed visually on a

fluorescent screen, recorded on photographic plate or film or

even stored on video tapes for subsequent processing and recon-

struction by a digital computer (100).

Electron. microscopes are usually classified as conventional

transmission electron microscopes (TIM) or high voltage electron

microscopes (I-NIM) and scanning electron microscopes (SPN) The

.
SEMhas achieved, recently, a high resolving power and versatility.
Due to this, only the SENwas used in the microscopic investiga-

tion of the lime treated soils. The SEN is therefore considered

in detail.

8.2 - Basic principles and the design of the SEI

The basic purpose of scanning electron microscopy like most

other forms of microscopy is to render finer details in the

specimen visible to the naked eye. To achieve this in the SB!,

electrons from a heated filament are accelerated by a high voltage,

usually of the order of 5- 50 kV, and passed down the centre of

 169.

an electron optical column. The electrons are formed into a

probe by a set of"magnetic lenses and focused onto. the surface of

a solid specimen. ` The probe scans the surface of the specimen

in square tiny sections in the same way as a spot can scan a

commercial television screen (100).

Scanning is achieved by passing a current frort a generator

through'the scan coils (located between the magnetic lenses) and

through the corresponding coils of a cathode ray tube'so as-to,,

produce an identical but larger image of the surface on its

viewing' screen. The probe-that strikes the-specimen surface

causes electrons to leave the surface. These in turn strike a

collector and an electron current signal is. set up. The signal

is amplified and used to control the intensity of the cathode ray

tube. There is an exact point'to point match between the section

on the specimen surface and that on the cathode ray tube screen.

The intensity of each point on the, image on the cathode ray tube

screen is a`direct measure of the number of electrons collected

from the corresponding point of the specimen surface.

Specimens are usually mounted on stubs that fit directly

into the specimen stage in the specimen chamber. They can be'

moved laterally, tilted and rotated in the SDI while being

examined. Non-conducting specimens are usually coated with a

conducting material for proper electrical contact and to prevent

charging occuring when they are being examined in the SEM.

8.3 The, sEms used in the investigation

A Cambridge 'Stereoscan' 600 scanning electron microscope

was used in investigating some of the lime treated soils, while

some other soils were studied using a Joel SFN, model JSM 35
 170.

available in the Scientific Section, Central Electricity.

Generating Board at Harrogate.

8.4 Sample preparation

A sample of the lime treated soil to be examined was

fractured and dried in vacuum. The side remote from the fractured

surface was gently ground flat on a'sand paper. The flat surface

was mounted on a stub with Araldite.

The mounted specimen was placed on-the substrate holder of

a Hummer11-coating apparatus and given a five minute coat of

gold-palladium. In good working conditions, no substrate manipu-

lation is necessary when Homer 11 is used as the sputtering is

a diffuse process and allows deposition around corners and in

recesses to prevent spot charging in the SFM. The coater used

was not in perfect working conditions and sometime. left the edges

of the specimen uncoated. When this happened'the edges were given

a paint of silver to ensure electrical contact between the conduc-

ting coat and the support stub. The specimen was examined in the

SIlVI immediately or stored. in a `des icator until required for

examination. In the latter circumstance, the specimen was given

one to two minutes coat prior'to examination in the SEM.

8.5 Results and discussion

The electron micrographs obtained with the Cambridge

Stereoscan 600 and Joel J94 35 are shown in plates 8.1 to 8.10.

Plates 8.1 and 8.2 show the electron micrographs of soil

T II obtained with the Joel JSM 35. They reveal dense micro-

structures in which plates have striated and oriented appearance

with intersecting morphology. The close textured massive

 171.

structures of interlocking rods, long fibres and plates of well

crystallised C-S-H can be seen. Amorphous reaction products

could be seen to be mixed with some of the well crystallised

C-S-H in plates 8.1b, 8.2a and 8.2b. These micrographs are

to those presented by Abo-El-Enein et al (107) who

similar

studied the morphology of autoclaved clinker and slag-lime pastes

in presence and absence of silica sand. By analogy to the

of the micrographs obtained by Abo-El-Enein et al, it

results

appears that the C-S-H in the soil-lime systems are results of

reactions between lime and amorphous silica which are probably as

reactive, at ordinary temperature, as slag-lime and silica pastes

at elevated temperatures.

Plates*8.3 and 8.4 are electron micrographs of soil T II

obtained with Joel JSM 35. Plate 8.3a shows. how the reaction

products, whose close up is shown in plate 8.3b, bond two'adjacent

mineral particles together. Plate 8.3b shows the reaction product

to be well crystallised, dense and hexagonal plate like crystals.

They are similar to the micrographs of "stepped clusters of

imperfectly developed plate like crystals of gehlenite hydrate

foreshadowed hexagonal limitation" shown-and described by

with

Jambor (108) as reaction products between activated kaolin

(kaolin heated to 500°C) and lime. Gehienite hydrate (C2ASHn)

can possibly form due to the reaction between amorphous alumina,

silica and lime (68,69). Soil J II contains appreciable amounts

of amorphous silica and alumina. Plate 8.4 shows similar reaction

products. (to that shown in plate 8.3) cementing and binding a

particle to the surrounding soil minerals.

 E
;rýý

ii
,1

Plate 8.1 (a)

Electron micrograph of soil T II showing dense microstructure

with plates of C-S-H having striated and oriented appearance.

Plate 8.1 (b)

A close up of Plate 8.1 (a)

revealing some amorphous reaction
products in presence of the well crystallised C-S-H.
173
 Plate 8.2 (a)

Electron micrograph of soil T II showing interlocking rods

and plates of well crystallised C-S-H.

Plate 8.2 (b)

Plate 8.2 (a) at a higher magnification showing not only the

well crystallised plate-like structure but also rod like long

fibres.
 175

._` dý
 Plate 8.3 (a)

One of the ways through which cementitious reaction products

increase the strength of the soil-lime system is binding

adjacent soil particles. This fact is illustrated by Plate

8.3 (a), an electron micrograph for soil T II.

Plate 8.3 (b)

This is a close up of Plate 8.3 (a). It shows the reaction

products to be well crystallised, dense and hexagonal. These

are probably hexagonal crystals of gehlenite hydrate.

177
 Plate 8.4 (a)

In contrast to the function of the reaction products described

in Plate 8.3 (a), strength increase in a soil-lime system can

be a result of the unreacted material being embedded in the

reaction products just as aggregates are embedded in concrete

or sand in mortar. Plate 8.4 (a), a scanning electron

micrograph of soil T II illustrates this fact.

Plate 8.4 (b)

A close up of Plate 8.4 (a) reveals most of the reaction

products to be that described in Plate 8.3 (a).

179
 Plate 8. S (a)

Electron micrograph for soil A II showing the formation of

massive plates of C-S-11. Plate like calcium ferrites can be

seen on the left hand side and top right hand side of the

plate.

Plate 8.5 (b)

Plate 8.5 (a) at a higher magnification.

181
 Plate 8.6 (a)

This clearly shows well crystallised plates of calcium

alwnino-ferri-silicates formed by treating soil J II with

lime. This product was also detected by x-ray diffraction.

Plate 8.6 (b)

Plate 8.6 (a) at a higher magnification.

 183
ý, <ý+

."
1.. ý ý

\\
 Plate 8.7 (a)

This shows how the calcium alumina-ferri-silicates formed by

soil J II help to give strength to the soil-lime system by

bonding and interlocking the unreacted mineral particles of

the soil.

Plate 8.7 (b)

A close up of Plate 8.7 (a) showing the calcium alumino-

ferri-silicates.
 185

V.

1.
 Plate 8.8 (a)

This Plate shows that not all the reaction products are

crystalline even after curing for about two years at 22°C.

These poorly crystallised reaction products formed by soil

J II have cementitious properties.

Plate 8.8 (b)

A higher magnification of Plate 8.8 (a). Notice the presence

of relatively large voids revealed in the microstructure.

1
 187
ý.

WIN
or! 1

_1

.. of
.4
IL

,0

ýr:
,MIN

a.
 Plate 8.9 (a)

Electron micrograph for soil J II after treatment with lime

showing the presence of amorphous reaction products.

Plate 8.9 (b)

Energy-dispersive (elemental) map of Plate 8.9 (a) for

calcium. Calcium can be seen to be present over the whole

surface but highly concentrated in the middle.

Plate 8.9 (c)

Elemental map for silicon. That for aluminium is similar.

This map shows very little silica or alumina to be present in

the middle thus showing that most of the calcium in the middle

is present as unreacted lime.

189
 Plate 8.10 (a)

Electron micrograph for soil J II showing the presence of

amorphous reaction products to the left hand side of the

micrograph.

Plate 8.10 (b)

Energy-dispersive (elemental) map for calcium showing

the presence of calcium products on the left hand

clearly

side of the micrograph.

Plate 8.10 (c)

Elemental map for silicon or aluminium. Their distribution

is similar to that of calcium showing the .left hand side of

Plate 8.10 (a) to be probably amorphous reaction products.

Plate 8.10 (d)

Elemental map for iron. The uniform distribution of iron,

unlike the distributions for calcium, silicon and aluminium,

suggests that the iron is unreacted or probably substitutes

into the reaction products on the left hand side.
1 91
 192.

Plate 8.5 obtained using a Cambridge Stereoscan SEI 600

shows the formation of massive plates of C-S-H and plate like

calcium ferrites (left hand side and top right hand side, plate

81.5a) by the soil A II. The morphology of the left hand side and

top right hand side of plate 8.5a are very similar to the morpho-

logy of the calcium ferrite, C4F, as reported by Rogers and

Aldridge (109).

The rest of the plates i. e. 8.6'to 8.10 were obtained

using Cambridge Stereoscan 600 for the soil J II after treatment

with lime and curing for two years at 22°C. Plate 8.6 shows

clearly plate like altnnino-ferri-silicates with a morphology

to that presented by Uchikawa and Uchida (110) and Roy et

similar
al (ill). One of the ways by which cementitious reaction products

help to increase the strength of a lime treated soil is`by bond-

ing and interlocking the unreacted mineral particles of the soil.

Plate 8.7 illustrates this phenomenon and shows how'the plate like

alumino-ferri-silicates achieve this in the soil J II. Notice

the relatively large voids in the microstructure.

Not all the reaction products can be crystalline even after

for about two years at 22°C. This fact is illustrated in

curing

plate 8.8 which shows supposedly poorly crystallised reaction

exhibiting cementing properties in their state.

products

Relatively large voids also characterise the microstructure.

Plate 8.9 shows the micrograph and energy-dispersive

(elemental) map, of the fractured surface, for calcium, silica

It
11,4
r
and alumina. The calcium element distribution map, Plate 8.9b,
h
"a

yS

S
shows that except at the middle, calcium compounds are more or
Hill

ý;x. '
ý.
ý:
r.
F:
ý.
j{
3
 193,

less uniformly distributed over the surface. Plate 8.9c shows

the elemental map for silica. The elemental map for alumina is

similar to that for silica. It can be seen in Plate 8.9c that

there is little or no silica or alumina at the centre (where

there is a high concentration of calcium). Thus most of the

calcium compound at the centre of Plate 8.9a is unreacted or

carbonated lime. Apart from the centre, silica and alumina are

almost as uniformly distributed as calcium which suggests that

reaction products without recognizable morphology are present.

This supports the fact that poorly crystallised or amorphous

reaction products are still present after about two years curing.

While such products could be uniformly distributed over an area

of the lime treated soil, it is not necessarily true to say that

such products are uniformly distributed over all the surfaces*of

the lime treated soil as suggested by Ormsby and Kinter (65).

Plate 8.10 highlights this fact and stresses the need for thorough

mixing of the soil and lime. The element map (Plate. 8.10 (d))

shows that iron is either unreacted or substitutes into the

calcium alumino-silicate (left hand side, Plate 8.10 (a)).

8.6 Chapter ' Su mary

a) Red tropical soils form reaction products which include.

calcium silicates, alumino-silicates and alumino-ferri-silicates.

b) While some of the reaction products have discrete morphology,

after about two years curing, others still remain in amorphous

states.

c) The C-S-H are associated with dense and close textured

massive microstructure whereas the alumino-silicates and aluminO-

ferri-silicates usually occur with relatively large voids in the

 194-.

microstructure. This suggests that the microstructure is

function of both the volume and type of the reaction products.

 Iss.

