USF Libraries
USF Digital Collections

Effect of tricalcium silicate content on expansion in internal sulfate attack

MISSING IMAGE

Material Information

Title:
Effect of tricalcium silicate content on expansion in internal sulfate attack
Physical Description:
Book
Language:
English
Creator:
Whitfield, Troy T
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla
Publication Date:

Subjects

Subjects / Keywords:
Ettringite
C3S
Alkali
Heat cured
Mortar
Dissertations, Academic -- Civil Engineering -- Masters -- USF
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Abstract:
ABSTRACT: The purpose of this study was to determine the cementitious parameters and placement temperature that impact internal sulfate attack in concrete. Concrete structures make up a large percentage of the infrastructure and multifamily housing. Durability is very important. Cements can be formulated to reduce the impact of external environmental exposure such as high salinity from marine environments or high sulfate levels from soils or surface waters. Concrete is also subject to internal attack such as alkali aggregate reaction, (AAR), and delayed ettringite formation, (DEF). This study focused on some of the cement chemistry issues that determine susceptibility of cement to DEF. Expansion due to DEF can weaken the concrete matrix resulting in microcracks that in some cases may progress to severe matrix cracking. The end result is loss of load carrying capacity and costly repairs.In this study, mortar bars were made with the as received cement chemistry and using additions of sulfate, and alkalis. The bars were then heat cured at various temperatures and stored in a saturated lime solution at room temperature. Measurements were made at predetermined time intervals. The series of mixes were made to determine the effect of varying sulfate levels, heat curing temperature, and alkali content in order to isolate the effect of these constituents. The cements were selected on the basis of tricalcium aluminate, alkali content, sulfate levels, C3S levels and fineness. The results indicate that a relationship exists between the rate and level of expansion experienced by the mortar bars and cementitious parameters, namely, alkali content, sulfate content, C3S levels and heat curing temperature.
Thesis:
Thesis (M.A.)--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 Troy T. Whitfield.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 77 pages.

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 - 001797663
oclc - 156997625
usfldc doi - E14-SFE0001631
usfldc handle - e14.1631
System ID:
SFS0025949:00001


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 001797663
003 fts
005 20070726142617.0
006 m||||e|||d||||||||
007 cr mnu|||uuuuu
008 070726s2006 flu sbm 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0001631
040
FHM
c FHM
035
(OCoLC)156997625
049
FHMM
090
TA145 (ONLINE)
1 100
Whitfield, Troy T.
0 245
Effect of tricalcium silicate content on expansion in internal sulfate attack
h [electronic resource] /
by Troy T. Whitfield.
260
[Tampa, Fla] :
b University of South Florida,
2006.
3 520
ABSTRACT: The purpose of this study was to determine the cementitious parameters and placement temperature that impact internal sulfate attack in concrete. Concrete structures make up a large percentage of the infrastructure and multifamily housing. Durability is very important. Cements can be formulated to reduce the impact of external environmental exposure such as high salinity from marine environments or high sulfate levels from soils or surface waters. Concrete is also subject to internal attack such as alkali aggregate reaction, (AAR), and delayed ettringite formation, (DEF). This study focused on some of the cement chemistry issues that determine susceptibility of cement to DEF. Expansion due to DEF can weaken the concrete matrix resulting in microcracks that in some cases may progress to severe matrix cracking. The end result is loss of load carrying capacity and costly repairs.In this study, mortar bars were made with the as received cement chemistry and using additions of sulfate, and alkalis. The bars were then heat cured at various temperatures and stored in a saturated lime solution at room temperature. Measurements were made at predetermined time intervals. The series of mixes were made to determine the effect of varying sulfate levels, heat curing temperature, and alkali content in order to isolate the effect of these constituents. The cements were selected on the basis of tricalcium aluminate, alkali content, sulfate levels, C3S levels and fineness. The results indicate that a relationship exists between the rate and level of expansion experienced by the mortar bars and cementitious parameters, namely, alkali content, sulfate content, C3S levels and heat curing temperature.
502
Thesis (M.A.)--University of South Florida, 2006.
504
Includes bibliographical references.
516
Text (Electronic thesis) 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 77 pages.
590
Adviser: Abla M. Zayed, Ph.D.
653
Ettringite.
C3S
Alkali.
Heat cured.
Mortar.
690
Dissertations, Academic
z USF
x Civil Engineering
Masters.
773
t USF Electronic Theses and Dissertations.
4 856
u http://digital.lib.usf.edu/?e14.1631



