USF Libraries
USF Digital Collections

Effect of Alkalis and Sulfates on Portland cement systems

MISSING IMAGE

Material Information

Title:
Effect of Alkalis and Sulfates on Portland cement systems
Physical Description:
Book
Language:
English
Creator:
Halaweh, Mahmoud
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla
Publication Date:

Subjects

Subjects / Keywords:
Ettringite
Expansion
Durability
Strength
Alite content
Dissertations, Academic -- Civil Engineering -- Doctoral -- USF   ( lcsh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Summary:
ABSTRACT: The effect of the sulfates and alkalis on the durability of Portland cement systems was investigated through a series of cube and prism mixes. Durability was assessed using expansion of mortar prisms and the compressive strength of mortar cubes. The study covered a large range of both alkali and sulfate contents using 5 different Portland cements. The alkali contents ranged from 0.27 to 3.8%, the sulfate content (as SO3) ranged from 2.54 to 5%. Doping was done using Terra Alba gypsum and potassium hydroxide. In addition to physical measurements, SEM, XRD, chemical analysis and heat of hydration calorimetry were used for further analysis. Mixing, curing and testing were done at room temperature. The results show that sulfate contents up to the levels used in this study, at low alkali contents and ambient temperature curing, did not adversely affect durability of Portland cement mortars up to 360 days. A correlation was established between expansion and ettringite formation. Increasing the alkali content always resulted in loss of compressive strength, and in some cases, excessive expansion. Excessive expansion was only experienced at the 3.8% level. Alkali levels of up to 2% and sulfate levels of 5% did not result in excessive expansion at room temperature-cure up to the ages reported here. The effect of alkali depended on thecement mineralogical composition, especially C3S content. The addition of alkalis seems to impact the nature of the microstructure and the nature of other hydration products. The addition of sulfates seems to counteract the effect of alkalis, especially on the loss of compressive strength. However, these sulfates may result in other problems as they may be available at any time to form ettringite which may, under certain conditions, result in excessive expansion. It was concluded that sulfate levels on the order of 3-3.6%, did not pose any major durability drawbacks under normal curing temperatures and low alkali contents( < 1% Alkali levels above 1% will adversely affect the durability of Portland cement systems.
Thesis:
Dissertation (Ph.D.)--University of South Florida, 2006.
Bibliography:
Includes bibliographical references.
System Details:
System requirements: World Wide Web browser and PDF reader.
System Details:
Mode of access: World Wide Web.
Statement of Responsibility:
by Mahmoud Halaweh.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 197 pages.
General Note:
Includes vita.

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001935453
oclc - 226061678
usfldc doi - E14-SFE0002120
usfldc handle - e14.2120
System ID:
SFS0026438:00001


This item is only available as the following downloads:


Full Text
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001935453
003 fts
005 20080423142247.0
006 m||||e|||d||||||||
007 cr mnu|||uuuuu
008 080423s2006 flu sbm 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0002120
035
(OCoLC)226061678
040
FHM
c FHM
1 100
Halaweh, Mahmoud.
0 245
Effect of Alkalis and Sulfates on Portland cement systems
h [electronic resource] /
by Mahmoud Halaweh.
260
[Tampa, Fla] :
b University of South Florida,
2006.
520
ABSTRACT: The effect of the sulfates and alkalis on the durability of Portland cement systems was investigated through a series of cube and prism mixes. Durability was assessed using expansion of mortar prisms and the compressive strength of mortar cubes. The study covered a large range of both alkali and sulfate contents using 5 different Portland cements. The alkali contents ranged from 0.27 to 3.8%, the sulfate content (as SO3) ranged from 2.54 to 5%. Doping was done using Terra Alba gypsum and potassium hydroxide. In addition to physical measurements, SEM, XRD, chemical analysis and heat of hydration calorimetry were used for further analysis. Mixing, curing and testing were done at room temperature. The results show that sulfate contents up to the levels used in this study, at low alkali contents and ambient temperature curing, did not adversely affect durability of Portland cement mortars up to 360 days. A correlation was established between expansion and ettringite formation. Increasing the alkali content always resulted in loss of compressive strength, and in some cases, excessive expansion. Excessive expansion was only experienced at the 3.8% level. Alkali levels of up to 2% and sulfate levels of 5% did not result in excessive expansion at room temperature-cure up to the ages reported here. The effect of alkali depended on thecement mineralogical composition, especially C3S content. The addition of alkalis seems to impact the nature of the microstructure and the nature of other hydration products. The addition of sulfates seems to counteract the effect of alkalis, especially on the loss of compressive strength. However, these sulfates may result in other problems as they may be available at any time to form ettringite which may, under certain conditions, result in excessive expansion. It was concluded that sulfate levels on the order of 3-3.6%, did not pose any major durability drawbacks under normal curing temperatures and low alkali contents( < 1% Alkali levels above 1% will adversely affect the durability of Portland cement systems.
502
Dissertation (Ph.D.)--University of South Florida, 2006.
504
Includes bibliographical references.
516
Text (Electronic dissertation) in PDF format.
538
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
500
Title from PDF of title page.
Document formatted into pages; contains 197 pages.
Includes vita.
590
Adviser: Abla M. Zayed, Ph.D.
653
Ettringite.
Expansion.
Durability.
Strength.
Alite content.
690
Dissertations, Academic
z USF
x Civil Engineering
Doctoral.
049
FHMM
090
TA145 (ONLINE)
773
t USF Electronic Theses and Dissertations.
4 856
u http://digital.lib.usf.edu/?e14.2120



PAGE 1

Effect of Alkalis and Sulfate on Portland Cement Systems by Mahmoud A. Halaweh A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Civil and Environmental Engineering College of Engineering University of South Florida Major Professor: Abla M. Zayed, Ph.D. Rajan Sen, Ph.D. Ram Pendyala, Ph.D. A. Sunol, Ph.D. Hana Y. Ghorab, Ph.D. Date of Approval: December 8, 2006 Keywords: ettringite, expansion, dur ability, strength, alite content Copyright 2007, Mahmoud A. Halaweh

PAGE 2

DEDICATION This work is dedicated in loving memory to my parents. Wi thout their knowledge, wisdom and guidance I would not have the goals I have, to stri ve and be the best to reach my dreams. This and other achievements I may realize are dedicated to my loving wife and my wonderful family for their support and encouragement.

PAGE 3

ACKNOWLEDGEMENTS The author would like to express his a ppreciation to Dr. Abla M. Zayed, major professor, for her support and guidance throug hout this study. The au thor would like to extend his thanks to the supervisory committee for their help and support. In addition, the author would like to thank Dr. A. Sunnol, Dr. Thomas Pichler and Mr. Anthony Greco for their assistance during the course of the study.

PAGE 4

TABLE OF CONTENTS LIST OF TABLES iv LIST OF FIGURES v LIST OF SYMBOLS AND ABREVIATIONS xv ABSTRACT xvi CHAPTER 1 INTRODUCTION 1 1.1 Background on Sulfate Attack 1 1.2 Expansion Theories 4 1.2.1 Swelling Theory 4 1.2.2 Crystal Growth Theory 5 1.3 Factors Affecting Ettringite Formation and Expansion Potential 6 1.3.1 Effect of Curing Temperature 9 1.4 Research Objective 11 1.4.1 Research Significance 13 1.5 Outline of Dissertation 13 CHAPTER 2 LITERATURE REVIEW 15 2.1 Role of SO 3 15 2.2 Role of C 3 A 19 2.3 Role of C 3 S 21 2.4 Role of Alkalis 22 2.4.1 Effect of Alkalis on Ettringite Formation 22 2.4.2 Effect of Alkalis on Hydration Process and Mechanical Properties 23 2.4.2.1 Effect of Alkalis on the Hydration of Clinker Phases 23 2.4.2.2 Effect of Alkalis on Ions Availability in Solution 25 2.4.2.3 Effect of Alkalis on Mechanical Properties and Microstructure 25 2.4.3 Effect of Alkalis on Expansion Potential 37 CHAPTER 3 EXPERIMENTAL METHODS AND PROCEDURE 39 3.1 Pure Clinker Minerals 39 3.2 X-Ray Diffraction (XRD) 39 3.3 Scanning Electron Microscopy (SEM) 40 3.3.1 Fractured Surfaces Preparation 40 i

PAGE 5

3.3.2 Polished Surfaces Preparation 41 3.4 Cement Analysis 41 3.4.1 Blaine Fineness 41 3.4.2 Oxide Chemical Analysis 42 3.4.3 Mineralogical Composition 42 3.4.3.1 Bogue Equations 42 3.4.3.2 Internal Standard Method (Calibration Curves) 42 3.4.3.3 The Rietveld Method 47 3.5 Chemical Analysis of Hydrating Cement 48 3.6 Porosity Measurements (BET) and Degree of Hydration 49 3.7 Mortar Preparation 49 3.7.1 Mortar Bars 49 3.7.2 Mortar Cubes 49 3.8 Paste Preparation 50 3.9 Heat of Hydration 50 CHAPTER 4 MATERIALS SELECTION AND EXPERIMENTAL PLAN 52 4.1 Materials Used for Doping 52 4.2 Sand 52 4.3 Cement Selection 52 4.4 Experimental Plan and Philosophy 55 CHAPTER 5 ROLE OF SULFATES 63 5.1 Expansion Results 63 5.2 Compressive Strength 67 5.3 XRD Results 70 5.4 Heat of Hydration 77 5.5 Ionic Species Concentrations 79 5.6 Scanning Electron Microscopy (SEM) 82 5.6.1 Fractured Surfaces 82 5.6.2 Polished Surfaces 83 CHAPTER 6 ROLE OF ALKALIS, PART I: EFFECT OF ALKALI SOURCE 84 6.1 Expansion Results 84 6.2 Compressive Strength 86 6.3 XRD Results 87 6.4 Heat of Hydration 93 6.5 Ionic Species Concentrations 95 6.6 Scanning Electron Microscopy (SEM) 101 6.6.1 Fractured Surfaces 101 6.6.2 Polished Surfaces 101 CHAPTER 7 ROLE OF ALKALIS, PART II 103 7.1 Expansion Results 103 7.2 Compressive Strength 108 7.3 XRD Results 112 ii

PAGE 6

7.4 Heat of Hydration 125 7.5 Ionic Species Concentrations 128 7.6 Scanning Electron Microscopy (SEM) 138 7.6.1 Fracture Surfaces 138 7.6.2 Polished Surfaces 142 CHAPTER 8 CEMENTS MH3, MH4 AND ERD07 143 8.1 Expansion Results 143 8.2 Compressive Strength 148 CHAPTER 9 DISCUSSION 153 CHAPTER 10 CONCLUSIONS 165 REFERENCES 166 APPENDICES 174 Appendix A: Cement Dissolution Methods 175 A.1 Salicylic Acid-Methanol Extraction (SAM) 175 A.2 Potassium Hydroxide/Sugar Extraction (KOSH) 175 Appendix B: XRD Scans 176 Appendix C: Schematic Representation of C 3 S Hydration 197 ABOUT THE AUTHOR End Page iii

PAGE 7

iv LIST OF TABLES Table 1: Mix Proportions for As-Received Cements (AR) 50 Table 2: Blaine Fineness 53 Table 3: Oxide Chemical Composition of As-Received Cements (weight %) 53 Table 4: Bogue Mineralogical Content 53 Table 5: Phase Composition of As-Received Cements Based on Calibration Curves 54 Table 6: Phase Composition of As-Received Ce ments (Rietveld Method) 54 Table 7: Mixes 58 Table 8: Major Phases of All Cements 59 Table 9: Mix Proportions for Cement C 61 Table 10: Mix Proportions for Cement E 61 Table 11: Mix Proportions for Cement ERD07 61 Table 12: Mix Proportions for Cement MH3 62 Table 13: Mix Proportions for Cement MH4 62 Table 14: Mortar Flow Values for Cements C and E 128

PAGE 8

v LIST OF FIGURES Figure 1: Calibration Curve for C 3 A 44 Figure 2: Calibration Curve for C 4 AF 44 Figure 3: Calibration Curve for MgO 45 Figure 4: Calibration Curve for Alite 46 Figure 5: Effect of Sulfate Content on Expans ion Behavior for Cement E 64 Figure 6: Effect of Sulfate Content on Expans ion Behavior for Cement C 64 Figure 7: Effect of Sulfate C ontent on Expansion Behavior for Cement ERD07 65 Figure 8: Effect of Sulfate C ontent on Expansion Behavior for Cement MH3 65 Figure 9: Effect of Sulfate C ontent on Expansion Behavior for Cement MH4 66 Figure 10: Effect of Sulfate Conten t on Strength Gain for Cement E 67 Figure 11: Effect of Sulfate Content on Stre ngth Gain for Cement C 68 Figure 12: Effect of Sulfate Content on Stre ngth Gain for Cement ERD07 68 Figure 13: Effect of Sulfate Content on Stre ngth Gain for Cement MH3 69 Figure 14: Effect of Sulfate Conten t on Strength Gain for Cement MH4 69 Figure 15 : XRD of Hydration products @ 24 hours for Cement E 70 Figure 16 : XRD of Hydration Products @ 24 h ours for Cement C 70 Figure 17 : Effect of Sulfate Content on Ettringite Formation in Cement E 72

PAGE 9

vi Figure 18: Effect of Sulfate Content on Ettri ngite Formation in Cement C 72 Figure 19: Effect of Sulfate Content on C 3 S Hydration in Cement E 73 Figure 20: Effect of Sulfate Content on C 3 S Hydration in Cement C 74 Figure 21: Effect of Sulfate Content on C 4 AF Hydration in Cement E 74 Figure 22: Effect of Sulfate Content on C 4 AF Hydration in Cement C 75 Figure 23: Effect of Sulfate Conten t on Formation of Calcium Hydroxide in Cement E 76 Figure 24: Effect of Sulfate Conten t on Formation of Calcium Hydroxide in Cement C 76 Figure 25: Effect of Sulfate Content on Formation of Amorphous Content for Cement E 77 Figure 26: Effect of Sulfate Content on Formation of Amorphous Content for Cement C 77 Figure 27: Effect of Sulfate Content on Heat of Hydration for Cement E 78 Figure 28: Effect of Sulfate Content on Heat of Hydration for Cement C 78 Figure 29: Effect of Sulfate Content on SO 4 -2 Concentration for Cement E 80 Figure 30: Effect of Sulfate Content on SO 4 -2 Concentration for Cement C 80 Figure 31: Effect of Sulfate Content on Calcium ions Concentration for Cement E 81 Figure 32: Effect of Sulfate Content on Calcium ions Concentration for Cement C 81 Figure 33: SEM Micrographs for Cement E 82 Figure 34: SEM Micrographs for Cement C 83 Figure 35: SEM Micrograph on a Polished Section for Cement E, Case AR-AR 83 Figure 36: Effect of Alkali Source on Expansio n Behavior for Cement E 85

PAGE 10

vii Figure 37: Effect of Alkali Source on Expansio n Behavior for Cement C 85 Figure 38: Effect of Alkali Source on Strengt h Gain for Cement E 86 Figure 39: Effect of Alkali Source on Strengt h Gain for Cement C 87 Figure 40: Effect of Alkali S ource on Ettringite Formation for Cement E 88 Figure 41: Effect of Alkali S ource on Ettringite Formation for Cement C 88 Figure 42: Effect of Alkali Source on C 3 S Hydration for Cement E 89 Figure 43: Effect of Alkali Source on C 3 S Hydration for Cement C 89 Figure 44: Effect of Alkali Source on C 4 AF Hydration for Cement E 90 Figure 45: Effect of Alkali Source on C 4 AF Hydration for Cement C 91 Figure 46: Effect of Alkali Source on Formation of Calcium Hydroxide for Cement E 91 Figure 47: Effect of Alkali Source on Formation of Calcium Hydroxide for Cement C 92 Figure 48: Effect of Alkali Source on Formation of Amorphous Content for Cement E 92 Figure 49: Effect of Alkali Source on Formation of Amorphous Content for Cement C 93 Figure 50: Effect of Alkali S ource on Heat of Hydration for Cement E 94 Figure 51: Effect of Alkali S ource on Heat of Hydration for Cement C 94 Figure 52: Effect of Alkali Source on SO 4 -2 Concentration for Cement E 96 Figure 53: Effect of Alkali Source on SO 4 -2 Concentration for Cement C 96 Figure 54: Effect of Al kali Source on Aluminum Ions Concentration for Cement E 97 Figure 55: Effect of Al kali Source on Aluminum Ions Concentration for Cement C 97

PAGE 11

viii Figure 56: Effect of Al kali Source on Calcium Ions Concentration for Cement E 98 Figure 57: Effect of Al kali Source on Calcium Ions Concentration for Cement C 99 Figure 58: Potassium Ions Concentration for Cement E 100 Figure 59: Potassium Ions Concentration for Cement C 100 Figure 60: SEM Micrographs, Cement E, A) Case KH-CS, B) Case KS 101 Figure 61: SEM Micrograph on a Polished Se ction and EDX Spectra 102 Figure 62: Effect of Alkali on E xpansion Behavior for Cement E-SO 3 =AR 104 Figure 63: Effect of Alka li on Expansion Behavior for Cement E-SO 3 =3.6% 104 Figure 64: Effect of Alka li on Expansion Behavior for Cement E-SO 3 =5.0% 105 Figure 65: Effect of Alkali on E xpansion Behavior for Cement C-SO 3 =AR 105 Figure 66: Effect of Alka li on Expansion Behavior for Cement C-SO 3 =3.6% 106 Figure 67: Effect of Alka li on Expansion Behavior for Cement C-SO 3 =5.0% 106 Figure 68 : Cracked Mortar Bar of Case 5-3.8 fo r Cement E at 180 days 108 Figure 69: Effect of Alkali on Strength Gain for Cement E-SO 3 =AR 109 Figure 70: Effect of Alkali on Strength Gain for Cement E-SO 3 =3.6% 109 Figure 71: Effect of Alkali on Strength Gain for Cement E-SO 3 =5.0% 110 Figure 72: Effect of Alkali on Strength Gain for Cement C-SO 3 =AR 110 Figure 73: Effect of Alkali on Strength Gain for Cement C-SO 3 =3.6% 111 Figure 74: Effect of Alkali on Strength Gain for Cement C-SO 3 =5.0% 111

PAGE 12

ix Figure 75 : XRD of hydration products for Ca se E-5-3.8 112 Figure 76: Effect of Alkali on Ettr ingite Formation for Cement E-SO 3 =AR 113 Figure 77: Effect of Alkali on Ettringite Formation for Cement E-SO 3 =3.6% 114 Figure 78: Effect of Alkali on Ettringite Formation for Cement E-SO 3 =5.0% 114 Figure 79: Effect of Alkali on Ettr ingite Formation for Cement C-SO 3 =AR 115 Figure 80: Effect of Alkali on Ettringite Formation for Cement C-SO 3 =3.6% 115 Figure 81: Effect of Alkali on Ettringite Formation for Cement C-SO 3 =5.0% 116 Figure 82: Effect of Alkali on C 3 S Hydration for Cement E-SO 3 =AR 117 Figure 83: Effect of Alkali on C 3 S Hydration for Cement E-SO 3 =3.6% 117 Figure 84: Effect of Alkali on C 3 S Hydration for Cement E-SO 3 =5.0% 118 Figure 85: Effect of Alkali on C 3 S Hydration for Cement C-SO 3 =AR 118 Figure 86: Effect of Alkali on C 3 S Hydration for Cement C-SO 3 =3.6% 119 Figure 87: Effect of Alkali on C 3 S Hydration for Cement C-SO 3 =5.0% 119 Figure 88: Effect of Alkali on Ca lcium Hydroxide Formation for Cement E-SO 3 =AR 120 Figure 89: Effect of Alkali on Ca lcium Hydroxide Formation for Cement E-SO 3 =3.6% 120 Figure 90: Effect of Alkali on Ca lcium Hydroxide Formation for Cement E-SO 3 =5.0% 121 Figure 91: Effect of Alkali on Ca lcium Hydroxide Formation for Cement C-SO 3 =AR 121 Figure 92: Effect of Alkali on Ca lcium Hydroxide Formation for Cement C-SO 3 =3.6% 122

PAGE 13

x Figure 93: Effect of Alkali on Ca lcium Hydroxide Formation for Cement C-SO 3 =5.0% 122 Figure 94: Effect of Alkali on Amorphous Content Formation for Cement E-SO 3 =AR 123 Figure 95: Effect of Alkali on Amorphous Content Formation for Cement E-SO 3 =3.6% 123 Figure 96: Effect of Alkali on Amorphous Content Formation for Cement E-SO 3 =5.0% 124 Figure 97: Effect of Alkali on Amorphous Content Formation for Cement C-SO 3 =AR 124 Figure 98: Effect of Alkali on Amorphous Content Formation for Cement C-SO 3 =3.6% 125 Figure 99: Effect of Alkali on Amorphous Content Formation for Cement C-SO 3 =5.0% 125 Figure 100: Effect of Alkali on Heat of Hydration for Cement E-SO 3 =AR 126 Figure 101: Effect of Alkali on He at of Hydration for Cement E-SO 3 =5.0% 126 Figure 102: Effect of Alkali on He at of Hydration for Cement C-SO 3 =AR 127 Figure 103: Effect of Alkali on Heat of Hydration for Cement C-SO 3 =5.0% 128 Figure 104: Effect of Alkali on SO 4 -2 Concentration for Cement E-SO 3 =AR 129 Figure 105: Effect of Alkali on SO 4 -2 Concentration for Cement E-SO 3 =3.6% 130 Figure 106: Effect of Alkali on SO 4 -2 Concentration for Cement E-SO 3 =5.0% 130 Figure 107: Effect of Alkali on SO 4 -2 Concentration for Cement C-SO 3 =AR 131 Figure 108: Effect of Alkali on SO 4 -2 Concentration for Cement C-SO 3 =3.6% 132 Figure 109: Effect of Alkali on SO 4 -2 Concentration for Cement C-SO 3 =5.0% 132

PAGE 14

xi Figure 110: Effect of Alkali on Calcium Ions Concentration for Cement E-SO 3 =AR 133 Figure 111: Effect of Alkali on Calcium Ions Concentration for Cement E-SO 3 =3.6% 133 Figure 112: Effect of Alkali on Calcium Ions Concentration for Cement E-SO 3 =5.0% 134 Figure 113: Effect of Alkali on Calcium Ions Concentration for Cement C-SO 3 =AR 134 Figure 114: Effect of Alkali on Calcium Ions Concentration for Cement C-SO 3 =3.6% 135 Figure 115: Effect of Alkali on Calcium Ions Concentration for Cement C-SO 3 =5.0% 135 Figure 116: Effect of Alkali on Aluminum Ions Concentration for Cement E-SO 3 =AR 136 Figure 117: Effect of Alkali on Aluminum Ions Concentration for Cement E-SO 3 =3.6% 136 Figure 118: Effect of Alkali on Aluminum Ions Concentration for Cement E-SO 3 =5.0% 137 Figure 119: Effect of Alkali on Aluminum Ions Concentration for Cement C-SO 3 =AR 137 Figure 120: Effect of Alkali on Aluminum Ions Concentration for Cement C-SO 3 =3.6% 138 Figure 121: Effect of Alkali on Aluminum Ions Concentration for Cement C-SO 3 =5.0% 138 Figure 122: SEM Images on Paste samples for Cements C and E-Case 5-3.8 140 Figure 123: SEM Images for Paste 141 Figure 124: SEM Images for Mortar Bars 141

PAGE 15

xii Figure 125: SEM Polished and EDX Spectra for Case 5-3.8 of Cements C and E 142 Figure 126: Effect of Alkali on Expansion Behavior for Cement MH3-SO 3 =AR 144 Figure 127: Effect of Alkali on Expansion Behavior for Cement MH3-SO 3 =3.6% 145 Figure 128: Effect of Alkali on Expansion Behavior for Cement MH3-SO 3 =5.0% 145 Figure 129: E Effect of Alka li on Expansion Behavior for Cement MH4-SO 3 =AR 146 Figure 130: Effect of Alkali on Expansion Behavior for Cement MH4-SO 3 =3.6% 146 Figure 131: Effect of Alkali on Expansion Behavior for Cement MH4-SO 3 =5.0% 147 Figure 132: Effect of Alkali on Expansion Behavior for Cement ERD07-SO 3 =AR 147 Figure 133: Effect of Alkali on Expansion Behavior for Cement ERD07-SO 3 =5.0% 148 Figure 134: Effect of Alkali on St rength Gain for Cement MH3-SO 3 =AR 148 Figure 135: Effect of Alkali on St rength Gain for Cement MH3-SO 3 =3.6% 149 Figure 136: Effect of Alkali on St rength Gain for Cement MH3-SO 3 =5.0% 149 Figure 137: Effect of Alkali on St rength Gain for Cement MH4-SO 3 =AR 150 Figure 138: Effect of Alkali on St rength Gain for Cement MH4-SO 3 =3.6% 150 Figure 139: Effect of Alkali on St rength Gain for Cement MH4-SO 3 =5.0% 151 Figure 140: Effect of Alkali on Strength Gain for Cement ERD07-SO 3 =AR 151 Figure 141: Effect of Alkali on Strength Gain for Cement ERD07-SO 3 =5.0% 152

PAGE 16

xiii Figure 142: Pore Size Distribution for Cement C 158 Figure 143: Photograph under Light Microscope for Cement C 159 Figure 144: XRD Pattern for Cement C-5-3.8, Case A in Fig. 143 159 Figure 145: XRD of Hydrat ion Products, Case E-5-3.8 176 Figure 146: XRD of Hydrat ion Products, Case E-3.6-3.8 177 Figure 147: XRD of Hydrat ion Products, Case E-AR-3.8 178 Figure 148: XRD of Hydrat ion Products, Case E-5.0-AR 179 Figure 149: XRD of Hydrat ion Products, Case E-3.6-AR 180 Figure 150: XRD of Hydrat ion Products, Case E-AR-AR 181 Figure 151: XRD of Hydrat ion Products, Case E-AR-1.5 182 Figure 152: XRD of Hydrat ion Products, Case E-AR-2.0 183 Figure 153: XRD of Hydrat ion Products, Case E-5.0-2.0 184 Figure 154: XRD of Hydrat ion Products, Case C-5.0-3.8 185 Figure 155: XRD of Hydrat ion Products, Case C-3.6-3.8 186 Figure 156: XRD of Hydrat ion Products, Case C-AR-3.8 187 Figure 157: XRD of Hydrat ion Products, Case C-5.0-AR 188 Figure 158: XRD of Hydrat ion Products, Case C-3.6-AR 189 Figure 159: XRD of Hydr ation Products, Case C-AR-AR 190 Figure 160: XRD of Hydrat ion Products, Case C-5.0-1.5 191 Figure 161: XRD of Hydrat ion Products, Case C-3.6-1.5 192 Figure 162: XRD of Hydrat ion Products, Case C-AR-1.5 193 Figure 163: XRD of Hydrat ion Products, Case C-5.0-2.0 194

PAGE 17

xiv Figure 164: XRD of Hydrat ion Products, Case C-3.6-2.0 195 Figure 165: XRD of Hydrat ion Products, Case C-AR-2.0 196 Figure 166 : Schematic Representation of C 3 S Hydration 197

PAGE 18

xv LIST OF SYMBOLS AND ABBREVIATIONS ASTM American Society of Testing and Materials SEM Scanning Electron Microscope XRD X-Ray Diffraction Cement Chemistry Nomenclature A Alumina, Al 2 O 3 C Calcium Oxide, CaO F Ferric Oxide, Fe 2 O 3 H Water, H 2 O S SiO 2 S Sulfur Trioxide, SO 3 C 3 A Tricalcium Aluminate, 3CaO. Al 2 O 3 C 4 AF Tetracalcium AluminoFerrite, 4CaO. Al 2 O 3 Fe 2 O 3 C 2 S Dicalcium Silicate, 2CaO. SiO 2 C 3 S Tricalcium Silicate, 3CaO. SiO 2 CH Calcium Hydroxide, Ca(OH) 2 C S H 2 Gypsum, Ca 2 SO 4 .H 2 O C-S-H Calcium Silicate Hydrate Na 2 O e Alkali Content as Na 2 O

PAGE 19

xvi EFFECT OF ALKALIS AND SULFATES ON PORTLAND CEMENT SYSTEMS Mahmoud A. Halaweh ABSTRACT The effect of the sulfates and alkalis on the durability of Portland cement systems was investigated through a series of cube and prism mixes. Durability was assessed using expansion of mortar prisms and the compre ssive strength of mortar cubes. The study covered a large range of both alkali and sulf ate contents using 5 different Portland cements. The alkali contents ranged from 0.27 to 3.8%, the sulf ate content (as SO 3 ) ranged from 2.54 to 5%. Doping was done us ing Terra Alba gypsum and potassium hydroxide. In addition to physical measurements SEM, XRD, chemical analysis and heat of hydration calorimetry were used for further analysis. Mixing, cu ring and testing were done at room temperature. The results show that sulfate co ntents up to the levels used in this study, at low alkali contents and ambient temperature curing, did not adversely affect durability of Portland cement mortars up to 360 days. A correlati on was established between expansion and ettringite formation. Increasing the alkali content always resulted in loss of compressive strength, and in some cases, excessive expansion. Excessive expansi on was only experienced at the 3.8% level. Alkali levels of up to 2% and sulfate levels of 5% did not result in excessive expansion at

PAGE 20

xvii room temperature-cure up to the ages reported here. The effect of alkali depended on the cement mineralogical composition, especially C 3 S content. The addition of alkalis seems to impact the nature of the microstructure and the nature of other hydration products. The addition of sulfates seems to counteract the eff ect of alkalis, especially on the loss of compressive strength. However, these sulfates may result in other problems as they may be available at any time to form ettringite which may, under certain conditions, result in excessive expansion. It was concluded that sulfate levels on the order of 33.6%, did not pose any major durability drawbacks under normal curing temper atures and low alkali contents (<1%). Alkali levels above 1% will adversely affect the durability of Portland cement systems.