. ',CHAPTER' 9

*OF'RID 'TROPICAL 'SOIL-LIME SYS IBS

OTHERCONSTITUENTS

9.1 Introduction

Apart from the main cementitious reaction products already

discussed in previous chapters, there are other constituents in a

red tropical soil-lime mixture. These may be other reaction

products or unreacted'residues. They include calcium carbonate,

unreacted lime, unreacted soil and voids. They also influence

the behaviour of the system. This chapter then reviews the

available methods for the determination of calcium carbonate,

unreacted lime and porosity of the systems studied. It presents

separately a review of thermogravimetric analysis and finally

presents the results together with the discussion of the findings.

9.2 Literature review mid methods

While engineers recognise the presence of unreacted lime,

carbonated lime and voids in lime-pozzblana mixtures, they have

made little efforts-to-. quantify them and thus quantitatively

.
assess their influence on the system. Valuable information such

as the efficiency of curing methods can be obtained from the

quantitative determination of unreacted and carbonated lime.

9.2.1 Calcium carbonate

In practice, lime is mixed with the soils in presence of

air. While some of the lime reacts with the soil minerals,

another fraction of the added lime reacts with the carbon dioxide

of the air to form calcium carbonate according to equation 9.1.

Ca(OH)2 + C02 -º CaCO3+ 1120 (9.1).
... ... ...
 196.

The estimation of carbon dioxide in cement materials is not a

difficult problem and this is the basis "for the estimation of

carbonated lime in soil-lime mixtures. Standard gas volumetric,

gravimetric and chemical methods are well established.

Gravimetric and chemical methods are hazardous, time consuming

and cumbersome (90). Gas volumetric methods are therefore

favoured and considered in more details.

Carbon dioxide in silicate "rocks" is usually determined by

the, decomposition of the carbonate by an acid. Early methods

relied on qualitative tests for carbon dioxide (112). The tests

involved the observation of bubbles when hydrochloric acid is

added to a test'tube'containing the ground sample and hot water.

Later these methods were modified to a quantitative one. This

involved absorbing and weighing the carbon dioxide evolved when

the sample is decomposed by acid. Methods of this type required

about 45 minutes for a single determination. Shapiro and

Brannock (112) developed a rapid method for quantitative determi-

nation of carbon dioxide in silicate "rocks". They used a carbon

dioxide tube with a side arm for catching and measuring the gas

evolved. Their method is described in detail in appendix B. 1.

Ormsby and Kinter (65) used this method to measure carbonated

lime in lime-montmorillonite-water systems.

9.2.2 Linreacted lime

The detection of free calcium hydroxide in soil-lime systems

is potentially very important since it can be used as a parameter

to measure the reactivity of the pozzolanic materials which are

contained in a soil.

Classically, free or uncombined lime in ordinary Portland

 1sß,

cement, clinkers and in presence of calcium silicate hydrates is

determined by selective dissolution of the lime from such mixtures.

The organic solvents employed vary from glycerol-alcohol through

ethylene-glycol to acetoacetic ester-isobutyl alcohol mixtures.

One such method is the ASTM method C114-77 (113) for the deter-

mination of free lime, in Portland. cement and clinker. The method

does not differentiate,. between free calcium oxide and free

hydrated lime. It is based on the-solution of free lime in a hot

solution of glycerol and alcohol and subsequent titration of the

dissolved lime with an alcoholic solution of anmnium acetate.

The method is described in Appendix B. 3.

Methods that employ organic solvents such as the ASTM method

C114-77 invariably attack calcium silicate hydrates and aluminates

when they occur with. free'lime"*(114). Pressler et al (114) found

that hydrates prepared'at lower temperatures and with higher molar

C/S ratio-were more vulnerable to attack. Of the calcium silicate

hydrates they prepared by hydrothermal reaction between CaO and

silica gel at 120°, 180° and 240°C, the worst case of organic

solvent attack was that prepared with a molar C/S ratio of 1.5

at 120°C. To overcome this problem, they modified an acetoacetic

ester-isobutyl alcohol method said to be originally due to Franke.

The details of their modified' method is given in Appendix B. 4.

They found their method to give accurate results for calcium

hydroxide in presence of calcium silicate hydrates in which

calcium was firmly bound, but the solvent attacked the calcium

when it was weakly bound.

While their method gave satisfactory results, it was time

consuming, requiring at least three hours for each determination.

 198.

Presselerr-et al (115) proposed two further modifications of their

method in order to shorten the time of testing and to reduce the

amount of sample required. In their modified methods, details of

which are given in Appendix B. 5, a single extraction is performed

on each 0.1 gm portion of the sample.

In one case of the two modifications proposed, the ratio of

the solvent volume to sample weight was varied from 500 to 3000

ml/gm. They found it convenient to use six 0.1 gm portions of a

sample for analysis. The time of each extraction was thirty

minutes and all six extractions could be done simultaneously.

They 'called this method the solvent variation method (SVM).

Longer but equal times'such as one hour may be used in extracting

each of the six sample-solvent ratios. In the second case of

modification, the ratio of the solvent volume to sample weight

was held constant at 1000 mi/gm while the time of extraction was

varied from one portion of the sample to another. This they

called the time variation method ('NM). In all other respects,

the TVM was the same as the SVM. In both methods, the number of

extractions can be varied according to need. While the TVM uses

less solvent than the SVM, the SVM requires less time than the

TVM.

In any extraction, all, or at least a large part of the

free lime may be removed and simultaneously a small amount of the

combined calcium (in the reaction products). Let the free lime,

which is easy to extract, be designated by Ce and the small

amount of combined lime, which is difficult to extract, by Cn.

It is for the latter that a correction should be made. The rate

of reaction of Cn'with ethyl acetoacetate is given by

 199..

-dCn (9.2)
dt = kiCna ..... ...

where a is the amount of ethyl acetoacetate and k is the rate

constant.
Under the conditions of the SVM, the solvent-sample ratio

is so large that the amount of Cn extracted with the lower ratios

and those extracted by higher ratios can. be assumed constant in

equation 9.2 without introducing significant errors. For equal

extraction times, equation 9.2 becomes

LCn = k2a (9.3)

... ... ... ...

where a is proportional to the'vvolume of solvent. per gram of

sample (Ns).
Hence equation 9.3 can be'written as

pCn = k2 Vs (9.4)
... .... ...
Because of the high solvent-sample weight ratios used'in SVM, all

the free lime is removed in one extraction. The total calcium

in an extraction, Ct is given by
removed ,
Ct = Ce + QCn (9.5)
... ... ...

Substituting. for ACn:

Ct = Ce + k2 Vs (9.6)
... ... ...
Equation 9.6 shows that a plot of Ct against Vs should give a

straight line with the intercept, Ce, equal to the free lime. If

all Ce is not removed in one extraction, probably for a very low

value of VS, Ct is low and the point falls below the line defined

by the other points. If this happens, this point is disregarded

and the straight line drawn using other points obtained for

higher values of Vs. In all the samples they (115) tested, Ce

was removed by Vs ratio of 1000 or greater. However if the

 200.

contrary should happen they recommend the use of the highest

values of Vs or the extraction of the samples for one hour

instead of thirty minutes.

The TVMmakes use of equation 9.2 in another way. The

average of the initial and final amounts of ethyl acetoacetate

(ai, of respectively) used is, as assumed, equal to a in equation

9.2. For constant solvent volume-sample weight ratio (Vs) but

variable time, equation 9.2 becomes

AC11 = kit (9.7)

... ... ...

where t is the time and k3 is a constant. Combining equation 9.7

with equation 9.5 gives

Ct = Ce + k3t (9.8)
... ...
Equation 9.8 indicates a straight line relationship between Ct

and t whose intercept is Ce, free lime.

9.2.3 Porosity

Porosity, n, is defined as the fraction of the total volume

that is air and evaporable water. Because in using lime treated

soils in construction attempts are made to exclude water and the

fact that rate of loading of such structures is so slow that there

is no pore water pressure build up, the inclusion of evaporable

water in the definition of porosity is not unreasonable. Hence

porosity as defined is used as a measure of voids in red tropical

soil-lime systems studied in this work. Total porosity will

therefore be calculated using the dry densities of the soil-lime

systems.
Many investigators have either directly or indirectly

related the strength of hardened cement pastes to porosity.

Powers'and Brownyard (95) found that:

 201.

Vm
S= 827.4 W- 24.8 (9.9)
o ... ... ...

where S= strength of hardened paste

Vm = the constant in the BET surface area equation.

Vm is proportional to the surface area of-the cement gel which in

turn is proportional to the amount of cement gel. Wo is the

original water content after bleeding.

Schiller (96) after studying the strength of highly brittle

materials found that

P=k log ek (9.10)

e ... ... ...

where ek = critical porosity for which the strength, 'P =0

I= coefficient independent of overall porosity but

dependent on pore shape and nature of crystal

contacts.

The definition of k in Schiller's equation suggests that strength

of highly brittle materials do not depend on the overall porosity

but on pore structure. '

Recently, Hoff (116) showed that for cellular concrete,

(1 (9.11)
o=ao n)b
- ... 0.0.0.

where a= strength

vo = theoretical strength at zero porosity '"

n= overall porosity
b=a power exponent.

He was able to determine ao and b for cellular concretes made

using only Portland cement, water and preformed foam which

provided the air contents.

Most relationships between strength and volume of voids can
be written in`the form of equation 9.11. If a relationship
 202,

similar to. that expressed by equation 9.11 can be found for red

tropical soil-lime systems, it would then be possible to relate a

desired design strength to the porosity of the systems which in

turn is a function of the moisture content of compaction, soil-

lime. density. and specific gravity of the soil particles. This

would provide some freedom in the selection of a soil-lime design

criteria for a particular requirement.

9.3 Thermogravimetry

9.3.1 Introduction

Thermal analysis is a technique which determines some

parameter of a system or material as a function of temperature.

This implies that the system or material must be reactive to, heat

and that such a reaction should be measurable. The property. of

the material used as. a measure of the reaction classifies the

different methods of thermal analysis.

Such properties of the material measured in thermal investi-

gations have been described by Mackenzie (117). They include:

a) Mass changes usually measured in thermogravimetry (TG),

derivative thermogravimetry (DTG), isobaric and isothermal

weight curves.

b) Energy changes usually measured in differential thermal

analysis, (DTA), differential scanning calorimetrY (DSC) and

heating curves.

c) Dimensional changes

d) Mechanical changes

e) Electrical changes

f) Magnetic changes

g) Optical changes
 203.

h) Emitted sound

i) Evolved volatiles.

The measurements can-be carried out under static, quasistatic or

dynamic temperature conditions depending on how the temperature

of the heat source changes with time.

The most widely used techniques applied for the analysis

of soils are thermogravimetry (including derivative thermogravi=

metry) and differential thermal analysis. The former is

considered in more details.

9.3.2 Basic principles

When a material is heated from room, temperature to 1000 -

1500°C, reactions occur which can lead to either a weight loss or

a weight gain. Such changes can be made to take place either in

air or in a chosen environment. Thermogravimetry allows to.

constantly measure such changes in weight as a function of time

or temperature.

Someof the important changes that take place when a soil

or clay is heated are (39,118):

a) Dehydration: Water lost from a heated soil sample can be:

(1) Pore water: this is normally held in the pores and is

usually lost at about 105°C.

(ii) Adsorbed water: this is water adsorbed by the surfaces

of the clay or other minerals in the soil düe to

hydrogen bonding between the clay minerals and water

molecules. It is lost at up to 2400C and is a

reversible process.
(iii) Hydroxyl water: this water is lost when the OH ions in

the crystal lattice of a mineral are disrupted and

 204.

transformed, to H20. This results in a complete des-

truction of the mineral structure.

b) Decomposition: carbonate materials in the soils decompose at

characteristic temperatures to give off carbon dioxide.

c) oxidation and reduction: some components of a soil can be

made to oxidise or reduce in a controlled atmosphere.

The temperature ranges at which some of the changes can

take place are-characteristic of minerals undergoing the changes.

The temperature at which the event is maximum is used to identify

and possibly quantify the mineral. Thus the use of this technique

in soil analysis requires an arrangement to measure the weight

changes. This is provided by the thermobalance.

9.3.3 Instrument design

A thermobalance has three essential components: the balance,

the thermal and recording components (118). The balance component

makes use of the same weighing arrangements as are used in a

laboratory balance. The change in weight can be measured either

by deviation (deflection balances) or by the value of the resto-

ring force required to return to the original null-position

(compensation or null-point balances). More often than not, it

is the thermal resistance of the sample holder which limits. the

temperature of thermogravimetric measurements (88). Temperature

regulation can be effected either by heating at a constant rate

or according to a definite thermal cycle. Electric ovens are

usually used for high temperatures with the heating elements made

of nichrome resistor or platinum or alloy of platinum or graphite;

the type used depending on the desired maximum temperature. The

recording component records the temperature and changes in weight.