PAGE 1

Effect of Tricalcium Silicate Content on Expansion in Internal Sulfate Attack by Troy T. Whitfield A thesis submitted in partial fulfillment of the requirements for the degree of Masters of Science in Civil Engineering Department of Civil and Environmental Engineering College of Engineering University of South Florida Major Professor: Abla M. Zayed, Ph.D. Rajan Sen, Ph.D. Gray Mullins, Ph.D. Date of Approval: June 6, 2006 Keywords: ettringite, C 3 S, Alkali, heat cured, mortar Copyright 2006, Troy T. Whitfield

PAGE 2

i TABLE OF CONTENTS LIST OF TABLES ii LIST OF FIGURES iv LIST OF SYMBOLS AND ABBREVIATIONS x ABSTRACT xi CHAPTER 1. INTRODUCTION 1 1.1 Objective 1 1.2 History of the Manufacture of Portland Cement 2 1.3 Chemistry of Portland Cement 3 1.4 Hydration of the Major Components of Cement 6 1.5 Recent Developments 7 1.6 Review of Previous Investigations 8 CHAPTER 2. EXPERIMENTAL PROCEDURES 13 2.1 Composition Determination 13 2.2 Materials 14 2.3 Procedure 15 2.3.1 Mix Design 15 2.3.2 Procedure for Mixing Mortars 16 2.3.3 Slump Test 17 2.4 Casting of Bars 17 2.5 Heat Curing Cycle 18 2.6 Length Change on Mortar Bars 19 2.6.1 Measurement Cycle 19 2.6.2 Measurement Procedure 20 CHAPTER 3. RESULTS AND DISCUSSION 21 3.1 Length Measurement Results 21 3.2 XRD Results Using Reitveld Analysis 61 CHAPTER 4. CONCLUSIONS AND RECOMMENDATIONS 72 REFERENCES 77

PAGE 3

ii LIST OF TABLES Table 1: ASTM Portland cement types and uses taken from FHWA website 5 Table 2: Bogue calculations, fineness and C3S/C2S ratio 14 Table 3: Chemical analysis of th e two cements used in the study 15 Table 4: Analysis of cements us ing the internal standard method and Reitveld analysis 15 Table 5: Spreadsheet used in mix design 16 Table 6: XRD results from cement C with 5.0% sulfate content and 1.5% alkali co ntent heat cured at 90C after 102 days storag e in a saturated lime solution 62 Table 7: XRD results from ceme nt C with 5.0% sulfate content and 1.5% alkali content heat cured at 80C after 116 days storage in a saturated lime solution 62 Table 8: XRD results from ceme nt C with 3.6% sulfate content and 1.5% alkali content heat cured at 90C after 122 days storage in a saturated lime solution 63 Table 9: XRD results from cement MH-3 with 5.0% sulfate content and 1.5% alkali cont ent heat cured at 90C after 123 days storage in a saturated lime solution 63 Table 10: XRD results from cement MH-3 with 5.0% sulfate content and 1.5% alkali co ntent heat cured at 80C after 121 days storage in a saturated lime solution 64 Table 11: XRD results from cement MH-3 with 5.0% su lfate content and 2.0% alkali co ntent heat cured at 90C after 125 days storage in a saturated lime solution 64

PAGE 4

iii Table 12: XRD results from cement MH-3 with 5.0% su lfate content and 2.0% alkali co ntent heat cured at 80C after 120 days storage in a saturated lime solution 64 Table 13: XRD results from ceme nt E with 5.0% sulfate content and 1.5% alka li content heat cured at 90C after 285 days stor age in a saturated lime solution 65 Table 14: XRD results from cement C with as received sulfate levels 20 minutes after the initial hydration 66 Table 15: XRD results from cement C as doped to a 1.5% alkali and 5.0% sulfate level 30 minutes after initial hydration 66 Table 16: XRD results from cement C with as received sulfate levels 75 minutes after the initial hydration 67 Table 17: XRD results from cement C as doped to a 1.5% alkali and 5.0% sulfat e level 75 minutes afte r initial hydration 67 Table 18: XRD results from cement C with as received sulfate levels 17 hours and 15 minutes after the initial hydration 68 Table 19: XRD results from cement C as doped to a 1.5% alkali and 5.0% sulfate leve l 17 hours and 15 minutes after initial hydration 68 Table 20: XRD results from cement C with as received sulfate levels 28 days after the initial hydration 69 Table 21: XRD results from cement C with as received sulfate levels 60 days after the initial hydration 69 Table 22: XRD results from cement C as doped to a 1.5% alkali and 5.0% sulfate level 28 days after initial hydration 70 Table 23: XRD results from cement C as doped to a 1.5% alkali and 5.0% sulfate level 60 days after initial hydration 70