PAGE 21

1 CHAPTER 1 INTRODUCTION 1.1 Background on Sulfate Attack As concrete becomes increasingly used as a construction material, durability issues become more a nd more important. Its availability and relative low cost compared to other building materials made it more versatile. With less skilled labor required as compared to other construction techniques, the improvement in workmanship and quality control can only go so far, if the quality of the material used is not up to par, and as a one engineer once sai d, The principles of engineering have not changed throughout the history, but we can always improve the materials. One of the durability problems that may occu r during the service life of a concrete structure is sulfate attack. It may result in expansion, cr acking, spalling and eventually reduction of the strength of concrete. It can also subject the structure to other forms of attack; corrosion, etc. The ordinary Portland cement used in the ma nufacture of concrete consists mainly of calcium silicates and aluminates; tricalcium silicates (C 3 S), dicalcium silicates (C 2 S),

PAGE 22

2 tricalcium aluminates (C 3 A), and tetracalcium aluminoferrites (C 4 AF) and of course other minor phases. ASTM classifies Portland cement into five types according to the relative amounts of these compounds. It should be noted, however, that although these classification are based on chemical co mposition, the ASTM specifications are performance based. That is, cements have to meet certain physical requirements. Tricalcium silicate phase is responsible for the early strength gain in concrete and the dicalcium silicate contributes to the late strength gain (beyond 28 days). Tricalcium aluminate phase is very reactive and at the addition of water it woul d hydrate quickly and causes what is called flash set. This phenomenon will make it hard to place and finish the concrete in timely manner. To control th e hydration, calcium sulf ates (in one or more forms) are added, usually in the form of gypsum (C S H 2 ), which reacts with C 3 A according to equation (1) and form an ettring ite (tricalcium aluminat e tricalcium sulfate hydrate) coat around C 3 A grains. C 3 A + 3C S H 2 +26H C 6 A S 3 H 32 Ettringite formed during early hydration is referred to as Primary Ettringite. The primary ettringite formed during the initial hydration does not cause damage, because of its occurrence in a plastic matrix and the volume changes will be accommodated within the matrix. The ettringite that forms afte r the concrete has hardened is called Secondary, Delayed or Late ettringite, and this could be detrimental to concrete.

PAGE 23

3 In Type I Portland cement (ASTM specification), the Al 2 O 3 /SO 3 ratio is mostly equal to ~ 1.6 and contains a maximum of 3.5 to 4% SO 3 (US and European standards (1)) so at the beginning of the hardening stage, pr imary ettringite will be converted to monosulfoaluminate according to equati on (2) since ettringite requires a Al 2 O 3 /SO 3 ratio of 3 to remain stable. C 6 A S 3 H 32 + 2 C 3 A +4H 3C 4 A S H 12 .2 On subsequent exposure to a sulfate source in excess of that present in the MS, the monophase may reconvert to ettr ingite following equation (3) 3C 4 A S H 12 + 2C S H 2 +16H C 6 A S 3 H 32 .3 On the other hand, the monosulfate hydrate phase (C 4 A S H 12 ) is known to be unstable at room temperature and lower temperatures (1,2). Therefore, even in the absence of an extern al sulfate source, the monosulfate hydrate may accept sulfate ions released from an internal source to re form ettringite. The later phenomenon is called internal sulfate attack (ISA) which occurs usually in heat treated concrete. The issue of internal sulfate attack started to receive more attention in the mideighties when damage was reported in heat-cur ed prestressed concrete railway ties; it was believed that this was due to the elevated heat-cure followed by ambient cure. However, ISA is not limited to heat-cured concretes. ISA potential under ambient temperature cure exists under certain circumstances; this not ion was emphasized by Collepardi (3). Collepardi proposed a model for the ISA to occur. The author called it a holistic

PAGE 24

4 approach. According to the author, the phenomenon of ISA under ambient temperature cure is plausible. The model stated that ISA could occur if the following is present: microcracks, delayed release of sulfate and mo isture. The microcracking is a problem that existed and will always exist, whether due to loading or defects in the microstructure. The delayed release of sulfates is definitely worth careful consid eration. This is not necessarily due to slowly-soluble phases, but could be due to chemical changes occurring in the matrix. These changes do not require high temperature. One possible way to define the possibility of DEF (delayed ettringite formation) due to ISA (internal sulfate attack) is as follows : The formation of ettringite after the cementitious material has substantially hardened and in which none of the sulfate comes from outside the concrete or mortar. 1.2 Expansion Theories From what has been reported in the lite rature, possible damage mechanisms due to ettringite formation in normal as well as in heat treated concrete can be drawn through two major theories, the crysta l growth and the swelling: 1.2.1 Swelling Theory If the formed ettringite, either primary or delayed, inside the microstructure is micro-crystalline in nature, then it may cause expansion pressure in hardened concrete due to adsorption of water. Mehta (4) reported that the ettringite formed in the presence of calcium hydroxide is colloidal in nature it has a high surface area and exhibit a net

PAGE 25

5 negative charge. On exposure to water, this type of ettringite attract a large number of water molecules that surround th e ettringite crystals and cau ses inter-particle repulsion that results in expansion. Famy (5) added to that by mentioning that the high negative charge is due to the high pH (resulting from the presence of the calcium hydroxide), and in the absence of calcium hydroxide the ettrin gite is well-crystallized and its formation does not result in expansion. Min and Mingshu (6 ) stated that the concentration of the hydroxyl ions is what determines the mode of ettringite formation, so this means in a high alkaline pore solution, even with low calcium hydroxide solu tions, the same result would be expected due to the high pH. 1.2.2 Crystal Growth Theory The recrystallization of ettringite in th e hardened concrete, due to the moisture changes and accumulation of reactants, may lead to damages because of the crystallization pressure and the increase in vo lume. According to this theory, not all the ettringite formed causes expansion, since a porti on of the ettringite wi ll be deposited in available voids and cracks, so only the exce ss will cause expansion. However, there are two different arguments regarding this theory (1 ). The first argument indicates that for an expansion pressure to occur, supersaturation must be achieved and this does not happen in hydrated cement solution. Another argument poi nts out that this theory does not take into account the nature of conc rete being a semi-brittle material and a partial saturation might be enough to exert pressu re above the concrete tensile strength. A tensile stress applied at the tip of a crack might be magnified depending on the length and the geometry of the crack. Also, due to the he terogeneous nature of concrete a uniform

PAGE 26

6 expansion is not always likely, so the preferential crystal formation and the heterogeneous nature may result in crack s due to non-uniform expansion. This mechanism could also take place if ettringite forms in harden ed concrete from less-sulfate containing compounds either dela yed due to temperature gradie nt which might occur in a massive concrete or to the heat treatment in prefabricated concrete. A delayed ettringite formation might occur, as well, as a result of sulfates present in the concrete ingredients such as the aggregate or from an external sulfate source wh ich reach the aluminate phases through diffusion. These conditions should take place at room temperature in cements with an A/S ~1.6 in which the monosulfate hydrate has already exis ted and transformed to ettringite as a result of the above mentioned factors. All these s hould lead to nucleation, crystal growth a nd to the increase in volume causing stresses which should exceed the tensile strength of concrete. The tr ansformation of monosulfate to ettringite is well known to cause 2.3 times increase in volume. 1.3 Factors Affecting the Fo rmation of Ettringite and Expansion Potential So far, it is believed that the secondary ettringite formation c ould result in damage in hardened concrete, as a result of ISA. The expansive nature of ettringite, coupled with heterogeneous nature of concrete can resu lt in non-uniform expa nsion. The expansion potential is influenced by tw o factors; ettringite formati on and the nature of the C-S-H gel.

PAGE 27

7 The following factors influence the phenomenon: 1. Cement composition a. Tricalcium silicates (C 3 S): This is the most a bundant mineral in Portland cement, and mostly responsible for strength by forming C-S-H gel upon hydration. C-S-H has an affinity to incorporate some of the sulfates present in cement (5, 7). However, the affinity for sulfates increases with increase in temperature and an increase in alkali c ontent as will be shown in subsequent paragraphs. Since the sulfate is a ma in ingredient in the formation of ettringite, it may be released at later ages and participate in the formation of ettringite after hardenin g (5). So, higher C 3 S content may require higher sulfate content for proper retardatio n, which may become problematic at later ages. b. Tricalcium aluminate (C 3 A): Since the control of the reaction of this mineral requires the addition of sulfate, higher C 3 A content means higher sulfate content in order to control the hydrati on. That is not all however, not all the sulfate will be bound as ettringite (as mentioned above) and even if so, the ettringite may decompose and reform again after hardening. So, higher sulfate additions may prove detrim ental to concrete eventually. c. Alkali content: The alkali influen ces the rate of hydration of C 3 A and ettringite stability (1), which in turn may impact the sulfate requirement. It also influences the nature of the hydr ation products includi ng ettringite and C-S-H gel. The presence of alkali hydroxi de may increase the rate of sulfate

PAGE 28

8 ion release into solution, coupled with ettringite instability, may lead to the inclusion of these sulfate in the C-S-H ge l (17). The effect of alkali will be expanded upon later. d. Sulfate content: As presented earlier in this document, the sulfates are added mainly to control the hydration of C 3 A. They participate in the formation of ettringite. The sulfate requirements are complicated, as mentioned above. So, balancing this requirement considering all other factors C 3 S, C 3 A, alkali content, and fineness is not an easy task. The key is to control the sulfate addition in order to ensure a durable concrete, with high strength and low drying shrinkage. Due to strict environmental regulati ons, and the need for less expensive fuels, modern cement clinkers contain higher sulfate contents. Cement producers are pushing and advocating rela xing the restrictions on sulfate contents. The complicated cement hydration mechanisms make it hard to just allow higher sulfate levels than currently acceptable without more understanding of the interaction of a ll these phases and their impact on long term durability. 2. Cement fineness: Fineness influences the rate of hydration; this in turn will influence the sulfate requirement in or der to control the rate of hydration. 3. Curing temperature: The temperature affects hydration rate and the nature of the gel formed. The higher temperature results in higher porosity and a coarser C-S-H structure (81) It also impacts the stability of ettringite (5).

PAGE 29

9 4. Sulfate forms: The late formation of ettri ngite means late release of sulfates. This could be due to the concentration of sulf ate in some silicate phases of the clinker such as dicalcium silicates, or due to th e presence of sulfates in slowly soluble forms, such as anhydrite or even in some sulfate-rich aggregates. While the idea of slowly-released sulfates was supported by several researchers (3, 8, 9) it was opposed by others (10, 11, 12, 13, and 14) This should not be a point of controversy; even if all the sulfates were to be very soluble, the late release may occur due to the other factors mentioned above. 1.3.1 Effect of Curing Temperature Ettringite stability As already mentioned, it is established in the literature (1, 2, 5, 15) that ettringite decomposes to monosulfate hydrate on e xposure to elevated temperature >80 C. The monophase is further converted to the sulfate free calcium al uminate hydrate at a long exposure time to boiling water temperature. Other published data offers a maximum curing temperature of ~70 C as a safe curing temperature. B ecause ettringite is the stable calcium sulfoaluminate hydrate phase at room temperature, it will then reform in the concretes stored at ambient temperature. One other thing to keep in mind, the temperature rise in concrete to levels where ettringi te becomes unstable and decomposes, does not have to be due to heat -curing. It was repo rted in the literature (91) that in large concrete sections with high cement content, the in ternal temperature c ould rise up to 85 C.

PAGE 30

10 The stability of ettringite depe nds on factors other than the te mperature. Heinz et. al. (16) detected the presence of ettri ngite after 12 hours of curing at 90 C when cement had 8.6% by weight SO 3. This could be due to the fact that the density of the matrix and the exposure time were not sufficient for its de composition It can be concluded that the exposure time has an effect, in addition to SO 3 content as well as cement composition, especially C 3 S. C-S-H is known to incorporate sulfat e ions (5,7). The C-S-H gel affinity for sulfate increases with the in crease in curing temperature (17). Ionic spices availability The high temperature seems to result in a lo wer pH (19); it is not known whether this affects the stability of ettringite directly. Ho wever, the drop in the pH results from the precipitation of the calcium hydroxide, whic h in a cement system will be supplied through the hydration of calcium silicate phases. Also, it has been established in the literature that the solubility of calcium hydroxide decreases with increasing temperature. This implies less calcium hydroxide available for the formation of ettringite. However, the decisive factor seems to be the availability of the sulfate in solution according to th e experiments conducted by Heinz (16). In reference (5), the author showed that the sulfate ion concentration increases with increasing curing temperatures; at the same time, no ettringite was detected under these conditions Nature of C-S-H gel Since high temperatures decrease the solubili ty of calcium hydroxide, it may result in less calcium available for the formation of C-S-H gel. This can result in a lower Ca/Si

PAGE 31

11 ratio. Also, it was reported in the literature that the nature of the gel formed at elevated temperature differs from that formed at am bient temperature (81). The gel formed at higher temperature has been reported as coarser with higher porosity. Many researchers (8-18, 20-39, 69, 72 and 75) have tackled the issue of DEF due to heattreatment offering different hypotheses on the mechanism of the attack. However, the published literature still offers conflicting theories. In reference (38) the author reports that certain cements did not expand even after pr olonged heat curing at 100 C. However, delayed ettringite due to ISA has been reported to occur in concrete which has not been subjected to heat treatment (3 ). This could be due to the decomposition or the lack of formation of the primary ettringite due to ch emical changes of the concrete pore solution, such as the change in alkalinity. While some investigators rejected the hypothe sis that the cause of damage was due to ISA in lieu of other causes, such as ASR (40) a nd freeze and thaw, as being the primary cause for damage in some heat-cured concrete da mage cases; nonetheless there seems to be an agreement that secondary ettri ngite will form without the pres ence of an external sulfate source. 1.4 Research Objective Role of alkalis at ambient cure temperatures As mentioned above, the cement compos ition greatly influences the phenomenon of the ettringite formation as well as its reformation. Increased fuel costs and

PAGE 32

12 tightened environmental regulations have led to increased levels of alkalis and sulfates in modern cements. The switch from wet to dry pr ocesses in order to conserve energy and the lack of low-alkali raw materials resulted in higher alkali levels (41 and 42). Since any attempt to remove the alkalis from the raw materials is not a cost-effective nor is a practical solution, the presence of alkali s in cements is inevitable. The ASTM specifications do not limit the alkali content of OPC where there is no potential for alkaliaggregate reaction. The limit im posed by ASTM C-150 is 0.6% Na 2 O e for potential ASR, however, the typical alkali content of Por tland cements range from 0.3%-1.3% expressed as Na 2 O e (42). In fact, since high alkalinity is fa vorably viewed to prevent the initiation of corrosion, high alkali-cements are being used as a measure to prevent corrosion. Recently, some DOTs received requests from concrete suppliers to uses alkalicontaining aggregates due to the lack of better aggregates in those localities. Higher alkalis are believed to have a detrimental effect on the durability of Portland cement concrete. The effect of alkalis on cement hydration and mechanical properties of concrete has been studied by many authors. However, the effect on the expansion potential due to internal sulfate attack, whic h is believed to be impacted greatly by alkalis, has not been studied thoroughly. In recent years however, this topic received more attention. This is due to the emergence of delayed ettringite formation especially in heat-cured concrete elements. Most of the re search conducted in this area considered the influence of alkalis on expansion of cem entitious systems exposed to elevated temperatures.

PAGE 33

13 The objective of this research is to study th e effect of alkali co ntent on the expansion potential of PC systems. While these co mpounds are not deliberately added to the cements, they may be present in the raw ma terials used in the manufacture of cement; namely, clays and limestone. 1.4.1 Research Significance As mentioned above, the alkalis greatly in fluence the properties of the Portland cement concretes. This study thoroughly inves tigates the influence of the addition of alkalis on the expansion potenti al of cements due to intern al sulfate attack (DEF) at ambient temperature-cure. This study also addr esses the exact influe nce of sulfates on expansion potential and verify the role of ettringite in expansion. From the data collected with analytical tec hniques used, more comp lete answers to the following questions will be addressed: 1. What is the influence of alkalis on expansion potential due to internal sulfate attack. 2. What is the influence of sulf ates on the expansion potential. 3. What is the effect of cement mineralogical composition on expansion potential and strength of concrete and strength. 4. What is the role of ettring ite as related to expansion. 1.5 Outline of Dissertation Chapter 2 presents a literature review of the studies performed on the effect of sulfates and alkalis at ambient temperature cure.

PAGE 34

14 Chapter 3 presents the methods and proce dures used in the experiments. Chapter 4 presents the materials used in this study and also details the experi mental plan and the philosophy adopted to achieve the objective. Chapters 5 through 8 present the results of the experiments described in chapter 4 and in clude a detailed analysis of all the data collected. Chapter 9 discusses the result s presented here and considers possible mechanisms for the effect of the alkalis on concrete durability. Chapter 10 concludes the main findings of this research.

PAGE 35

15 CHAPTER 2 LITERATURE REVIEW In chapter 1, it was mentioned that the expansion potential due to internal sulfate attack (ISA) exists. Collepardi (3) seems to believe that this phenomenon is possible under normal curing conditions in concretes th at have not been subjected to heat treatment. The factors that affect the expansi on potential were also listed. In this chapter we will examine what has been reported in the literature. Although the main focus of this study is the effect of alkalis, it is prudent to study its effect in terms of additional cementitious parameters because of the effects of the latter on the former. It is believed that one of the shortcomings of the previous studies is not thoroughly studying the alkali effect in those terms. Parameters that are believed to be of significance are: SO 3 content, C 3 A content, and C 3 S content. 2.1 Role of SO 3 Collepardi (3) seems to believe that the laterelease of sulfates is possible in concretes that were not subjected to heat-cure. This could be due to the use of sulfate-rich fuels in the kilns in recent years, this has results in higher sulfate content in modern clinker. G. Hime and others (8,89,33) stated that DEF due to ISA is possible due to the presence of slowly-soluble sulfate phases in modern clinkers. This hypothesis was rejected by several authors (10, 13, 14, 36, 85, 92, and 93).

PAGE 36

16 W.G. Hime (33) stated that in clinkers with high SO 3 levels, that are not balanced by high alkalis, sulfates do not occur as alkali sulfat es. This may result in sulfates encapsulated in silicate phases, or possibly as insoluble anhydrite These forms of sulfates may become available later resulting in DEF. M.D.A Thom as (11) indicated that the latter explanation is unlikely, and even if insoluble phases ex isted, they will be balanced by release of alumina, resulting in the formation of monosulfate rather than ettringite. Stark and Bollmann (1) stated that there is no clear connection be tween normal sulfate levels in Portland cements and the expansion of concrete, the occurrence of damage, or the degree of damage. Diamond (24) supported the idea of DEF due to ISA under ambient conditions. The explanation proposed in this work indicates that DEF and IS A are due to hi gher sulfate contents in modern cements (up to 4-5%) because of the higher clinker sulfate content (up to 3%). The work of Lawrence (20) and Kelham (38) showed no significant expansion to occur when mortars were cured at ambient temperatures, even with SO 3 content as high as 44.5%. In Kelhams work, the cement had 0.61% alkali eq uivalent, 65% C 3 S and 11.8% C 3 A. In Lawrences work, the cement had 0.99% alkali equivalent, 51% C 3 S and 5.2% C 3 A. The cement composition was assessed us ing Bogue method in both studies. The work of Heinz and Ludwig (16) showed no signi ficant expansion to occur in mortar bars

PAGE 37

17 when curing temperature is kept below 70 C, even with high sulfate content. This study did not provide details about the composition of these cements. Day (51) stated that there is no evidence that DEF is the principal cause of damage in non-heat-treated concrete. However, in the sa me literature review, the work of Al-Rawi showed an increase in the expansion with an increase in the SO 3 content of the cement that was added as gypsum. The expansi on was 0.7% at 23 weeks at the 7.5% SO 3 level, and about 0.14% at 5.7% leve l. The two cements used in this study had 62 and 63% C 3 S and 12 and 13% C 3 A contents, alkali levels were moderate (0.48%). The cement composition was assessed using Bogue method. The work of Odler and Gasser (84) showed an increase in the expansion with the increase of sulfate content (added as gypsum), the alka li content of the cements ranged form 0.67 to 0.8% alkali equivalent. Duncan Hertford et. al. (85) reported no si gnificant expansion to occur in mortars made from cements with SO 3 content as high as 6.5% when cured and stored at 20 C, even after 2 years, the expansion was almost nonexistent. In their work, the C 3 A content was about 6-8%, alkali was less than 0.6%. The author did not show a co rrelation between the expansion and the sulfate contents.

PAGE 38

18 Stark and Seyfarth (34) reported almost no e xpansion in mortar samples cured at 20 C after almost 2 years with SO 3 contents of 3.3% and 3.9%. C 3 A content was 8.6% however. K.M Alexander et. al. (87) reported very low expansion to occur when the SO 3 content was increased to 4%. Most of the expansion o ccurred in the first 7 days, with the highest value being less than 0.02%, however, the expa nsion increased with increasing sulfate content. This work was done on cements with moderate C 3 S (54-59%) and C 3 A (5-7%) content, the alkali contents range d from 0.57-0.75% as al kali equivalent. Chengsheng Ouyang (88) showed an increase in the expansion with sulfate content. In this work sulfates were increased using phosphogypsym. The expansion also increased as the C 3 A (Bogue) content was increased. For example, at SO 3 content of 7.1%, the expansion was less than 0.1% at 10 months for the cement with C 3 A content of 4.3%, while it was more than 0.6% at 3 months for the cement with C 3 A of 12% content. The C 3 S content was 63.3% for the first case and 59% for the second case. Taylor (89) seems to believe th at there is no danger of DE F in modern cements unless the SO 3 content exceeds 5-6%. It is believed that his finding needs to be redefined or specific to the mineralogical compositi on and alkali content of the cements. There is no doubt that SO 3 impacts the expansion potential due to ISA. The question is the exact role of the SO 3 content and how it is impacted by other parameters. SO 3 is

PAGE 39

19 definitely needed for the formation of ettringite, however, there is a competition for sulfate ions between C 3 A and C-S-H gel (31). These two reactions are also affected by the alkali content of cements. Base d on these facts, the effect of SO 3 has to be studied with respect to those phases. What the precedin g studies are lacking is the isolation of the effect of each variable, one at a time. It is true that some of the studies mentioned above varied the C 3 A content, or even compared two cements with different C 3 S contents, however, it is not enough to draw reliable conc lusions. There is also a lack of focus on the role of alkalis; the variations in alkali content were not enough or did not entail wide cement composition to allow for reliable conclusions. 2.2 Role of C 3 A Odler and Gasser (84) showed that the expa nsion due to the increase in sulfates decreased by decreasing the amount of C 3 A and increasing the C 4 AF content. Their conclusion was ettringite formed in cemen ts with high iron contents had different morphology. Chengsheng Ouyang (88) showed th at expansion due to internal sulfate attack increased with increase in C 3 A content at the same SO 3 level. Lerch (43) stated that cements with higher C 3 A require higher amounts of gypsum for proper setting. Taylor (15) seems to believe that there is no danger of DEF at ambient temperature cures in modern cements unless the SO 3 content exceeds 5-6%. He also added that the high safe limit increases with an increase in C 3 A content. This is not exclusively true, since as mentioned earlier, these factors are influe nced by the cement composition and alkali content.

PAGE 40

20 D. L. Kantro (90) showed increased expansion when the C 3 A content increased. The SO 3 content used was based on 7.59% gypsum The expansion increased when C 3 A content was increased form 2% to 13%. The C 3 S content was held at 60%. The primary concern of C 3 A content is its availability to participate in the formation of ettringite, which could be expansive. The C 3 A content has several implications; the most obvious one is the increase in the sulfate requirement for prope r setting. So, it has to be remembered that this could result in a potentia l problem later, depending on the amounts of sulfates and aluminates that will be available for ettringite formation into the hardening stage. Since the form ation of ettringite requires bo th sulfates and aluminates, it is expected that, and for the same SO 3 content, expansion will decrease if there is an increase in C 3 A content. This will be entirely true if all the aluminates and sulfates were available for reaction and formation of ettrin gite in the plastic stage, and whatever ettringite forms during this stage will remain stable. Unfortunately this is not the case. First of all, not all the aluminates and sulfates will be released during the early stages of hydration and become available for ettringite formation. Add to that the effect of other parameters. For example, it was mentioned that C-S-H gel competes fo r sulfate ions with C 3 A. These sulfate ions could be released at later ages and may participate in the formation of expansive ettringite. Also, the presence of alkalis a ffects the rate of C 3 A hydration, and is known to increase the affinity of C-S-H gel for sulfate ions (31). So, in studying the role of C 3 A on the expansion due to ISA, it is important to study the effect of C 3 S content, SO 3 content as well as alkali content of the cement.

PAGE 41

21 2.3 Role of C 3 S It was reported in the literature (31) that C-S-H gel competes for SO 4 -2 ions with C 3 A It was also reported that C-S-H gel incorp orates sulfate ions in its structure (5, 7). The impact of the adsorbed sulfate on the nature of the gel is unclear (5). These ions may be released later to form late ettringite (7). This was shown by the work of L. Divet et. al. (16). The uptake of sulfate ions by C-S-H ge l is impacted by several parameters, the most investigated parameter in the literature be ing the curing temperature. The increase in curing temperature resulted in an increase of the sulfate ions adsorbed by the C-S-H gel (5). The desorption rate was slower as the curing temperature increased. The adsorption also increased with an increase in the alkali content (17) at all curing temperatures. The desorption rate was slower than the adsorption rate; actually, the desorption rate was half that of the adsorption. The effect of alkali was not shown since this was only done for one alkali content. Since the C-S-H gel competes with C 3 A for sulfate ions, higher C 3 S means more sulfates are adsorbed, since there will be more C-S-H gel formed. Higher C 3 A may imply less sulfate will be adsorbed by the C-S-H gel; how ever, this will depend on the alkali content which impacts both the rate of C 3 A hydration and the stability of the calciumsulfoaluminate phases formed. It will also depend on the SO 3 content. There have not been any direct st udies as to the exact role of C 3 S on ISA at ambient temperature. Any study would have to addre ssed IAS in terms of the other parameters mentioned above to allow for meaningful conclusions.

PAGE 42

22 2.4 Role of Alkalis The effect of alkalis on ISA phenomenon are direct and indi rect. First, they affect the stability of ettringite, whether it will form or not and its morphology. Second, alkalis affect the hydration process as well as the micr ostructure and the qual ity of the formed CS-H gel. 2.4.1 Effect of Alkalis on Ettringite Formation Mehta (4) stated that the presence of cal cium hydroxide in th e solution affects the mode of ettringite formation and its morphology as to whether it forms as fine or coarse crystals. These two different types of ettringite behave differently, with the finer morphology (in presence of calcium hydroxide ) being expansive and the coarse r (formed in its absence) is not. Famy (5) stated that th e difference is due to the higher pH resulting from the presence of calcium hydroxide Min and Mingshu (6) stated that the concentration of the hydroxyl ions is what de termines the mode of ettringite formation. So, based on this, high pH values due to higher alkali content of the solution will result in the expansive type of ettringite. Stark and Bollmann (1) presented a summary of what has been reported in the literature about the stability domain of ettringite. Th e domain covered a wide range of pH. It seems, however, that the reason for this wide range is the dependence of this phenomenon on other factors, such as cement composition (SO 3 and C 3 S).

PAGE 43

23 This indicates that even at high pH values if there is enough s upply of sulfate ions, ettringite may form. So, we ca nnot just consider the pH of the solutions without other factors, such as sulfate content. The formation of ettringite and its stability depend on several factors. These are the sulfate content and the availability of the sulfate ions, the pH of the solution, the availability of calcium hydroxide The higher alkali content in creases the availability of the sulfate ions in solution, but that does not mean that ettringite will form, it will depend on the solubility limits of ettringite. It al so depends on the amount of sulfates that are being adsorbed by the C-S-H gel as mentioned earlier (17). The C 3 S content is a definite concern. The calcium hydroxide solubility is decreased, but there is always enough supply of the calcium hydroxide in the hydra tion of Portland cement solution. Based on these facts, it is necessary to study the stab ility of ettringite in these terms. The controversy in the previous study stems from the lack of the systematic approach. It has to be mentioned that some of the studies mentioned above were carried out on pure phases which makes it difficult to draw the sa me conclusions for the Portland cement systems. 2.4.2 Effect of Alkali on Hydratio n Process and Mechanical Properties 2.4.2.1 Effect of Alkalis on the Hydration of Clinker Phases Lerch (43) reported the necessity to increase the amount of sulfate needed to control the setting of cemen t with high alkali content. This means the reactivity of C 3 A is altered A similar result was reported by Sprung and Rechenberg (44); they

PAGE 44

24 indicated an increase in th e rate of hydration of C 3 A. B. Osbaeck (60) also reported an increase of the reactivity of C 3 A and C 3 S with the increase in the total alkali content of cement. Odler and Wonneman (45) reporte d an increase of th e reactivity of C 3 A when the clinker was doped with K 2 O, but not when doped with Na 2 O, the hydration of C 3 S was not altered. They also reported (46) that adding the alkalis as sulfates did not affect the hydration of either phase. Johansen (47) also reported an increased rate of hydration when adding alkalis (mainly in the form of su lfates). Way and Shayan (48) reported an increase in the gel production with increase in alkali content; the increase was achieved through the addition of NaOH to the mixing water. Juenger and Jennings (49) reported that the addition of NaOH to the mixing wate r increases the initial hydration of Portland cement and retards that after the first day. Similar findings were reported by Bentz (50). Mori, according to ref. (42) reported an acceleration of C 3 S hydration in NaOH solution compared to pure water. Similar findings were reported by other researchers (42). The reports as to the effect of alkalis on ferrite and belite phases are scarce. Generally, the evidence points out to the increase of initial hydration. However, it is possible that the nature of the reaction products is impacted rather than just the reaction rate itself; also, the role of alkalis on the hydration proce ss greatly depends on the sulfate levels in cement. In all the studies done so far, there was no systematic way of investigating the interrelation of the alkalis and sulfates. The results on the effect of the source of alkalis (sulfates vs. hydroxide) were not consistent as well. Some studies (45,46) reported an increase in the rate of hydration when alkalis were increased as

PAGE 45

25 hydroxides but not as sulfates, wh ile another study (47) showed an increase in the rate of hydration when the alkalis were in creased using alkali sulfates. 2.4.2.2 Effect of Alkalis on Io ns Availability in Solution It is evident from the literature that the solubility of calcium hydroxide decreases with the increase of the concentra tion of alkali hydroxide. The effect of alkali sulfates on the solubility of lime is more co mplicated and depends on the concentration of each phase and the type of alkalis. This re sult was reported by Sprung and Rechenberg (44). Ghorab et al (77) reported an increase in the solubility of gyps um with the increase in NaOH concentration as well; this means hi gher sulfate availability in solution. The solubility of silica is known to increase with the increase in alkalinity and the respective pH value (52, 66). The high alkalin ity in this case was due to the addition of NaOH to the mixing water. Alkali sulfates did not show the same effect (55). 2.4.2.3 Effect of Alkalis on Mech anical Properties and the Microstructure Up to this point, it is very clear from the literature that alkalis impact the hydration and kinetics of hydratio n of Portland cements. This effect also extends to the nature of the hydration products formed, especially the C-S-H gel. The results reported in the literature are either incomplete or somewh at contradicting. That is, the dependence of the hydration process on many f actors, especially sulfate co ntent, and the dependence of this on alkali content as well as the cement composition appear not to be clearly considered .This makes it hard to really generalize a role for alkalis.