 205.

The type of balance and method of temperature regulation influen-

ces the type or types of recorders to be used.

Modern thermobalances have. facilities for working either

or in a controlled atmosphere. The gases evolved

under vacuum

during reaction-, may be collected and analysed. In some thermo-

balances, facilities also exist which permit simultaneous DTA and

TG experiments on the-same chart thereby offering a further

to further clarify the thermal events that occur in the

picture

system studied.

Temperature measurements in TG are chiefly inluenced by the

position of the collector, the crucible or the sample container

being studied (88). Such material properties as

and the material

thermal conductivity, heat of reaction, fineness, density

weight,

can start a temperature gradient thereby disturbing

and sintering

temperature measurement. Measurement of weight changes are

influenced by those factors which exert actions on the balance of

the weighing devices. Such factors are gaseous buoyancy (a

function of temperature exerted on the. weighing devices

pressure

by the gaseous atmosphere) and convection. These factors that

TG measurements must either be taken into account or

affect

suppressed in the instrument design.

9.3.4 Thermogravimetric analysis of'soils'and clays

One of the problems with chemical analysis of soils and

clays is that the solvents employed are not strictly selective;

hence in-situ methods of analysis such as x-ray diffraction and

TG are potentially powerful. Not much attention has been given

to the use of TG in qualitative and quantitative analysis of soils

and clays in the past. Theoretically, any constituents of soils

 206.

and clays that react thermally can be identified and possibly

quantified. Such constituents are organic matter, hygroscopic

or pore water, adsorbed water, clay minerals, calcium carbonate,

lime, goethite and gibbsite.

Schnitzer et al (119) used the thermobalance to study a

number of soils differing widely in origin and compositon. They

were able to detect and in some cases quantify hygroscopic

moisture, organic matter, inorganic carbonate and lattice water

in clays. They pointed out the presence of hydrous oxides of

iron and aluminium as possible sources of error in quantifying

of the soils. For example, it is

some of the other constituents

difficult to quantify organic matter when goethite or gibbsite is

present. Similarly allophanes and amorphous minerals will cause

high values for hygroscopic moisture while the presence of

kaolinite will interfere and cause absolute positive error of

1.5% in carbonate measurements.

Longuet (88) and Biffen (120) have used the thermobalance to

determine free calcium hydroxide and calcium carbonate in a

number of calcium silicate hydrates with a good accuracy. The

systems or mixtures Biffen (120) studied were not complex. because

the thermal events associated with calcium hydroxide and calcium

carbonate were fairly distinct thus eliminating the problem of

overlaps. Quantitative interpretation of overlapping thermal

events is not easy and to effectively utilise the full potentials

of the TG, this problem calls for resolution.

9.4 Experimental methods

Under this sub-chapter, the methods chosen and used to

quantitatively determine the carbonated and unreacted lime are

 207.

described including the modifications made. The methods of cal-

culation of the porosity and unreacted. soil are also given. The

thermobalance used is described and the experimental conditions

stated. All lime treated samples used were vacuum dried-for

seven days.
9.4.1 Calcium carbonate

Calcium carbonate in the lime treated soils was determined

by the gas volumetric method od Shapiro and Brannock (112).

Carbonated lime was also determined by thermogravimetric analysis.

The method of Shapiro and Brannock was slightly modified owing to

availability of materials.

Shell Tellus 27 oil was used in place of motor oil S. A. E.

No. 10. In calibrating the carbon dioxide tube, a one gram

mixture of an untreated red tropical soil and varying amounts of

calcium carbonate (analytical grade) was used because the National

Bureau of Standards standard sample No. 79 was not available.

1: 1 hydrochloric' acid and 3% mecuric chloride were used. The

other steps as outlined for the "calibration of the carbon

dioxide tube" were unchanged.

9.4.2 Unreacted lime

The ASIM method for the determination of free lime in

clinkers and cements, C114-77 (113) was tried for the measurement

of unreacted lime in the soil-lime systems. This method was

found to be unsatisfactory. The slight pink colour which is an

indication of the end point of the titration was impossible to

determine because of the red or reddish-yellow colour of the soils.

Because of this, a method which provides for the titration of the

clear filtrate of the solvent after lime has been solubilized in

 208.

it was preferred. The solvent variation method (SVM) of Pressler

et al (115) was therefore used. SVMwas preferred to TVMbecause

time was considered more essential than the volume of the solvent.

The multiple extraction heating mantle with temperature regulating

controls used for the extraction of the free lime in the lime-soil

mixtures is shown in Plate 9.1. Along side the mantle in Plate

9.1 is the equipment used for vacuum filtration.

9.4.3 Porosity

The overall porosity, n, of each lime treated soil at a

moisture content for maximum dry density and strength was calcu-

lated using its dry density and the specific gravity of the soil

particles.
(1 - i) (9.12)
n= x 100 ... ... ...
Gs

where 7d = dry density

Gs = specific gravity of the soil particles

n= porosity in percentage.

No attempt was made to measure or determine the pore size dis-

tribution of each lime treated and compacted soil.

9.4.4 Unreacted soil

The unreacted soil was qualitatively assessed by TG and

quantified by the difference between 100% and the sum of the

volumes of reaction products, overall porosity, carbonated and

unreacted lime.

9.4.5 Thermogravimetric analysis of lime treated red tropical soils

The TG analysis of the soil treated with lime for maxiinmt

strength was carried out using a Stanton TR 02 thermobalance to

ted
quantify carbonated and i nreac. lime and to identify the
 Plate 9.1

Apparatus used for the multiple extraction of unreacted lime.

Alongside the heating mantle is the set up for vacuum filtration.

i
 210

10
 211.

unreacted minerals in the soils.

The instrument is a beam form of the deflection type ther-

mobalance consisting of a capacitance-follower plate which

measures the beam movement and records the change in weight for

a pre-determined beam movement (39). The furnance wall tempera-

ture, change in weight and time are simultaneously recorded on the

same output. Three different linear heating rates are obtainable

while temperature measurement is effected by a platinum rhodium-

platinum thermocouple with the hot junction on the furnance wall.

The furnance is a vertical 50 mmbore with nichrome bifilar

windings capable of reaching a maximum temperature of 1000°C.

Samples were ground to pass B. S. sieve 75um and heated in

static air using silica crucibles. The calculated heating rate

was 40C/minute. 0.5 gm of the ground sample of each lime treated

soil was used.

9.5 Results and discussion

9.5.1 Calcium carbonate by gas evolution

As can be seen from Table 9.1, the carbonated lime as

obtained from gas evolution is in general lower than the values

obtained from TG measurements. This may be due to the fact that

despite the relative insolubility of carbon dioxide in oil when

compared with the use of sodium chloride solution or water as the

supernatant liquid in the carbon dioxide tube, some carbon dioxide

evolved may have solubilized in oil. In using the method of

Shapiro and Brannock (112), care must be taken to avoid foaming

of the soil-acid solution while at the same time ensuring that

the temperature is high enough to allow the decomposition of the

carbonate by acid.
Table 9.1 Constituents of red tropical soil-lime systems
other than cementitious reaction products
 213.

LIME Over- LIME UNRE-

UNREACTED
f or max a ll L ACTED
stren: 7 . poro- GAS T. G a- +c SOIL
9t h s i ty VOLUME a
SOIL ' s. G % Wt. o%vol. % Wt. % vol. % % vol.
Q % by , `vt.
o
(a) vol. (b) (c)
JPA 4.5 2.99 36.42 0.57 0.26 1.35 1.51 0.46 44.9 59.3

JPM 6.0 3.10 37.29 0.35 0.15 1.15 1.28 0.44 62.5 55.2:,

JPB 4.5 3.04 38.16 1.0 0.45 0.90 0.68 0.26 62.7 56.6

CT 4.5 2.94 35.24 1.05 0.47 1.35 1.81 1.06 46.7 61.4
AI 4.5 2.95 38.58 0.90 0.40 1.30 1.31 1.00 50.9 55.4

All 4.5 3.09 42.14 0.87 0.39 1.25 1.14 0.92 55.3 52.0

SIA 4.5 2.95 34.98 1.85 0.83 1.45 1.60 1.59 23.3 61.6

SIB 6.0 2.92 33.22 1.70 0.76 2.20 2.31 1.42 33.3 61.0

SII 4.5 2.72 28.82 0.95 0.43 1.175 1.28 1.11 50.0 64.3

NF 4.5 3.05 35.54 0.45 0.20 1.675 1.85 1.10 48.9 59.5

JI 4.5 2.86 33.99 0.80 0.36 1.45 1.62 1.00 46.10 60.8

JII 6.0 3.16 35.25 1.10 0.49 1.80 1.21 0.72 44.80 60.5

TI 6.0 2.75 35.20 1.17 0.52 1.20 1.28 0.76 59.20 57.8

TH 6.0 2.82 39.04 0.55 0.25 2.20 2.42 0.79 50.50 56.0

RC 6.0 2.81 33.10 0.50 0.20 2.15 2.28 0.99 53.70 60.1

USM 4.5 2.69 36.43 0.70 0.31 1.95 1.68 1.00 47.1 60.0

SB 4.5 2.69 31.19 0.80 0.27 1.25 1.38 0.92 51.6 62.0

PH 4.5 2.80 36.54 0.40 0.18 1.70 1.82 1.08 50.70 58.8

SM 6.0 2.99 37.62 0.60 0.27 1.60 1.81 1.02 59.80 56.1
 214,

9.5.2 Unreacted lime'by extraction

Figures 9.1 and 9.2 show the results of the extraction

plotted according to equation 9.6. It can be seen that apart

from possibly Cr a solvent volume-soil weight ratio of 500 is

usually sufficient to take all the unreacted lime into solution.

Each point was obtained from the average of two titrations, the

values of which were equal in most cases. The values of the

intercepts i. e. the unreacted lime are given in Table 9.1. It

can be seen that after about two years, more of. the lime that has

not reacted with the soil has carbonated than remain as free lime.

Carbonation is a continuing process as long as there is unreacted

or free lime and air in the pores.

Figure 9.3 shows the variation between strength. and the

reacted lime as a percentage of the ammmt of lime added. It can

be seen from the figure that soils in which more of the added lime

reacted have higher strengths. A similar relationship exists

between lime reactivity and the amount of reacted lime as shown in

Figure 9.4. These relationships are not linear but are similar

to those found to exist between volume of reaction products and

strength (Figure 6.5). This suggests that the percentage of lime

reacted may be related to the volume of reaction products. This

possibility was investigated by plotting the percentage of lime

reacted Vs volume of reaction products. This is shown in

Figure 9.5.

Using a stepwise multiple regression analysis (Statistical

package for Social Sciences, SPSS 5) a quantitative relationship

between the strength, lime reactivity, activity index (each

considered as a separate dependent variable) and the percentage

Fig. 9.1 Weight of lime Vs volume of solvent used in
extraction
 I

216.

I.
20
SM
NF

1 Ti
2.1 11

PH

N
1o
Zl

v SB
T11
O 1'0
E
rn
I-
a,
a

as
E

0 JPB
rn RC
3

0
2.0
SII

10

0.5
500 1000 1500 2000 2500
Volumeof solventper gram of soil, V5
Fig. 9.2 Weight of lime Vs volume of solvent used in

extraction
 218.

20
1 11

'SIB

2"
Al
JPA
1.i

2-C CT

All
V

S1A
N

rn

a- 2.0
CA.

d
E
JPM
0
r
O
rn

20
us

1.0

0.5
5000 1000 1500 2000 2500
Volumeof solventper gramof soil, Vs
Fig. ' 9.3 Relation between strength and reacted

lime as % of added lime

Fig. 9.4 Relation between lime reactivity and reacted

lime as % of added lime
 7

220.

NE

VN
V

01 1111f

10 20 30 60 50 60 70
Reactedtime as /. of added lime

N3
E

ci
d
V

CL.)
E

10 20 30 - 60 50 60 70
time
-Reacted as °/o of ridded lime
Fig. 9.5 Relation between volume of reaction products and

reacted lime as % of added lime

 222.

'6

4
O

e
ö

y3

0
o.
-2
d

0l 30 t0 50 60 7u
10 20
Reactedlime as °/o of added lime

0
 223.

of reacted lime as the independent variable was investigated.