PAGE 5

iv LIST OF FIGURES Figure 1: Heat curing cycle adopted in this study 19 Figure 2: Mortar bar showing bending 20 Figure 3: All three cements as received and heat cured at 90C 22 Figure 4: All three cements as received and heat cured at 80C 23 Figure 5: All three cements as received and heat cured at 60C 23 Figure 6: All three cements as received and heat cured at 23C 24 Figure 7: Cement E showing the effect of temperature on as received chemistry 25 Figure 8: Cement MH-3 showing the effect of temperature on as received chemistry 25 Figure 9: Cement C showing th e effect of temperature on as received chemistry 26 Figure 10: The effect of alkali co ntent after 90C heat cure cycle for cement E 26 Figure 11: The effect of alkali cont ent after 80C heat cure cycle for cement E 27 Figure 12: The effect of alkali cont ent after 60C heat cure cycle for cement E 27 Figure 13: The effect of alkali conten t after 23C heat cure cycle for cement E 28 Figure 14: The effect of alkali co ntent after 90C heat cure cycle for cement MH-3 29 Figure 15: The effect of alkali cont ent after 80C heat cure cycle for cement MH-3 29

PAGE 6

v Figure 16: The effect of alkali co ntent after 60C heat cure cycle for cement MH-3 30 Figure 17: The effect of alkali co ntent after 23C heat cure cycle for cement MH-3 30 Figure 18: The effect of alkali co ntent after 90C heat cure cycle for cement C 31 Figure 19: The effect of alkali co ntent after 80C heat cure cycle for cement C 31 Figure 20: The effect of alkali co ntent after 60C heat cure cycle for cement C 32 Figure 21: The effect of alkali co ntent after 23C heat cure cycle for cement C 32 Figure 22: Expansion of cement E with 1.5% alkali content after 90C heat cure 33 Figure 23: Expansion of cement MH-3 with 2.0 % alkali content after 90C heat cure 34 Figure 24: Expansion of cement MH-3 with 1.5% alkali content after 90C heat cure 34 Figure 25: Expansion of cement C with 1.5% alkali content after 90C heat cure 35 Figure 26: Expansion of cement E with 1.5% alkali content after 80C heat cure 35 Figure 27: Expansion of cement MH-3 with 2.0% alkali content after 80C heat cure 36 Figure 28: Expansion of cement MH-3 with 1.5% alkali content after 80C heat cure 36 Figure 29: Expansion of cement C with 1.5% alkali content after 80C heat cure 37 Figure 30: Expansion of cement E with 1.5% alkali content after 60C heat cure 37

PAGE 7

vi Figure 31: Expansion of cement MH-3 with 2.0% alkali content after 60C heat cure 38 Figure 32: Expansion of cement MH-3 with 1.5% alkali content after 60C heat cure 38 Figure 33: Expansion of cement C with 1.5% alkali content after 60C heat cure 39 Figure 34: The effect of cement composition on expansion at constant al kali and sulfate levels af ter 90C heat cure 40 Figure 35: The effect of cement composition on expansion at constant al kali and sulfate levels af ter 80C heat cure 40 Figure 36: The effect of curi ng temperature on cement with 1.5% alka li and 5% sulfate afte r 60C heat cure 41 Figure 37: The effect of curi ng temperature on cement with 1.5% alka li and 5% sulfate afte r 23C heat cure 41 Figure 38: The effect of curi ng temperature on cement with 1.5% alkali and 3.6 % sulfate after 90C heat cure 42 Figure 39: The effect of curi ng temperature on cement with 1.5% alkali and 3.6 % sulfate after 80C heat cure 42 Figure 40: The effect of curi ng temperature on cement with 1.5% alkali and 3.6 % sulfate after 60C heat cure 43 Figure 41: The effect of curi ng temperature on cement with 1.5% alkali and 3.6 % sulfate after 23C heat cure 43 Figure 42: The effect of heat cure temperature on cement E with alkali = 1.5% and SO3 = 5.0% 44 Figure 43: The effect of heat cure temperature on cement MH-3 with alkali = 1.5% and SO3 = 5.0% 44 Figure 44: The effect of heat cu re temperature on cement C with alkali = 1.5% and SO3 = 5.0% 45 Figure 45: The rate of expansion for cements with SO3 = 5% and Alkali =1.5% cured at 90 C 46