PAGE 46

26 Inam Jawed and Jan Skalny (42) conducted a thorough literature review on the influence of the alkali on the properties of concrete. Ge nerally, most of the literature reported that higher alkali content, present in the form of hydr oxide or sulfate, result in quick setting. The alkali hydroxide favors the formati on of crystalline monophase around the C 3 A grains, and the sulfate, mostly in the form of potassium, leads to the precipitation of syngenite. These effects are strongly dependent on the conten t of these ions. For strength development, most of the evidence pointed to an increase in the early strength and a decrease in the 28-day strength. This is due to lowering of the gel in the presence of alkalis compared to that of calcium which enhances the gelatinous properties of the hydrates. Other researchers, however, showed a positive effect on strength development according to ref. (42). Alexander and Davis (53) show ed a 56% decrease in 28-day strength in cement pastes when increasing alkali conten t from 0.15 to 2.8% (as Na 2 O e ). The increase was done through the addition of alkali hydroxides. The authors indicated that the decrease occurred regardless of the type of the alka li cation (sodium or potassium). The study was done on five different cements. The C 3 S content of the cements varied form 31-51% (Bogue) and the C 3 A content varied from 4-14% (Bogue ). There was no mention of the sulfate contents of these cements. The sulfate content is believed to impact the effect of the alkalis. There was no explanation for the reduction in the strength. Water J. McCoy and Ottomar L. Eshenour (54) reported that the alkali that is watersoluble varies from 10% to above 60% of the total alkalis pres ent in clinker. The

PAGE 47

27 amount of total water-soluble alkali in cemen t has very little effect on the pH of the aqueous extract from cement pastes; th is was done on w/c ratios of 0.5 at hour hydration and w/c ratio of 2 at hour and 24 hours hydration. The pH ranged from 12.312.7 for cements with alkali contents ranging from .03%-1.19% Na 2 O e The authors also added that the alkali in clinker has a si gnificant effect on the compressive strength development. The absence of alkalis resulted in an abnormally low early strength, and the presence of alkalis resulted in higher early strength (1-3 days) and lower strength at 28 days. The study included several cases; in one case industrial cemen ts with low and high watersoluble alkali contents. In another cas e, an industrial clinker was obtained and it was modified in the lab to remove alkalis. A third case, KOH was added to the no-alkali clinker. B. Osabaeck (55) reported that increasing the alkali content (through the addition of K 2 SO 4 ) increased the early strength and decreased the late strength. The data reported showed an increase of the ai r content in the cases with hi gher alkali content. It was concluded that the effect of the alkalis on late strength is caused by something happened during the early hydration and no t governed by the conditions of the pore liquid at later ages. This study was done on one cement, the cement had C 3 S content of 75% (Bogue) and C 3 A content of 4% (Bogue). No details were given about the hydration process or the hydration products. S. Sprung and W. Rechenberg (44) reported higher reactivity of C 3 A in higher alkali contents with higher pH values of the solution. If there is a lack of gypsum,

PAGE 48

28 mainly aluminate hydrates form which cause rapid setting; however, in presence of high sulfate supply, large amounts of ettringite form which causes rapid setting. The authors also added that with higher amounts of alkali s in clinker, particul arly alkali sulfates, lower calcium hydroxide concentrat ion exist in solution, which in turn increases the rate of hydration of C 3 A prior to the dormant period whic h may lead to rapid setting. They also added that the composition of the solu tion in the early hydrat ion, not only impacts the setting, but the strength development as we ll. This study did not investigate the effect of the alkali on the ra te of hydration of C 3 S, which could impacts th e setting behavior as well. V. Johansen (47) indicated that the switch from wet to dry processes resulted in higher alkali content in the cement. Also, the use of cheap sulfur-rich fuels, led to higher SO 3 content in clinker. The author stated that th is change has resulted in a 10% decrease in the 28-day strength. The author also reported that an increase in K 2 SO 4 results in 28day strength decrease, although there was an increas e in the early strength (1 and 3 days). The decrease occurred regardless of the source of alkali sulfates, wh ether coming from the clinker or externally added. The author attrib uted this decrease in strength to something happening in the early hydrati on stages. It was shown that the 28-day strength correlated negatively with amount of combined water after 3 minutes of hydration (chemicallybound water). This implies that this has some thing to do with early hydration of the C 3 S and the nucleation of the gel early on. This st udy did not provide more details about the hydration process, however.

PAGE 49

29 Bhatty and Greening (56) reported that C-S-H gels with low C/S ratio retain more alkalis than high C/S hydrates when these hydrates co me in contact with alkali solutions of various concentrations. The authors mentioned that this change, de pending on the alkali concentration of the solution as well as the type of alkali cation, will change the nature of the C-S-H formed as indicated by XRD pattern. Th is change was attributed to a change in the crystal structure. The study did not menti on the physical implications of this change. This work was done on pure C 3 S and C 2 S hydrates and the focus of the study was on the effect of increased alkalis on the alkali-aggregate reaction. Way and Shayan (48) reported an increase in the rate of the production of cement gel and CH when increasing the concen tration of sodium hydroxide in mixing water from 0 to 1M. At a concentration of 2M and higher, th e formation of aluminosulfates was retarded and sodium-substituted monosulfate formed inst ead of ettringite. It was also noted that the formation of a sodium-containing C-S-H gel at these higher concentration occurred. This study did not provide firm analytical data to support the increase in the rate of gel production. Also, the study was only done on one cement which makes it not suitable to infer a role for the cement composition or the sulfate content. The cement used in the study had 52.7% C 3 S and 5.9% C 3 A (Bogue) and 3.3% SO 3 Jelenic et al (57), studied the effect of gypsum on the strength development of two different cements differing only in the alkali content. The cements used in the study had a high alite content, approximately 70% (XRD), the C 3 A content was about 8-10% (XRD), the alkali contents were 0.16 and 1.06%. The study showed an optimum sulfate

PAGE 50

30 content for strength gain for each of the cements. The optimum was higher for the low alkali cement at the same age, the optimum for the high-alkali cement shifted to a higher value with age. The maximum strength was higher for the low alkali cement up to 47 days. However, at 90 days the difference in the strength was not noticeable. This was explained by the reaction of the adsorbed su lfate to form sulfoaluminate phases and the redistribution of the sulfate in the gel. Higher strength was attributed to better quality gel and in turn higher strength according to th e authors. The authors did not present any evidence to support this statem ent. Also, there was no mention of the expansion behavior. This study indicates the eff ect of alkalis and sulfates on the hydration process and strength gain. The alite hydration was sligh tly higher for the low alkali cement at all sulfate contents (1-5%) up to 47 days. However, this could not explain the difference in compressive strength. The study lacks more data such as the expansion behavior and the quantification of the C-S-H gel, which may have helped explain the difference in the strength. The study makes it hard to draw a conclusion on the effect of cement composition since it was done on the same alite and C 3 A contents. J. Gebauer and M. Kristman (58) reported an increased reactivity in higher alkali clinkers. These alkalis were mainly sulfate alkalis. The ea rly hydration rate increased with increase in alkali-sulfate content of the clinker. Mortar early compressive strength (2d) increased while the 28d strength decreas ed. There was no effect on the 28d flexural strength. There was also no clear influence on the 28d compressive strength in concrete.

PAGE 51

31 In conclusion, it is not easy to generalize a specific trend from the presented data due to a number of influencing factors; that is, cem ent composition, fineness, sulfate content, etc. B. Osbaeck et al (59) reporte d a negative effect of the tota l and soluble alkali content on the 28 day strength. This influence depended, as expected, on the total SO 3 content. The study showed an increase in the rate of hydration of certain phases (C 3 S, C 3 A, and C 4 AF) at early ages. This was not necessarily responsible for the increased strength. However, the general conclusion reflected on the nature of the hydrat ion products. No specific details were given about the hydration products. B. Osbaeck (60) reported an increase in the early age strength and a decrease in the 28day strength when soluble alkali content was increased through doping with potassium sulfates. Doping levels were 0.5% and 1.0% K 2 O. It was indicated that the increase in early strength might diminish if a constant slump is maintained through increasing w/c ratio, as the addition of alkalis increased the water demand. I. Odler and R. Wonnemann (45) conducted a study on the influence of the alkalis added to the raw feed of clinker preparation. They prepared two different clinkers through the addition of Na 2 O and K 2 O. In both cases, the alkalis were incorporated preferentially into the lattice st ructure of C 3 A modifying it from cubic to orth orhombic. It was reported that the hydration of C 3 A was accelerated in the potassiumdoped clinker and slowed down in sodium-doped one. The hydration of C 3 S was not altered. The strength development was

PAGE 52

32 not affected by the presence of alkalis, yet the setting time was moderately extended in the sodium case and shortened in the potassium case. I. Odler and R. Wonnemann (46) conducted a study on the influence of the alkalis added to the clinker as su lfates of sodium and potassium (the doping levels were 0.72% Na 2 O as Na 2 SO 4 1.26% Na 2 O as Na 2 SO 4 0.88% K 2 O as K 2 SO 4 and 1.48% K 2 O as K 2 SO 4 ). It was reported that these additions did not signi ficantly alter the hydration process of either C 3 S or C 3 A. The setting time was shortened in both cases, but significantly in the case of potassium sulfates. This was attributed to the formation of syngeni te, but there was no explanation for the case of the sodium sulfate doping. The co mpressive strength decreased in both cases at a ll hydration ages. Though the cause of this decrease was not obvious, it was postulated that changes in the structure an d intrinsic properties of the hydrates was the reason. This occurred due to changes in the liquid phase caused by dissolution of the alkali sulfat es. These two studies (45 an d 46) were conducted on one laboratory-prepared clinker having the following mineralogical composition: C 3 S=70%, C 2 S=10%, C 3 A=10%, C 4 AF=10% (based on Bogue). The SO 3 content for the control case was maintained at 3.0% using interground gypsum and the Bl aine fineness was 300 m 2 /kg. This was a good detailed study, however it was only done on one clinker with one sulfate content. No role for the cement co mposition could be drawn from these results. The XRD analysis mentioned lacks more details as far as the methods used in the analysis.

PAGE 53

33 K. Suzuki et al (61) reported a lower Ca/S i mole ratio for C-S-H in systems of Ca-SiO 2 NaCl-NaOH. The authors suggested an adsorp tion of Na ions by the C-S-H gel. This results in a lower exothermal temperature in the DTA curve. After dispersion with water, this temperature rises to the same value obt ained by the pure C-S-H gel. This study was done on pure compounds. N. Smaoui et al (62) reported a reducti on in mechanical properties of concrete (compressive strength, splitting, di rect tensile, and flexure, but not modulus of elasticity) by increasing the Na 2 O e from 0.6% to 1.25% through the addition of NaOH. The authors attributed that to a more reticular porous cement paste as observed under SEM, even though there was no reported difference in th e hydration products in both cases. Both cases resisted freeze/thaw well. It was also concluded that the additional alkalis did not modify the air-void system. The followi ng values were calculated for the cement composition based on the chemical oxide analysis provided in the study: 55.5% C 3 S, 7.3% C 3 A, C 3 A (based on Bogue ca lculations). The SO 3 content was 2.94%. There was no mention of the Blaine fineness value. Th e study did not provide any details about the hydration products other than the SEM result s. The study was also done on one case where there was no variation of the cement composition. Again, it is believed that the effect of alkalis on cements is influenced by the cement mineralogical composition. Vivian (63) conducted a flexure test on mo rtars made from reactive and non-reactive aggregates and increased the alkali content of the cement through the addition of NaOH. The alkali levels (Na 2 O e ) were 0.59%, 0.98%, 2.14%, and 4.08%. The author did

PAGE 54

34 not observe any reduction in the tensile strength of the mortars containing non-reactive aggregates, except in the case of 4.08% Na 2 O e This is an unrealistic amount, and also the w/c ratio was higher than the other cases, which according to (62) could explain the decrease in strength. The intention of this study was to show the effect of reactive aggregates, however, there was no details abou t the cement composition in this report. Shayan and Ivanusec (64) conducted a study on the influence of alkalis through the addition of NaOH in the mixing water. The alkali levels ranged from 0.8 to 10.5% Na 2 O e An increase in expansion with alkali level was observed. The study also noted a delayed formation of ettringite and enhanced formation of calcium hydroxide. However, as the alkali level was increased (2 and 4.5 M NaOH), neither ettr ingite nor monosulfate were formed; rather, the formation of a new phase (sodium-substituted monosulfate, referred to as U-Phase) and a crystalline form of C-S-H gel was observed. There was also a decrease in the compressive strength as well as the m odulus of rupture with an increase in alkali level. It was attributed to the formation of the new phase and to the substitution of sodium in the C-S-H gel structure, a (Na 2 O/SiO 2 ) ratio of 0.17 was reported. It was also observed, under SEM, that high-alkali cement pastes had a less dense microstructure as opposed to low-alkali cement pa stes. This would possibly cont ribute to loss of strength. The study did not provide the cement mineral ogical composition, nor the sulfate content which may have influenced the effect of th e alkalis on strength and expansion. The XRD analysis did not provide any quantitative data on the hydration process and the amorphous content.

PAGE 55

35 M.P. Luxan et al (65) conducted a study on the potential expansion of Portland cement with the addition of alkali salts (Na 2 CO 3 or K 2 SO 4 ). Aside from the expansion results, it was found that the addition of these alkali salts at 1, 2, an d 3 % by weight reduced the expansion of Portland cement mortars. The au thors found that the a ddition of the sodium carbonate altered the hydration reaction by causing abnormal sett ing (false or quick set). This problem was corrected by the additions of natural pozzolans. For the potassium sulfate, it was reported that the addition of 1, 2 or 3% did not cause abnormal setting; however, it did accelerate the se tting of cements. The addition of <1% of this salt delayed both initial and final set. Th e addition of this salt caused a decrease in compressive strength at early and late ages; this reduction was attributed to modification of the hydration of clinker minerals due to the presence of the alkalis in the liquid phase. Maria C. G. Juenger and Hamlin M. Jennings (49) reported that increasing the alkali content through the addition of NaOH (1M) accelerated the initial hydration; however, after the first day the high alka linity retarded the hydration pr ocess. The total surface area obtained from nitrogen absorption decreased; this was attributed to the decrease of the gel pores of the radii of 1-4 nm. It was also noted that the presence of NaOH led to the preferential formation of denser C-S-H gel, which resulted in hete rogeneous structure. A slower rate of drying shrinkage was also observed. The authors offered the following explanation: since NaOH-containing samples had heterogeneous structure, localized stresses caused larger cracks. These cracks helped reduce the measured shrinkage. Also the samples continued to shrink without wate r loss, suggesting that this is due to microstructural rearrangement. The total al kali content for this study was 1.2% as

PAGE 56

36 Na 2 O e The study was done on one cement which had 54% C 3 S and 10% C 3 A (Bogue). The SO 3 content was 2.84%, and the Blaine Fineness was 368 m 2 /kg. This cement was used for all the results reported in the study except for the heat of hydration data which was done on a similar cement having a C 3 S content of 64%, and C 3 A content of 8% (Bogue). Although the cements are similar, the higher C 3 S content may impact the results of this test. There was no variation in the cement mineralogical composition or the sulfate content which may impact the effect of the alkali on th e properties studied. Dale P. Bentz (50) reported that the incr eased alkali content (using both sodium and potassium sulfates and hydroxides) of cement paste increased the early age hydration and retarded the hydration at later ages. It was also concluded that the presence of alkalis results in depercolated capillary pores. The aut hor attributed that to the nature of the gel formed, being plate lathlike with higher crystallinty rather than random. This morphology might result in slower diffusion rate and slower hydration at later ages. The depercolation may be advantageous since it re sults in less freezable water in the pores. Considering the studies presented above, it can be seen that there is a lack of systematic approach. The effect of alkalis strongly de pends on sulfates; theref ore one cannot isolate the effect of the alkalis alone on the propertie s of concrete without addressing the levels of sulfate studied. This is believed to be one of the major reasons for conflicting findings in the literature. Additionally, cement composition and the mineralogical phase content need to be considered when addressing the role of alkali content on concrete durability.

PAGE 57

37 2.4.3 Effect of Alkalis on Expansion Potential A. Shayan and I. Ivanusec (64) reported an increase in the expansion of mortar bars using sodium hydroxide doping. The expansion was reported to be approximately 0.04% at 12 weeks at 3.8% Na 2 O e This study did not provide details about the cement composition or the sulfate content. M.P. Luxan et al (65) reported that incorporation of al kali carbonate and sulfate salts (Na 2 CO 3 1 and 2% by weight; K 2 SO 4 1.2 and 3% by weight) reduced the expansion of Portland cement mortars, mainly when C 3 A content is low (<5%). It was concluded that there is an optimum concentra tion for the alkali salt for ea ch cement. Again, this study did not provide any details about the hydration products, or any details on the Na 2 O e levels. Vivian (63) showed no signi ficant expansion to occur when the alkali content of the cement was increased to 4.08% using NaOH in the mixing water. The study was done on mortar prisms (1 in x 1in x 10 in). The e xpansion did not exceed .007% at 196 days. The study did not show any detail s about cement composition. As can be seen from this review, the findings in the literature about th e exact effect of the alkalis on the progress of hydr ation and the expansion behavi or is complicated and needs further investigation.

PAGE 58

38 This research was initiated in order to addr ess the role of alkali content on concrete durability. In order to address this objectiv e in a thorough and systematic way, several cements with variable mineralogical and chemical composition were selected. The following chapters will present experiment al techniques, methodology, results, discussion and conclusions.

PAGE 59

39 CHAPTER 3 EXPERIMENTAL METHODS AND PROCEDURE 3.1 Pure Clinker Minerals Pure clinker minerals, monoclinic C 3 S (Alite), C 2 S (Belite), cubic and orthorhombic C 3 A, C 4 AF, MgO were obtained from Construction Technology Laboratory (CTL) of Skokie, Il linois. These pure phases were used in the pr eparation of the calibration curves for the mineralogical phase quantificat ion of Portland cements used in this study. These phases had particles finer than 45 micron. TiO 2 (rutile) powder was obtained from Aldrich Chemical Company; it has a particle size smaller than 5 microns with 99.9% purity 3.2 X-Ray Diffraction (XRD) All XRD scans were collected using th e following settings, unless otherwise noted. The diffractometer is Phill ips XPert PW3040 Pro with Cu K radiation. The samples were scanned from 2 of 5 to 60. Step size was 0.02, counting time is 4 seconds per step. The tension was set at 45 kV and the current at 40 mA. Divergence slit was fixed at 1, receiving slit was set at 0.2 mm and the anti -scatter slit was fixed at 1. The diffractometer was periodi cally aligned using a silicon wafer standard supplied by the machine manufacturer.

PAGE 60

40 3.3 Scanning Electron Microscopy (SEM) SEM, and sometimes coupled with EDS, was used to study the morphological changes in the chemistry of the cement hydr ation products. The instrument used was a Hitachi S-350N variable pr essure scanning electron micr oscope. Energy Dispersive Spectrometry (EDS) was performed in a spot mode using Princeton Gamma Tech prism light element detector. The software used to analyze EDS data was Princeton Gamma Tech IMIX. 3.3.1 Fractured Surfaces Preparation At selected ages, a cross section of mo rtar or paste specimen was cut with a Buehler Isomet slow speed saw. The thickne ss of the cross section was approximately 5 mm; one of the faces of the cross-sec tion was a fractured surface exposing the morphology of the hydration products. The sample was submerged in acetone overnight to stop hydration and to dry the sample through displacing the water with acetone. The samples were then placed in desiccat or under vacuum of 25 in. of Hg. The samples were then mounted onto the sa mple holder with double-stick tape and the sample was attached to the holder with coppe r tape. The samples were then placed in a Hummer 6.2 sputter coater with a vacuum pump. The samples were pumped down to 30 millitorre to ensure total drying. The samples were then coated with a 40 nm of AuPd. The samples were then loaded into the SEM instrument for viewing.

PAGE 61

41 3.3.2 Polished Surfaces Preparation A cross section of mortar or paste specimen was cut using a Buehler Isomet slow speed saw. The cross-section was then soaked in Ethanol for several days; this was done to help displace the water out of the pores. The samples were then placed in a vacuum desiccator and vacuum-pumped for several days. The samples were then mounted in a epoxy resin, the epoxy used was Epotek 301, 2-part epoxy. The samples were placed in a vacuum desiccator to get rid of any air bubbl es trapped in the epoxy. The samples were left to cure for 24 hours at room temperature. The samples were then demolded, the botto m face in the mold was cut using the a diamond blade. The samples were then polis hed at 320, 400 and 600 gr it sand papers, to remove the saw marks. Then, the samples were polished on a Textmet cloth using diamond paste with the following sizes, 15 m 9 m, 3 m, 1 m, and 0.25 m, for 30 seconds each (78), using diethy l glycol. The samples were then cleaned in an ultra-sound box. 3.4 Cement Analysis 3.4.1 Blaine Fineness Cement fineness was determined according to ASTM C-204-00 Standard Test Method for Fineness of Hydraulic Cement by Air-Permeability Apparatus. The Blaine air-permeability apparatus, obtained from Humboldt Manufacturing Company,

PAGE 62

42 was calibrated prior to use with three separate ly prepared beds of standard cement SRM 114 obtained from the National Institute of Standards and Technology (NIST). 3.4.2 Oxide Chemical Analysis The oxide analysis was performed by an external laboratory, CTL of Skokie, Illinois. Cements were fused at 1000 C with Li 2 B 4 O 7 and analyzed by x-ray fluorescence (XRF) in accordance with th e precision and accuracy requirements of ASTM C-114-99 Standard Test Method for Chemical Analysis of Hydraulic Cement. 3.4.3 Mineralogical Composition 3.4.3.1 Bogue Equations ASTM C-150-00 Standard Specificati on for Portland Cement provides the guidelines for using the Bogue equations to determine the theoretical mineralogical cement composition; the equations are based on the chemical oxide composition. 3.4.3.2 Internal Standard Method (Calibration Curves) This method involves the construction of calibration curves for the pure clinker phases mentioned above, mixed with an internal sta ndard, the internal standard used here was rutile (TiO 2 ). Calibration curves for and orthorhombic C 3 A, C 4 AF, MgO were prepared according to ASTM C 1365-98 Standard Test Method for Determination of the Proportion of Phases in Portland Ce ment and Portland-Cement Clinker Using Xray Powder Diffraction Analysis. The XRD scans were collected according to section 2.2.

PAGE 63

43 The collected scans were analyzed using Profit software to determine the selected peak areas. The peak areas were used rather than peak heights since they provide more reliable data (70). The background was determined ma nually, and then the program automatically searched for the peaks. The user determin ed if all the peaks were identified; and necessary corrections were made. The peaks at 2 of 24.3 for C 4 AF, at 27.4 and 36.1 for TiO 2 and at 42.9 for MgO had to be corrected for the K -2 contribution by entering 0.5 in the 1/ 2 column for the appropriate peak. After that, the profile fitting command was executed, and the peak areas were calculated. The 27.4 for TiO 2 was used in the construction of these curves. The peak at 2 of 21.8 was used in the case of C 3 A, at 24.3 for C 4 AF and at 42.9 for MgO. The curves we re checked against Standard Reference Materials (SRM) provided by NIST. Thes e clinkers were SRM 2686, SRM 2687 and SRM 2688. The curves yielded good data except for the orthorhombic C 3 A, so the data for that curve was not used. The curv es are shown in Figures 1, 2 and 3.

PAGE 64

y = 0.0102xR2 = 0.96760.000.100.200.300.400.500.600.700.8001020304050607Weight % of SampleArea Ratio 0 Figure 1: Calibration Curve for C 3 A y = 0.0102xR2 = 0.96760.000.100.200.300.400.500.600.700.8001020304050607Weight % of SampleArea Ratio 0 Figure 2: Calibration Curve for C 4 AF 44

PAGE 65

y = 0.0921xR2 = 0.97290.000.501.001.502.002.503.003.504.0005101520253035Weight % of SampleArea Ratio Figure 3: Calibration Curve for MgO 45 The calibration curve for alite was prepared by grinding 10 g of C 3 S in a Norton ceramic Jar mill operated at 150 RPM. The jar had a volume of approximately 300 ml and was filled with 68 cylindrical beads. Alite was ground for 90 minutes. Five drops of ethylene glycol were added for lubrication to minimize the impact of the beads. After grinding, 2 different mixtures were prepared. The total sample weight was 0.8533 g; the internal standard (TiO 2 ) was 10% of the total mixture. The first mixture contained 0.7680 g of ground alite, and 0.0853 of internal standard. This sample containing 90% alite, and 10% TiO2 was called % sample. The second mixture contained 0.384 g of alite, 0.384 g of CaF 2 and 0.0853 g of TiO 2 This sample was called the % sample. Each sample was repeated 3 times. The samples were weighed out, placed in a glass vial and mixed for 5 minutes to ensure that the standard is intimately mixed with the other materials. Two drops of cyclohexane were added to each sample before mixing to facilitate the

PAGE 66

process. After mixing, the mixture was placed in the sample holder and an XRD scan was collected per the procedure outlined in section 2.4. The scans were analyzed using the Profit software again as explained above. The area under the peak at 2 of 30.1 for alite was divided by that at 36.1 for the TiO 2 The ratio was plotted against the weight percent of alite. The curve is shown in Figure 4. Again, the accuracy of the curve prepared was checked against the SRM 2686, 2687 and 2688. y = 0.0232xR2 = 0.94950.000.501.001.502.002.503.003.504.00020406080100120140C3S Weight %Area Ratio Figure 4: Calibration Curve for Alite After that, cement samples were prepared. Approximately 10 g of the as-received cement were ground in a Norton ceramic jar as described earlier. Cement was ground for 30 minutes with ethylene glycol. The cement sample was collected carefully to prevent contamination and hydration, then stored in a vial in a dessicator. 0.768 g of the cement 46

PAGE 67

47 and a 0.0853 g of TiO 2 were weighed out (total sample weigh of 0.8533g), the mixture was mixed intimately, and cyclohexane was used to facilitate the process. XRD scans were collected and then analyzed using the Pr ofit software. The ratios of the areas under the peaks to the area under the TiO 2 peak were calculated. The factors derived from the calibration curves were used to determine the weight % of the phases of interest. 3.4.3.3 The Rietveld Method The Rietveld method is a full-patt ern fit method. The measured profile (XRD patterns) and a calculated profile are compared. By the variation of many parameters, the difference between the two pr ofiles is minimized. In order to perform a Rietveld refinement, structure data for all phase s present in the sample are needed. This is a classical Rietveld refinement. It requires cr ystal structure data base. Peak positions and intensities are calculated from crystal structur es. Scale factors, cell and profile parameters (and more parameters) are varied to minimize the differences between observed data and the calculated profile. This method is based on very complex mathematical algorithms. It requires a lot of accuracy and care in the prep aration of the samples and collecting scan data. Due to the complexity of the method and the complications experienced in the material we are dealing with (cement and cement hydration pr oducts), this is considered as a backup method to the internal standard method. Despite all that, this method yielded very good results, and it was used for the anal ysis of the hydration products, as will be shown later in the text. A great deal of time and effort was sp ent in mastering this method of analysis, and the results were confirme d by independent operators when necessary.

PAGE 68

48 Since the method relies on the cr ystal structure of the phases pr esent, the crystal structure inserted in the program for analysis had to be as accurate as possible. One of the problems of analyzing cements is significant peak overlap (70). So, to overcome this problem two dissolution techniques were adopted. The procedure adopted was that detailed in references (70) and (71). The first dissolution method used salicylic acid and methanol to dissolve the silicate phases and free lime; the method is called SAM for short. The second method used solution of potassium hydroxide (KOH) and sucrose. This is used to dissolve th e interstitial phases (aluminate and ferrite ), and the method is called KOSH for short. More about the procedure foll owed in these techniques is presented in Appendix A. These dissolution methods prove d valuable in cement mineralogical analysis. After the extraction, XRD patterns were collect ed on the KOSH extract, the SAM extract, and the ground cement as a whole. Rietveld refinement was performed using HighScore Plus software. 3.5 Chemical Analysis of Hydrating Cements The chemical analysis of the hydrating cement solution was done as follows: for each doping case, the ingredients were mixed w ith water, at w/c rati o of 7:1. The mixture was continuously agitated for 7 days. The solution was extracted by centrifuging at 30 minutes, 1 hours, 4 hours, 8 hours, 24 hours, 3 days, and 7 days. The solution was then analyzed using Perkin Elmer, Optima 2000 DV inductively coupled plasma optical emission spectrometer.