The following equations and statistics were obtained for their

lines of best fit:

a) Between strength and percentage of reacted lime:

" St =-0.0013 12 - 0.0394 1+0.4885 (9.13)

...

multiple R- 0.572, confidence level = 32.7%, error of

estimate = 1.149.

b) Between lime reactivity and percentage of 'reacted lime:

LR = 0.0014 12, - 0.0663 1+0.8295 (9.14)
...
multiple R=0.6038, confidence level = 36.4%, error of

estimate = '0.7818.

c) Between activity index and percentage of reacted lime:

Ai'= 0.0003 12 + 0.0237 1-0.6064 (9.15)

...

multiple R=0.385, confidence level = 15%, error of

estimate = 1.196

where
St = total strength of the soil-lime system
1'= percentage of reacted lime

LR = lime reactivity

Ai = activity index

Equations 9.13,9.14 and 9.15 are not significant. The

poor correlation may be due to the fact that the percentage of

lime reacted does not take'into account the type of reaction

products formed. This parameter has been shown to influence the

strength of the soil-lime system and was considered in defining

the'activity index.

9.5.3 Unreacted soil

The relation between strength and volume of unreacted

1 224.

in the lime treated soils is shown in Fig. 9.6. The

material
sinn of the volumes of pores, unreacted lime and carbonated lime in

is denoted by t. Where a figure is written beside

each specimen

a point, this represents the value of t applicable to that point.

This is written beside those points whose t do not lie within

the ranges indicated by either of the three lines.

When t is kept 'constant'. (varying within a certain limit

e. g. t= 39.47 ± 0.60 or 36.40 ± 1.0 or 31.47 ± 1.11), the

strength decreases with increasing volume of unreacted material.

This finding is similar to that, reached by Jambor (63) for

unreacted residues of pozzolanic admixtures who concluded that

do not help-the strength, of the system. The
unreacted residues
in Fig. 9.6 can be explained by the fact that
relationships
increase in the volume of the unreacted soil (when the volume of

other phases is 'constantis accompanied by a reduction in the

volume of the reaction products. This shows that when only the

unreacted soil and reaction products are allowed to vary in a red

tropical soil-lime system, more strength is generated if more

products are produced than if less reaction products

reaction
(resulting in more unreacted soil) are produced.

This explanation does not mean that unreacted soil is

harmful to the system; rather it can be said that it is an

inferior imparter of strength than the reaction products. Infact

in practice, an amount of unreacted soil is indispensable in the

system if a red tropical soil-lime mix is to satisfy other

other than strength. Ringsholt (121) writing on how

requirements

to select samples of the most suitable red tropical soils for the

production of "Latorex" (products of lime treated laterites)

 225.

building elements noted that while the clayey "laterite" gives

a high strength, it is difficult to pressure mould into blocks

and can easily develop cracks. A well graded red tropical soil

with adequate amount of clay seems ideal.

9.5.4 Porosity vs strength

Figure 9.7 shows a plot of overall porosity vs strength

according to equation 9.11. No relationship of the form in

equation 9.11 exists between the overall porosity of the systems

and their strengths. Some other investigators (63) have found

similarly for hardened lime pozzolana pastes. Rather, strength

is found to be related to the pore structure which can be

approximated by average pore size. The smaller the average pore

size, the higher the strength of the system.

Pore size distributions and/or the average. pore sizes of

hardened lime-pozzolana pastes depend on the type and volume of

the newly formed reaction products contained in the paste as well

as on its overall porosity (63). Materials that contain the same

type of reaction products have similar pore size distributions.

The differences in their average pore sizes are more or less pro-

portional to the volume of reaction products and overall porosity

with the average pore size decreasing with increasing volume of

reaction products. Pore size distribution differs with different

types of reaction products, even when the soil-lime system

contains the same volume of reaction products and total porosity.

9.5.5 Thermogravimetry and Derivative Thermogravimetry

Typical thermogravimetry (TG) and the derivative thermogra-

vimetry (DTG) curves of. the soils are shown in Figs. 9.8 and 9.9

for two of the soils. The TG and the DTG of the rest of the soils
 Fig. 9.6 Relation between strength and volume of

unreacted soil

Fig. 9.7 Relation between strength and overall porosity

hL
I

227.,

5-

CAD

4
ö ý%r

3
3420
,. "
.ý,
N
E " 0
0
.w"
fl)
C-, "

43.45
"""
"
""
37.80 "
"
oý 52 54 56 5B 60 62 64
Unreactedsoil-°/. vol

3.8

3-4
A

4-2 3.0
°'
0

2-6

2.2

1.5 1.6 1.7 1.8 1.9 2.0 2.1

-log10 (I-n)
 Fig. 9.8 TG and DTG curves of lime treated soils

Fig. 9.9 TG and DTG curves of lime treated soils

k
 229.

Temperature,
°C
200 400 600 Boo 1000
0 o
.....

10 z

20 4

30 6

a0 6

i 50 -10

N
d

NF
TG ;
....... 0TG `z
00E
BC

10-

20 4

"30 6

co e
SII
TG
50 """""""" DTG
 230.

are presented in Appendix B. 6.

The TG curves were obtained by plotting the cu mnulative

weight loss vs temperature while the DTG curves were obtained

from the TG curves by plotting weight changes in five minute

intervals. The period of five minutes was chosen for convenience

since a time marker pen incorporated in the recording mechanism

marks off five-minute periods in the TG traces. This method of

obtaining the DTG appears to be a simple and satisfactory way of

obtaining a derivative curve from the TG result produced by a

Stanton balance (39).

Five weight loss'peaks numbered A to E are present in all

the lime treated soils, although they are present in varying

quantities.

Peak A occurs at 100° - 120°C. This is due to the loss of

combined (hygroscopic and adsorbed) water in the specimen.

Peak B occurs at 280° - 320°C. The peaks give a dip on the
TG curve. The possible thermal events taking place can be either

the burning of organic matter contained in the soils or the

dehydration of goethite or the structural water loss in gibbsite.
The soils contain very little organic matter and not even traces

of gibbsite, hence the latter event i. e. dehydration of goethite

accounts for peak B. Goethite is dehydrated to form haematite.

This event is more pronounced in some soils than in others showing

that such soils contain more goethite than others even after the

soils have been treated with lime.

Peak C occurs at 4000 - 440°C and is due to the decomposi-

tion of the unreacted lime to give off water. The event appears

as a small peak in the DTG of some samples while in others it is

 231.

"swallowed" up by event D thereby conferring a skew on peak D.

The decomposition temperature depends on the amount of lime

present. Biffen (120) showed that for large samples of calcium

hydroxide (1.223 gms) the range of the decomposition temperature

is between 395° to 649°C while for small samples (0.104 gm) it is

between 425° to 534°C. This suggests that the smaller the sample

of calcium hydroxide, the less the range of decomposition tempera-

ture. The peak temperature and size of peak C shows that the

unreacted lime in the soils are very small indeed. The use of

decomposition peak temperature as a possible parameter for

quantifying lime in complex systems as the ones considered in this

study may be worthy of investigation.

The DTG curves in Figs. 9.8 and 9.9 and the temperature

ranges quoted by Biffen (120) show that the decomposition of lime

is overlapped by the thermal event having peak D in the samples.
This is confirmed by the fact that at the temperatures where

peaks C occur in the DTG, no dips can be observed in the TG

curves. This is unlike peaks A, B, D, E which show dips in the

TG curves where they occur. Biffen (120) also showed that the

water of hydration driven off calcium silicates comes off very

slowly in the region where calcium hydroxide would give off

water. When lime occurs in presence of calcium silicate hydrate,

the dip on the TG curve due to the presence of lime could readily

be observed. By taking the vertical distance from the point at

which the straight-line curve (of the TG) due to the evolution of

the combined water from the calcium silicate hydrates starts to

change to the point where the straight line resumed and calcula-

ting the calcium hydroxide from the loss in weight of water

 232,

equivalent to this vertical distance, he obtained a good estimate

of the amount of calcium hydroxide in all cases. The estimation

of calcium hydroxide in the samples studied from the TG curve-was

out because the small quantities involved are over-

not carried
lapped by peak D.
Peak D occurs 4800 °C
500. is due to the dehydroxy-
at - and

llation of the unreacted kaolinite. The curves show that the

kaolinite vary in quantity from sample to sample.

unreacted

Peak E which occurs at 6700 - 700°C is due to the decomposi-

tion of the carbonated lime in the samples. A greater part of

this event is not overlapped. This is subtantiated by the dip it

in the TG curve. The carbonated lime in each sample was

produces
by converting the weight of carbon dioxide given off
calculated

to equivalent weight of calcium hydroxide. This was done by

the weight of carbon dioxide by 1.6818. The figure

multiplying

of 1.6818 was obtained from the stoichiometry of the reactions:

CaC03 CaO + CO2 (6.3)

-} ... ... ...
100 44

(from 100
'. ymg of CO2 TG) = y CaC03
. 44

Ca(OH)2 + C02 CaC03 + H2O (9.1)

...
74 100

10.01 74 100 Ca(OH) 74

y CaCO3 __ y 2= y Ca(OH)2
' 44 100 44 44

i. e. ymg of C02 (from TG) = 1.6818 y Ca(OH)2-

The measured values of the carbonated lime from the TG are shown

in Table 9.1 and compare very well with values obtained by gas

evolution method.
 233.

9.6 Chapter Summary

a) Carbonated lime measured by carbon dioxide evolution is

usually less than that obtained by TG measurements in the soil-

lime systems.
b) Because of the colour of red tropical soils, the ASIM

method for determining free lime in Portland cement and clinkers

cannot be applied to determine the unreacted lime in the soil-

lime systems. On the other hand, the solvent variation method

proposed by Presseler et al (115) can be used. In this method, a

solvent volume - soil weight ratio of 500 is usually enough to

extract all the unreacted lime.

c) The accumulation of unreacted soil at the expense of

reaction products is not desirable for increase in strength.

However, an amount of unreacted material is required to overcome

cracking of the soil-lime to some extent and to give optimal

working conditions in practice.

d) Strength does not depend on the overall porosity but

probably depends on the microstructure of the soil-lime system.

The microstructure can be represented by pore structure which in

turn can be approximated by the average pore size. The average

pore size depends on the type and volume of the reaction products.

e) The unreacted material consists of goethite, kaolinite,

quartz and haematite. Quartz and haematite do not show on the TG

which offers a potential in quantifying the unreacted kaolinite

and goethite.
f) Unreacted lime is difficult to quantify from TG results

because it is overlapped by the dehydroxylation''. of kaolinite

whereas, carbonated lime can be quantified with reasonable accuracy.

 234..

r ApTm 1n

CONCLUSIONS

The following general conclusions regarding the reactions

that take place, the reaction products formed and their influences

on strength of red tropical soil-lime systems emerge by collect-

ively considering the Chapter summaries.

When lime is mixed with the soils in presence of water,

rapid initial reaction takes place between the non-crystalline

constituents of the soils and lime. This reaction which could

be responsible for amelioration is'followed by a slower reaction

between lime and the crystalline and non crystalline clay

minerals; the less crystalline the clay, the more likely it is to

react. In totality, the reaction is mainly a diffusion controlled

process and most of the reaction takes place within seven days.

Depending on the nature of the amorphous constituents of the

soil, the rapid and slower reactions produce cementitIous reaction

products of various types and the strength of the soil-lime

system is increased with increasing crystallinity of the reaction

products. The cementing abilities of the reaction products-are

sensitive to varying degrees-to ionic substitution within the

lattice of the reaction products. Strengths of the systems

depend mainly on the type and volume of the reaction products.

A well graded red tropical soil with excess of fines seems to
be ideal for lime treatment.

In addition to the broad conclusions given above, the

following specific conclusions drawn from the evidences
 235.

presented in this thesis are given in detail:

1. Red tropical soils are pozzolanas and hence react with lime.

Their activity is associated mainly with their clay size

fraction.

2. Nbst of the reactions that take place occur during the

initial period of five to seven days. This implies that most

of the reaction products are generated within this period;

therefore the gain in strength beyond this period is due

mostly to the increased degree of crystallinity of the

reaction products formed during the initial five to seven days

period. In practice therefore, new methods of curing which

accelerate the crystallinity of the reaction products are

required to bring about rapid strength in the soil-lime

Systems.

3. While some components of the clay size fraction could react

through dissolution of the minerals, the mechanism of reaction

between the soils and lime is mainly a diffusion controlled

process which can be expressed by Jander's diffusion equation

for solid state reactions. The reactions could be multiple

i. e. where more than one mineral in the clay size fraction

react with lime to form two or more different, distinct

products or consecutive, in which case one reaction product

initially formed changes from one form to the other.