PAGE 8

vii Figure 46: The rate of expansion for cements with SO3 = 5% and alkali =1.5% cured at 80C 46 Figure 47: The rate of expansion for cements with SO3 = 5% and alkali =1.5% cured at 60C 47 Figure 48: The rate of expansion for cements with SO3 = 5% and alkali =1.5% cured at 23C 47 Figure 49: The rate of expansion for cements with SO3 = 3.6% and alkali = 1.5% cured at 90C 48 Figure 50: The rate of expansion for cements with SO3 = 3.6% and alkali = 1.5% cured at 80C 48 Figure 51: The rate of expansion for cements with SO3 = 3.6% and alkali = 1.5% cured at 60C 49 Figure 52: The rate of expansion for cements with SO3 = 3.6% and alkali = 1.5% cured at 23C 49 Figure 53: The rate of expansion for cements with as received SO3 and alkali = 1.5% cured at 90C 50 Figure 54: The rate of expansion for cements with as received SO3 and alkali = 1.5% cured at 80C 50 Figure 55: The rate of expansion for cements with as received SO3 and alkali = 1.5% cured at 60C 51 Figure 56: The rate of expansion for cements with as received SO3 and alkali = 1.5% cured at 23C 51 Figure 57: The expansion rate for cements with as received SO3 and alkali cured at 90C 52 Figure 58: The expansion rate for cements with as received SO3 and alkali cured at 80C 52 Figure 59: The expansion rate for cements with as received SO3 and alkali cured at 60C 53 Figure 60: The expansion rate for cements with as received SO3 and alkali cured at 23C 53

PAGE 9

viii Figure 61: Expansion exhibited by cement E at one-hundred and twenty days 54 Figure 62: Expansion exhibited by cement E at one-hundred and eighty days 55 Figure 63: Expansion exhibited by cement E at two-hundred and seventy days 55 Figure 64: Expansion exhibited by cement MH-3 at one-hundred and twenty days 56 Figure 65: Expansion exhibited by cement MH-3 at one-hundred and fifty days 56 Figure 66: Expansion exhibited by cement C at one-hundred and twenty days 57 Figure 67: Expansion exhibited by cement C at one-hundred and eighty days 58 Figure 68: Expansion exhibited by cement C at two-hundred and seventy days 58 Figure 69: Comparison of expansion exhi bited by all cements with a sulfate content of 5. 0% and alkali content of 1.5% at 120 days 59 Figure 70: Comparison of expansion exhi bited by all cements with a sulfate content of 5. 0% and an alkali content of 1.5% at 180 days 59 Figure 71: Comparison of expansion exhi bited by all cements with a sulfate content of 3. 6% and an alkali content of 1.5% at 120 days 60 Figure 72: Comparison of expansion exhi bited by all cements with a sulfate content of 3. 6% and an alkali content of 1.5% at 180 days 61 Figure 73: Ettringite and portlandite formation over time from XRD results of cement C with 1.5% alkali and 5.0% sulfate level 71 Figure 74: Effect of C3S level upon expansion of ba rs with 5.0% sulfate and 1.5% alkali at 180 days in limewater 74

PAGE 10

ix LIST OF SYMBOLS AND ABBREVIATIONS AASHTO American Association of St ate Highway and Transportation Officials ASTM American Society of Testing and Materials ISA Internal Sulfate Attack SEM Scanning Electron Microscope XRD X-Ray Diffraction Cement Chemistry Abbreviations A Alumina, Al2O3 C Calcium Oxide CaO F Ferric Oxide, Fe2O3 H Water, H2O S Silica, SiO2 Š Sulfur Trioxide, SO3 C3A Tricalcium Aluminate, 3CaO•Al2O3 C4AF Tetracalcium Aluminoferrite, 4CaO•Al2O3•Fe2O3 C2S Dicalcium Silicate, 2CaO•SiO2 C3S Tricalcium Silicate, 3CaO•SiO2 CH Calcium Hydroxide, Ca(OH)2 C Š H2 Gypsum, Ca2SO4•2H2O C-S-H Calcium Silicate Hydrate, nCaO•SiO2•mH2O