PAGE 69

49 3.6 Porosity Measurements (BET) and Degree of Hydration The porosity measurements were perfor med using a NOVA series instrument by QUANTACHROME. The paste samples were taken out of the curing solution. The samples were then washed thoroughly with me thanol to remove al l the excess lime on the surface. The samples were ground to particle sizes in the range between 600 and 1180 m (No. 30 and No. 16 sieve). The samples were then D-Dried. The D-drying apparatus was set-up according to Copeland and Hayes (80). It consisted of vacuum drying to the vapor pressure of water at the temperature of dry ice (5x10 -4 torr). Particles smaller than 600 m were used for loss on ignition testing to determine the degree of hydration ( ) following the procedure adopted in the same reference. 3.7 Mortar Preparation 3.7.1 Mortar Bars Mortar bars were prepared, cured and st ored in accordance with ASTM C305 and ASTM C1038. Some minor adaptations were ma de in the doped cases, in the cases where gypsum or potassium sulfates were used; th ese materials were mixed for 30 seconds in the mixing water prior to adding the cement. In the cases where KOH was used, it was dissolved in the mixing water. Length m easurement were taken in accordance with ASTM 490-00a. 3.7.2 Mortar Cubes Mortar cubes were mixed in accordance with ASTM C 305-99 Standard Practice for Mechanical Mixing of Hydraulic Cement Pa stes and Mortars of Plastic Consistency and molded and tested according to ASTM C 109-99 Standard Test Method for

PAGE 70

50 Compressive Strength of Hydraulic Cement Mortars. Compressive strength was determined per ASTM C 109 using MTS 809 Axial/Torsional Test System. 3.8 Paste Preparation Cement mixing followed the procedures in ASTM C 305-99. Paste discs were prepared to use for the XRD work and SEM wo rk, to eliminate the interference of the sand on mortar samples. Table 1: Mix Proportions for As-Received Cements (AR) Bars Cubes Paste Discs Cement (g) 500 240 250 Sand (g) 1375 2035 N/A Water (ml) 242 359 121 w/c ratio 0.485 0.485 0.485 3.9 Heat of Hydration The heat of hydration was measured using an 8 channel calorimeter by TAM Air. TAM Air is an 8-channel is othermal heat conduction calorimeter for heat flow measurements in the milliwatt range. The operating temperature range is 5-60C. All calorimetric channels are of twin type, consis ting of a sample and a reference vessel, each with a volume of 20 ml. The thermostat uses circulating air and an advanced temperature regulating system to keep the temperature ve ry stable within 0.02 K. The high accuracy and stability of the thermost at makes the calorimeter well suited for heat flow measurements over extended peri ods of time, e.g. weeks. A ch emical or physical process may be exothermic (heat is evolved) or e ndothermic (heat is absorbed). Consequently

PAGE 71

51 there is a change in temperature of th e sample during reaction. The difference in temperature between the sample and the surrou ndings (that are kept constant) results in heat flow between the sample and the surr oundings which is monitored continuously. Inbuilt calibration heaters are used for calibration of the calorimetric units. Pico LogTM which is a commercial software package, wa s used for data collection and analysis. Sample preparation Samples are usually prepared by external mi xing by hand or in a mixer to achieve a homogeneous sample. Alternatively, the dry constituents of the cement sample can be loaded into a micro reaction system with stirring facilities, positi oned in a channel of TAM Air. A known amount of water is then a dded by the use of a sy ringe and the sample is stirred inside the calorimeter in order to in itiate the hydration process. As a result of the hydration process, heat is formed and the rate of heat production is continuously monitored as a function of time. We used the second method in order to be able to record the first peak, w/c ratio used was 0.485.

PAGE 72

52 CHAPTER 4 MATERIALS SELECTION AND EXPERIMENTAL PLAN 4.1 Materials Used for Doping Gypsum used to increase the sulfate c ontent of cements, was Terra Alba grade conforming to ASTM C-563, obtained from US G, 125 South Franklin Street, Chicago, Illinois, 60606 All other chemicals were ASC grade chemicals obtained from Fisher Scientific. 4.2 Sand All sand used in this study was obtained from the U.S. Silica Company, 701 Boyce Memorial Drive, Ottawa, Illinois 61350. It is Graded sand conforming to ASTM C 778-00. 4.3 Cement Selection As mentioned in chapter 1, the influence of the sulfate and alkalis content of cement on concrete durability is greatly influenced by the cement mineralogical composition and its fineness. Careful consider ation was made in selecting the subject cements in order to be able to isolate the e ffect of each phase. Great effort was placed in obtaining and characterizing these cements. Th e methods of cement characterization were described in detail in Chapter 2. All cements had to have similar fineness, since its

PAGE 73

53 almost impossible to manipulate this propert y in the lab, unlike ot her minerals where it could be doped if needed. The results of the analysis are presented in Tables 2-6. Table 2: Blaine Fineness Cement C E ERD07 MH3 MH4 Blaine Fineness (m 2 /kg) 384 380 408 395 427 Table 3: Oxide Chemical Compositio n of As-Received Cements (weight %) Analyte C E ERD07 MH3 MH4 SiO 2 20.52 21.15 20.08 20.20 20.99 Al 2 O 3 4.92 4.78 4.98 4.02 4.92 Fe 2 O 3 3.70 3.76 2.22 2.78 2.24 CaO 64.31 64.41 62.21 64.02 63.93 MgO 1.71 0.95 3.65 2.47 2.14 SO 3 2.81 2.58 3.47 3.09 2.55 Na 2 O <001 0.18 0.3 0.21 0.43 K 2 O 0.41 0.34 1.12 1.10 0.67 TiO 2 0.27 0.33 0.21 0.22 0.21 P 2 O 5 0.03 0.07 0.2 0.15 0.16 Mn 2 O 3 0.04 0.03 0.1 0.06 0.04 SrO 0.04 0.12 0.34 0.04 0.13 Cr 2 O 3 <0.01 <0.01 0.01 <.01 <.01 ZnO <0.01 0.02 0.01 0.04 <.01 L.O.I (950 C ) 1.08 1.15 0.66 1.40 1.36 Total 99.83 99.84 99.57 99.79 99.77 Alkalis as Na 2 O 0.27 0.4 1.03 0.93 0.87 Table 4: Bogue Mineralogical Content Cement Compound (%) C E ERD07 MH3 MH4 C 3 S 60 57 54 67 57 C 2 S 14 18 17 7 17 C 3 A 17 6 9 6 9 C 4 AF 11 11 7 8 7

PAGE 74

54 Table 5: Phase Composition of As-Recei ved Cements Based on Calibration Curves Cement Compound (%) C E ERD07 MH3 MH4 C 3 S 70 58 59 67 62 C 3 A (Cubic) 3 4 9 3 2 C 4 AF 17 10 5 8 6 MgO 1 0 6 2 3 Table 6: Phase Composition of As-Received Cements (Rietveld Method) Cement Compound (%) C E ERD07 MH3 MH4 C 3 S 67 54 59 72 65 -C 2 S 15 25 15 12 14.5 Cubic 2 4 9 2 2 C 3 A Ortho 8 C 4 AF 14 13 4 9 3.3 Gypsum (CaSO 4 .2H 2 O) 2 0.3 1.0 Bassanite (CaSO 4 .0.5H 2 O) 1.5 1.6 2.4 1.4 1.6 Anhydrite (CaSO 4 ) 1.3 Potassium Sodium Sulfate 1.6 0.7 Syngenite 1.4 Arcanite (K 2 SO 4 ) 1.0 MgO 0.6 4 1.0 1 Since the Bogue method is only an approximation of the values of the phases present, the values obtained from the XRD analysis (Rietveld) will be considered in interpreting the data. From the above analysis results, the following can be concluded: 1. Cement C is high in C 3 S, low in C 3 A and alkali contents. 2. Cement E has a moderate C 3 S, low C 3 A and alkali contents.

PAGE 75

55 3. Cement ERD07 has a moderate C 3 S, high C 3 A and alkali contents, high MgO content, and slightly higher SO 3 content. 4. Cement MH3 has a very high C 3 S content, high alkali content, and low C 3 A content. 5. Cement MH4 is high in C 3 S, C 3 A and alkali contents. It has a low SO 3 content. 4.4 Experimental Plan and Philosophy As mentioned in chapter 1, there is a lack in studies dedicated to investigate the influence of alkalis on the durability of c oncrete. Since the influence of alkalis is dependent on sulfates as well, it is imp erative to study both t ogether in thorough and systematic fashion. First, the influence of the source of alkalis (that is sulfates versus hydroxide) had to be addressed. For this purpose, the following regime was employed in order to vary the alkali and sulfate content and form of the as -received cement (case AR). Three different doping regimes were adopted: 1. Case KS : increasing the sulfate and alkali level using potassium sulfate (K 2 SO 4 ). SO 3 content was increased to 5% which brought Na 2 Oe content to 2.32%. 2. Case 5.0-AR : increasing sulfate leve l using gypsum (CaSO 4 .2H 2 O). SO 3 content was increased to 5%.

PAGE 76

56 3. Case KH-CS: increasing sulfate level using gypsum (CaSO 4 .2H 2 O) and alkali level using potassium hydroxide (KOH). SO 3 content was increased to 5% and Na 2 Oe content to 2.0 %. This experiment was done for cements C an d E only. The results of this experiment would reveal the influence of the alkali source, but, since using potassium sulfate as a source of alkali will change both sulfate and alkali content at the same time; it will not be possible to isolate the effect of each para meter. So, it was decided to use potassium hydroxide (KOH) to vary the alka li content from this point on. The following summarizes the experi mental plan for this study: 1. Vary the alkali content (hydr oxide) and study its effect. 2. Vary the sulfate content (usi ng gypsum) and study its effect. 3. Study the combined effect of alkalis and sulfates. As mentioned earlier, the effect of ISA (DEF ) manifests in volume changes, cracking, and eventual loss of strength. Experimental methods used in assessing each effect are listed in the following paragraphs: 1. Volume changes: As a result of ettr ingite formation, rate of formation and particle morphology. A dditionally, the nature of the matrix (C-S-H gel) can be a contributor. Experiment al techniques used to assess these changes are.

PAGE 77

57 a. Expansion of mortar bars (ASTM C-1038/C-305). b. XRD: to study the presence and the amount of ettringite as well as rate of formation. c. SEM: to study the nature of gel and morphology of ettringite. d. Porosity: gel porosity was studied using nitrogen absorption method. 2. Loss of strength: Since the expansio n potential will be exacerbated by a weaker gel, the strength must also be assessed. The following is used to assess the phenomenon: a. Mortar cubes (ASTM C-109): used to assess the compressive strength. b. SEM: to study the nature of gel, using fractured sections. c. SEM: used to study the chemical make-up of gel using polished samples; this will give us an idea about the Ca/Si ratio as well as the presence of ionic substitution in the gel. 3. Reaction kinetics: a. Heat of hydration: used to study the timing and rate of hydration of cement phases. b. Cement solution chemistry: used to study the rate of dissolution of cement phases, and availability of certain ionic species in the solution and its depletion.

PAGE 78

58 The above techniques were used in studying the following cases: 1. As-received cements. This case was done for cements C, E, MH3, MH4 and ERD07. 2. Cements doped with gypsum to study the influence of the sulfates on expansion, cements C, E, MH3 and MH 4 were doped to levels of 3.6% and 5% by weight of cement. Some case s for cement ERD07, especially at the 3.6% level were not mixed. This is due to the shortage in this cement, and the higher sulfate level of the cement (3.47%). This range encompasses all sulfate contents available in the market. 3. Cements doped with potassium hydroxide, three different doping levels were used; 1.5% 2.0%, and 3.8% (Na 2 O e ). The 3.8% case will be done for cements C and E only; this is an extreme case that will be used for comparison reasons, since some of the studies in the literatu re have used this case. For all other cements (MH3, MH4 and ERD07), only 1.5% and 2.0% doping levels were adopted. A summary of the different cases studied here is presented in table 7. Table 7: Mixes SO3 Alkalis as Na 2 O e AR AR 1.5% 2.0% 3.8%* 3.6% AR 1.5% 2.0% 3.8%* 5% AR 1.5% 2.0% 3.8%* For cements C and E only.

PAGE 79

59 While cements C and E were initially selected to serve as a baseline, cements ERD07, MH3 and MH4 selected because of their alkali content (about 1%). ERD07 has low C 3 S content and high C 3 A content, MH3 has High C 3 S and low C 3 A, and MH4 has high C 3 S and high C3A. ERD07 will be comp ared to cement E, MH3 will be compared with cement C and MH4 will be compared with both to study the effect of high C 3 S and high C 3 A content. Table 8 shows the mineralogical composition of all cements using different analytical methods. In conclusion, cements selected for this st udy encompass a wide variance of chemical composition. This is considered to be a criti cal step so that the findings of this study can be used for a wide range of cement compositions. Table 8: Major Phases of All Cements C 3 S C 3 A SO 3 Alkalis Cement Bogue Rietveld Curves Bogue Rietveld Curves Na 2 O K 2 O Na 2 O e Blaine m 2 /Kg ERD07 54 59 58.82 9 9 9 3.47 .30 1.12 1.03 408 MH3 67 72 67 6 2 3.3 3.09 .21 1.10 .93 395 MH4 57 65 63 9 10 N/A* 2.55 .43 .67 .87 427 C 60 67 70 7 2 3 2.81 <.01 .41 .27 384 E 57 54 58 6 4 4 2.58 .18 .34 .40 380 Labeling and mix identification method: The following labeling method was adopted fo r this study. It is based on the doping levels of sulfate and alkalis. For example C-3.6-1.5 means cement C with 3.6% SO 3 and 1.5% alkali as Na 2 Oe. The following is a listing of diffe rent cases studied in this research. 1. Case AR-AR : As-received cements. No doping.

PAGE 80

60 2. Case 3.6-AR: Sulfate content increased to 3.6% using gypsum (CaSO 4 .2H 2 O), and alkali contents reflec ts as-received condition 3. Case 5-AR : Sulfate content increased to 5.0% using gypsum (CaSO 4 .2H 2 O), and alkali content reflect s as-received condition. 4. Case AR-1.5:Sulfate content reflects as-recei ved condition, and alkalis content increased to 1.5% using potassium hydroxide (KOH). 5. Case AR-2 : Sulfate content reflects as-recei ved condition, and alkalis content increased to 2.0% using potassium hydroxide (KOH). 6. Case AR-3.8: Sulfate content reflects as-r eceived condition, and alkalis content increased to 3.8% using potassium hydroxide (KOH). 7. Case 3.6-1.5: Sulfate content incr eased to 3.6% (CaSO 4 .2H 2 O), alkalis content increased to 1.5% using potassium hydroxide (KOH). 8. Case 3.6-2.0: Sulfate content increased to 3.6% (CaSO 4 .2H 2 O), and alkalis content increased to 2.0% using potassium hydroxide (KOH). 9. Case 3.6-3.8: Sulfate content increased to 3.6% (CaSO 4 .2H 2 O), and alkalis content increased to 3.8% using potassium hydroxide (KOH). 10. Case 5-1.5: Sulfate content increased to 5.0% (CaSO 4 .2H 2 O), and alkalis content increased to 1.5% using potassium hydroxide (KOH). 11. Case 5-2: Sulfate content incr eased to 5.0% (CaSO 4 .2H 2 O), and alkalis content increased to 2.0% using potassium hydroxide (KOH). 12. Case 5-3.8: Sulfate content increased to 5.0% (CaSO 4 .2H 2 O), and alkalis content increased to 3.8% using potassium hydroxide (KOH). Tables 9-13 present the mix proporti ons for all cases studied here.

PAGE 81

61 Table 9: Mix Proportions for Cement C Cement (g) Added Gypsum (g) Added KOH (g) Alk SO 3 AR 1.5 2 3.8 AR 1.5 2 3.8 AR 1.5 2 3.8 AR 500 486.90 482.23 462.75 0 0 0 0 0 13.10 17.77 37.25 3.6 490.96 476.78 472.43 452.06 9.04 9.62 9.8 10.54 0 13.50 17.77 37.40 Bars 5 474.94 461.13 457.22 437.21 26.06 25.19 25.22 25.29 0 13.50 17.77 37.50 AR 740 720.61 713.7 684.87 0 0 0 0 0 13.39 26.3 55.13 3.6 726.62 705.78 699.2 669.05 13.38 14.24 14.5 15.60 0 19.98 26.3 55.35 Cubes 5 702.91 682.74 676.38 647.07 37.09 37.28 37.32 37.43 0 19.98 26.3 55.50 AR 250 243.45 241.12 231.38 0 0 0 0 0 6.55 8.88 18.63 3.6 245.48 238.44 236.22 226.03 4.52 4.81 4.90 5.27 0 6.75 8.88 18.70 Paste Nuggets 5 237.47 230.66 228.51 218.60 12.53 12.59 12.51 12.65 0 6.75 8.88 18.75 Table 10: Mix Proportions for Cement E Cement (g) Added Gypsum (g) Added KOH (g) Alk SO 3 AR 1.5 2 3.8 AR 1.5 2 3.8 AR 1.5 2 3.8 AR 500 489 483.25 464.75 0 0 0 0 0 11.0 16.75 35.25 3.6 488.39 476.98 471.10 452.03 11.61 11.97 12.15 12.67 0 11.05 16.35 35.30 Bars 5 472.45 461.11 455.93 436.93 27.55 27.54 27.52 27.42 0 11.35 16.55 35.65 AR 740 723.72 715.21 687.83 0 0 0 0 0 16.28 24.79 52.17 3.6 722.81 705.92 697.23 669.01 17.19 17.72 17.98 18.75 0 16.35 24.79 52.24 Cubes 5 699.23 682.44 674.77 646.66 40.77 40.76 40.74 40.58 0 16.80 24.49 52.76 AR 250 244.5 241.63 232.38 0 0 0 0 0 5.5 8.38 17.63 3.6 244.19 238.49 N/A* 226.02 5.81 5.99 N/A* 6.33 0 5.53 N/A* 17.65 Paste Nuggets 5 236.22 230.56 227.96 218.47 13.78 13.77 13.76 13.71 0 5.68 8.28 17.83 *Not mixed Table 11: Mix Proportions for Cement ERD07 Cement (g) Added Gypsum (g) Added KOH (g) Alk SO 3 AR 1.5 2 AR 1.5 2 AR 1.5 2 AR 500 495.17 N/A* 0 0 N/A 0 5.83 3.6 N/A N/A N/A N/A N/A N/A N/A N/A N/A Bars 5 482.22 477.07 N/A 17.78 17.98 N/A 0 4.95 AR 740 N/A N/A 0 N/A N/A 0 N/A N/A 3.6 737.26 729.96 721.53 2.24 2.79 3.38 0 7.25 15.10 Cubes 5 713.69 705.52 N/A 26.31 26.84 N/A 0 7.84 Mixes were not carried out due to shortage in material.

PAGE 82

62 Table 12: Mix Proportions for Cement MH3 Cement (g) Added Gypsum (g) Added KOH (g) Alk SO 3 AR 1.5 2 AR 1.5 2 AR 1.5 2 AR 500 494.09 488.89 0 0 0 5.91 11.11 3.6 494.13 487.67 482.35 5.87 6.23 6.51 0 6.10 11.14 Bars 5 478 471.86 466.54 22 22.14 22.25 0 6 11.21 AR 740 731.26 723.56 0 0 0 8.74 16.44 3.6 731.31 721.76 713.63 8.69 9.21 9.65 0 9.03 16.72 Cubes 5 707.44 698.35 690.47 32.56 32.77 32.93 0 8.88 16.60 Table 13: Mix Proportions for Cement MH4 Cement (g) Added Gypsum (g) Added KOH (g) Alk SO 3 AR 1.5 2 AR 1.5 2 AR 1.5 2 AR 500 493.48 488.30 0 0 0 0 6.52 11.70 3.6 488.05 481.07 475.87 11.95 12.16 12.32 0 6.77 11.81 Bars 5 472.13 465.37 460.20 27.87 27.86 27.84 0 6.77 11.95 AR 740 730 722.53 0 0 0 0 9.70 17.46 3.6 722.32 711.86 704 17.68 18 18.24 0 10.14 17.76 Cubes 5 698.75 688.33 680.59 41.25 41.23 41.21 0 10.43 18.2

PAGE 83

63 CHAPTER 5 ROLE OF SULFATES This chapter presents the results on the effect of increasing the sulfate content (SO 3 ) of the cements on expansion potential and compressive strength. The effect of sulfate is characterized through different analytical methods; such as XRD, SEM, chemical analysis of the hydrating solution (Ionic species concentration) and heat of hydration calorimetry. 5.1 Expansion Results Figures 5 through 9 depict the effect of increasing the sulfate content on the expansion behavior of all cements at as-receiv ed alkali content. As mentioned in chapter 3, the mortar prisms used for the expansion results were prepared in accordance with ASTM C-1038. All samples were stored in lime solution for th e duration of the experiments.

PAGE 84

0.000.020.040.060.080306090120150180Age (days)Expansion (%) 5-AR 3.6-AR AR-AR Figure 5: Effect of Sulfate Content on Expansion Behavior for Cement E It is quite evident that there is an increase in expansion with the increasing sulfate content. For the AR cases, where sulfate content was between 2.58% and 3.47%, expansion did not exceed 0.03%, with the highest reported for cement ERD07. This could be due to the fact that cement ERD07 has the highest SO 3 content, although the values for all cements were within close range. 0.000.020.040.060.080306090120150180Age (days)Expansion (%) 5.0-AR 3.6-AR AR-AR Figure 6: Effect of Sulfate Content on Expansion Behavior for Cement C 64

PAGE 85

0.000.020.040.06 0 08 0306090120150180Age (days)Expansion (%) 5.0-AR AR-AR Figure 7: Effect of Sulfate Content on Expansion Behavior for Cement ERD07 0.000.020.040.060.080306090120150180Age (days)Expansion (%) 5.0-AR 3.6-AR AR-AR Figure 8: Effect of Sulfate Content on Expansion Behavior for Cement MH3 65

PAGE 86

0.000.020.040.060.080306090120150180Age (days)Expansion (%) 5.0-AR 3.6-AR AR-AR Figure 9: Effect of Sulfate Content on Expansion Behavior for Cement MH4 The increase in the expansion values from the AR case to the 3.6% case was almost negligible. There was no 3.6% case for cement ERD07 since the AR SO 3 content was close to that value. The increase in expansion values in the 5% case is clear; the values are almost double those of the AR case for cements C, E, MH3 and MH4. However, all the values were under .05%, up to the ages reported, in the cases of C and E and ERD07, and just slightly above that in case of cement MH3. The increase from the AR case to the 5% case in cement ERD07 was not much; this could be due to the fact that the original cement has high sulfate content. Cement MH3 recorded the highest expansion, followed by cement C. These two cements have high C 3 S content and low C 3 A content. Cement ERD07 showed the lowest expansion value at 5% sulfate. This could be due to its high C 3 A content. MH4 has a high C 3 A content as well, but the C 3 S content is high. It appears that cements with high C 3 A content expand less for the same sulfate contents (for the sulfate levels studied here). 66

PAGE 87

5.2 Compressive Strength Figures 10 through 14 depict the effect of increasing the sulfate content on the compressive strength, without increasing the alkali content. 0200040006000800010000050100150200Age (days)Strength (psi) 5.0-AR 3.6-AR AR-AR Figure 10: Effect of Sulfate Content on Strength Gain for Cement E Increasing the sulfate content to 3.6% did not result in any major changes in strength gain for cements E, C, MH3 and MH4. At the 5% level, the strength gain was slowed down in the case of these four cements; however, beyond 7 days, the strength values seem to be close for all three sulfate levels up to the ages reported here. In the case of cement ERD07, we can see that the 5% case showed a slightly higher strength up to 60 days. Beyond this age, it appears that the sulfate content of 3.6% generated highest strength. 67

PAGE 88

0200040006000800010000050100150200Age (days)Strength (psi) 5.0-AR 3.6-AR AR-AR Figure 11: Effect of Sulfate Content on Strength Gain for Cement C 0200040006000800010000050100150200Age (days)Strength (psi) 5.0-AR 3.6-AR AR-AR Figure 12: Effect of Sulfate Content on Strength Gain for Cement ERD07 68

PAGE 89

0200040006000800010000050100150200Age (days)Strength (psi) 5.0-AR 3.6-AR AR-AR Figure 13: Effect of Sulfate Content on Strength Gain for Cement MH3 02000400060008000 10000 050100150200Age (days)Strength (psi) 5.0-AR 3.6-AR AR-AR Figure 14: Effect of Sulfate Content on Strength Gain for Cement MH4 In order to go about explaining the observations, two of the cements, namely C and E were chosen for further analysis using various analytical methods. These methods are: XRD, Heat of hydration Calorimetry, Chemical analysis of cement solution and SEM. 69

PAGE 90

5.3 XRD Results 10 20 30 40 50 0 100 400 900 0 100 400 900 0 100 400 900 C2S E-ARC3S E-3.6 E-5EBMECHEETiO2BMCHC3SC3SCHTiO2TiO2BMCHC3STiO2C2SGGCounts Diffraction Angle (2) Figure 15: XRD of Hydration Products @ 24 hours for Cement E C 3 S= Tricalcium Silicates, C 2 S=Dicalcium Silicates, TiO 2 ,=Rutile CH=Ca (OH) 2 ,, BM=C 4 AF, E=Ettringite, G=gypsum 10 20 30 40 50 0 100 400 900 0 100 400 900 0 100 400 900 GBMEECHETiO2BMCHCountsCHC3STiO2C2SCHBMTiO2TiO2C3SC2SC3SCHC-ARC-3.6 C-5 Diffraction Angle (2) Figure 16: XRD of Hydration Products @ 24 hours for Cement C C 3 S= Tricalcium Silicates, C 2 S=Dicalcium Silicates, TiO 2 ,=Rutile CH=Ca (OH) 2 ,, BM=C 4 AF, E=Ettringite, G=gypsum 70

PAGE 91

71 In order to explain the findings on expa nsion and strength, phase analysis was performed using XRD. The hydration and hydration products were monitored through XRD. The XRD work was performed on the pa ste samples. Figures 15 and 16 show the hydration products at 24 hours for all three su lfate levels (AR-AR,3.6-AR and 5-AR) for cements E and C respectively. It can be s een that the gypsum p eak persisted up to 24 hours for 5% mix; however, it disappeared by three days. It can also be seen that ettringite was present as ear ly as 8 hours in all three ca ses. Additionally, the x-ray patterns reveal differences in other phases such as C 3 S and CH. Certain phases were selected for further quantification. Selecti on was based on the significance of those phases to ISA phenomenon. Quantification analysis for ettringite, alite, C 4 AF, CH and amorphous contetnt. This is presented in Figures 17-26. Ettringite Formation: Ettringite was quatinfied usi ng the two different methods of quantification; namely, internal standard (semi-quant ification using rutile as an internal strandard), and the Rietveld method. The results of both methods exhibited the same trend.The results of the internal standard method are presented here.