4. Based on the quantity of the active components and the rate

constant of reaction obtained from Jander's equation, an

activity index which is statistically related to the uncon

fined compressive strengths of lime treated soils can be

obtained. A highly significant statistical linear correlation

 236.

links the empirical reactivity parameter of Thompson and the

activity index. This finding is of significance since it

allows the use of either of the two values for assessing'the

pozzolanic properties of red tropical soils. The advantages

of the activity index are that it is a parameter based on a'

theoretical model and it can be obtained faster than the

reactivity value which requires not only time but great input

of labour.

S. The cementitious reaction products formed by the reactions

vary from calcium silicates of the tobermorite-like type

through calcium alumino-silicates to calcium alumino-ferri-

silicates.

6. X-ray diffraction is a suitable technique for identifying the

reaction products. Identification is greatly aided by the

use of a semi-automated computer search program which helps

in eliminating the many closely related calcium silicates,

alumino-silicates and alumino-ferri-silicates that can be

formed when the soils are treated with lime.

7. Iron substitutes into the lattice of the silicates and

alumino-silicates. The strengths of the silicate forming

soils are more sensitive to ionic substitution than those of

the soils that form alumino-silicates especially at small

volumes of the reaction products. This results in the

calcium alumino-silicates and alumino-ferri-silicates impart-

ting more strength than calcium silicates at low volumes.

A
8. The strengths depend on the type and volume of the reaction

products. Where silicates and alumino-silicates are formed,

 237.

the strength can be expressed as a cubic polynomial of the

volume of reaction products whereas a quadratic relation

gives the best fit for strength vs volume of reaction products

for alumino-ferri-silicate forming soils.

9. Nhile some of the reaction products have discrete morphology

after about two years curing at 220C, others still remain in

amorphous states.

10. Electron microscopy study shows that the calcium silicate

hydrates formed by some of the soils have dense and close

textured massive structures while the alumino-silicates and

alumino-ferri-silicates usually occur with relatively large

voids in the microstructure.

11. Thermogravimetry offers a quick, simple and neat method for

the determination of carbonated lime in the systems while

unreacted lime can be accurately determined by extraction

methods that provide for the analysis of clear filtrates.

12. Increase in the quantity of unreacted material at the expense

of the reaction products is not desirable. On the other hand,

an amount of unreacted material may be necessary for

requirements other than high strength.

13. Unlike other cemented systems, strength is not directly

related to overall porosity in the red tropical soil-lime

systems studied.
 238.

CHAPTER' 11

FORFUTURE
SUGGESTIONS 'RESEARCH

This thesis is a pioneer research in this field of red

tropical soils and as such no attempt has been made to solve some

of the problems which were identified'and recognised during the

study. Results obtained in this work can be improved upon if the

problems are resolved. It is therefore suggested that investi-

gations should be made into the following: -

1. The improvement of the activity index proposed in this work.

This can be done by measuring the effect of the varying volume

of reaction products with-time on the determination of the

rate constant of reaction. The reactions in the soil-lime

system can be arrested at predetermined times using any

suitable technique and the volume of the reaction product

measured. Using these measurements and the corresponding

fractions of lime reacted, the rate constant of reaction can

be determined.

2. Any method that can overcome the izse of standard minerals in

quantifying the reaction products will be invaluable. The use

of thermogravimetryin a controlled atmosphere is promising.

This is worthy of investigation as it seems useful in quanti-

fication and identification of the reaction products.

3. The use of TG to determine the reacted and unreacted consti-

tuents of the soil. This should be investigated for it offers

an indirect method of measuring some of the reactive

 239.

constituents in the clay size fraction. For example, if the

reacted kaolinite can be*determined from TG and the amount of

calcium silicate and/or alumino silicate formed is quantified,

it may be possible to quantify the amorphous silica and

alumina. The advantages of such in-situ methods over selec-

tive extraction methods are obvious.

4. Separation of the overlapping peaks in the DTG.
This will enable the quantification of unreacted lime in the

soil-lime mixture thus offering an alternative method of

measuring the fraction of lime reacted in order to determine

the activity" index'.

S. The classification of the reaction products into different

types should be improved*. The groupings should contain those

soils that form known' quantities' of specific. reaction products

or known' fractions of the different phases in the mixture that

forms a group. This will help to"improve the strength -

volume of reaction products relationships which will lead to

better equations for forecasting purposes.

6. There is room for improving the semi-automated computer

search program. The output can be'improved to provide such

information as the name and chemical formula of the ASTM

standard minerals in the data base, an abstract showing the

standard minerals with.. high "percentage of strong peaks

present" in their descending order etc.

7. Porosimetric investigation of the soi. 1-lime systems is

recommended. This can be'linked to strength and moisture

content thus enabling more latitude in the soil-lime mix

design using different criteria applicable for different

uses of the soil-lime systems.

 240.

REFERENCES

1. Sivarajasingham, S., Alexander, L. T., Cady, J. G. and Cline, M. G.

'Laterite", Advanced Agronomy, New York, Vol. 14, pp. 1- 60,
1962.

2. Buchanan, F. "A journey from Madras through the countries of

Mysore, Kanara and Malbar, " Vol. 2, pp. 436 - 461, London, 1807.

3. Nanda, R. L. "Laterite as a highway construction material, "

Indian Road Congress Journal, Vol. 36, pp. 371 -389,1975.
e
4. Townsend, F. C., Manke, P. G. and Parcher, J. V. "The influence of
sesquioxides on lateritic soil properties, " Highway Research
Record, No. 374, pp. 80 -91,1971.

S. Townsend, F. C., Manke, P. G. and Parcher, J. V. "Effect of

remoulding on properties of lateritic soils, " Highway Research
Record, No. 284,1969.

6. Bawa; K. S. "Laterite soils and their engineering characteristics, "

Journal Soil Mechanics and Foundation Division, A. S. C. E.,
Vol. 83, Nov. 1957.

7. McFarlane, M. J. Laterite and Landscape, London, Academic Press,

1977.

8. De Graft-Johnson. "Engineering Characteristics: General-Report",

7th Int. Conf. S. M. F. E., Speciality Session on Engineering
Properties of Lateritic Soils, Mexico, Vol. 1, pp. 1- 43,1969.

9. Lyon Associates, Inc. U. S. A. "Laterite and Lateritic Soils and

Other Problem Soils of Africa; An Engineering Study for Agency
for International'Developnent, " Report AID/CSd - 2164, June 1971.

10. Lyon Associates, Inc. U. S. A. "Laterite, Lateritic Soils and

Expansive Soils of South America; An Engineering Study for
Agency for International Developnent, 1977.

11. Faniran, A. 'Mghemite in the Sydney duricrusts, " American

Mineralogist, Vol. 55, pp. 925 - 933,1970.

12. Kitagawa, Y., Kyuma, K. and Kawaguchi, K. "Clay mineral compo-

sition of some volcanogenous soils in Indonesia and the
Philippines, " Soil Science and Plant Nutrition, Vol. 19, No. 3,
pp. 147 - 159,1973.

13. Yoshinaga, N. and Aomine, S. "Allophane in some ando soils, "

Soil Science and Plant Nutrition, Vol. 8, No. 2, pp. 6- 13,
1962.
 241.

14. Yoshinaga, N. "Chemical composition and some thermal data. of

eighteen allophanes from ando soils and weathered pumices, "
Soil Science and Plant Nutrition, Vol. 12, No. 2, pp. 47 - 54,
1966.

15. Kitagawa, Y. "A short discussion on the chemical composition

of allophanes based on the data of Yoshinaga in 1966, " Soil
Science and Plant Nutrition, Vol. 19, No. 4, pp. 321 - 324,
1973.

16. De Kimpe and Fripat, J. J. "Kaolinite crystallization from

H-exchanged zeolites, " American Mineralogist, Vol. 53, pp. 218,
1968.

17. Thomas, M. F. Focal Problems in Geography: Tropical Geomorpho-

logy, London, Macmillan Press, 1974.

18. De Graft-Johnson, J. W.S., Bhatia, H. S. and Yeboa, S. L.

"Influence of geology and physical properties on strength
characteristics of lateritic gravels for road pavements, "
Highway Research Record, No. 405, pp. 87 - 104,1972.

19. Harder, H. "Clay mineral formation under lateritic weathering

conditions, " Clay Minerals, Vol. 12, pp. 281 - 287,1977.

20. Martin, -F. J. and Doyne; H. C. "1: Laterite and lateritic soils
in Sierra Leone, " Journal of Agricultrual Science, Vol 17,
pp. 530 - 547,1927.

21. Dury, G. H. "Rational descriptive classification of duricrusts, "

Earth Science'Journal, Vol. 3, No. 2, pp. 77 - 86,1969.

22. D'Hoore, J. L. Soil Map of Africa and Explanatory Monograph,

Conununication for Technical Co-operation in Africa, Publication
93,1964.

23. Moh, Z. C. and Mashar, F. M. "The effects of methods of prepara-

tion on the index properties of lateritic soils, " Speciality
Session on the Engineering Properties of Lateritic Soils,
Proc. 7th Int. Conf. S. M. F. E., Mexico, Vol. 1, pp. 23 - 35,
1969.

24. Terzaghi, K. "Design'and performance of the Sasumuadam,"

Proc. Institution of Civil Engineers, Vol. 9, pp. 369 - 394,
1958.

25. Clare, K. E. and O'Reilly, M. P. "Road construction over tropical

red clays, " Conf. on Civil Engineering Problems Overseas:
Institution of Civil Engineers, Paper No. 12, pp. 243 - 256,
1960.

26. Newill, D. "A laboratory investigation of two red clays from

Kenya, " Geotechnique, Vol. 11, No. 4, pp. 302 - 318,1961.
 242.

27. Gidigasu, M. D. and Yeboa, S. L. "Significance of pretesting

preparations in evaluating index properties of laterite
materials, " Highway Research Record, No. 405, pp. 105 - 114,
1972.

28. Wesley, L. D. "Some basic engineering properties of halloysite

and allophane clays in Java, Indonesia, " Geotechnique, Vol. 23,
No. 4, pp. 471 - 494,1973.

29. Gidigasu, M. D. "Parameters for classification of fine grained

laterite soils of Ghana, " Highway Research Record, No. 374,
pp. 57 - 76,1971.

30. Schofield, A. N. "Lime stabilization- of nodular clayey pea-

laterite in Nyasaland, " Dept. of Science and Industrial Research
(U. K. ), Road Research Laboratory, Overseas Bulletin, No. 3,
pp. 1- 15,1957.

31. Quinones, P. J. "Compaction characteristics of tropically

weathered soils, " Ph. D. thesis, University of Illonis, 1963.

32. Baldovin, G. "The shear strength of lateritic soils, " Proc.

Speciality. Session on Engineering Properties of Lateritic Soils.,.
7th ?nt. Conf. S. M. F. E., Mexico, Vol. 1, pp. 129 - 142,1969.

33. Sherwood, P. T. "The properties of cement stabilized materials, "

Road Research Laboratory, Report LR 205,1968.

34. Fieldes, M. "Clay mineralogy of New Zealand Soils: 11 -

Allophane and related colloidal materials, ". New Zealand Journal
of Science and Technology, Vol. 37, pp. 337 - 350,1955.

35. Matyas, E. L. "Some engineering properties of Sasumua Clay, "

Proceedings Speciality Session on the Engineering Properties of
Lateritic Soils, 7th Int. Conf. S. M. F. E., Mexico, Vol. 1,1969.

36. Hirashima, K. B. "Highway experience with thixotropic volcanic

clay, " Proceedings Highway Research Board, Washington, Vol. 28,
pp. 481 - 496,1948-

37. Novais-Ferreira, H. and Correia, J. A. "The hardness of

lateritic concretions and its influence in the performance of
soil mechanics tests, " Proceedings 6th Int. Conf. S.M. F. E.,
Montreal, Vol. 1, pp. 82 - 86,1965.

38. Brand, E. W. and Hongsnoi, M. "Effects of method of preparation

on compaction and strength characteristics of lateritic soils, "
Proceedings Speciality Session on Engineering Properties of
Lateritic Soils, 7th Int. Conf. S. M. F. E., Mexico, Vol. 1,1969.

39. Malomo, S. S. "Me nature and engineering properties of some red

soils from N. E. Brazil, " Ph. D. thesis, Dept. of Civil
Engineering, Leeds University, 1977.
 243.