PAGE 92

00.20.4060120180Age (days)Ettringte-Ratio TiO2 AR-AR 3.6-AR 5-AR Figure 17: Effect of Sulfate Content on Ettringite Formation in Cement E 00.20.4060120180Age (days)Ettringte-Ratio TiO2 5-AR 3.6-AR AR-AR Figure 18: Effect of Sulfate Content on Ettringite Formation in Cement C It is clear by looking at Figures 17 and 18 that there is an increase in the amount of formed ettringite as the SO 3 content increased in both cements. The increase is believed to be due to the delayed release of the sulfate, as can be recalled from Figures 15 and 16, that the gypsum peak persisted longer in the 5% case. However, the amount of ettringite seemed to have leveled off after 28 days. It can be seen that the amount of ettringite formed, as well as the expansion levels, in case of cement C are higher than those in case 72

PAGE 93

of cement E, for all three sulfate levels. It is to be noted that the main mineralogical difference between these two cements is C 3 S content (see Chapter 3). It seems, based on the results presented here, that higher C 3 S content for the same C 3 A contentmay result in the formation of higher amounts of ettringite. Hydration of Alite Figures 19 and 20 show the effect of increasing the sulfate content on hydration of alite (C 3 S) for cements E and C, respectively. The rate of C 3 S hydration appears to be affected by the sulfate content of the cement. This appears to be true for cements E and C as depicted in Figures 19 and 20. Without any sulfate addition, the amounts of unreacted C 3 S was smallest, for both cements. Upon increasing the sulfate contents to 3.6% and 5%, this unreacted amount appears to increase. In case of cement C, the highest rate of hydration occurred in case AR-AR, and case 5.0-AR was the slowest. For cement E, case 3.6-AR was the slowest, however, this will be verified by other methods later (heat of hydration). The difference in the rate of alite hydration between all three sulfate cases was evident first 24 hours. Beyond 24 hours, differences appeared to be minimal. 010203040506070 80 01234567Age (days) C3S % Wt Ritveld 5.0-AR 3.6-AR AR-AR Figure 19: Effect of Sulfate Content on C 3 S Hydration in Cement E 73

PAGE 94

010203040506070 80 01234567A g e ( da y s ) C3S % Wt Ritveld 5.0-AR 3.6-AR AR-AR Figure 20: Effect of Sulfate Content on C 3 S Hydration in Cement C Brownmillerite Hydration (C 4 AF) : Figures 21 and 22 show the rate of hydration of C 4 AF for both cements (E and C) respectively. No trend could be established for the collected data. Both cements yielded differnt results, also, the results could not be correlated with any physical measurements; such as expansion or strength. 0.002.004.006.008.0010.0012.0014.0016.0001234567Age (days) BM % Wt-Rietveld 5-AR 3.6-AR AR-AR Figure 21: Effect of Sulfate Content on C 4 AF Hydration in Cement E 74

PAGE 95

024681012141601234567Age (days) BM % Wt-Rietveld 5.0-AR 3.6-AR AR-AR Figure 22: Effect of Sulfate Content on C 4 AF Hydration in Cement C Formation of Calcium Hydroxide: Figures 23 and 24 show the formation of calcium hydroxide. It is evident that the rate of formation decreased with increasing the sulfate content, for both cements. These results are in agreement with the data collected on alite hydration as discussed previously. The difference was more clear in the 5.0-AR case. As can be seen, the amount of the formed calcium hydroxide is lower at 7 days. Though the amount of alite hydrated at this time was almost the same in all three cases, calcium hydroxide formation was significantly less for the 5% sulfate case. This difference could be due the higher amount of ettringite formed in this case. Also, the amounts formed was higher in cement C than E. This is expected since cement C has a higher alite content. 75

PAGE 96

0.005.0010.0015.0020.0001234567Age (days) CH % Wt-Rietveld 5.0-AR 3.6-AR AR-AR Figure 23: Effect of Sulfate Content on Formation of Calcium Hydroxide in Cement E 0510152001234567Age (days) CH % Wt-Rietveld 5.0-AR 3.6-AR AR-AR Figure 24: Effect of Sulfate Content on Formation of Calcium Hydroxide in Cement C Formation of Amorphous Content: The formation of the amorphous content is attributed to the formation of the C-S-H gel, which is responsible for the strength. Figures 25 and 26 show the formation of amorphous content of cements E and C, respectively. It is clear that the rate of formation was slower in the 5.0-AR case, in both cements. The difference persisted up to 7 days; 76

PAGE 97

however, the difference at 7 days was more pronounced in cement E than cement C. This is consistent with compressive strength data shown in Figures 10 and 11. 01020304050607001234567Age (days)Amorphous Cont % Wt-Rietveld 5.0-AR 3.6-AR AR-AR Figure 25: Effect of Sulfate Content on Formation of Amorphous Content for Cement E 01020304050607001234567Age (days)Amorphous Cont % Wt-Rietveld 5.0-AR 3.6-AR AR-AR Figure 26: Effect of Sulfate Content on Formation of Amorphous Content for Cement C 5.4 Heat of Hydration Figures 27 and 28 show the heat of hydration curves for both cement E and C. The effect of sulfate content is quite clear. An increase in the sulfate content is 77

PAGE 98

accompanied by a delay in the first peak, which is attributed to C 3 A hydration. The delay was about 4 minutes. The same trend was observed in both cements; however differences existed in the amount of heat generated; with cement C lower than E. This could be due to difference in C 3 A content. 051015202530354000.10.20.30.40.50.6Time (hours)Rate of Heat Evolution (cal/g.h) 5.0-AR AR-AR Figure 27: Effect of Sulfate Content on Heat of Hydration for Cement E 051015202530354000.10.20.30.40.50.6Time(hrs)Rate of Heat Evolution (cal/g.h) 5.0-AR AR-AR Figure 28: Effect of Sulfate Content on Heat of Hydration for Cement C 78

PAGE 99

79 5.5 Ionic Species Concentrations The chemistry of the solution gives an id ea about the rate of release and the depletion of certain ions from the hydrating cement solution. This will be related to the formation of certain phases; for example th e rate of sulfate ion depletion will be correlated with the formation of ettringite. Sulfate Ions (SO 4 -2 ): Figures 29 and 30 show the su lfate ion concentrations. The same trend was observed in both cements. The values appear to be almost the same for all 3 cases between 30 minutes and 8 hours, although there is a difference in the sulfate content. This is due to the fact that the excess sulf ate in the 5% cases did not dissolve immediately, this is consistent with XRD patterns presented a bove in Figures 15 and 16; where the gypsum peak persisted up to 24 hours. The behavior after 8 hours is consistent with sulfate content; as the sulfate content increased the depletion times increased as well. The ettringite formation as was shown earlier, it did not follow exactly the trend of the sulfate ion depletion, the ettringite formation continue d for some times after the depletion of the sulfate from the solution, it continued up to 28 days in the 5.0-AR case for example. This may imply that the sulfate ions removed fr om the solution were concentrated in a different phase, i.e. C-S-H gel, which acted as a sulfate source subsequently.

PAGE 100

010002000300040005000600070000326496128160Age (hours)Conc. (mg/l) 5.0-AR 3.6-AR AR-AR Figure 29: Effect of Sulfate Content on SO 4 -2 Concentration for Cement E 010002000300040005000600070000326496128160Age (hours)Conc. (mg/L) 5.0-AR 3.6-AR AR-AR Figure 30: Effect of Sulfate Content on SO 4 -2 Concentration for Cement C Calcium ions (Ca +2 ): Figures 31 and 32 show the calcium ions concentrations for cements E and C. The trend is consistent with the CH formation trends as shown above by XRD. Comparing the AR-AR case to 5.0-AR case, the difference is clear, especially after 24 hours, where the values were lower in the 5.0-AR case. The difference persisted up to 7 days, where the 80

PAGE 101

values converged for all 3 cases. It is important to note that the higher values in the 5.0 case, could be in part due the higher gypsum content as well. 04008001200160020000326496128160Age (hours)Conc. (mg/l) 5.0-AR 3.6-AR AR-AR Figure 31: Effect of Sulfate Content on Calcium ions Concentration for Cement E 04008001200160020000326496128160A g e ( hours ) Conc. (mg/l) 5.0-AR 3.6-AR AR-AR Figure 32: Effect of Sulfate Content on Calcium ions Concentration for Cement C The curves for other ionic species ( such as potassium, sodium, silica and aluminum) were low and did not show any specific trends; therefore it was decided not to include them. 81

PAGE 102

5.6 Scanning Electron Microscopy (SEM) The SEM was used to compliment and corroborate some of the results obtained using the other analytical methods mentioned above. Using the fractured surfaces (in the secondary electron mode) givers an idea about the phases present, it also sheds some light on the morphology of certain phases; such as ettringite and the C-S-H gel. Studying polished-surfaces samples (in the back-scattered mode) gives an idea about the chemical make-up of certain phases. This sheds some light about the inclusion of some ions in certain phases. 5.6.1 Fractured Surfaces A B 2.5m Case 5.0-AR Case AR-AR Figure 33: SEM Micrographs for Cement E Figures 33 and 34 show some SEM micrographs for fractured surfaces of paste samples. It can be generally seen that there is no major differences in the appearance of the gel. 82

PAGE 103

A B 2.5m Case AR-AR Case 5.0-AR Figure 34: SEM Micrographs for Cement C 5.6.2 Polished Surfaces Back-scattered images coupled with EDS were used to study the chemical make-up of the C-S-H gel. The results were not conclusive, no trend could be established. Figure 35: SEM Micrograph on a Polished Section for Cement E, Case AR-AR 83

PAGE 104

84 CHAPTER 6 ROLE OF ALKALIS, PART I: EFFECT OF ALKALI SOURCE In this chapter the results of increasing the sulfate and the alkali levels using different sources are presented ; alkali sulfat e and alkali hydroxide (The experiments are outlined in Sec. 4.4). The effect will be characterized through different analytical methods; such as XRD, SEM, and chemistr y of the hydrating solution and heat of hydration. The results will be presented as to compare the effect of increasing both the alkali and sulfate using potassium sulfate (K 2 SO 4 )-case KS (alkali content=2.32% and sulfate content=5.0% by weight of cement ), increasing the alkali using potassium hydroxide (KOH) and increasing the sulfate using gypsum-case KH-CS (alkali content=2.0%, and sulfate content= 5.0% by weight of cement). Case 5.0-AR, increasing the sulfate level (only) to 5.0% using gypsum will be used as reference. 6.1 Expansion Results Figures 36 and 37 show the expansion be havior of all three cases mentioned above, for cements E and C respectively. It is clear that there is no significant difference in expansion values between all three cases. The expansion values ve ry close, with one exception in the case of cement E; the case KH-CS is slightly lower. These values are true however, up to the ages reported here. The expansion will continue to be monitored for any changes.

PAGE 105

0.000.010.020.030.040.050.060.070.0804080120160200240280320360Age (days)Expansion (%) 5.0-AR KS KH-CS Figure 36: Effect of Alkali Source on Expansion Behavior for Cement E 0.000.010.020.030.040.050.060.07 0 08 04080120160200240280320360Age (days)Expansion (%) 5.0-AR KS KH-CS Figure 37: Effect of Alkali Source on Expansion Behavior for Cement C As can be seen, at this level of sulfate content (5%), whether increasing the sulfate content using K2SO4 or gypsum, the result is the same, and increasing the alkali level to 2%, whether the source was potassium sulfate or potassium hydroxide, the result is the same. This is true for the ages reported here. 85

PAGE 106

6.2 Compressive Strength The effect of varying alkali source on compressive strength is presented in this section. Figures 38 and 39 indicate the influence of alkali source on compressive strength for cements E and C. The addition of alkali (at 5% sulfate level) resulted in an increase in the 7-day strength (compared to the 5.0-AR case). However, the values in cases KS and KH-CS did not seem to increase much after that, it almost leveled off. In the case of 5.0-AR, the value almost doubled between 7 days and 56 days, where the values seemed to have leveled off in the other two cases. The results indicate that increasing the alkali content decreases the late compressive strength for the two cements studied here, irrespective of the source of alkalis. 0200040006000800010000050100150200250300350400Age (days)Strength (psi) 5.0-AR KS KH-CS Figure 38: Effect of Alkali Source on Strength Gain for Cement E 86

PAGE 107

0200040006000800010000050100150200250300350400Age (days)Strength (psi) 5.0-AR KS KH-CS Figure 39: Effect of Alkali Source on Strength Gain for Cement C 6.3 XRD Results QXRD was used to study and monitor the hydration process, and the hydration products. Appendix B contains some of the XRD scans collected for the cases mentioned here. Some phases are considered in detail below. Ettringite Formation: As mentioned in chapter 4, the ettringite was quantified using the internal standard method. The results are shown in Figures 40 and 41. In the previous chapter a correlation established between the ettringite formation and the expansion values. Figures 40 and 41 show the amount of ettringite formed. The values are within same range for all three cases in both cements. There was a slight delay in the appearance of ettringite in case KH-CS in cement C; no ettringite was observed at 8 hours. The expansion trends correlate well with the expansion values. However, there was a drop in the compressive strength values in cases KS and KH-CS. This would be expected to result in larger expansion in these two cases. At this point, this has not happened; the values will be monitored at further ages for any changes in expansion behavior. 87

PAGE 108

00.20.4060120180Age (days)Ettringte-Ratio TiO2 5.0-AR KS KH-CS Figure 40: Effect of Alkali Source on Ettringite Formation for Cement E 00.2 0 4 060120180Age (days)Ettringte-Ratio TiO2 5.0-AR KS KH-CS Figure 41: Effect of Alkali Source on Ettringite Formation for Cement C Hydration of Alite: Looking at Figures 42 and 43, the effect of increasing the alkalis content is clear. The hydration of alite increased as compared to the 5.0AR case. The increase in the rate of hydration mostly occurred in the first 24 hours, the difference disappeared at 7 days. There is a difference in the trend between these cements; in cement C case KH-KS 88

PAGE 109

showed the highest hydration, where case KS showed the highest hydration in cement E. The reason for this difference is not clear. 010203040506070 80 01234567Age (days) C3S (% Wt)Ritvel d 5.0-AR KS KH-CS Figure 42: Effect of Alkali Source on C 3 S Hydration for Cement E 0102030405060708001234567Age (days) C3S (% Wt)-Ritveld 5.0-AR KS KH-CS Figure 43: Effect of Alkali Source on C 3 S Hydration for Cement C 89

PAGE 110

Brownmillerite Hydration (C 4 AF) : 0.002.004.006.008.0010.0012.0014.0016.0001234567Age (days) BM (% Wt)-Rietveld 5-AR KS KH-CS Figure 44: Effect of Alkali Source on C 4 AF Hydration for Cement E Figures 44 and 45 depict the hydration of Brownmillerite. Again, the trend was not clear. In cement E, the hydration of this phase was more rapid in cases KS and KH-CS than the 5.0-AR case, however, the difference was not very clear in cement C. This trend could not be established. No conclusive effects could be related to the expansion or strength behavior. 90

PAGE 111

024681012141601234567Age (days) BM (% Wt)-Rietvel d 5.0-AR KS KH-CS Figure 45: Effect of Alkali Source on C 4 AF Hydration for Cement C Formation of Calcium Hydroxide: 0.005.0010.0015.0020.0001234567Age (days) CH (% Wt)-Rietvel d 5.0-AR KS KH-CS Figure 46: Effect of Alkali Source on Formation of Calcium Hydroxide for Cement E Cases KH and KH-CS both showed higher calcium hydroxide formation for the first 7 days. This is consistent with hydration of alite as shown in Figures 42 and 43. Another possible reason for the higher calcium hydroxide is the faster precipitation due to the higher alkalinity of these two cases as compared to 5.0-AR case. 91

PAGE 112

051015 20 01234567Age (days) CH (% Wt)-Rietveld 5.0-AR KS KH-CS Figure 47: Effect of Alkali Source on Formation of Calcium Hydroxide for Cement C Formation of Amorphous Content: 01020304050607001234567Age (days)Amorphous Cont (% Wt)-Rietvel d 5.0-AR KS KH-CS Figure 48: Effect of Alkali Source on Formation of Amorphous Content for Cement E The amorphous content is an indication of the amount of C-S-H gel formed, which is responsible for the strength. The trends observed in Figures 48 and 49, could not be exactly correlated with the hydration of alite. 92

PAGE 113

0102030405060 70 01234567Age (days)Amorphous Cont (% Wt)-Rietveld 5.0-AR KS KH-KS Figure 49: Effect of Alkali Source on Formation of Amorphous Content for Cement C 6.4 Heat of Hydration Figures 50 and 51 show the heat of hydration profile for cements E and C for the cases studied here. Increasing the alkali level in both cases (KS and KH-CS) resulted in acceleration in the first peak as compared to case 5.0-AR. In cement C, the peak occurred at 13 min in 5.0-AR case, at 4.38 minutes in KH-KS case, and at 3.24 minutes in KS case. 93

PAGE 114

051015202530354000.10.20.30.40.50.6Time ( hours ) Rate of Heat Evolution (cal/g.h) 5.0-AR KS KH-KS Figure 50: Effect of Alkali Source on Heat of Hydration for Cement E 051015202530354000.10.20.30.40.50.6Time ( hrs ) Rate of Heat Evolution(cal/g.h) 5.0-AR KS KH-CS Figure 51: Effect of Alkali Source on Heat of Hydration for Cement C 94 The same trends were observed in cement E, the peak occurred at 5.76 minutes in case 5.0-AR, at 3.24 minutes in case KH-CS, and at 2.67 minutes in case KS. Also, there was an increase in the rate of heat generated with the increase in alkali content in both cases. The increase in case KH-CS was not dramatic, however, the rate almost doubled in case KS, although both cases had almost the same alkali content. This could be due to the fact that syngenite formed and persisted for at least the first 8 hours of hydration in case

PAGE 115

95 KH-CS, which could have contributed to slowing the reaction down. The amount of heat was higher in cement E than C, again this could be due to higher C 3 A content. 6.5 Ionic Species Concentrations The chemistry of the solution gives an idea about the rate of release and the depletion of certain ions from the hydrating cement solution. This will be related to the formation of certain phases; for example th e rate of sulfate ion depletion will be correlated with the formation of ettringite. Sulfate Ions (SO 4 -2 ): Sulfate ions concentration pr ofiles are shown in Figures 52 and 53 for cements E and C respectively. The trends observed in both cemen ts are identical. A lthough, all three cases had the same sulfate content, it is clear that the concentration was highest in KS case, this is expected since K 2 SO 4 is readily soluble. In case KH-CS, the added gypsum reacted with KOH to form syngenite, this could have delayed the release of sulfate ions into solution. This was confirmed by XRD (see A ppendix B). The sulfate ions were released faster and depleted faster in case KS than the other two cases. The concentration in case 5.0-AR had always the lowest va lue. This implies th at ions were rel eased into solution and depleted at a close rate. The amount of ettringite formed did not reflect this huge difference in the availability of sulfate ions. This could be due to two factors; first, the determining factor was the availability of the aluminum ions, which was the same in all three cases (See Figures 54 and 55). Second, there was a compet ition for the sulfate ions from another phase; (C-S-H gel), which is believ ed to form at a higher rate (at least early on in the reaction) due to higher alkalinity. It is therefore expected that in cases KS, and

PAGE 116

KH-CS, the gel will contain larger amounts of sulfates. It appears that the observed difference in sulfate ions concentration is due to combination of both effects; however, the second effect could not be verified. 010002000300040005000600070000326496128160Age (hours)Conc. (mg/l) 5.0-AR KS KH-CS Figure 52: Effect of Alkali Source on SO 4 -2 Concentration for Cement E There are some minor differences in the SO 4 -2 concentrations recorded between the same cases for cements E and C, although the sulfate contents are the same in both. This could be due to the difference in the solubility of the original sulfates for each cement. 010002000300040005000600070000326496128160Age (hours)Conc. (mg/l) 5.0-AR KS KH-CS Figure 53: Effect of Alkali Source on SO 4 -2 Concentration for Cement C 96

PAGE 117

0.00.51.01.52.00326496128160Age (hours)Conc. (mg/l) 5.0-AR KS KH-CS Figure 54: Effect of Alkali Source on Aluminum Ions Concentration for Cement E 0.00.51.01.52.00326496128160Age (hours)Conc. (mg/l) 5.0-AR KS KH-CS Figure 55: Effect of Alkali Source on Aluminum Ions Concentration for Cement C Calcium Ions Concentration: Figures 56 and 56 show calcium ion concentration for cements E and C, respectively. The trend observed is the same in both cements. As we can see the values increase between 30 minutes and 4 hours, and then decrease until 8 hours. This is an indication of the dormant period 97

PAGE 118

040080012001600 2 000 0326496128160Age (hours)Conc. (mg/l) 5.0-AR KS KH-KS Figure 56: Effect of Alkali Source on Calcium Ions Concentration for Cement E The values start to increase again after 8 hours, during the acceleration stage of hydration. The values start to decrease after that, indicating setting and precipitation of calcium hydroxide. Based on these two facts, the trends are consistent with the data presented earlier (heat of hydration and calcium hydroxide formation). It can be seen that case KS showed the lowest calcium ion concentration with case KH-CS being close, case 5.0-AR showed higher values. As mentioned elsewhere, the added alkalis appear to accelerate the hydration process, and possibly gel nucleation. It also reduces the solubility of calcium hydroxide, which was evident from the calcium hydroxide formation shown earlier. 98

PAGE 119

04008001200160020000326496128160Age (hours)Conc. (mg/l) 5.0-AR KS KH-CS Figure 57: Effect of Alkali Source on Calcium Ions Concentration for Cement C The implication of this effect on strength gain, and eventually, durability of such system will be discussed in a later chapter. Potassium Ions Concentration: Potassium ions concentrations are shown in Figures 58 and 59 for cements E and C respectively. As expected, the values are higher for cases KS and KH-CS than AR. This was really a test for the preparation and analytical techniques. Case KS for cement E is a slightly higher than Case KH-CS, unlike the values for cement C. A possible explanation is that cement E has lower as-received sulfate content, so in order to achieve the 5% level, more K 2 SO 4 had to be added. 99

PAGE 120

010002000300040005000600070000326496128160Age (hours)Conc. (mg/l) 5.0-AR KS KH-CS Figure 58: Potassium Ions Concentration for Cement E 010002000300040005000600070000326496128160Age (hours)Conc. (mg/l) 5.0-AR KS KH-CS Figure 59: Potassium Ions Concentration for Cement C The results here indicate that potassium ions are not really tied in any phase. They are adsorbed possibly by the gel. However, potassium ions were not detected in the SEM/EDS work. 100

PAGE 121

6.6 Scanning Electron Microscopy (SEM) 6.6.1 Fractured Surfaces Figure 60 shows SEM micrographs for fractured surfaces of paste samples. Comparing these two micrographs to the micrograph shown earlier in Fig. 34-A, it can be generally indicated that there is no major differences as far as the appearance of C-S-H gel. The general conclusion from SEM work on fractured surfaces of paste samples is that at this sulfate level of 5% and alkali level of 2%, regardless of the source of alkalis, gel appeared to be similar. A B Figure 60: SEM Micrographs, Cement E, A) Case KH-CS, B) Case KS 6.6.2 Polished Surfaces Back-scattered imaging with EDS was used to study the chemical make-up of the gel. Again, the results were not conclusive. Figure 61 shows a micrograph for a polished section for case 5.0-AR for cement C. 101

PAGE 122

A. Cement C, Case 5.0-AR B. Cement E, Case 5.0-AR Figure 61: SEM Micrograph on a Polished Section and EDX Spectra 102

PAGE 123

103 CHAPTER 7 ROLE OF ALKALIS, PART II Chapter 5 presented the role of increas ing the sulfates. Chapter 6 presented the effect of increasing of the alkali content us ing two different forms of alkalis; namely potassium hydroxide and potassium sulfates. Th e purpose of this study is to investigate the influence of the alkali form, and as fr om the results presented, and up to the age studied here, the form of alkali did not resu lt in significant difference. However, variation in alkali level (was also accompanied by variation in sulfate level as well). In this chapter potassium hydroxide (KOH) will be used to vary the alkali content of cements, and gypsum will be used to vary the sulfate level. This will allow the st udy of a wide range of both contents at constant valu es. A detailed list of the case s studied and th eir labeling method can be found in chapter 4. As in the two preceding chapters, the effect of both variables on durability will be characterized through different analytical methods; XRD, SEM, heat of hydration, chemical analys is, surface area measurements and optical microscopy when necessary. 7.1 Expansion Results Figures 62 through 67 show the expansi on patterns for both cements E and C at all doping levels. It was shown in chapter 5 that, with no al kali addition, the expansion increased with increasing sulfate content. Fo r cement E, the expansion for the 5-AR case was just above .04%, and about .016 for the AR-AR case, case 3.6-AR showed an

PAGE 124

expansion of just above .018%. Expansion values for cement C for the same cases were slightly higher; however, it was within the same range. These maximum values were attained at 56 days, and leveled off thereafter 0.000.200.400.600.80 1 00 04080120160200240280320360Age (days)Expansion (%) AR-3.8 AR-2 AR-1.5 AR-AR Figure 62: Effect of Alkali on Expansion Behavior for Cement E-SO 3 =AR 0.000.200.400.600.801.0004080120160200240280320360Age (days)Expansion (%) 3.6-3.8 3.6-2.0 3.6-1.5 3.6-AR Figure 63: Effect of Alkali on Expansion Behavior for Cement E-SO 3 =3.6% 104

PAGE 125

Increasing the alkali level to 1.5% did not really impact the expansion behavior much. The values, for the three sulfate levels for both cements, were almost the same for ages up to 360 days and beyond. As a matter fact, the values showed some decrease, the decrease was more noticeable in the case of cement E than cement C, but for all practical purposes, they will be considered unchanged. 0.000.200.400.600.801.0004080120160200240280320360Age (days)Expansion (%) 5.0-3.8 5.0-2.0 5.0-1.5 5.0-AR Figure 64: Effect of Alkali on Expansion Behavior for Cement E-SO 3 =5.0% 0.000.200.400.600.801.0004080120160200240280320360Age (days)Expansion (%) AR-3.8 AR-2.0 AR-1.5 AR-AR 105 Figure 65: Effect of Alkali on Expansion Behavior for Cement C-SO 3 =AR

PAGE 126

0.000.200.400.600.801.0004080120160200240280320360Age (days)Expansion (%) 3.6-3.8 3.6-2.0 3.6-1.5 3.6-AR Figure 66: Effect of Alkali on Expansion Behavior for Cement C-SO 3 =3.6% 106 0.000.200.400.600.801.0004080120160200240280320360Age (days)Expansion (%) 5.0-3.8 5.0-2.0 5.0-1.5 5.0-AR Figure 67: Effect of Alkali on Expansion Behavior for Cement C-SO 3 =5.0% Increasing the alkali level to 2%, again, did not result in any dramatic changes to the expansion behavior, up to the ages reported here. Again, the expansion at this alkali level is to be considered unchanged as compared to the no-alkali cases.

PAGE 127

107 The situation is completely different when th e alkali level is incr eased to 3.8%. At 5% sulfate level, for cement E, the expansion values were almost the same as the other alkali cases, 5-AR, 5-1.5, and 5-2, up to 56 days, howev er, the expansion started to increase rapidly after that, to reach values close to 1.0%, and the bars s howed cracking by 180 days (Figure 68). The same trend occurred in cement C, with one exception; the expansion values started to depart from the ot her cases at an earlier age; just after 28 days. It reached a value close to 1.0% at 91 days, where the bars started to show some cracking. The expansion behavior followed th e same trend at the 3.6% sulfate level, except the damage was slightly delayed. In the case of cement E, the expansion values were close to the other cases up to an age of 120 days; then rapid expansion occurred. Expansion values of 1.0% accompanied by mortar cracking were reported at180 days. For cement C, the expansion started to increas e rapidly at 91 days. It reached a value of about 0.7% by 120 days and started to show some cracks. For AR condition, similar trends to the 3.6% and 5% sulfate conten ts were observed in cement C, except the damage did not occur until 270 days. The expansion value in this case reached approximately 0.05% at 56 days, and remained constant between 56 and 180 days. It started to increase rapidly af ter that, with eventual crack ing at 270 days. In case of cement E, the expansion value leveled at a value of just above .04% after 150 days. The expansion values for cement E (for the ages reported) corresponds we ll with the values reported by Shayan et al (45).

PAGE 128

For case AR-3.8 in cement C, the expansion behavior is quite different from cement E. The difference between cements C and E is the C 3 S content, where it is much higher in cement C; this is believed to be the cause of the observed trends. Figure 68: Cracked Mortar Bar of Case 5-3.8 for Cement E at 180 days 7.2 Compressive Strength The effect of the alkali additions on the compressive strength was more pronounced and more immediate, at all alkali levels. As can be seen from the results presented (Figures 69-74); increasing the SO 3 content (without any alkali additions) did not greatly impact the ultimate strength of the mortar cubes (at least up to the ages reported here). The most discernable difference is the delay in strength gain in the case 5-AR, in both cements. This was established in chapter 5, and it is reiterated here for comparison. Increasing the alkali content, and at each sulfate level resulted in reduction in the compressive strength at all ages. An exception is for the increase in 7-day strength in the cases 5-1.5 and 5-2 as compared to case 5-AR, for both cements. 108

PAGE 129

0200040006000800010000050100150200250300350400Age (days)Strength (psi) AR-3.8 AR-2 AR-1.5 AR-AR Figure 69: Effect of Alkali on Strength Gain for Cement E-SO 3 =AR 0200040006000800010000050100150200Age (days)Strength (psi) 3.6-3.8 3.6-2 3.6-1.5 3.6-AR Figure 70: Effect of Alkali on Strength Gain for Cement E-SO 3 =3.6% It was reported in the literature (54) that increasing the alkali content resulted in an increase in early strength, and a decrease in ultimate strength. In the current investigation this was only true in the 5% sulfate case. In this case, as presented earlier, the sulfate addition delayed the strength gain. Addition of the alkalis counter that effect, by possibly increasing the rate of the hydration at early ages, and increasing the gel nucleation. 109

PAGE 130

0200040006000800010000050100150200250300350400Age (days)Strength (psi) 5.0-3.8 5.0-2 5.0-1.5 5.0-AR Figure 71: Effect of Alkali on Strength Gain for Cement E-SO 3 =5.0% 0200040006000800010000050100150200250300350400A g e ( da y s ) Strength (psi) AR-3.8 AR-2 AR-1.5 AR-AR Figure 72: Effect of Alkali on Strength Gain for Cement C-SO 3 =AR 110

PAGE 131

0200040006000800010000050100150200250300350400Age (days)Strength (psi) 3.6-3.8 3.6-2 3.6-1.5 3.6-AR Figure 73: Effect of Alkali on Strength Gain for Cement C-SO 3 =3.6% 0200040006000800010000050100150200250300350400A g e ( da y s ) Strength (psi) 5.0-3.8 5.0-2 5.0-1.5 5.0-AR Figure 74: Effect of Alkali on Strength Gain for Cement C-SO 3 =5.0% Strength loss was about 25% in the 1.5% and 2% alkali cases, and more than 50% in the 3.8% alkali cases. Furthermore, there was a regression in the strength values in the 3.8 alkali cases. In case 5-3.8 for cement C, the cubes showed cracks at 91 days, the strength at that time dropped by 1/3 of the value it attained at 60 days. The same phenomenon was observed in cement E, except the strength regression occurred at 180 days. 111

PAGE 132

7.3 XRD Results QXRD was used to assess the hydration products. Ettringite formation was studied and quantification results reported here were collected using internal standard method. Other phases were studied and quantified using the Rietveld method. Figure 75 shows XRD scans for case 5-3.8 for cement E at different ages, note the absence of ettringite at 8 hours. Appendix B contains some of the XRD scans collected for the cases mentioned here. 10 20 30 40 50 ts 0 100 400 900 0 100 400 900 0 100 400 900 ECHTiO2 A B CKSC3SKSCHBMEETiO2CHC2STiO2CHC3SCHC3SC2SKSBMCounts Diffraction Angle (2) Figure 75: XRD of hydration products for Case E-5-3.8 A) @ 8 hours, B) @ 3 days, C) @ 56 days C 3 S= Tricalcium Silicates, C 2 S=Dicalcium Silicates, TiO 2 ,=Rutile CH=Ca (OH) 2 ,, BM=C 4 AF, E=Ettringite, G=gypsum, KS= Arcanite Ettringite Formation: Figures 76 through 81 depict the amount of ettringite formed for cements E and C. It is evident that the amount of ettringite formed increases with sulfate content of cement. 112

PAGE 133

Increasing the alkali content did not affect the total amount of ettringite formed ultimately. The addition of the alkalis, however, and depending on the sulfate content, resulted in delay of ettringite formation as detected by XRD. This implies ettringite instability at higher alkalinity. The delay was most in the AR-3.8 cases, ettringite instability at higher alkalinity was reported in the literature (1). This phenomenon depends on the sulfate content. At 5% SO 3 ettringite was detected as early as 8 hours in cases 5-3.8. The instability of ettringite has two implications; first: late formation of ettringite, would result in its formation after the setting and hardening of concrete. Second, the concentration of sulfate in other phases will be affected. These two effects and their impact on concrete durability will be discussed later. 00.20.4060120180Age (days)Ettringte-Ratio TiO2 AR-3.8 AR-2 AR-1.5 AR-AR Figure 76: Effect of Alkali on Ettringite Formation for Cement E-SO 3 =AR 113

PAGE 134

00.20.4060120180Age (days)Ettringte-Ratio TiO2 3.6-3.8 3.6-1.5 3.6-AR Figure 77: Effect of Alkali on Ettringite Formation for Cement E-SO 3 =3.6% 00.20.4060120180Age (days)Ettringte-Ratio TiO2 5-3.8 5-2 5-1.5 5-AR Figure 78: Effect of Alkali on Ettringite Formation for Cement E-SO 3 =5.0% 114

PAGE 135

00.20.4060120180Age (days)Ettringte-Ratio TiO2 AR-3.8 AR-2 AR-1.5 AR-AR Figure 79: Effect of Alkali on Ettringite Formation for Cement C-SO 3 =AR 00.20.4060120180Age (days)Ettringte-Ratio TiO2 3.6-3.8 3.6-2 3.6-1.5 3.6-AR Figure 80: Effect of Alkali on Ettringite Formation for Cement C-SO 3 =3.6% 115

PAGE 136

00.20.4060120180Age (days)Ettringte-Ratio TiO2 5.0-3.8 5.0-2 5.0-1.5 5.0-AR Figure 81: Effect of Alkali on Ettringite Formation for Cement C-SO 3 =5.0% Hydration of Alite: Figures 82-87 show the hydration of C 3 S using the Rietveld method on cement paste. The addition of alkali (without any sulfate addition) seems to slow the rate of hydration of C 3 S during early stages of reaction; as evident from the data depicted in Figures 82 and 85. The trend is not entirely clear when considering all cases of sulfate and alkalis additions. However, considering the extreme doping cases for both cements. It can be seen that the addition of alkalis counteract the effect of the sulfates. This is evident from the increased rate of hydration in the 5-3.8 cases as compared to the 5.0-AR cases. This is true for both cements, see Figures 84 and 87. The observation here will be verified by the heat of hydration curves (to be shown later). 116

PAGE 137

0102030405060708001234567A g e ( da y s ) C3S % Wt Ritveld AR-3.8 AR-2 AR-1.5 AR-AR Figure 82: Effect of Alkali on C 3 S Hydration for Cement E-SO 3 =AR 0102030405060708001234567A g e ( da y s ) C3S % Wt Ritveld 3.6-3.8 3.6-1.5 AR-AR Figure 83: Effect of Alkali on C 3 S Hydration for Cement E-SO 3 =3.6% 117

PAGE 138

0102030405060708001234567A g e ( da y s ) C3S % Wt Ritveld 5.0-3.8 5.0-2 5.0-1.5 5.0-AR Figure 84: Effect of Alkali on C 3 S Hydration for Cement E-SO 3 =5.0% 0102030405060708001234567Age (days) C3S % Wt Ritveld AR-3.8 AR-2 AR-1.5 AR-AR Figure 85: Effect of Alkali on C 3 S Hydration for Cement C-SO 3 =AR 118

PAGE 139

0102030405060708001234567Age (days) C3S % Wt Ritveld 3.6-3.8 3.6-2 3.6-1.5 3.6-AR Figure 86: Effect of Alkali on C 3 S Hydration for Cement C-SO 3 =3.6% 0102030405060708001234567Age (days) C3S % Wt Ritveld 5.0-3.8 5.0-2 5.0-1.5 5.0-AR Figure 87: Effect of Alkali on C 3 S Hydration for Cement C-SO 3 =5.0% Formation of Calcium Hydroxide: Figures 88-93 show calcium hydroxide formation, as quantified on cement paste samples using the Rietveld method. It is clear that calcium hydroxide amounts increased with the increase in alkali content, for all sulfate levels. This trend was definitely evident in the first 24 hours. The trend is somewhat different after 24 hours. It has to be remembered that the amounts shown here may contain some errors due to leaching of calcium 119

PAGE 140

hydroxide during samples preparation. The data presented here does not imply a faster hydration rate for higher alkali cases. From previous results, there was no clear evidence that higher alkali cases increased C 3 S hydration. On the contrary, the conclusion was that alkalis did slow the hydration process. 0.005.0010.0015.0020.0001234567Age (days) CH % Wt-Rietveld AR-3.8 AR-2 AR-1.5 AR-AR Figure 88: Effect of Alkali on Calcium Hydroxide Formation for Cement E-SO 3 =AR 0.005.0010.0015.00 20 00 01234567Age (days) CH % Wt-Rietveld 3.6-3.8 3.6-1.5 3.6-AR Figure 89: Effect of Alkali on Calcium Hydroxide Formation for Cement E-SO 3 =3.6% The reason for higher calcium hydroxide quantities is believed to be reduced solubility of calcium hydroxide in the high alkali solution (19). 120

PAGE 141

0.005.0010.0015.0020.0001234567Age (days) CH % Wt-Rietveld 5.0-3.8 5.0-2 5.0-1.5 5.0-AR Figure 90: Effect of Alkali on Calcium Hydroxide Formation for Cement E-SO 3 =5.0% 0.005.0010.0015.0020.0001234567Age (days) CH % Wt-Rietveld AR-3.8 AR-2 AR-1.5 AR-AR Figure 91: Effect of Alkali on Calcium Hydroxide Formation for Cement C-SO 3 =AR 121

PAGE 142

0.005.0010.0015.0020.0001234567Age (days) CH % Wt-Rietveld 3.6-3.8 3.6-2 3.6-1.5 3.6-AR Figure 92: Effect of Alkali on Calcium Hydroxide Formation for Cement C-SO 3 =3.6% 0.005.0010.0015.0020.0001234567Age (days) CH % Wt-Rietveld 5.0-3.8 5.0-2 5.0-1.5 5.0-AR Figure 93: Effect of Alkali on Calcium Hydroxide Formation for Cement C-SO 3 =5.0% Formation of Amorphous Content: Figures 94-99 show the amorphous content as quantified by the Rietveld analysis. There are a lot of variations, especially after the first 8-24 hours of hydration. From Figures 94 and 97, it can be seen that increasing alkali content resulted in less amorphous content. Other researcher (29) reported that alkalis increase gel production. This will be visited in a later chapter. However, considering the cases with high sulfate contents (Figures 106 122

PAGE 143

and 109), it can be seen that alkalis result in higher amorphous content early on. The reason for this is the added sulfates retard rate of hydration, as was established in chapter 5. 01020304050607001234567Age (days)Amorphous Cont % Wt-Rietveld AR-3.8 AR-2.0 AR-1.5 AR-AR Figure 94: Effect of Alkali on Amorphous Content Formation for Cement E-SO 3 =AR 01020304050607001234567Age (days)Amorphous Cont % Wt-Rietveld 3.6-3.8 3.6-1.5 3.6-AR Figure 95: Effect of Alkali on Amorphous Content Formation for Cement E-SO 3 =3.6% 123

PAGE 144

01020304050607001234567Age (days)Amorphous Cont % Wt-Rietveld 5.0-3.8 5.0-2.0 5.0-1.5 5.0-AR Figure 96: Effect of Alkali on Amorphous Content Formation for Cement E-SO 3 =5.0% 0102030405060708001234567Age (days)Amorphous Cont % Wt-Rietveld AR-3.8 AR-2.0 AR-1.5 AR-AR Figure 97: Effect of Alkali on Amorphous Content Formation for Cement C-SO 3 =AR 124

PAGE 145

01020304050607001234567Age (days)Amorphous Cont % Wt-Rietveld 3.6-3.8 3.6-2.0 3.6-1.5 3.6-AR Figure 98: Effect of Alkali on Amorphous Content Formation for Cement C-SO 3 =3.6% 01020304050607001234567Age (days)Amorphous Cont % Wt-Rietveld 5.0-3.8 5.0-2.0 5.0-1.5 5.0-AR Figure 99: Effect of Alkali on Amorphous Content Formation for Cement C-SO 3 =5.0% 7.4 Heat of Hydration Looking at the heat of hydration results (Figures 100-103), it is suggested that at 3.8% alkalis content, the rate of C 3 A hydration is accelerated. This is evident when considering the first peak in cement E for example. The peak appeared at 2.5 minutes in case AR-3.8 as opposed to 4.5 minutes in case AR-AR. Additionally, the rate of heat generation in the first case was more than double that in the second case. In case AR-2.0 125

PAGE 146

for cement E, the behavior is not much different from the case AR-AR. These values are consistent with the flow values recorded. The flow values ( Table 14) were highest in AR-AR case, and lowest in AR-3.8 case. Similar trends were observed for cement C, with the exception of case AR-2.0, where the peak occurred earlier than the AR-AR case. 051015202530354000.10.20.30.40.50.6Time (hours)Rate of Heat Evolution (cal/g.h) AR-3.8 AR-2.0 AR-AR Figure 100: Effect of Alkali on Heat of Hydration for Cement E-SO 3 =AR 051015202530354000.10.20.30.40.50.6Time (hours)Rate of Heat Evolution(cal/g.h) 5.0-3.8 5.0-2.0 5.0-1.5 5.0-AR Figure 101: Effect of Alkali on Heat of Hydration for Cement E-SO 3 =5.0% Increasing the sulfate content (without the addition of alkalis) retards the appearance of the first peak. This is evident when comparing 5-AR case to AR-AR case, for both 126

PAGE 147

cements The peak was delayed by about 3 minutes in cement E. Adding the alkali seems to counteract this effect, when considering case 5-3.8. The peak appeared at almost the same time as AR-AR case. It can be concluded that for high alkali content (3.8%) the early rate of reaction increases. This will result in the quick hydration of the C 3 A. An increase of C 3 A reactivity due to alkali additions was also reported by S. Sprung and W. Rechenberg (25). The hydration cases of 1.5% and 2% alkali did not vary much from the no-alkali addition cases (at all sulfate levels), with some minor differences. As a matter of fact, the reaction of the first peak seems to have been slowed down slightly as a result of theses alkali additions 051015202530354000.10.20.30.40.50.6Time ( hours ) Rate of Heat Evolution (cal/g.h) AR-3.8 AR-2.0 AR-1.5 AR-AR Figure 102: Effect of Alkali on Heat of Hydration for Cement C-SO 3 =AR 127

PAGE 148

051015202530354000.10.20.30.40.50.6Time ( hours ) Rate of Heat Evolution (cal/g.h) 5.0-3.8 5.0-2.0 5.0-1.5 5.0-AR Figure 103: Effect of Alkali on Heat of Hydration for Cement C-SO 3 =5.0% Table 14: Mortar Flow Values for Cements C and E Cement C Cement E Alk SO 3 AR 2 3.8 AR 2 3.8 AR 111 110 92 103.5 90.5 63 3.6 115.5 104 90 113 96 87.5 5 116 79 106.5 104.5 102 96 7.5 Ionic Species Concentrations The chemistry of the solution gives an idea about the rate of release and the depletion of certain ions from the hydrating cement solution. This will be related to the formation of certain phases; for example the rate of sulfate ion depletion will be correlated with the formation of ettringite. Sulfate Ions (SO 4 -2 ): Figures 104-109 show the sulfate ions concentrations profiles. It is clear that for the same sulfate content, there was a higher sulfate concentration as the alkali level increased. 128

PAGE 149

The same trend was observed for both cements. The sulfate ions persisted up to 3 days in the AR and 3.6 cases, and up to 7 days for sulfate content of 5%. 010002000300040005000600070000326496128160Age (hours)Conc. (mg/l) AR-3.8 AR-2.0 AR-1.5 AR-AR Figure 104: Effect of Alkali on SO 4 -2 Concentration for Cement E-SO 3 =AR Now, considering Figures 104 and 107, it can be seen that there was an ample supply of sulfate ions for AR-3.8 cases, yet ettringite was not detected by XRD until 14 days in cement E, and not until 28 days in cement C. It is beleived that the sulfate ion concentrations availble in AR-3.8 cases were below the solubility limit of ettringite in high alkaline solutions. Though, exact values for the solubility limits of ettringite were not determined at different sulfate ion concentrations, the formation of ettringte in cases 3.6-3.8 and 5.0-3.8 as early as 8 hours corraboates this conclusion. 129

PAGE 150

010002000300040005000600070000326496128160Age (hours)Conc. (mg/l) 3.6-3.8 3.6-2.0 3.6-1.5 3.6-AR Figure 105: Effect of Alkali on SO 4 -2 Concentration for Cement E-SO 3 =3.6% 010002000300040005000600070000326496128160Age (hours)Conc. (mg/l) 5.0-3.8 5.0-2.0 5.0-1.5 5.0-AR Figure 106: Effect of Alkali on SO 4 -2 Concentration for Cement E-SO 3 =5.0% Furthermore, considering Figures 106 and 109. The sulfate ions concentration in case 5-3.8 was almost double that of cases 5.0-AR, yet the same amount of ettringite was recorded by XRD. This is indicative of the persistence of sulfate ions in solution for longer periods of time. The depletion times were almost the same for all alkali cases at the same sulfate level, with some variations. 130

PAGE 151

In the high alkali cases (3.8%), sulfate ions persisted longer in solution, longer than the no-alkali cases. However, amounts of ettringite detected are the same, and no other sulfate-bearing phases were detected beyond 24 hours in the AR cases and 3 days in the 5% cases. The following scenarios could have taken place: alkali increases the solubility of sulfates, so it is released into solution at a faster rate. Yet, it is not removed from solution fast enough. For the no alkali-cases, the sulfates are being removed at the same rate they are released. Also, in the high alkali cases, the sulfates persisted in solution and were not available for use in the formation of C-S-H gel. 010002000300040005000600070000326496128160Age (hours)Conc. (mg/l) AR-3.8 AR-2.0 AR-1.5 AR-AR Figure 107: Effect of Alkali on SO 4 -2 Concentration for Cement C-SO 3 =AR 131

PAGE 152

010002000300040005000600070000326496128160Age (hours)Conc. (mg/l) 3.6-3.8 3.6-2.0 3.6-1.5 3.6-AR Figure 108: Effect of Alkali on SO 4 -2 Concentration for Cement C-SO 3 =3.6% 010002000300040005000600070000326496128160Age (hours)Conc. (mg/l) 5.0-3.8 5.0-2.0 5.0-1.5 5.0-AR Figure 109: Effect of Alkali on SO 4 -2 Concentration for Cement C-SO 3 =5.0% Calcium Ion Concentration: The calcium ion concentration is shown in Figures 110-115. It is clear that the values decreased as the alkali concentration increased. This is consistent with calcium hydroxide quantification curves shown earlier. The implications of this is as follows: low solubility of calcium hydroxide in these high alkali solutions will cause calcium hydroxide to precipitate and nucleate in available space. This will impede and interfere with further 132

PAGE 153

nucleation of the C-S-H gel. That could have been one of the factors for the poor quality gel, and the low strength attained by the high alkali cases. Another implication is the Ca/Si ratio of the C-S-H gel, since less calcium ions are available in solution in the high alkali cases, this may have resulted in lower ratio, but this could not be verified. 04008001200160020000326496128160A g e ( hours ) Conc. (mg/l) AR-3.8 AR-2.0 AR-1.5 AR-AR Figure 110: Effect of Alkali on Calcium Ions Concentration for Cement E-SO 3 =AR 04008001200160020000326496128160A g e ( hours ) Conc. (mg/l) 3.6-3.8 3.6-2.0 3.6-1.5 3.6-AR Figure 111: Effect of Alkali on Calcium Ions Concentration for Cement E-SO 3 =3.6% 133

PAGE 154

04008001200160020000326496128160Age (hours)Conc. (mg/l) 5.0-3.8 5.0-2.0 5.0-1.5 5.0-AR Figure 112: Effect of Alkali on Calcium Ions Concentration for Cement E-SO 3 =5.0% 04008001200160020000326496128160Age (hours)Conc. (mg/l) AR-3.8 AR-2.0 AR-1.5 AR-AR Figure 113: Effect of Alkali on Calcium Ions Concentration for Cement C-SO 3 =AR 134

PAGE 155

04008001200160020000326496128160Age (hours)Conc. (mg/l) 3.6-3.8 3.6-2.0 3.6-1.5 3.6-AR Figure 114: Effect of Alkali on Calcium Ions Concentration for Cement C-SO 3 =3.6% 04008001200160020000326496128160Age (hours)Conc. (mg/l) 5.0-3.8 5.0-2.0 5.0-1.5 5.0-AR Figure 115: Effect of Alkali on Calcium Ions Concentration for Cement C-SO 3 =5.0% Aluminum Ions: The aluminum ions concentration remained low. No trend could be established. None of the properties discussed above could be correlated with the values shown in the Figures below. 135

PAGE 156

0.00.51.01.52.00326496128160Age (hours)Conc. (mg/l) AR-3.8 AR-2.0 AR-1.5 AR-AR Figure 116: Effect of Alkali on Aluminum Ions Concentration for Cement E-SO 3 =AR 0.000.501.001.502.002.500326496128160Age (hours)Conc. (mg/l) 3.6-3.8 3.6-2.0 3.6-1.5 3.6-AR Figure 117: Effect of Alkali on Aluminum Ions Concentration for Cement E-SO 3 =3.6% 136

PAGE 157

0.00.51.01.52.02.53.00326496128160Age (hours)Conc. (mg/l) 5.0-3.8 5.0-2.0 5.0-1.5 5.0-AR Figure 118: Effect of Alkali on Aluminum Ions Concentration for Cement E-SO 3 =5.0% 0.00.51.01.52.00326496128160Age (hours)Conc. (mg/l) AR-3.8 AR-2.0 AR-1.5 AR-AR Figure 119: Effect of Alkali on Aluminum Ions Concentration for Cement C-SO 3 =AR 137

PAGE 158

0.000.501.001.502.002.503.003.504.000326496128160Age (hours)Conc. (mg/l) 3.6-3.8 3.6-2.0 3.6-1.5 3.6-AR Figure 120: Effect of Alkali on Aluminum Ions Concentration for Cement C-SO 3 =3.6% 0.000.501.001.502.002.503.003.504.000326496128160Age (hours)Conc. (mg/l) 5.0-3.8 5.0-2.0 5.0-1.5 5.0-AR Figure 121: Effect of Alkali on Aluminum Ions Concentration for Cement C-SO 3 =5.0% 7.6 Scanning Electron Microscopy (SEM) 7.6.1 Fractured Surfaces Figure 122 shows some SEM micrographs for pastes for case 5-3.8 for cements C and E. In Figure 122-A and C are taken at 3 days. Visible ettringite needles dispersed throughout the matrix can be observed. Also, arcanite and syngenite crystals were identified but disappeared after 3 days as confirmed by XRD. Figure 122-B and D 138

PAGE 159

139 (cements E and C respectively), show the presence of well-crystalline and thicker ettringite needles in both cases than what wa s observed at 3 days. This could be explained in two ways; first, the crystal growth theory by Mehta (4) and Charlo tte (5), and by the space availability occupi ed previously by the syngenite and arcanite as was explained by Stark (78). This observation was only made for high alkali, high sulfate case. This is believed to be one of the reasons for exces sive expansion observed in cases 5-3.8 and 3.6-3.8. The total amount of ettringite formed ultimately, as quantified by XRD, was the same as the other cases with same sulfate content, though displaying different degrees of expansion. Figure 123 shows micrographs on paste sample s for cement C. The difference in gel quality can be seen. It is clear that in cas e AR-3.8 (Fig. 123-A) the gel is disintegrated, more discrete, less continuous. The difference was more pronounced in the case of mortar bars as shown in Figure 124 B. The differen ce in gel quality of AR-AR and AR-3.8 cases is quite apparent. Figure 124A shows SEM micrographs for mortar bars for cases AR3.8 and AR-AR for cement E. The difference is not dramatic, if it is noticed at all.

PAGE 160

A B C D Figure 122: SEM Images on Paste samples for Cements C and E-Case 5-3.8 A) E-5-3.8@ 3 days, B)E-5-3.8 @ 91 days, C) C-5-3.8@ 3 days and D) C-5-3.8 @ 360 days 140

PAGE 161

A C-AR-3.8 B C-AR-AR Figure 123: SEM images for Paste E-AR-3.8 E-AR-AR A C-AR-3.8 C-AR-AR B Figure 124: SEM Images for Mortar Bars 141

PAGE 162

7.6.2 Polished Surfaces Figure 125 shows some SEM micrographs and the EDX spectra for cases 5-3.8 for cements C and E. Polished samples were prepared for certain cases. The purpose was to study the C/S ratio and how it is affected by the increase of the alkali content of the cement. It was mentioned in the literature that the addition of alkalis results in lower ratio. This could not be verified in the current investigation. Foreign ions inclusion in the C-S-H gel was always observed, but no particular trend could be established. Cement C, Case 5-3.8 Cement E, Case 5-3.8 Figure 125: SEM Polished and EDX Spectra for Case 5-3.8 of Cements C and E 142

PAGE 163

143 CHAPTER 8 CEMENTS MH3, MH4 AND ERD07 In the previous chapter, detailed study on the effect of alkalis and sulfates for cements C and E was presented. That study included ample of analytical methods and aimed at understanding the effect of both variables on the durability and the hydration process of cements. In this chapter, the results of increasing the alkalis and sulfates on durability using additional three cements (MH3, MH4 and ERD07) will be pres ented. The alkali level will be increased up to 2%. Furthermore, since these cements are already higher in alkali content than cements C and E, they will be compared to cements C and E based on the AR-AR cases to see what is the effect of higher alkalis. 8.1 Expansion Results Figures 126, 127 and 128 show the effect of increasing the alkalis on expansion at three different sulfate levels. Increasing the alkali level at the AR sulfate level (Figure 126) did not result in major changes. All valu es are within the same range for the ages shown here (the trend is not expected to change). AR-2.0 case shows the highest expansion. This is the same trend experi enced by cement C. At 3.6% sulfate level (Figure 127), it can be seen that that all values are within a close range; however, the

PAGE 164

expansion increased with the increase in alkali content. This is the opposite of the trends experienced by cements C and E. But the values are all within a close range. At 5% sulfate level (Figure 128), it can be seen that 5-1.5 cases show the highest expansion, the 5-2.0 case started higher than the 5-AR case, then by 91 days the values are real close. This again, is not similar to the trends observed for cements C and E, although, there is no dramatic increase due to the addition of alkalis, up to the ages shown here. For all practical purposes, all expansion values at each sulfate level, are considered unchanged. The minor differences experienced here could be due to experimental errors. Cement MH3 generally showed higher expansion values than cements C and E. This cement is very high in C 3 S and low in C 3 A which implies a lower tolerance for sulfate. 0.000.020.040.060.080306090120150180Age (days)Expansion (%) AR-2.0 AR-1.5 AR-AR Figure 126: Effect of Alkali on Expansion Behavior for Cement MH3-SO 3 =AR 144

PAGE 165

0.000.020.040.060.080306090120150180Age (days)Expansion (%) 3.6-2.0 3.6-1.5 3.6-AR Figure 127: Effect of Alkali on Expansion Behavior for Cement MH3-SO 3 =3.6% 0.000.020.040.060.080306090120150180Age (days)Expansion (%) 5.0-2.0 5.0-1.5 5.0-AR Figure 128: Effect of Alkali on Expansion Behavior for Cement MH3-SO 3 =5.0% Figures 129-131 show the expansion behavior of cement MH4 at three sulfate levels. Again, increasing the alkali level did not result in any major differences, at all three sulfate levels. There was, however, some minor changes due to the addition of alklais. The most obvious was the reduction in expansion in the 5% case due to the addition of alkalis. Also, at the 3.6% level, the expansion incresaed due to the alkali addition. 145

PAGE 166

This cement is high in C 3 A; this means it will have higher tolerance for sulfate. This tolerance seems to increase due to the alkali addition. 0.000.020.040.060.080306090120150180Age (days)Expansion (%) AR-2.0 AR-1.5 AR-AR Figure 129: Effect of Alkali on Expansion Behavior for Cement MH4-SO 3 =AR 0.000.020.040.060.080306090120150180Age (days)Expansion (%) 3.6-2.0 3.6-1.5 3.6-AR Figure 130: Effect of Alkali on Expansion Behavior for Cement MH4-SO 3 =3.6% 146 Figures 132 and 133 show the expansion behavior for cement ERD07. There is no 3.6% sulfate case for this cement. Again, no major differences due to the addition of alkalis. The values decreased slightly. This cement is higher in C 3 A, and has a moderate C 3 S content, comparable to cement E.

PAGE 167

0.000.020.040.060.080306090120150180Age (days)Expansion (%) 5.0-2.0 5.0-1.5 5.0-AR Figure 131: Effect of Alkali on Expansion Behavior for Cement MH4-SO 3 =5.0% 0.000.020.040.06 0 08 0306090120150180Age (days)Expansion (%) AR-1.5 AR-AR Figure 132: Effect of Alkali on Expansion Behavior for Cement ERD07-SO 3 =AR 147

PAGE 168

0.000.020.040.060.080306090120150180Age (days)Expansion (%) 5.0-1.5 5.0-AR Figure 133: Effect of Alkali on Expansion Behavior for Cement ERD07-SO 3 =5.0% 8.2 Compressive Strength Figures 134-141 show the compressive strength for all three cements at all sulfate levels. The general trend is what was observed in cements C and E; that is, strength decreases with the addition of alkalis. 0200040006000800010000050100150200Age (days)Strength (psi) AR-2.0 AR-1.5 AR-AR Figure 134: Effect of Alkali on Strength Gain for Cement MH3-SO 3 =AR 148

PAGE 169

0200040006000800010000050100150200Age (days)Strength (psi) 3.6-2.0 3.6-1.5 3.6-AR Figure 135: Effect of Alkali on Strength Gain for Cement MH3-SO 3 =3.6% 0200040006000800010000050100150200Age (days)Strength (psi) 5.0-2.0 5.0-1.5 5.0-AR Figure 136: Effect of Alkali on Strength Gain for Cement MH3-SO 3 =5.0% 149

PAGE 170

02000400060008000 10000 050100150200Age (days)Strength (psi) AR-2.0 AR-1.5 AR-AR Figure 137: Effect of Alkali on Strength Gain for Cement MH4-SO 3 =AR 02000400060008000 10000 050100150200Age (days)Strength (psi) 3.6-2.0 3.6-1.5 3.6-AR Figure 138: Effect of Alkali on Strength Gain for Cement MH4-SO 3 =3.6% 150

PAGE 171

0200040006000800010000050100150200Age (days)Strength (psi) 5.0-2.0 5.0-1.5 5.0-AR Figure 139: Effect of Alkali on Strength Gain for Cement MH4-SO 3 =5.0% 0200040006000800010000050100150200Age (days)Strength (psi) AR-1.5 AR-AR Figure 140: Effect of Alkali on Strength Gain for Cement ERD07-SO 3 =AR 151

PAGE 172

0200040006000800010000050100150200Age (days)Strength (psi) 5.0-1.5 5.0-AR Figure 141: Effect of Alkali on Strength Gain for Cement ERD07-SO 3 =5.0% 152

PAGE 173

153 CHAPTER 9 DISCUSSION Chapter 5 presented the effect of increasing sulfate content (SO 3 ) on durability, which was assessed in terms of expansion and compressive strength. It was evident that increasing the SO 3 content resulted in an increase in expansion, with some variation depending on the cement mineralogical composition. The cause of the expansion was tied to the fo rmation of ettringite. This was done using XRD. Figures (25, and 26) show these results where the expansion increased with the amount of ettringite. The microstructure was not impacted by increasing the SO 3 content. This is evident from the compressi ve strength values pres ented in Figures (1822). And also evident from the SEM microgr aphs. So, the microstructure could be excluded as a contributing factor for expansion in these cases. Increasing the SO 3 content resulted in slowing the hyd ration process. This is evident from the delay of the appearance of the firs t peak on the heat of hydration curve (Figures 111 and 113). It also resulted in slowing down the hydration of C 3 S (Figures 92-97). This is contrary to some reports in the literatu re (67 and 83).

PAGE 174

154 As to the effect of the mineralogical co mposition of cements on expansion, the results indicate that Cement MH3 and Cement C expanded the most. These two cements are high in C 3 S. Cement MH3 recorded the highest values. This cement has a C 3 S content of over 70% (Rietveld). It is clear that a high C 3 S content result in high expansion. It was reported in the literature (31) that the C-S-H gel has the affinity to adsorb sulfate ions. These ions are to be released later and become available for ettringite formation and potential expansion. This scenario, however, is more conducive to heat-cured mortars and concretes. In the current investigations, at room temperature, ettringite formation was evident as early as 30 minutes and continued th ereafter to level off at about 28 days. This was the case in both cements, C and E where cement C has about 67% C 3 S and cement E about 54%. The sulfate adsorption and its la ter release by C-S-H gel could not explain the different behavior be tween the high and low C 3 S cements. There are two other possible explanations. First, the quality of the gel, since it is known that the formation of C-S-H gel requires a certain amount of sulfate for wellrefined uniform gel. This may imply that cements higher in C 3 S may require excess sulfates to ensure good quality gel, otherwise there will be a lesser quality, w eaker matrix But from the compressive strength results assessed in this study, it appear s that this is not the case (at least up to the ages studied here). The sec ond explanation is related to the microstructure. The higher C 3 S will result in more C-S-H gel, and more calcium hydroxide. It is believed that there will be more sites for the formation and nuclea tion of ettringite within the matrix. Now, these localized formations, will not affect the total amount of ettringite formed, but will result in more expansion since there are more of them spread throughout the matrix.