40. Diamond, S. and Kinter, E. B. Mechanisms of soil-lime

stabilization: an interpretative review, " Highway Research
Record, No. 92, pp. 83 - 102,1965.

41. Stocker, P. T. "Diffusion and diffuse cementation in lime and

cement stabilized clayey soils - summary of an interpretative
review, " Australia Road Research, Vol. 3, No. 9, p.. 9,1969.

42. Transportation and Road Research Board, "State of the art;.

Lime stabilization, " Transportation Research Circular, No. 180,
pp. 1- 12,1976.

43. Hilt, H. G. and Davidson, D. T. "Lime fixation in clayey soils, "

Highway Research Board, Bulletin No. 262, pp. 20 - 32,1960.

44. Ho, C. and Handy, R. L. "Characteristics of lime retention by

mL)ntmorrillonitic clays, " Highway Research Record, No. 29,
pp. 55 - 69,1963.

45. Stocker,, P. T. "Diffusion and diffuse cementation in lime and

cement stabilized clayey soils, " Australia Road Research Board,
Special Report, No. 8,1972.

46. Jackson, M. L. "Aluminium bonding in soils: a unifying principle

in soil science, " Soil Science Society of America, Vol. 27,
p. 1,1963.
47. Diamond, 'S and Kinter, E. B. "Adsorption of calcium hydroxide
by montmorillonite and kaolinite, " Journal of Colloid and
Interface Science, Vol. 22, pp. 240 - 249,1966.

48. Kelley, W.P. and Page, J. B. "Criteria for the identification

of the constituents of soil colloids, " Proceedings Soil Science
Society of America, Vol. 7, pp. 175 - 181,1943.

49. Tanada, T. "Certain properties of the inorganic colloid

fraction of Hawaiian soils, " Journal Soil Science, Vol. 2,
pp. 83 - 96,1951.

50. Krishna Murti, G. S. R., Sarma, V. A. K. and Rengasamy, P.

"Amorphous ferri-alumino silicates in some tropical ferruginous
soils, " Clay Minerals, Vol. 11, pp. 137 - 145,1976.
51. Fieldes, M., Swindale, L. D. and Richardson, J. P. "Relation of
colloidal hydrous oxides to the high cation exchange capacities
of some tropical soils of the Crook Islands, " Soil Science,
Vol. 74, pp. 197 - 205,1952.

52. Birrell, K. S. and Gradwell, M. "Ion-exchange in some soils

containing amorphous mineral constituents, " Journal of Soil
Science, Vol. 7, No. 1, pp. 103 - 143,1956.
 244.

53. Van Olphen, H. 'Meetings:. Amorphous clay materials, " Science,

Vol. 171, pp. 91 - 92,1971.

54. Aomine, S. and Jackson, M. L. "Allophane determination in ando

soils by cation exchange capacity delta value, " Soil Science
Society of America, Proceedings, Vol. 23, pp. 210 - 214,1959.

55. Diamond, S. "Rapid reaction of lime with hydrous alumina, "

Nature, Vol. 204, pp. 183,1964.

56. De Carvalho, J. B. Q. "The soil properties affecting lime

stabilization of tropical red soils from N. E. Brasil, " Ph. D.
thesis, Department of Civil Engineering, Leeds University, 1979.

57. Thompson, M. R. "Lime reactivity of Illinois soils, " Proceedings

American Society of Civil Engineers, S. M. F. E. Division, Paper
SM 5 (4907), pp. 67 - 92,1966.

58. Fades, J. L., Nicholas, F. P. Jr. and Grim, R. E. "Formation of

new minerals with lime stabilization as proven by field experi-
ments in Virginia, " Highway Research Board, Bulletin No. 335,
pp. 31 - 39,1963.
59. Fades, J. L. and Grim, R. E. "Reaction of hydrated lime with pure
clay minerals in soil stabilization, " Highway Research Board,
Bulletin No. 262, pp. 51 - 63,1960.

60. Marion, G.M., Hendricks,, D. M., Dutt, G. R. and Fuller, W.H.

"Aluminium and silica solubility in soils, " Soil Science,
Vol. 121, No. 2, pp. 76 - 85,1976.

61. Willoughby, D. R., Gross, K. A., Ingles,; O. G., S. R. and

Spiers, V. M. "The identification _Silva,
of reaction products in alkali-
stabilized clays by electron microscopy, x-ray and electron
diffraction, Australia Road Research Board, Proceedings, Vol. 4,
Part 2, Paper No. 430, pp. 1386 - 1407,1968.

62. Barret, par P., Menetrier, B. and Cottin, B. "Study of silica-

lime solution reaction, " Cement and Concrete Research, Vol. 7,
pp. 61 - 68,1977.

63. Dambor, J. "Relation between phase constitution, overall

porosity and strength of hardened lime-pozzolana pastes, "
Magazine of Concret Research, Vol. 15, No. 45, pp. 131 - 142,
1963.

64. Glenn, G. R. and Handy, R. L. "Lime-clay mineral reaction

products, " Highway Research Record, No. 29, pp. 70 - 82,1963.

65. Ormsby, W.C. and Kinter, E. B. "Strength development and

reaction products in lime-montmorillonite-water systems, "
Public Roads, Vol. 37, No. 4, pp. 136 - 148,1973.
 245.

66. Diamond, S., White, J. L. and Dolch, W.L. "Transformation of

clay minerals by calcium hydroxide attack, " Proceedings 12th
National Conference on Clay and Clay Minerals, Vol. 12,
pp. 359 - 379,1964.

67. Ruff, C. G. and Hb, C. "Time-temperature strength-reaction

product relationships. in lime-bentonite-water mixtures, "
Highway Research Record, Vol. 139, pp. 42 - 60,1966.

68. Ariizumi, A., Maki, T. and Oba, M. "On stabilization of Kanto.

loam by means of lime gypsum admixtures, " Annual Report of Roads,
Japan Roads Association, Tokoyo, pp. 22 - 31,1962.

69. Benton, E. J. "Cement-pozzolana reactions, " Highway Research

Board, Bulletin No. 239, pp. 56 - 65,1960.

70. Harty, J. R. and Thompson, M. R. "Lime reactivity of tropical and

sub tropical soils, " Highway Research Record, No. 442,
pp. 102 - 112,1973.

71. British Standards Institution: BS 1377, Methods of Testing Soils

for Civil Engineering Purposes, London, The Institution, 1975.

72. Jackson, M. L., Soil Chemical Analysis - Advanced Course,

Published by author, Department of Soils, University of
Wisconsin, Madison 6, pp. 233 - 237,1956.

73. Cabrera, J. G. and De Carvalho, J. B. Q. "The crystallinity of

kaolinite in red tropical soils, " Presented for publication to
Soil Science, 1979.

74. Fratini, N. "Lime content in cement mixtures: 11 evaluation of

pozzolana cements, " Annalytical Chemistry (Rome), Vol. 39,
p. 41,1949.

75. Rio, A. "Ion-exchange activity of natural pozzolanas, "

Annalytical Chemistry (Rome), Vol. '42, p. 526,1952.

76. Mackenzie, J. D. and Banerjee, R. K. "Formation kinetics of

Portland cement clinker phases: 1- Tricalcium aluminate, "
Transcations and Journal of British Ceramic Society, Vol. 77,
pp. 88 - 92,1978.
77. Taylor, H. F. W. "Hydrate calcium silicates. Part 1-Compound
formation at ordinary temperatures, " Journal of Chemical Society,
Vol. 726, pp. 3682 - 3690,1950.

78. Greenberg,. S. A. "Reaction between silica and calcium hydroxide

solutions. 1, Kinetics in the temperature range 300 to 85°C, "
Journal of Physical Chemistry, Vol. 65i, pp. 12 - 16,1961.
 246.

79. Raask, E. and Bhasker, M. C. "P6zzolanic activity of pulverised

fuel ash, " Cement and Concrete Research, Vol. 5, No. 4,
pp. 363 - 376,1975.
80. Jander, W. "Reactions in solid-state at high temperatures: 1,
Rate of reaction for an endothermic reaction, " Ceramic Abstracts,
Vol. 7, No. 9, pp. 644,1928.

81. Ginstling, A. M. and Brounshtein, B. I. "Concerning the diffusion

kinetics of reaction in spherical particles, " Journal Applied
Chemistry, U. S. S. R., Vol. 23, pp. 1327 - 1338,1950.

82. Carter, R. E. "Kinetic model for solid state reactions, "

Journal Chemical Physics, Vol. 34, No. 6, pp. 2010 - 2015,1961.

83. Hansen, W.C. "Basic chemistry of reactions of aggregates in

Portland cement concrete, " Journal of Materials, Vol. 2,
pp. 408 - 431,1967.

84. Fripiat, J. J. "Surface properties of alumino-silicates, "

Proceedings 12th National Conference on Clays and Clay Minerals,
pp. 327,1964.

85. De Mumbrum, L. E. and Chester, G. "Isolation and characterization

of some soil allphanes, " Soil Science'Society of America,
Proceedings, pp. 355 - 359,1964.

86. Clare, K. E. and Mohan, D. "An examination of five soils and an

aggregate from Nigeria, " Road Research Laboratory, Road Note,
No. 882,1947.

87. Lea, F. M., The Chemistry of Cement and Concrete, Revised Edition
of Lea and Desch, Arnold, p.. 569,1970.

88. Longuet, P. "The use of thermogravimetric measurements in

cement chemistry, " Proceedings Sth International Symposium on
the Chemistry of Cement, Tokyo, Part 1,1968.

89. Steopoe, A. "Investigation concerning reactivity of Roumanian

trasses in trass-lime and trass-cement mortars, " International
Association for Testing Materials, Zurich, Vol. 1, p. 918,1931.

90. Gray, M.N. "Soil stabilization with lime and secondary chemical
additives, " Ph. D. thesis, Department of Civil Engineering,
Leeds University, U. K., 1972.

91. Jambor, J. Institute of Construction and Architecture of the

Slovak Academy of Sciences, Bratrslava, Czechoslovakia. In
private communication, July 7,1977.
 247,

92. Tuncer, E. R., Paulson, S.'K., Demirel, T. and Lohnes, R. A.

"A method for quantitative determination of soil minerals by
x-ray diffraction, " Transportation Research Record, Vol. 560,
pp. 75 - 80,1976.

93. Kalousek, G.L. and Prebus, A. F. "Crystal chemistry'of hydrous

calcium silicates III, morphology and other properties of
tobermorite and related phases, " Journal of the American
Ceramics Society, Vol. 41, No. 4, pp. 124 - 133,1958.

94. Powers, T. C. "Structure and physical properties of hardened

Portland cement pastes, " Journal of the American Ceramic Society,
Vol. 41, No. 1, pp. 1-6,1958.

95. Powers, T. C. and Brownyard, T. C. "Studies of the physical

properties of hardened Portland cement paste, " Journal of the
American Concrete Institute, Proceedings, Vol. 18, Nos.. 2 - 8,
1947.

96. Schiller, K. "Strength of highly brittle porous materials, "

Nature, Vol. 180, pp. 862 - 863,1957.

97. Jambor, J. "Written discussion, " Proceedings 5th International

Symposiumon the Chemistry of Cement, Tokyo, Vol. 111,
pp. 541 - 557,1968.

98. Alexander, K. M., Taplin, J. H. and Wardlaw, J. "Correlation of

strength and hydration with composition of Portland cement, "
Proceedings Sth International Symposium on Chemistry of Cement,
Tokyo, Vol. 111, pp. 152 -166,1968.

99. Kalousek, G. L. "High temperature curing-of concrete under high

pressure, " Proceedings 5th International Symposium on Chemistry
of Cement, Tokyo, Vol. 111, pp. 523 - 540,1968.

100. Diamond, S. "Tobermorite and tobermorite-like calcium silicate

hydrates: their properties and relationships to clay minerals, "
Ph.D. thesis, Purdue University, Lafayette, 1963.

101. Diamond, S., White, J. L. and Dolch, W.L. "Effects of is. omorphous
substitution in hydrothermally synthesized tobermorite, " The
American Mineralogist, Vol. 51, pp. 388 - 401,1966.

102. Copeland, L. E., Bordor, E., Chang, T. N. and Weise, C. H.

'Reactions of tobermorite gel with aluininates, ferrites, and
sulphates, " Journal of Portland Cement Associations, Research -
and Development Laboratory, Vol. 9, pp. 61 - 67,1967.

103. Taylor, H. F. N. The Chemistry of Cement Hydration, Progress in

Ceramic Science, Pergamon Press, New York, Vol. 1, pp. 89 145,
-
1961.
 248.