PAGE 175

155 Cement MH4 is high in C 3 S, yet it did not follow the same behavior as cements MH3 and C. This is believed to be due to the high C 3 A content. Now, MH3 and C both are very low in C 3 A (2% -Rietveld). Consequently, there will be more sulfates available in solution. These sulfates are possibly adsorbed by C-S-H gel, and may later be released to form ettringite at numerous sites throughout the matrix as mentioned above. In cements with high C 3 A contents, even with high C 3 S, more sulfates are used by the aluminates to form ettringite or other aluminosulfate phases. This means less sulfates are available to form ettringite and cause more expansion as that which occurred in C and MH3 at the 5% sulfate level. Since there is an interrelationship between the SO 3 and C 3 A contents, it may be better to relate their effect as a ratio of the two. One way is to use the S /A molar ratio as has been used in the literature. It seems like keeping this ratio equal to or less than one will not result in excessive expansion. Even higher ratios up to 1.6 did not result in excessive expansion. The highest expansion value ( in the current investigation) was for cement MH3 at just above 0.05%. This cement had a S /A close to 1.0. It is imperative to indicate that this ratio should not be considered in isolation, since other factors such as C 3 S could affect the expansion. To emphasize, considering the behavior of cements MH3 and E, both had the same S /A molar ratio (approximately 1.0) yet cement MH3 expanded higher, this is be lieved to be due to the higher C 3 S content.

PAGE 176

156 As was presented in previous chapters, incr easing the alkali content to 1.5% and 2% did not result in a major impact on the expansion be havior for all cements, and for all sulfate levels. The expansion behavior was dramatical ly altered when the alkali level increased to 3.8% for all sulfate levels. A correlation between the expansion behavi or and the ettringite formation was established earlier, the e xpansion increased as the am ounts of ettringite formed increased. However, that relationship holds en tirely true under low alkali contents, as was previously presented in Figures 76-81 in chapter 7. It is eviden t that the amount of ettringite formed increases as we increase the sulfate content. Increasing the alkali content did not affect the total amount of ettringite formed ultimately. The addition of the alkali, howev er, and depending on the sulfate content, resulted in delay of the appearance of the ettringite as detected by XRD. This implies ettringite instability at highe r alkalinity and the delay wa s at most in AR-3.8 cases, a finding that is consistent with what was reported previously in the literature (1). This phenomenon, of course, depends on the sulfate c ontent; at 5% SO3 ettr ingite appeared as early as 8 hours in cases 5-3.8. Evidently, the expansion behavior observed could not be totally explained by the amount of ettringite formed and de tected by XRD. For example, the amount of ettringite was almost the same in the 5% sulfate cases, yet, only the 5-3.8 case showed excessive expansion and the eventual damage. Again, for the 3.6% sulfate cases, at all tested alkali

PAGE 177

157 contents, showed the same amount of ettrin gite, yet the 3.6-3.8 case is the only case that showed excessive expansion. Using SEM, th e nature of ettringite was examined. Figure 122 shows the ettringite needles observed. In cases of E-5-3.8, and C-5-3.8 at 3 days the ettringite consisted of fine needles disper sed throughout the gel. At 91 days the needles appeared well crystallized and much thicker. This could be explained in two ways, first the crystal growth theory by Mehta (4) and Charlotte (5), and by the space availability occupied previously by the syngenite and ar canite as was explained by Stark (76). This only occurred in the high alkali, high sulfate case. SEM investigation showed the syngenite and arcanite crystals to persist for at least 3 days in the 5-3.8 cases. One other factor to consider, the space cr eated by the less quality gel due to the alkali addition. This may have exacerbated the effect of the ettringite on the expansion potential. It seems however, that the key action, besides the ettringite formation here is the strength of the mortar and the quality of the gel. By looking at the st rength curves, it could be seen that the strength of the 3.8% alkali case was almo st the same in all sulfate levels. Yet, the damage occurred in the 5% case, then 3.6% cas e, and did not occur yet in the AR case in cement E, but it did occur in cement C. It has to be remembered that cases AR-3.8 expanded the most in the first 28 days; yet, ther e was no ettringite det ected up to that age. The higher expansion could be due to the mois ture absorption by the porous structure as a result of the high alkali content. The total porosity and surface area were m easured for cases C-AR-AR and C-AR-3.8 at almost 98% hydration. The porosity is an i ndication of the quality of the gel. The

PAGE 178

hydration of the C 3 S and, hence, the formation of C-S-H gel goes through different stages (See Appendix C). Each part of the formed gel has a typical pore size associated with its formation. The normal hydration process of C 3 S is as follows: hydrolysis phase during the initial reaction, the induction phase follows, where the first layer of gel forms around the unhydrated grains, at the onset of the acceleration stage, this layer of gel begins to break down allowing further hydration of inner grain and more layers of gel forms, this layer becomes thicker as the hydration continues forming a barrier, through which water must flow to allow further hydration of the most inner grains. The gel formed as a result of the last stages of hydration is typically denser than the gel formed on the outer layers, and has pore sizes in the range of 1-4 nm in diameter (49). The results of the pore size distribution on the above mentioned samples are presented in Figure 142. It can be seen how the number of these small pores (and hence total surface area) was higher for the AR-AR case. These results suggest more complete hydration of the inner grains in this case. This is consistent with the results reported by M.C.G. Juenger and H. Jennings (49). 1.E-061.E-051.E-041.E-031.E-021.E-011.E+001.E+01Total Small (1-4nm) Med (4-20nm) Large (>20 nm)Pore Diamete r Pore Volume (mL/g) C-AR-AR C-AR-3.8 Figure 142: Pore Size Distribution for Cement C 158

PAGE 179

A Cement C-5-3.8 B Cement C-5-AR Figure 143: Photographs under Light Microscope for Cement C It is apparent that increasing the alkali content results in strength reduction, at all levels. Investigation of the microstructure showed clear differences between alkali-doped cements and the cases of as-received alkali contents. Figure143 shows a photograph taken under light microscope. It could clearly be seen that there are dark regions. These dark regions were noticed in all alkali-doped cases, and were more pronounced as the alkali content was increased. These spots were investigated using XRD and SEM. 20 30 40 50 un t s 0 25 100 225 0 25 100 CountsCHCSHCHCHCHCCSHCHLIGHTDARK Figure 144: XRD Pattern for Cement C-5-3.8, Case A in Figure 143 CH= Calcium Hydroxide, CSH= Calcium Silicate Hydrate Gel, C=Calcite 159

PAGE 180

160 Examination using XRD showed that the dark regions contained crystalline C-S-H gel. The crystalline nature of the gel formed in higher alkali conten t was reported by other researchers (50). This could have resulted in preferentially weak planes and a lower compressive strength. The SEM investigation on polished sections did not reveal major differences between the dark regions and th e remainder of the gel. Also, there were clearly identified white spots that appeared under the light microscope (Figure 143 A) for case C-5-3.8. XRD analysis showed that thes e white spots where CH. CH quantification using Rietveld analysis on th e hydrated paste (Figure 93) revealed these differences, especially in the first 24 hours. This also was confirmed by the chemical analysis on the liquid phase, where calcium ions concentration was lowest in the higher alkali cases (Fig.110-115). The conclusion was that in high alkali cases, CH prec ipitated faster. CH deposits early on (within the fi rst few hours) would have in terfered with further gel nucleation and resulted in a less continuous gel structure. Also, it is believed that the alkali accelerates the stiffening of the cement pa ste. This is evident from the shortening of the dormant period. The presence of the al kalis is believed to accelerate the nucleation around the cement grains prior to the dormant period thus preventing the diffusion and further hydration of the inner cement particle s later on. This will result in a less refined, less continuous microstructure. It has been reported in the literature that hi gher alkalinity results in a lower Ca/Si ratio (48, 52, 61 and 66). This could not be verified in the current investigation using polished sections under SEM/EDS. Calcium ion con centration in the liqui d phase was always

PAGE 181

161 lower as the alkali content increased; this ma y have resulted in a lower Ca/Si gel, but it could not be verified. Inclusion of other ions within the gel seems to impact the Ca/Si ratio. Other ions (S, Fe, Mg, and Al) were always present in the gel with some variations; however, a trend could not be established. The effect of these ions inclusion, in minor quantities, on the quality of the gel is not quite clear. It was mentioned in the l iterature (17) that the gel has an affinity to adsorb sulfate ions, and this affin ity increases at higher al kalinity. It was also mentioned elsewhere (67) that sulfate ions coul d substitute up to 1/6 of the Si in the gel. This may result in a weaker structure. Again, this could not be verified in the current investigation. Other researchers (64) have me ntioned that in cases where the mixes were doped with alkalis (NaOH), the sodium ions subs tituted for the Si ions in the C-S-H gel, a similar notion was reported by K. Suzuki et al (61) which resulted in a lower Ca/Si ratio. In the current investigation, the potassium was not detected in any of the cases. Figures 123 and 124 show some SEM micrographs on fractured surfaces for certain cases of cements C and E. Figure 123 is for past e samples of cement C. Difference between cases C-AR-3.8 and C-AR-AR is apparent. In th e former, it can be seen that the gel being disintegrated and less continuous. The effect was more eviden t in the case of the mortar bars (Figure 124). Figure 124-A also shows mortar bars for case E-AR-AR. In this case the bars did not experience the degradation that was experienced by cement C. This is believed to be due to the higher C 3 S content of cement C. The effect of the C 3 S lies in the production of higher amounts of C-S-H gel early on, and higher amounts of calcium

PAGE 182

162 hydroxide. The effect of C 3 S is evident from the expansion results of cement MH3 presented in chapter 8. The expansion results were higher for that cement. Cements C, MH3 and E, all have very low C 3 A content. The difference in behavior between cements MH3 and C as compared to cement E was explained by the C 3 S content. But, cement MH4 has a high C 3 S content, yet the expansion values were lower compared to cement MH3. This is believed to be due to the difference in the C 3 A content The C 3 A content is about 10% (Rietveld) for ceme nt MH4 as compared to 2% for cement MH3. It seems that cements higher in C 3 A content are more tolerant to sulfate additions. This conclusion is not inde finitely true however. The S /A molar ratio mentioned earlier still holds true for alkali levels up to 2%, at 360 days, the expansion values did not exceed .05% (except for cement MH3) keeping the S /A equal to or less than one, or even a little higher than that. This value, however, is not a measure of cement performance at higher alkali contents. At 3.8% alkali level, and for the same S /A molar ratio, th e behavior was totally different. Increasing the alkali content seems to increa se the rate of hydration early on, especially the rate of hydration of C 3 A. This was evident at 3.8% al kali content from the heat of hydration curves. This effect was not as clea r at 1.5% and 2% alkali levels. As shown earlier, increasing the sulfate content resulte d in delaying the app earance of the first

PAGE 183

163 peak, which implies that sulfates delay the hydration of the C 3 A. Increasing the sulfate level at the higher alkali level resulted in slowing down the re action, although it did not delay the appearance of the first peak. It resu lts in less heat generated (cases 5-3.8). It was also shown that increasing the sulfate levels to 3.6% and 5% resulted in higher strength when the alkali levels were increased. The conclusion was that sulfate counteract the effect of alkalis on the hydration process. Increasing the alkali content seems to increase the availability of the sulfate ions in solution. This makes these ions unavailable to participate in the formation of ettringite to slow the hydration of C 3 A down. Additionally, it would prevent the sulfates from participating in the formation of the gel. It is apparent that certain amount of sulfates are necessary for the formation of the gel. Now, increasing the sulfate content (while keeping the alkali level constant) will make more sulf ate ions available to participate in these reactions. But, the excess sulf ate will be available and ma y prove disastrous later on. From the results presented here, it seems an SO 3 /Alkali molar ratio of 2 or above is needed to prevent excessive expansion or se vere loss of strength. One thing to keep in mind though, increasing sulfates and alkalis inde finitely to maintain this ratio is not a viable option since these increases will have adverse effects on long-term durability. Proposed Mechanism of Alkalis Effect: The presence of high alkali concentration seem s to interrupt the growth of the C-S-H gel fibers as these crystals grow on the surface of the C 3 S grains. This will prevent the

PAGE 184

164 bridging action and the interlocking of th ese fibers. This will result in a weak microstructure. High alkali c oncentration would also resu lt in early precipitation of calcium hydroxide which may interfere with the growth of the C-S-H gel fibers. Without proper retardation, the alkalis seem to impact the first layer formed around the C 3 S grains. Typically this laye r will breakdown and become more permeable at the end of the dormant period, allowi ng diffusion and more water to help the hydration of the inner unhydrated grains. But the presence of alka lis results in an early stiffening of this layer and this will hinder the further di ffusion and further hydration of the inner particles. This may explain the lower portion of the gel pores (1-4 nm) in the higher alkali cases. These pores are typical of the denser phenograins. Also, the ettringite formed in the presence of high OH is the expansive type as mentioned before. The combination of the porous microstructure and expansive ettr ingite resulted in high expansion, reduction in compressive strength and eventual damage.

PAGE 185

165 CHAPTER 10 CONCLUSIONS Based on the preceding results and discussion, we could draw the following conclusions: 1. Increasing the SO 3 contents results in an increase in expansion. 2. The increase in expansion seemed to be the result of an increase in the formation of ettringite. 3. Increasing the SO 3 contents did not dramatically alter the hydration process. However, it seems to slow down the process. 4. There was a delay in strength gain at the 5% level; the strength at 180 days was not altered. 5. Increasing the alkali levels up to 2% (Na 2 O equivalent), at all sulfate levels, did not impact the expansion behavior much, at room temperature cure. 6. Increasing the alkali level to 3.8% (Na 2 O equivalent), resulted in the eventual damage of mortar prisms. 7. Increasing the alkali content, decr eases the compressive strength. 8. Alkalis influence the kinetics of hydration in the early stages. 9. Sulfates counteract the effect of the alkalis on hydration process. 10. Sulfate content of 3-3.6% did not seem to adversely affect the Portland cement systems of low alkali content and moderate fineness cured under ambient conditions.

PAGE 186

166 REFERENCES 1. Jochen Stark, Katrin Bollmann, Delayed Ettringite Formation in Concrete, Bauhaus-University Weimar/Germany. 2. H.Y. Ghorab, D. Heinz, U. Ludwig, T. Meskendahl and A. Wolter, On the Stability of Calcium Aluminate Sulfat e Hydrates in Pure Systems and in Cements, 7 th International Congress on Chemis try of Cement, Paris (1980), pp.496-503. 3. M. Collepardi, A State-of-the art Re view on Delayed Ettringite Attack on Concrete, Cement and Concre te Composites 25 (2003), pp 401-407. 4. P. K. Mehta, Mechanism of Expansion Associated with Ettringite Formation, Cement and Concrete Research Vol. 3 (1973), pp.1-6. 5. C. Famy, Expansion of Heat-Cured Mort ars, PhD Thesis, Imperial College of Science, Technology and Medicine (University of London) (1999). 6. Deng Min and Tang Mingshu, Formation and E xpansion of Ettringite Crystals, Cement and Concrete Research, Vol. 24 (1994), pp. 119-126,. 7. I. Odler, Interaction Between Gypsum and C-S-H Phase Formed in C3S Hydration, 7 th international Congress on Chemis try of Cement, Vol. 4 (1980), pp. 439-495. 8. W. G. Hime and S.L. Marusin, Delayed Ettringite Formation: Many Questions and Some Answers, ACI SP177-13, pp. 199-206. 9. J. Stark and K. Bollmann, Laboratory a nd Field Examinations of Ettringite Formation in Pavement Concrete, ACI SP177-12, pp. 183-198. 10. Charlotte Famy and Hal F. W. Taylor, Ettringite in Hydration of Portland Cement Concrete and its Occurrence in Mature Concretes, ACI Materials Journal July-August 2001, pp 350-356. 11. M.D.A Thomas, Ambient Temperature DEF: Weighing the Evidence, Fifth Canmet/ACI International Conference on Durability of Concrete, Barcelona, Spain 2000, pp 43-56.

PAGE 187

167 12. H.F.W. Taylor, C. Famy and K.L Scri vener, Review: Delayed Ettringite Formation, Cement and Concrete Research 31 (2001), pp 683-693. 13. Paul D. Tennis, Sankar Bhattacharja, Wa ldemar A. Klemm, and F. MacGregor Miller, Assessing the Distri bution of Sulfate in Portland Cement and Clinker and its Influence on Expansion in Mortar, The Symposium on Internal Sulfate Attack on Cementitious Systems, Dec. 7-12, 1997, pp 212-216. 14. F.M. Miller and F.J. Tang, The Distributi on of Sulfur in Present-Day clinkers of Variable Sulfur Content, Cement and Concrete Research, Vol. 26, no. 12 (1996), pp 1821-1829. 15. H.F.W. Taylor, Ettringite in Cement Paste and Concrete, Concrete; From Material to Structure; Arles, France, September 11-12, 1996; Proceedings of the RILEM International Conference. 16. L. Divet and R. Randriambololona, Delayed Ettringite Formation: The Effect of Temperature and Basicity on the Interacti on of Sulfate and C-S-H phase, Cement and Concrete Research, Vol. 28, No. 3 (1998), pp. 357-363. 17. D. Heinz and U. Ludwig, Mechanism of Secondary Ettringite Formation in Mortars and Concretes, ACI SP100-105, pp. 2059-2071. 18. Yan Fu, Ping Xie, Ping Gu, J.J. Beaudoin, Effect of Temperature on Sulfate Adsorption/Desorption by Tricalcium Sili cate Hydrates, Cement and Concrete Research, Vol. 24, No. 8 (1994), pp. 1428-1432. 19. H. Y. Ghorab, E. A. Kishar, and S.H. Abou Elfetouh, Studies on the Stability of Calciumsulfoaluminate Hydrates.Part II: Effect of Alite, Lime, and Monocarboaluminate Hydrate, Cement a nd Concrete Research, Vol. 28, No. 1 (1998) pp. 53-61. 20. C. D. Lawrence, Long-Term Expansion of Mortars and Concrete, EttringiteThe Sometimes Host of Dest ruction, ACI SP177, pp 106-123. 21. Yan Fu and James J. Beaudoin, Mechanism s of Delayed Ettringite Formation in Portland Cement Systems, ACI Mate rials Journal July-August 1996, pp 327-333. 22. Zhaozhou Zhang, Jan Olek, Sidney Diam ond, Studies on Delayed Ettringite Formation in Early-age, Heat-Cured Mortars. I. Expansion Measurements, Changes in Dynamic Modulus of Elasti city, and Weight Gains, Cement and Concrete Research 32 (2002), pp 1729-1736.

PAGE 188

168 23. Zhaozhou Zhang, Jan Olek, Sidney Diam ond, Studies on Delayed Ettringite Formation in Early-age, Heat-Cured Mort ars. II. Expansion Characteristics of Cements that may be Susceptible to DEF, Cement and Concrete Research 32 (2002), pp 1737-1742. 24. S. Diamond, Delayed Ettringite Form ation-Processes and Problems, Cement and Concrete Composites 18 (1996), pp 205-215. 25. F.P Glasser, The Role of Sulfate Mine ralogy and Cure Temperature in Delayed Ettringite Formation, Cement a nd Concrete Composites 18 (1996) 187-196. 26. H.F.W. Taylor, C. Famy and K.L. Scri vener, Review: Delayed Ettringite Formation, Cement and Concrete Research, 31 (2001), pp 683-693. 27. C.D. Lawrence, Mortar Expansion Due to Delayed Ettringite Formation. Effect of Curing Period and Temperature, Cement and Concrete Research 25 (1995), pp 903-914. 28. Yan Fu, Jian Ding, and J.J Beaudoin, Expa nsion of Portland Cement Mortar Due to Internal Sulfate Attack, Cement and Concrete Research, Vol. 27, No. 9 (1997), pp 1299-1306. 29. Yan Fu and J.J. Beaudoin, Microcracks as a Precursor to Delayed Ettringite Formation in Cement Systems, Cement and Concrete Research, Vol. 26, No. 10 (1996) pp 1493-1498. 30. Oscar R. Batic, Carlos A. Milanesi, Pedro J. Maiza, Silvina A. Marfil, Secondary Ettringite Formation in Concrete Subj ected to Different Curing Conditions Cement and Concrete Research 30 (2000) 1407-1412. 31. Yan Fu, Ping Gu, Ping Xie, J.J Beadoin, A kinetic Study of Delayed Ettringite Formation in Hydrated Portland Cement Paste, Cement and Concrete Research, Vol. 25, No. 1 (1995), pp. 63-70. 32. D.W. Hobbs, Expansion and Cracking in Concrete Associated with Delayed Ettringite Formation, ACI Sp177-11, pp.159-181. 33. W.G. Hime, Clinker Sulfate: A Cause fo r Distress and a Need for Specification, Concrete for Environment Enhancement and Protection, 1996. 34. J. Stark and K. Seyfarth, Ettringite Formation in Hardened Concrete and Resulting Destruction, ACI SP177-9, pp. 125-140. 35. C. D. Lawrence, J. A. Dalziel and D. W. Hobbs, Sulfate Attack Arising from Delayed Ettringite Formation, BCA Interim Technical Note, May 1990.

PAGE 189

169 36. Waldemar A. Klemm and F. MacGregor Miller, Internal Sulfate Attack: A Distress Mechanism at Ambient and Elevated Temperatures?, PCA R&D Serial No. 2125. Prsented at the 1997 Spring Convention, ACI, Seattle. 37. Sharon L. Tracy, Stephen R. Boyd and James D. Connolly, Effect of Curing Temperature and Cement Chemistry on the Potential for Concrete Expansion Due to DEF, PCI Journal Vol. 49, No. 1, January-February 2004, pp. 46-57. 38. S. Kelham, The Effect of Cement Composition and Fineness on Expansion Associated with Delayed Ettringite Formation, Cement and Concrete Composites 18 (1996), pp-171-179. 39. David McDonald, Technical Note: Delaye d Ettringite Formation and the Heat Curing-Implication of the Work of Kelham, Cement and Concrete Research, Vol. 28, No. 12 (1998), pp 1827-1830. 40. A. Shayan and Ivanusec, An Experimental Clarification of the Association of Delayed Ettringite Formation with Alkali-Aggregate Reaction, Cement and Concrete Composites 18 (1996), pp 161-170. 41. Inam Jawed and Jan Skalny, Alkalis in Cement: A review I. Forms of Alkalis and Their Effect on Clinker Formation, Cement and Concrete Research, Vol. 7 (1977), pp. 719-730. 42. Inam Jawed and Jan Skalny, Alkalis in Ce ment: A Review, II. Effects of Alkalis on Hydration and Performance of Port land Cement, Cement and Concrete Research, Vol. 8 (1978), pp. 37-52. 43. William Lerch, The Influence of Gypsum on the Hydration and Properties of Portland Cement Pastes, Presented at the Forty Ninth Annual Meeting of the Society (ASTM), June 24-28, 1946. 44. S. Sprung and W. Rechenberg, Influ ence of Alkalis on the Hydration of Cement, Effects of Alkalis on Properties of Concrete, Proceedings of a Symposium held in London, 1976. 45. I. Odler and R. Wonnemann, Effects of Al kalis on Portalnd Cement Hydration, I. Akalai Oxides Incorporated into the Crystalline Lattice of Clinker Minerals, Cement and Concrete Research, Vol. 13 (1983), pp. 477-482. 46. I. Odler and R. Wonnemann, Effects of Alkalis on Portalnd Cement Hydration, II. Akalis Present in Form of Sulfates, Cement and Concrete Research, Vol. 13 (1983), pp. 771-777.

PAGE 190

170 47. Vagn Johansen, Influence of Alkalis on the Strength Development of Cements, Proceedings of the Symposium on the Effects of Alkalis on Properties of Concrete, London, Sept. 1976, p 81. 48. S. J. Way and A. Shayan, Early Hydrat ion of a Portland Cement in Water and Sodium Hydroxide Solutions: Compositions of Solutions and Nature of Solid Phases, Cement and Concrete Research, Vol. 19 (1989), pp. 759-769. 49. Maria C. Garci Juenger and Hamlini M. Jennings Effects of High Alkalinity Cement on Pastes, ACI Materials Jour nal, V. 98, No. 3 May-June 2001, pp. 251255. 50. Dale P. Bentz, Influence of Alkalis on Porosity Percolation in Hydrating Cement Pastes, Submitted for CCR 2005. 51. Robert L. Day, The Effect of Secondary Ettringite Formation on the Durability of Concrete: A Literature Anal ysis, PCA R&D Bulletin RD108T. 52. Sujin Song, Hamlin M. Jennings, Pore So lution Chemistry of Alkali-Activated Ground Granulated Blast-Furnace Slag, Cement and Concrete Research 29 (1999), pp. 159-170. 53. K. M. Alexander and C.E.S. Davis, Eff ect of Alkali on the Strength of Portland Cement, Australian Journal of A pplied Sciences, V. 11, No. 1, 1960, pp. 146156. 54. Walter J. McCoy and Ottomar L. Eshenour, Significance of Total and WaterSoluble Alkali Content of Cement Proceedings of the 5 th International Symposium on the Chemistryu of Ce ment, Vol II, (1968), pp.437-443, Tokyo. 55. B. Osbaeck, On the Influence of Alka lis on Strength Development of Blended Cements, pp. 375-383. 56. M. S. Y. Bhatty and N. R. Greening, I nteraction of Alkalis with Hydrating and Hydrated Calcium Silicates, Port land Cement Association, pp. 87-111. 57. A.I. Jelenic, A. Panovic, R.Halle and T. Gacesa, Effect of Gypsum on the Hydration and Strength of Commercial Portland Cements Containing Alkali Sulfates, Cement and Concrete Research, Vol. 7 (1977), pp. 239-246. 58. J. Gebauer and M. Kristman, The Infl uence of the Composition of Industrial Clinker on Cement and Concrete Propert ies, World Cement Technology, Match 1979, pp. 46-51.

PAGE 191

171 59. B. Osbaeck and E. S. Jons, The influence of the content and Distribution of Alkalis on the Hydration Propert ies of Portland Cement, 7 th International Congress on the Chemistry of Cement, Volume 2 (1980), pp. 135-140. 60. B. Osbaeck, Alkalis and Cement Strength, Proceedings of the 6 th International Conference on Alkalis in Concrete; Research and Practice, Copenhagen, Denmark, 1983, pp.93-100. 61. K. Suzuki, T. Nishikawa, H. Ikenaga and S. Ito, Effect of NaCl or NaOH on the Formation of C-S-H, Cement and Concrete Research, Vol. 16, pp. 333-340. 62. N. Smaoui, M. A. Berube, B. Fournier B. Bissonnette, B. Durand, Effecets of Alkali addition on the Mechanical Propert ies and Durability of Concrete, Cement and Concrete Research (Articles in Press) 2004. 63. H. E. Vivian, The Effect of Added S odium Hydroxide on the Tensile Strength of Mortar, CSIRO (Australia ), Bulletin No. 299,256, 1950, pp. 48-52. 64. A. Shayan and I. Ivanusec, Influen ce of NaOH on Mechanical Properties of Cement Paste and Mortar with and without Reactive Aggregate, Proceedings of 8 th International Conference on Al kali-Aggregate Reaction, 1989, pp. 715-720. 65. M.P. Luxan, M. Frias and F. Dorrego, Potential Expansion of Cement Mortars in the Presence of K 2 SO 4 and Pozzolan, Cement and Concrete Research, Vol. 24, No. 4 (1994), pp. 728-734. 66. S. Song, D. Sohn, H. M. Jennings, T. O. Mason, Hydration of Alkali-Activated Ground Granulated Blast Furnace Slag, J ournal of Material s Science 35 (2000) pp. 249-257. 67. Howard M. Kanare and Ellis M. Gartner, Optimum Sulfate in Portland Cement, Cement Research Progress 1984, pp. 214-250. 68. W. Rechenberg and S. Sprung, Composition of the Solution in the Hydration of Cement, Cement and Concrete Research, Vol. 23 (1983), pp. 119-126. 69. Wieslaw Kurdowski, Role of Delayed Release of Sulfates From Clinker in DEF, Cement and Concrete Research 32 (2002), pp 404-407. 70. Stutzman, P. E Guide For X-ray Powder Diffraction Analysis of Portland Cement and Clinker, NIST Internal Report 575, (1996). 71. Gutteridge, A.A. On the Dissolution of the Interestitial Phases in Portland Cement. Cement and Concrete Research, Vol. 9 (1979), pp. 319-324.