104. Johnson, G.G. Jnr. and Vand, V. "A computerized powder diffrac-
tion identification system, " Industrial and Engineering Chemistry,
Vol. 59, No. 8, pp. 18 - 31,1967.

105. Hughes, T. E. "Experience in the use of the AS"IMdiffraction

search program, " Lecture to the Crystallography Group of the
Institute of Physics at the University of Hull, 6th April, 1972.
106. O'Connor, B. H. and Bagliani, F. "A semi-automated system for
identifying crystalline materials with powder diffraction data, "
Journal of Applied Crystallography, Vol. 9, pp. 419 - 423,1976.

107. Abo-El-Fnein, S. A., Gabr, N. A. and Mikhail, R. Sh. 'Morphology

and microstructure of autoclaved clinker and slag-lime pastes in
presence and absence of silica sand, " Cement and Concrete
Research, Vol. 7, No. 3, pp. 231 - 238; No. 4, pp. 363 368,
-
1977.

108. Jambor, J. "Influence of phase: composition of hardened binder

pastes on its pore structure and strength, " Proceedings of the
International Symposium on Ibre Structure and Properties of
Materials, Prague, Part 11, pp. D75 - D96,1973.

109. Rogers, D. E. and Aldridge, L. P. "Hydrates of calcium ferrates

and calcium alumino ferrites, " Cement and Concrete Research,
Vol. 7, No. 4, pp. 399 - 410,1977.

110. Uchikawa, H. and Uchida, S. "The hydration of llCa0 . 7A1203

CaF2 at 20öC, Cement and Concrete Research, Vol. 2, No. 6,
pp. 681 - 695,1972.

111. Roy, D. M., Gouda, G. R. and Bobrowsky, A. "Very high strength

cement pastes prepared by hot pressing and other high pressure
techniques, " Cement and Concrete Research, Vol. 2, No. 3,
pp. 349 - 366,1972.

112. Shapiro, L. and Brannock, W.W. 'Rapid determination of

carbon dioxide in silicate rocks, " Analytical Chemistry, Vol. 27,
No. 11, pp. 1796 - 1797,1955.

113. American Society for Testing and Materials, Annual Book of ASTM
Standards, Chemical analysis of hydraulic cement, ASTMC 114 77,
-
"Free Calcium Oxide in Portland Cement and Clinker, " ASTM
part 13, pp. 110 - 112,1977.
114. Pressler, E. E., Brunauer, S. and Kantro, D. L. "Investigation of
the Franke method of determining free calcium hydroxide and
free calcim oxide, " Analytical Chemistry, Vol. 28, No. 5,
pp. 896 - 902,1956.

115. Pressler, E. E., Brunauer, S., Kantro, D. L. and Weise, C. H.

"Determination of the free calcium hydroxide contents of
hydrated Portland cements and calcium silicates, " Analytical
Chemistry, Vol. 33, No. 7, pp. 877 - 882,1961.
 249.

116. Hoff, G. C. "Porosity-strength considerations for cellular

concrete, " Cement and-Concrete Research, Vol. 2, pp. 91 - 100,
1972.

117. Mackenzie, R. C. "Highways and byways in thermal analysis, "

Analyst, Vol. 99, pp. 900 - 912,1974.

118. Wendlant, W.W. Thermal Methods of Analysis, 2nd Edition,

J. Wiley, New York, 1974.

119. Schnitzer, M., Wright, J. R. and Hoffman, I. "Use of the

thermobalance in the analysis of soils and clays, " Analytical
Chemistry, Vol. 31, No. 3, pp. 440 - 444,1959.

120. Biffen, F. M. "Determination of free lime and carbonate in

calcium silicate hydrates by*thermobalance, " Analytical
Chemistry, Vol. 28, pp. 1133 - 1136,1956.

121. Ringsholt, T. "Short guidance on the selection of suitable

lateritic soils for production of LATOREX," Building Materials
Laboratory - Technical Information, The Technical University of
Denmark, 1975.
 250.
"

i.

APPENDIX A1

X-ray diffraction traces for some of the soils

showing the lime treated and natural soils.

Shaded areas represent positions of new peaks.

 251.

=Y

t.7

_
ýV V.
iIi.
36 34 32 30 28 26 2a
ation 29 16 14 12 10
MöK4 rad --

l9

Mo Km radiation ! 29
,
 I

252,

r w
= S

se
N

I
Y
2
Ill
Y

Mo,kd radicNor. ý-- 29 f° :i. 1Z 10 86

36 34 31 30 18 26 24 j` 1
22 20 t8 16
Mo K.cradiatton 4-----28' 12-'_ 10
 1

253.

Yv s
Z

YL
Y1M

36 34 32 30 28 26 24
Mo,Ka rodintfion_ 2 e" 10 e64

MajCecrodiation -- 29'
 254.

v N
x
u
t7 Y Til

vv

22 20 18 16 10 6
t o, Ks radiation 28'

JiL
w r t7
Z

Y tm

wT

jY

36 34 32 30 28 26 24 22 20 '8 16 14 12 10 6 to
Mo, Ka radiation +-29*
36 34 32 30 28 26 Zc 20 ,euu
Mo Kdrudintion 4-28"
 256.

PREHHITE
IC. AS%H)

ýi
EPIDOTE
(Coe(A(,%Syge0H)

OKENITE
(CS1H1)

t -+ -+ -4---t- 1 F- -i - -l Fi1-"i "--t 4

-"' "f+ .i-` t'.. t1
34 32 30 28 26 24 22 20 18 16 114 12 10
36 . 864
f- 1 .'

Mo. KK radiation
 257.

APPENDIX A2

ASIM powder x-ray data for calciun salts concordant

with new lines observed in the lime treated soils
 258.

r-,
O v
r-I .
CO n t0 r1 M
W Ln \O al. lo 0o
L N N r-1

xM
00
LD co LD
rl
a1 ýp I- Lfl O Ln
N\O "
U r-q N N ri

O Oý
H
5N
M
1 ,. 1 C 00 -t M ,--1 O
u 14 N N N N N
N 14 14 14

00
[ N
"
O 00 D
%. Lf)
00
r- mt O
U) if N O
ý7j c- 00 (71
C/) to MI M M C,; C4 N N r-i 1-4

r1

r4
4)
r
H " 4-º N
to
U 0 p
0
10 r1 ý O Co Co
M N N r--1 e-4

X
0

w
a O
rj v
üf
00 r-ý Ö 00
r-1 %
"
N M N N r-1 1- r-1
U

L [O Co
0 0
144,
00 00 ýo
T` " M N r-1 ri 1;
H
v

9
O II
8 co M N ýo Q
r- q [ý M OO Vl
H %0 -1
U) M M N N N 1; 14
C7i

C C 01 O
.Ol N Co r--1 N Co

O QQi R+ Q ýflu N r-1

M
0
M N
CO
N
'ý? N
N
HI
N
O
N
[ý M
r- {
ºý 1{ r-1

Hz
a

ýi rn pI N M ,-1
- rn t%
U ºw-ý O O rn oo
V) M fV N e1 rý

i-i M tý Ln M 00
ýO N ý
H n M aoý ýr O Co
E-4 H N M M N N N r-1 r-4 r-1
W ý--1
 259.

.. N .... ..
ýj
N MO
C)
Cl) '-4

UN N d Mý M

\
ÖO M p C) %0
65' C. N Co
Co ýt 0Ln
?
VM
Ln
M N N N 1-; -l r-1 r-1

.. .. .. ..
n
0
O v
L N. 0
r"1 Co CMO
Üv
Pr) N N r" 1 r-1

"N ý--1
n Q 0 n
%D 00

U pry M N N "-^I r--1

rl

O
"
xb
ýN 1 0 N.
%0 0 . -1
1 M M N
- c - et

r1 \
P4 00N
rli r'i ; C
N. %9 G1 00 Lf1L/1

cýa V Gd UQ'r1 ri ; ; (14 Nc. N' r1 r"1 i1, -1

W
«-a NO
N.... O
to r-i ooh 03 N.
"n OO 0000 0 03
N) M NN N' r-1
v cio- .

Z O
0
Q
1 M r-4,
O 1 ...
OD"rl OHO O (O 10
" r1
o " " "
M N N r1 r-1
U "

R I NM
co
rN-I 000 C) 000
" " . " " " ""
" "a N) N) N i-f r"1 r-1 i- f

p . ... .
HH N
lfý M "t
OMO Ql N
" " "
00
N M N N rI r"i ry

H M O)
ýp N 00 10 '.O 00 Q1

Cý %10f-AN) M K) N 00
r"1
ýo
. .. ... ... ... ...
Lni tl!
O 00 WJ-"
Ln tn
(n C,
% 00
00 r- s
'
,-ý M N N N . -1 r-1
ýW r-1, "-1
 260.

NOl
Mý M rn 01 00 r-ý N.
oo 0 0 0 cc cc
uM OO M M N N r-1 H r""1

M
-
0 '1:
tÖ
[\
04
1-1
%0 C14 m r- M 0
r-4
44 1 01 Ln N O) OO '.O %0 'et
cd
"
"rq cd %D M N N N r-1 r-4 ri r-1 r-i '-i
U V] U r-i

r-. O
M 00 v
1.4 C13 Q N
$-4 N N. 00 e-4 N r-1
"" U) O a OO %0 Lfl
r-4 "o I r-i
M V, M M
`J -2. ýt r-i r-1 e-i ý-1

I- M
Q o
I e, H O0 V '00 ' cl "-4 Ql \0
Üa
C4
qtd* M N N N "4 r-1 r--ý

U)

O
U) 4n
IRT cn 0 -ti- CA
:t t-- Ln
dO
C71
M N N N N r-1
U

U 91

Q- C
wN\ N 0v-I
ýI V to Lne-ý C H N-
N 1 tý qo ýn
r4 r-4 \0 N N r-I N N
U UrH r-1 e-1

Z o

12 0
rl M
0 0
I r-I v-i
v v
1 00 l- . -1 n M N. O
T 00 O CO
ý--1 %9
° U
UH N N r"1 ri r1

C'.
^ rý
O O
O r-I
ý v
H
00 O
j 2R 0 00 kc v
Uff' M N rl rl 1

W %0 N{ '.0 n U) N "-i '.O C71 '.O M

il t N Ql \S LJS N
p, N 0 al t
M M" N
l N N r-1 14 "-1 , "ý rý
C
ý". ý

HH Ln ýt U)
H N \O O 0 \0 i
"H4
ýt Ni M N --1 r1 r-1 r'i

00 O
%. 00 -t tl. 00
HH [ý Q Oý CA 00 u

M M N N N N . -1 ri r-1 ý..ý

lý N O n 00 ýO et
pýý N L N r. l r. l rl
 261.

ý ýQ O
r9 ley
N Q1 O0
l
M Mü N N N e--1r-1
U Ln

^ O
ýO n

`
N Q1 lý
OO
Fn O ao rý
oo M O rn
00 M M N N r-I e-1
UM

r-1
r"ý DO "

ý
Qj o0 ýO

(U ý/ 1 M N r-1 r4
. C; ri
C/)
r1

v
lý M Q 00 In ' Q1 11ý %0 Q
O)
rl O 'O 1r r-I C ' ýO
_. oo
M N N' N . --1 r-4 r-;
V
V

M O N
ýA vrý
o n1 lý ýO ýt uý ýO et
00 ýi N Q ýO et
Ü N N N N
M r1 r-1
A
W
0.4 O

.5 r,
r o n Q
iiý
u Cý ýO N e-i ýO
HO
0
" " " " " "
M N N N N N ý--ý

ý
n
z O
r4
ö
ýý
"
II M
ö'
co O °° O
t r-I CC 000 .
ý Ln" " 4 "
U M N N rl r-1

I
M

s C/ý 2 ý-1
Cý W r o F
ha M ö I v
0) C" . -4 Chi
M M M N N N N r-1 r-1 ý4 r-i

Q H tý
(ý 00 Q ý7 . -I %0 et et
f:
v% to " " "" " " " ."
.ý Q i VY M M NN N ý r.. ý ýý
O
C)

W
H r- 00 Q
ý, ý [w cm rý Ol N r-1 Ol "D ýp
ý} M N N N r-ý r-i r-1

Pr) CC o0 ý0 r-1. M o0 I n 01 ýf
HH OO 00l n v: N H CO if *}

n MI N N N N N r-1 r-1 r-1

W
H
 262.

00 O
cn r-I
Ü v O
Q 0 M
M
Co
Co
Co
p O N
M
M M N N N' ri r'I b
9
fnFJ[-
10
C-i
Lr)
rn ýO ct N /o
i--I O
H l" ri M N N cl; C4 r-, 1
,

a)
r-I
N E
c, 0
O r-% oo
0 " 0
r-I 9
01 tý N- Co
C4
\
r--1
r"I
U `D
00

M N N N fV ri e-4
44
0

V, N v
.
r-f
qmr
00
a Ol ýO
C--ý I Ol 8
Urn Co w ýr -I o, ýo
co
to C'4r-I
N
N
W
W
a
O
I %. -I

r-4 tl-

o
r"
r-I
M
oo
N
O
N
00
r-I
ko
U r-4

rl .
00
o
II
0 .n CO
C
- " rýI
u M N r1 r1

I-A N
O
M t" 0

o M N 1"4 '1 ý"I rI

cri
ý
I
Ln (14
co 0) C*
O
ý, I: t
C)
r--i bC
HH
4ý
M N N N N rl r-1 r-1

ý
Oo N N M [- 01 n ýO it N
I-a co OI ýo v N O O) CO
o Cl)p 4
M N N N N r-4 r-1 " 4-4 -rq
u
r-1

h n M )
h p+ (
ýi 03 oo Ey
4
r M N r -1 ri f- r-1
 263

APPENDIX B. 1

Shapiro and Brannock's Method for Rapid Determination

of Carbon Dioxide in Silicate Rocks

 264.

APPENDIX B. 1

Shapiro and Brannock's 'Method 'for 'Rapid 'Determination

of Carbön 'Dioxide ' in ' Silicate Rocks

a) The carbon dioxide tube consists of a borosilicate glass

test tube, 180 mu diameter x 150 mm long to which they attached a

side arm of borosilicate glass 10 mm outside diameter x 200 mm

long, with a closed end. The side arm was attached at 45 mm from

the bottom of the test tube.

b) They used an electric heater with-. a regulator to heat the

sample and the liquid in the lower portion of the carbon dioxide

tube.

c) The carbon dioxide tube was calibrated as given in

appendix B. Z.

d) To determine the carbon dioxide content of a sample, 1.0 gm

of the sample powder'was transferred to the bottom of the test

tube and procedures'of steps 2-7 of the "Calibration of carbon

dioxide tube" in appendix B. 2 followed.

e) The percentage of carbonate in the sample was estimated

from the scale on the side arm..

After trying three different supernatant liquids viz.

distilled water, 20% sodium chloride solution and motor oil S. A. E.

No. 10, they found that distilled water and 20% sodium chloride

solution were usable but inferior to oil because of the greater

solubility of carbon dioxide in water and the sodium chloride

solution. Boiling in the lower'part of the tube was also easier

to control with oil. 200 mmside arm length was found best for
 265.

the recovery of evolved'gas while a boiling time of at least two

minutes was adequate. Mecuric chloride was used to eliminate the

interference of iron which would-otherwise react with hydrochloric

acid to produce hydrogen.

 266.

APPENDIX B. 2

Calibration of Carbon Dioxide Tube

 267.

APPENDIX B. 2

Calibration 'Of 'Carbon 'Dioxide Tube

1. Weigh 1.03gins of "National Bureau of Standards" standard

,
sample no. 79 and transfer, by means of a dry funnel to the

bottom of the carbon dioxide tube.

2. Add 2 ml. of the mecuric chloride solution and tap the tube

to free entrapped air bubbles.

3. Add oil to the oil-level mark, 110 mmfrom the bottom of

the tube.
4. Tilt the tube*so that the oil completely displaces the air

from the side arm. Then return the tube to a position such that
the main part of the tube'is vertical.
5. Add 2 ml. of 1: 1 hydrochloric acid, and tilt the tube so

that the side arm is vertical to. allow any carbon dioxide

produced to. enter the side ärm.

6. Mount the carbon dioxide tube'on a clamp attached to a

support so that the side arm is vertical, and insert the lower

part of the tube through the hole. in the cover of the heater with

the interface of the aqueous and oil phases just at the level of
the cover (the heater should be preheated so that the temperature

around the lower part of the tube can be maintained at about

185°C). Allow the aqueous phase to-boil for about 2.5 minutes.

7. Remove the tube from the heater and allow tap water to flow

down the outside of the side arm for 15 seconds (the water

temperature should be between 150 and 25°C).

8. Remove from the stream of tap water, hold the tube with the

side arm upright, and mark the position of the meniscus on the
 268.

side arm. This-is the 1% mark.

9. Repeat steps 1 to 8 using 0,0.206,0.412,1.54 and 2.06 gms

of the standard sample to obtain calibration marks equivalent to

0,0.2,0.4,1.5,2.0% carbon dioxide respectively.

10. The tube can now be marked off by interpolation to give

marks for each 0.1% carbon dioxide.

 269.

. ýýý

APPENDIX B. 3

AS1M C114-77 Method for the Determination

of Free Lime in Portland Cement and Clinker

 270.

APPENDDCB. 3

ASIM 'CII4=77 'Met}iöd för 'tlie :Defermiriätion

of Free 'I: inie 'in 'Pertland :Cement 'and 'Clinker

To one gram of the finely ground sample in a 200 or 250 ml

flat bottom flask is added 60 ml of the solvent (glycerol and

alcohol) and a few glass beads to agitate the solution while it

is boiling. A reflux condenser is, attached and the flask and its
contents boiled over a suitable source. of heat (use of open gas

flame for boiling the solvent-sample mixture presents a fire

hazard). The condenser is removed and the soil titrated while

still near boiling. A slight pink colour indicates the end point
of the titration.

The flask and contents. with. the condenser attached are

returned to the hot plate and boiled. The titration and boiling
cycles are repeated and may be as frequently as five minutes but
not exceeding twenty minutes-in the early stages. The final end
point of the titration is when the pink colour persists for about

ten seconds. The free lime present is calculated as equal to

EV x 100 where,

E= lime (CaO) equivalent of the ammoniumacetate solution

in gm/ml (1 ml = 0.005 gm CaO) and

V= ml of ammoniumacetäte solution required by the sample.

 271.

ýr

APPENDIX B. 4

Modified Franke Method for Determination

of Free Lime in Portland Cements

.0
 272.

APPENDIX ' B. 4

Modified 'Fra ike Methöd 'för 'Determinätion

"' of 'Free Lune:.. Ibrtland Cements

Reagents:

a) Acetoacetic ester (ethyl acetöacetate), anhydrous grade.

b) Isobutyl alcohol, anhydrous grade.

c) Thymol blue indicator; chemical name, thymol

sulfonphthalein.
d) 60% perchloric acid.

Solution preparations: -
a) Solvent: 450 ml of ethyl acetoacetate and 3 litres of
isobutyl alcohol.

b) Indicator: 0.1 gm of thymol blue indicator powder

dissolved in 100 ml of isobutyl alcohol.

c) Titration: to obtain approximately 0.2 N perchloric

acid, 21.8 ml of 60% perchloric acid is made up to

1000 ml with isobutyl alcohol.

Standardization of the acid solution:

This follows in principle, the ASIM method of standardizing

animoniumacetate solutions, C114-77 (113).

Equipment:

a) Water-cooled condensers, preferably with standard-taper

inner joints to fit 200 ml Erlenmeyer flasks; and upper

adapter tubes to fit absorption tubes containing soda-

lime and ascarite.

b) Buchner-type borosilicate glass filtration funnel of

 273.

350 ml capacity and F grade porosity.

c) Equipment for vacuum filtration.

Procedure:

Sieve and grind the sample of cement according to the

standard ASIM (113) procedure for free' lime determination.

Measure 70 ml of the prepared'solvent solution (acetoacetic ester

and isobutyl alcohol) and transfer into'a 200 ml Erlenmeyer flask.

Weigh accurately 1 gm of the prepared sample of cement and

transfer into the flask. Adjust the flask in position to a

water-cooled condenser, fitted with an upper adapter tube con-

taining soda-lime and ascarite and reflux at boiling temperature

on a hot plate for 3 hours. Remove the'flask, stopper and cool.

Filter with vacuum on a 350 ml filter funnel, receiving the
filterate in a. second flask. Wash the flask and residue in the

funnel with 50, ml of isobutyl alcohol, using a policeman to guide

,
the flow. Add 10 to 12 drops of the indicator solution to the
filterate and titrate. with-the standard 0.2 N perchloric acid

solution to a distinct reddish tinge.

 274.

APPENDDC B. 5

Solvent Variation and Time Variation Methods for

free lime determination

 275.

APPENDIX *B. 5

Solvent Variation Method

Reagents, solution preparations and equipments as in the

modified Franke method (Appendix B. 4). Procedure is as before

but with the following modifications:

Volumes of solvent used are 50,100,150,200,250 and

300 m1. Six 0.1 gm portions of sample are weighed out. Each

solvent-sample mixture is-refluxed for 30 minutes.

Time Väriatiöri method

As above, but volume of solvent used is always 100 ml.

Six 0.1 gm portions of sample are weighed out. Two solvent-

sample mixtures are refluxed'for 1 hour, two for 1.5 hours and

two for 2 hours.

 276.

APPENDIX B. 6

TG and DTG Curves of some of Lime Treated Soils

 277.
Temperature,
°C
200 400- 600 800
1000
10

10 2

20 4

30 6

40 8

c 50 PH 10
N
O!
. es

E
E_ E
TG u,
60 DTG
""""""" n,
CL.
W2
E 0 U cn,
Ei
''r
.

ö+ " 2ý
10 B
3 3

20

30 6

40 TI

50 --- TG
"""""""" 0 TG
 I

278.

Temperature C
0 0
201 400 600 800 1000
0 1 -- 1-.. f. -1 -ý 0
.....

10 ý

20 4

30 6

40 8

50 10

E c

60 E

Cn SM L'
Is
E =6

'"
' 0
""
0TG O E
Vf ".
"
" O
ýO.

Caw
10 *"0 : ".
E

20--4

30 6
SIA
TG
40 8
........ 0T6

I
 279.

Temperature,
°C
200 600 600 600
00

10

20 6

30 6

40 8

5u 1H

d
C

r TII E
Ln
60 TG
MM
e
a
C:
......... 0TG
c:n
UE
0

-- A s
an
3 10 '"E 23

20 4

30 6

60
SB
8

TG
"""""""" 0 TG
50
 280.

°C
Temperature,
200 400 600 800 1000
0

1o 2

20 4

30
B

40 ßH
JI E

50 --- TG 10 ä
""""""" OTG C75

0E N
N
O

"_ ý-
U
O1
cu
aj
3 10 13

20 4

30 ö

40 8
1II
TG
50 10
........ DIG
 281%
Temperature,
°C
200 400 600 800 1000
0

10 2

20

30 &

to a

50 10
n,
4J

60- USM 11 -
E
E
TG a
""""-"" OTG ä
E
0N
N
c2
0-
L C\ '' .G
D1 D1
cu - A : s cu

10 12
B E

20 4

0
30 6

RC
40 1G 8

50
 282.
Temperature,
°C
0 200 400 600 800 1000
00

10 2

20

1`
30 6

40 8

a,
so 10 d

E
E CT -E
v 60
jr ``
"""""""" DTG
E0 0 .
"y ", """ E
N

H A q-.

.C
_
10 1 ý3n!
3

20 4

30 6

40 SIB

--- TG
"""""""" 0 TG
50
 I

Temperature,°C 283.

0 200 400 - 600 800

Or
1000
T

10 2

30 6

40 8

50 10
a,
"C

E AI
E
60 Ln
TG Ca.

E0
"OE
N "N
H
N

S.
L ýO
Of
0)
E1ö,
3 3

20

30 6

40 8

50 10
All .
TG
""""""""" 0 TG
 I

284.
°C
Temperature,
200 400 600 800 1000 00
00

10

20 a

30 6

40 8

JPM
50 lV N

TG
d E
......... 0TG
E us
CLO
C36
ö,
E
NN

ö 10 22
B

a,
3
3 .: p
20 4

30 6
JPA

Lo 8
TG
""""""""" 0TG

50 10
 285.

°C
Temperature
,
200 400 600 800 1000
0
0
..:
C
2.0 0.2

4.0 0.4
n,
Aý
N
N

d
E 6.0 06
E
CLJ

C36
E
8 'o ö,
8.0 0$
O
N
N
O

w
t
Of
o,
3 10 ý=-
10.0

JPB
T6
120 0TG
""""""""