PAGE 192

172 72. Renhe Yang, Christopher D. Lawrence, Cyril J. Lynsdale, John H. Sharp, Delayed Ettringite Formation in Heat-c ured Portland Cement Mortars, Cement and Concrete Research 29 (1999), pp. 17-25. 73. Stutzman, Paul, Chemistry and Struct ure of Hydration Products, Cement Research Progress. Chapter 2, Ameri can Ceramic Society (1999), pp 37-69. 74. H.F.W. Taylor, Distribution of Sulfat e between Phases in Portland Cement Clinkers, Cement and Concre te Research 29 (1999), pp. 1173-1179. 75. Nikola Petrov and Arezki Tagnit-Hamou, Is Microcracking Really a Precursor to Delayed Ettringite Formation and Cons equent Expansion?, ACI Materials Journal, V. 101, No. 6, Nov/Dec. (2004), pp. 442-447. 76. J. Stark, B. Moser, F. Bellmann, New A pproaches to Ordinary Portland Cement Hydration in the Early Hardening Stage, Proceedings of the 5 th International Symposium on the Cement and Concre te, Shanghai, China, Oct. 28-Nov. 1,(2002), pp. 56-70. 77. H. Y. Ghorab and S.H. Abu El Fetouh, Factors Affecting the Solubility the Solubility of Gypsum, Part II, the E ffect of Sodium Hydroxide under Various Conditions, Journal of Chemical Technology and Biotechnology, 35A, (1985), pp. 36-40. 78. Stutzman, Paul, Sanning Electron Micros copy in Concrete Petrogrophy Material Science of Concrete Special Volume: Calcium Hydroxide In Concrete. Proceedings. J. American Ceramic Society. Nov. 1-3, (2000), pp. 59-72. 79. I. Odler and Yoaxin, On the Delayed Expansion of Heat Cured Portland Cement Pastes and Concretes, Cement and C oncrete Composites, 18 (1996), 181-185. 80. L.E. Copeland and J. C. Hayes, Determination of Non-Evaporable Water in Hardened Portland Cement Paste, ASTM Bulletin, December, (1953), pp.70-74. 81. Hewlett, Peter C, Leas Chemis try of Cement and Concrete, 4 th ed. John Wiley and Sons, Inc., New York, NY (1989). 82. L. E. Copeland and D. L. Kantro, Hydr ation of Portland Cement, Proceedings of the Fifth International Symposium on Chemistry of Cement Vol. II (1968), pp. 387-421. 83. F.W. Locher, The Theme II: Hydr ation of pure Portland Cements, 7 th International Congress on the Chemistr y of Cement, Volume IV (1980), pp. 4962.

PAGE 193

173 84. Ivan Odler and Michael Gasser, Mechani sm of Sulfate Expansion in Hydrated Portland Cement, J. American Ceram. Soc. 71,11(1998), pp.1015-1021. 85. Duncan Hertfort, Soren Rasmussen, Ebbe Jones, and Bjarne Osbaeck, Mineralogy and Performance of Ceme nt Based on High Sulfate Clinker, Cement, Concrete, and Aggregates CCAGDP, Vol. 21, No. 1, June 1999. 86. Richard C. Mielenz, Stella L. Marusin, William G. Hime, and Zvonimir T. Jgovic, Investigation of Prestressed Concrete Railway Tie Distress, Concrete International, December (1995), pp. 62-68. 87. K.M. Alexander, J. Wardlaw and I. Ivanusec, The Influence of SO3 Content of Portland Cement on the Creep and Other Physical Properties of Concrete, Cement and Concrete Research 9, (1979), pp. 451-459. 88. Cheng Sheng Ouyang, Antonio Nanni and We n F. Chang, Internal and External Sources of Sulfate Ions in Portland Ce ment Mortar: Two Types of Chemical Attack, Cement and Concrete Research, Vol. 18 (1988), pp. 699-709. 89. H.F.W. Taylor, Ettringite in Cement Paste and Concrete, Concrete; From Material to Structure; Arles, France, September 11-12, 1996; Proceedings of the RILEM International Conference. 90. D. L. Kantro, Sulfate Sp ecifications as a Constraint to Gypsum Addition to Cement and Possible Repalcement of G ypsum as an additive, PCA, 1979. 91. R. Barbarulo, H. Peycelon, S. Prene, J. Marchand, Delayed Ettringite Formation Symptoms on Mortars Induced By High Temperature due to Cement Heat of Hydration or Late Thermal Cycle, Ceme nt and Concrete Research, 35 (2005) pp. 125-131. 92. Fulvio J. Tang and Ellis M. Gartner Influence of Sulfate Source on Portland Cement Hydration, Advances in Cement Research, Vol. 1, No. 2 (1988), pp. 6774. 93. V. Michaud and R. Suderman, Sol ubility of Sulfates in High SO 3 Clinkers, Ettringite-The Sometimes Host of Destruction, ACI SP177, pp. 15-25.

PAGE 194

174 APPENDICES

PAGE 195

175 Appendix A: Cement Dissolution Methods A.1 Salicylic Acid-Methanol Extraction (SAM) This method is used to dissolve silicate phases and free lime, it leaves a residue of aluminates, ferrites, alkali sulfates, pericl ase, carbonates and doubl e-alkali sulfates. To prepare the extraction, 20 grams of salicyli c acid was added to 300 ml of methanol, the mixture was then mixed on a stirring plate. 5 g of cement then added and mixed for two hours in a stoppered flask The mi xture allowed to settle for about 15 minutes, then it was vacuum-filtered through 0.45 a polypropylene f ilter. The collected residue was then washed with 100 ml of methanol, th en dried at 95C for 30 minutes. A.2 Potassium Hydroxide/Sugar Extraction (KOSH) This method is used to dissolve the inte rstitial phases pf aluminate an ferrites, it leaves behind the silicate phases and some other minor phases such as gypsum and periclase. To prepare the extraction, 30 gm of sucrose were dissolved in 300 ml of distilled water in a flask. The mixture was then placed on stirring plate and heated up to 95C. 5 gm of the cement where then added and solution was mixed for one more minute, maintaining the temperature above. The suspen sion was allowed to se ttle, then it was vacuum-filtered through a 0.45 micron polypropylene filter. The collected residue was then washed with 50 ml of me thanol, then dried at 60 C.

PAGE 196

Appendix B: XRD Scans 10 20 30 40 50 ts 0 400 900 0 100 400 900 0 400 0 400 1600 0 400 1600 0 400 0 400 C3SC3SCHTiO2EBMKSEEC3STiO2C3SCHEC3SC3SC3SKSBMKSCHTiO2@ 24 hours@ 8 hours@ 30 minutes@ 3 days@ 7 days@14 days@ 28 daysSyng.Syng.CountsC3SCH 10 20 30 40 50 0 100 400 900 0 100 400 900 0 100 400 900 0 100 400 900 0 100 400 900 0 100 400 C3SC3SCHTiO2BMTiO2C3SCHEC3STiO2CountsC3SCHEEECHCE CE CHC2S@ 56 days@ 91 days@ 120 days@ 150 days@ 180 days@ 360 days Diffraction Angle 2 Figure 145: XRD of Hydration Products, Case E-5-3.8 C 3 S= Tricalcium Silicates, C 2 S=Dicalcium Silicates, TiO 2 ,=Rutile CH=Ca (OH) 2 ,, BM=C 4 AF, E=Ettringite, G=gypsum,C= Calcite, KS=Arcanite 176

PAGE 197

Appendix B: (Continued) 10 20 30 40 50 ts 0 100 400 900 0 100 400 0 100 400 0 100 400 0 100 400 900 0 100 400 900 0 400 C3SC3SCHTiO2BMKSC3STiO2C3SCHC3SC3SC3SKSBMKSCHTiO2@ 24 hours@ 8 hours@ 30 minutes@ 3 days@ 7 days@14 days@ 28 daysCHEECHC2SCH CountsE 10 20 30 40 50 unt s 0 100 400 900 0 100 400 900 0 100 400 900 0 100 400 900 0 100 400 0 100 400 C3SC3SCHTiO2BMC3STiO2C3SCHC3SC3SCHTiO2CHC2SCH Counts@ 91 days@ 56 days@ 120 days@ 150 days@ 180 days@ 360 daysEECHEC Diffraction Angle 2 Figure 146: XRD of Hydration Products, Case E-3.6-3.8 C 3 S= Tricalcium Silicates, C 2 S=Dicalcium Silicates, TiO 2 ,=Rutile CH=Ca (OH) 2 ,, BM=C 4 AF, E=Ettringite, G=gypsum,C= Calcite, KS=Arcanite 177

PAGE 198

Appendix B: (Continued) 10 20 30 40 50 unt s 0 100 400 900 0 100 400 900 0 100 400 0 100 400 0 100 400 0 100 400 900 0 100 400 C3SC3SCHTiO2BMC3STiO2C3SCHC3SC3SCHTiO2CHC2SCH CountsCH@ 8 hours@ 30 minutes@ 24 hours@ 3 days@ 7 days@ 14 days@ 28 daysKSC3SEEE 10 20 30 40 50 s 0 100 400 0 100 400 0 100 400 0 100 400 0 100 400 0 100 400 C3SC3SCHTiO2BMC3SC3SCHCHC2SCH CountsCHC ETiO2C2STiO2@ 91 days@ 120 days@ 150 days@ 180 days@ 360 days@ 56 daysEECHCC Diffraction Angle 2 Figure 147: XRD of Hydration Products, Case E-AR-3.8 C 3 S= Tricalcium Silicates, C 2 S=Dicalcium Silicates, TiO 2 ,=Rutile CH=Ca (OH) 2 ,, BM=C 4 AF, E=Ettringite, G=gypsum,C= Calcite, KS=Arcanite 178

PAGE 199

Appendix B: (Continued) 10 20 30 40 50 0 400 900 0 400 1600 0 100 400 0 400 900 0 400 0 400 0 400 C3SC3SCHTiO2BMEEC3STiO2C3SCHEC3SC3SC3SBMCHTiO2@ 24 hours@ 8 hours@ 30 minutes@ 3 days@ 7 days@14 days@ 28 daysEECHC2SEGCounts 10 20 30 40 50 ts 0 400 900 1600 0 100 400 900 0 400 900 1600 0 400 900 1600 0 100 400 900 0 100 400 900 C3SCHTiO2BMC3STiO2C3SCHEC3STiO2ECHC2SCounts@ 56 days@ 91 days@120 days@150 days@ 180 days@ 360 daysEECHCE Diffraction Angle 2 Figure 148: XRD of Hydration Products, Case E-5.0-AR C 3 S= Tricalcium Silicates, C 2 S=Dicalcium Silicates, TiO 2 ,=Rutile CH=Ca (OH) 2 ,, BM=C 4 AF, E=Ettringite, G=gypsum,C= Calcite 179

PAGE 200

Appendix B: (Continued) 10 20 30 40 50 0 100 400 900 0 100 400 0 100 400 0 100 400 900 0 100 400 900 0 100 400 900 C3SC3STiO2EC3STiO2C3SCHC3SC3SC3SBMCHTiO2@ 30 minutes@ 8 hours@ 24 hours@ 3 days@ 7 days@ 28 days BMGE ECHGEECHC2SBMCounts 10 20 30 40 50 ts 0 100 400 900 0 100 400 900 0 100 400 900 0 100 400 900 0 100 400 0 100 400 C3STiO2C3STiO2C3SCHC3S BME CHCHC2SCounts@ 56 days@ 91 days@ 120 days@ 150 days@ 180 days@ 360 daysEEECHTiO2 Diffraction Angle 2 Figure 149: XRD of Hydration Products, Case E-3.6-AR C 3 S= Tricalcium Silicates, C 2 S=Dicalcium Silicates, TiO 2 ,=Rutile CH=Ca (OH) 2 ,, BM=C 4 AF, E=Ettringite, G=gypsum,C= Calcite 180

PAGE 201

Appendix B: (Continued) 10 20 30 40 50 ts 0 100 400 900 0 100 400 900 0 100 400 900 0 100 400 900 0 100 400 900 0 100 400 900 C3SC3STiO2C3STiO2C3SCHC3SC3SC3STiO2@ 28 days BM CHEECHC2SCounts@ 8 hours@ 24 hours@ 3 days@ 7 days@ 14 daysECHE 10 20 30 40 50 0 100 400 900 0 100 400 900 0 400 1600 0 100 400 900 0 100 400 900 0 100 400 900 C3SC3STiO2TiO2C3SCHC3STiO2 BM CHEECHC2SCountsECHE @ 56 days@ 91 days@ 120 days@ 150 days@ 180 days@ 360 daysECCHCHC2S Diffraction Angle 2 Figure 150: XRD of Hydration Products, Case E-AR-AR C 3 S= Tricalcium Silicates, C 2 S=Dicalcium Silicates, TiO 2 ,=Rutile CH=Ca (OH) 2 ,, BM=C 4 AF, E=Ettringite, G=gypsum,C= Calcite 181

PAGE 202

Appendix B: (Continued) 10 20 30 40 50 unt s 0 100 400 900 0 100 400 0 100 400 900 0 400 900 0 400 0 400 0 400 900 C3SC3SCHTiO2BMC3STiO2C3SCHC3SC3SCHTiO2CHC2SCH Counts@ 8 hours@ 30 minutes@ 24 hours@ 3 days@ 7 days@ 14 days@ 28 daysKSC3SEEEEECH 10 20 30 40 50 unts 0 100 400 900 0 100 400 0 100 400 900 0 100 400 900 0 100 400 0 100 400 C3SC3SCHTiO2BMC3STiO2C3SCHC3SC3STiO2CHC2SCH CountsEEECH@ 91 days@ 56 days@ 120 days@ 150 days@ 180 days@ 360 days EECHC P i ti [ 2 T h t ] Diffraction Angle 2 Figure 151: XRD of Hydration Products, Case E-AR-1.5 C 3 S= Tricalcium Silicates, C 2 S=Dicalcium Silicates, TiO 2 ,=Rutile CH=Ca (OH) 2 ,, BM=C 4 AF, E=Ettringite, G=gypsum,C= Calcite KS=Arcanite 182

PAGE 203

Appendix B: (Continued) ts 0 100 400 0 100 400 0 100 400 0 100 400 0 100 400 0 100 400 0 100 400 900 C3SC3SC3STiO2C3SCHC3STiO2C2SCH Counts@ 7 days@ 28 daysEECH@ 8 hours@ 4 hours@ 24 hours@ 3 days@ 14 daysECHCHTiO2CH 10 20 30 40 50 t s 0 100 400 900 0 100 400 900 0 100 400 0 100 400 0 100 400 900 0 100 400 900 C3SC3STiO2TiO2C3SCHC3STiO2 BM CHEECHC2SCountsECHE @ 56 days@ 91 days@ 120 days@ 150 days@ 180 days@ 360 daysECCHCHC2S Diffraction Angle 2 Figure 152: XRD of Hydration Products, Case E-AR-2.0 C 3 S= Tricalcium Silicates, C 2 S=Dicalcium Silicates, TiO 2 ,=Rutile CH=Ca (OH) 2 ,, BM=C 4 AF, E=Ettringite, G=gypsum,C= Calcite KS=Arcanite 183

PAGE 204

Appendix B: (Continued) 10 20 30 40 50 ts 0 100 400 0 100 400 0 100 400 0 100 400 0 100 400 0 100 400 900 0 100 400 C3SC3SBMKSC3STiO2C3SCHC3SC3SC3SKSBMKSTiO2@ 24 hours@ 8 hours@ 30 minutes@ 3 days@ 7 days@14 days@ 28 daysSyng.Syng.EETiO2CHC2SEECHCounts 10 20 30 40 50 s 0 100 400 900 0 100 400 900 0 100 400 0 100 400 0 100 400 0 100 400 900 C3SC3SCHTiO2BMTiO2C3SCHETiO2CountsC3SCHEEECHEE CHC2S@ 56 days@ 91 days@ 120 days@ 150 days@ 180 days@ 360 daysEC Diffraction Angle 2 Figure 153: XRD of Hydration Products, Case E-5.0-2.0 C 3 S= Tricalcium Silicates, C 2 S=Dicalcium Silicates, TiO 2 ,=Rutile CH=Ca (OH) 2 ,, BM=C 4 AF, E=Ettringite, G=gypsum,C= Calcite KS=Arcanite 184

PAGE 205

Appendix B: (Continued) 10 20 30 40 50 nt s 0 100 400 0 400 900 0 100 400 900 0 400 900 0 100 400 900 0 400 1600 0 100 400 900 C3SC3SCHTiO2BMKSEEC3STiO2C3SCHEC3SC3SC3SKSKSCHTiO2@ 24 hours@ 8 hours@ 30 minutes@ 3 days@ 7 days@14 days@ 28 daysSyng.Syng.EC3SCountsC2SCHCH 10 20 30 40 50 0 100 400 900 0 100 400 900 0 100 400 900 0 100 400 900 0 100 400 0 100 400 900 C3SC3SCHTiO2BMTiO2C3SCHEC3STiO2CountsC3SCHEEECHCE CE CHC2S@ 56 days@ 91 days@ 120 days@ 150 days@ 180 days@ 360 days Diffraction Angle 2 Figure 154: XRD of Hydration Products, Case C-5.0-3.8 C 3 S= Tricalcium Silicates, C 2 S=Dicalcium Silicates, TiO 2 ,=Rutile CH=Ca (OH) 2 ,, BM=C 4 AF, E=Ettringite, G=gypsum,C= Calcite, KS=Arcanite 185

PAGE 206

Appendix B: (Continued) 10 20 30 40 50 ts 0 400 0 400 0 100 400 0 400 0 400 0 400 0 400 C3SC3SCHTiO2BMKSC3STiO2C3SCHC3SC3SC3SKSBMKSCHTiO2@ 24 hours@ 8 hours@ 30 minutes@ 3 days@ 7 days@14 days@ 28 daysCHEECHC2SCH CountsE 10 20 30 40 50 0 100 400 900 0 100 400 0 100 400 0 100 400 0 100 400 0 100 400 C3SCHTiO2BMTiO2C3SCHECountsC3SCHEEECHE CEC2S@ 56 days@ 91 days@ 120 days@ 150 days@ 180 days@ 360 daysCH CH C3STiO2E C Diffraction Angle 2 Figure 155: XRD of Hydration Products, Case C-3.6-3.8 C 3 S= Tricalcium Silicates, C 2 S=Dicalcium Silicates, TiO 2 ,=Rutile CH=Ca (OH) 2 ,, BM=C 4 AF, E=Ettringite, G=gypsum,C= Calcite KS=Arcanite 186

PAGE 207

Appendix B: (Continued) 10 20 30 40 50 un t s 0 100 400 900 0 100 400 0 100 400 0 100 400 0 100 400 0 100 400 0 100 400 C3SC3SCHTiO2BMC3STiO2C3SCHC3SC3SCHTiO2CHC2SCH CountsCH@ 8 hours@ 30 minutes@ 24 hours@ 3 days@ 7 days@ 14 days@ 28 daysKSC3SEEE 10 20 30 40 50 0 100 400 900 0 100 400 900 0 100 400 0 100 400 0 100 400 900 0 100 400 900 C3SC3SCHTiO2BMTiO2CHEC3STiO2CountsC3SCHEEECHEEC2S @ 56 da y s@ 91 days@ 120 days@ 150 days@ 180 days@ 360 daysCH C C H C Diffraction Angle 2 Figure 156: XRD of Hydration Products, Case C-AR-3.8 C 3 S= Tricalcium Silicates, C 2 S=Dicalcium Silicates, TiO 2 ,=Rutile CH=Ca (OH) 2 ,, BM=C 4 AF, E=Ettringite, G=gypsum,C= Calcite KS=Arcanite 187

PAGE 208

Appendix B: (Continued) 10 20 30 40 50 ts 0 100 400 0 100 400 0 100 400 900 0 100 400 900 0 100 400 900 0 100 400 900 0 400 C3SC3SCHTiO2EC3STiO2C3SCHEC3SC3SC3SBMCHTiO2@ 24 hours@ 8 hours@ 3 days@ 7 days@14 days@ 28 daysCHCH BMEEBM GGCounts@ 4 hours 10 20 30 40 50 s 0 100 400 900 0 100 400 900 0 100 400 900 0 100 400 900 0 100 400 0 100 400 900 C3SC3SCHTiO2BMTiO2C3SCHC3SCountsC3SCHEEECHE CC2S@ 56 days@ 91 days@ 120 days@ 150 days@ 180 days@ 360 daysECTiO2E CH Diffraction Angle 2 Figure 157: XRD of Hydration Products, Case C-5.0-AR C 3 S= Tricalcium Silicates, C 2 S=Dicalcium Silicates, TiO 2 ,=Rutile CH=Ca (OH) 2 ,, BM=C 4 AF, E=Ettringite, G=gypsum,C= Calcite 188

PAGE 209

Appendix B: (Continued) 10 20 30 40 50 ts 0 400 900 0 100 400 900 0 100 400 900 0 400 900 0 400 0 100 400 900 0 400 C3SC3SCHTiO2EC3STiO2C3SCHC3SC3SC3SBMCHTiO2@ 24 hours@ 8 hours@ 3 days@ 7 days@14 days@ 28 daysCHCH BMEBM GGCounts@ 4 hoursEE Pi i [ 2 T h ] 10 20 30 40 50 ts 0 100 400 900 0 400 1600 0 100 400 0 100 400 900 0 100 400 900 0 100 400 C3SCH BME CHCHC2SCounts@ 56 days@ 91 days@ 120 days@ 150 days@ 180 days@ 360 daysEEECH TiO2 CHCHTiO2TiO2CC Diffraction Angle 2 Figure 158: XRD of Hydration Products, Case C-3.6-AR C 3 S= Tricalcium Silicates, C 2 S=Dicalcium Silicates, TiO 2 ,=Rutile CH=Ca (OH) 2 ,, BM=C 4 AF, E=Ettringite, G=gypsum,C= Calcite 189

PAGE 210

Appendix B: (Continued) 10 20 30 40 50 ts 0 100 400 0 400 900 0 100 400 900 0 400 0 400 0 400 0 100 400 900 C3SC3SCHTiO2EC3STiO2C3SCHEC3SC3SC3SBMCHTiO2@ 24 hours@ 8 hours@ 3 days@ 7 days@14 days@ 28 daysCH BMBM Counts@ 4 hoursEECHC2S 10 20 30 40 50 ts 0 100 400 900 0 100 400 900 0 100 400 900 0 100 400 900 0 100 400 900 0 100 400 CH BME CHCHC2SCounts@ 56 days@ 91 days@ 120 days@ 150 days@ 180 days@ 360 daysEEECH TiO2 CHCHTiO2TiO2CC C2S Diffraction Angle 2 Figure 159: XRD of Hydration Products, Case C-AR C 3 S= Tricalcium Silicates, C 2 S=Dicalcium Silicates, TiO 2 ,=Rutile CH=Ca (OH) 2 ,, BM=C 4 AF, E=Ettringite, G=gypsum,C= Calcite 190

PAGE 211

Appendix B: (Continued) 10 20 30 40 50 nt s 0 100 400 0 100 400 0 100 400 900 0 100 400 0 100 400 900 0 100 400 0 100 400 C3SC3SCHTiO2BMKSEEC3STiO2C3SCHEC3SC3SC3SKSKSCHTiO2@ 24 hours@ 8 hours@ 30 minutes@ 3 days@ 7 days@14 days@ 28 daysSyng.EC3SCountsC2SCHCHEEE Pi i [ 2 T h ] 10 20 30 40 50 ts 0 100 400 0 100 400 0 100 400 0 100 400 0 100 400 900 0 100 400 C3STiO2CHTiO2CountsCHEECHE CE CHC2S@ 56 days@ 91 days@ 120 days@ 150 days@ 180 days@ 360 daysEEBMCHECCHTiO2 Diffraction Angle 2 Figure 160: XRD of Hydration Products, Case C-5.0-1.5 C 3 S= Tricalcium Silicates, C 2 S=Dicalcium Silicates, TiO 2 ,=Rutile CH=Ca (OH) 2 ,, BM=C 4 AF, E=Ettringite, G=gypsum,C= Calcite KS=Arcanite 191

PAGE 212

Appendix B: (Continued) 10 20 30 40 50 nt s 0 100 400 0 100 400 0 100 400 0 100 400 0 100 400 0 100 400 900 0 100 400 900 C3SC3SCHBMEEC3STiO2C3SCHEC3SC3SC3SCHTiO2@ 24 hours@ 8 hours@ 3 days@ 7 days@14 days@ 28 daysECountsC2SCHCHEEE@ 4 hoursTiO2 10 20 30 40 50 0 100 400 0 100 400 0 100 400 0 100 400 0 100 400 900 0 100 400 C3STiO2CHTiO2CountsCHECH CE CHC2S@ 56 days@ 91 days@ 120 days@ 150 days@ 180 days@ 360 daysEEBMCHECCHTiO2C2S Diffraction Angle 2 Figure 161: XRD of Hydration Products, Case C-3.6-1.5 C 3 S= Tricalcium Silicates, C 2 S=Dicalcium Silicates, TiO 2 ,=Rutile CH=Ca (OH) 2 ,, BM=C 4 AF, E=Ettringite, G=gypsum,C= Calcite 192

PAGE 213

Appendix B: (Continued) 10 20 30 40 50 un t s 0 100 400 0 100 400 0 100 400 0 100 400 900 0 100 400 0 100 400 900 0 100 400 C3SC3SCHBMEEC3STiO2C3SCHEC3SC3SCHTiO2@ 24 hours@ 8 hours@ 3 days@ 7 days@14 days@ 28 daysECountsC2SCHCHEEE@ 4 hoursTiO2BMCHCH 10 20 30 40 50 0 100 400 0 100 400 0 100 400 0 100 400 0 100 400 0 100 400 C3STiO2CHTiO2CountsCHECH CE CHC2S@ 56 days@ 91 days@ 120 days@ 150 days@ 180 days@ 360 daysEEBMCHECCHTiO2C2S Diffraction Angle 2 Figure 162: XRD of Hydration Products, Case C-AR-1.5 C 3 S= Tricalcium Silicates, C 2 S=Dicalcium Silicates, TiO 2 ,=Rutile CH=Ca (OH) 2 ,, BM=C 4 AF, E=Ettringite, G=gypsum,C= Calcite 193

PAGE 214

Appendix B: (Continued) 10 20 30 40 50 un t s 0 100 400 0 100 400 0 100 400 0 100 400 0 100 400 0 100 400 0 100 400 C3SC3SC3STiO2C3SCHC3SC3SC3SKSTiO2@ 24 hours@ 8 hours@ 30 minutes@ 3 days@ 7 days@14 days@ 28 daysSyng.CountsC2SCHBMEECHKSTiO2CHEESyng.KSCH 10 20 30 40 50 t s 0 100 400 0 100 400 0 100 400 0 100 400 0 100 400 900 0 100 400 C3SCHTiO2BMTiO2CHC3STiO2CountsC3SCHECHCE CEC2S@ 56 days@ 91 days@ 120 days@ 150 days@ 180 days@ 360 daysEEECH CH Diffraction Angle 2 Figure 163: XRD of Hydration Products, Case C-5.0-2.0 C 3 S= Tricalcium Silicates, C 2 S=Dicalcium Silicates, TiO 2 ,=Rutile CH=Ca (OH) 2 ,, BM=C 4 AF, E=Ettringite, G=gypsum,C= Calcite, KS=Arcanite 194

PAGE 215

Appendix B: (Continued) 10 20 30 40 50 un t s 0 100 400 0 100 400 0 100 400 0 100 400 0 100 400 0 100 400 0 100 400 C3SC3SC3STiO2C3SC3SC3SKSTiO2@ 24 hours@ 8 hours@ 30 minutes@ 3 days@ 7 days@14 days@ 28 daysCountsC2SCHBMTiO2Syng.CH CH KSEE CHC3SCH 10 20 30 40 50 ts 0 100 400 0 100 400 0 100 400 0 100 400 0 400 1600 0 400 1600 C3SCHTiO2BMTiO2CHTiO2CountsCHECHE CEC2S@ 56 days@ 91 days@ 120 days@ 150 days@ 180 days@ 360 daysEECH CHC2S CEE Diffraction Angle 2 Figure 164: XRD of Hydration Products, Case C-3.6-2.0 C 3 S= Tricalcium Silicates, C 2 S=Dicalcium Silicates, TiO 2 ,=Rutile CH=Ca (OH) 2 ,, BM=C 4 AF, E=Ettringite, G=gypsum,C= Calcite 195

PAGE 216

Appendix B: (Continued) 10 20 30 40 50 ts 0 100 400 0 100 400 0 100 400 0 100 400 0 100 400 0 100 400 900 0 100 400 C3SC3SC3STiO2C3SCHC3STiO2C2SCH Counts@ 7 days@ 28 daysECH@ 8 hours@ 24 hours@ 3 days@ 14 daysCHCHCH@ 30 minutesEEE 10 20 30 40 50 ts 0 100 400 0 100 400 0 100 400 0 100 400 0 100 400 900 0 400 1600 C3SCHTiO2BMTiO2CHTiO2CountsCHECHE CEC2S@ 56 days@ 91 days@ 120 days@ 150 days@ 180 days@ 360 daysEECH CHC2S CEE Diffraction Angle 2 Figure 165: XRD of Hydration Products, Case C-AR-2.0 C 3 S= Tricalcium Silicates, C 2 S=Dicalcium Silicates, TiO 2 ,=Rutile CH=Ca (OH) 2 ,, BM=C 4 AF, E=Ettringite, G=gypsum,C= Calcite 196

PAGE 217

Appendix C: Schematic Representation of C 3 S Hydration Portland Cement Hydration Portland Cement HydrationPortland Cement Hydration CC33SS Initial ReactionInitial Reaction Induction PeriodInduction Period AccelerationAcceleration CC33SS Initial ReactionInitial Reaction Induction PeriodInduction Period AccelerationAccelerationHydrolosisHydrolosis Gel CoatGel Coat Permeable Permeable coatcoatHydrationHydration OHOH-CaCa+2+2 SiOSiO--44 HH22OO Figure 166: Schematic Representation of C 3 S Hydration 197

PAGE 218

ABOUT THE AUTHOR Mahmoud Halaweh received a B.Sc. in civil engineering from the University of Jordan, Amman-Jordan in 1991 and a M.S in ci vil engineering from the University of South Florida, Tampa in 2000. Mr. Halaweh is a registered professi onal engineer. He worked as a structural design engineer for 10 years for several consulting firms until joining the Ph.D. program in 2003. While in the Ph.D. program, Mr. Halaweh worked as a research assistant for multiple projects. He also worked as a teach ing assistant for several classes.