PAGE 11

x C6A Š H32 Ettringite, 3CaO•Al2O3•3CaSO4•32H2O C4A Š H12 Monosulfoaluminate, 3CaO•Al2O3•CaSO4•12H2O

PAGE 12

xi EFFECT OF TRICALCIUM SILICATE CONTENT ON EXPANSION IN INTERNAL SULFATE ATTACK Troy T. Whitfield ABSTRACT The purpose of this study was to determ ine the cementitious parameters and placement temperature that impact internal sulfat e attack in concrete. Concrete structures make up a large percentage of the infrastructure and multifamily housing. Durability is very important. Cements can be formulated to reduce the impact of external environmental exposure such as high salinity from marine environmen ts or high sulfate levels from soils or surface waters. Concrete is also subject to internal attack such as alkali aggregate reaction, (AAR), and delaye d ettringite formation, (DEF). This study focused on some of the cement chemistry issues that determine susceptibility of cement to DEF. Expansion due to DEF can weaken th e concrete matrix resulting in microcracks that in some cases may progress to severe ma trix cracking. The end result is loss of load carrying capacity and costly repairs. In this study, mortar bars were made with the as received cement chemistry and using additions of sulfate, and alkalis. Th e bars were then h eat cured at various temperatures and stored in a saturated lime so lution at room temperature. Measurements were made at predetermined time intervals. The series of mixes were made to determine the effect of varying sulfate levels, heat curi ng temperature, and alkali content in order to

PAGE 13

xii isolate the effect of these constituents. The cements were selected on the basis of tricalcium aluminate, alkali content, sulfate levels, C3S levels and fineness. The results indicate that a relationship exists between th e rate and level of expansion experienced by the mortar bars and cementitious parameters, namely, alkali content, sulfate content, C3S levels and heat curing temperature.

PAGE 14

1 CHAPTER 1 INTRODUCTION 1.1 Objective There are a large number of concrete structures built every year for infrastructure, industry, and multi-family housing. Durability is a very important factor to consider in the design process. Environmental exposure can lead to durability issues. Most mix designs take this into account. Other durabi lity issues come from within the concrete members themselves. Examples of these include delayed ettringite formation, (DEF), and alkali aggregate reaction, (AAR). DEF or ISA, (internal sulfate attack), refers to ettringite formed in the cement after its in itial set and hardening. It was originally considered a problem for heat cured precast cement members such as railway sleepers. Initial research focused on determining a ma ximum temperature that the members could be exposed to without durability issues. The research then expanded to cement chemistry. The earliest research focused upon SO3 and C3A content. Present research has expanded the scope to include alkali content in the form of Na2Oeqivilant and C3S. It is common knowledge that C3A, C3S, Alkali content, heat curing temperature, and internally generated heat from thick cross-s ections work both separately and together to provide a mechanism for DEF. There is mu ch research linking alkali levels, sulfate levels and heat curing temperatures to the e xpansion due to delayed ettringite formation that is experienced by laborator y concretes. The ettringite formed in the cement after its

PAGE 15

2 initial set and hardening causes internal stre sses that lead to durability issues. Many researchers have explored the link between C3S levels and expansion due to ettringite formation. A common problem in these studie s is that the cement chemistry is not closely controlled enough to definitively state a correlation ex ists between the C3S level of a cement and its propensity to experience expansion due to ettringite formation over time. This is because the C3A levels, fineness, sulfate cont ent, and alkali content also have a large effect. In this study, there was an attempt to eliminate these variables to isolate the effect of higher levels of C3S on expansion. 1.2 History of the Manufacture of Portland Cement Portland cement was invented in Engla nd in 1824. It differs from lime based cements by the manufacturing method and the chem ical reactions that take place. In the 1870’s Portland cement began to be manufactured in the Unite d States. The process by which it is produced has changed little in con cept since that time. What has changed is the equipment used to manufacture the cement, better control of the chemistry of the raw material, and the ability to mon itor and control the temperature. The process to manufacture Portland cem ent requires the burning of a finely ground mixture of about 75% limestone for the calcium oxide, along with shale or clay to provide the needed silica, alumina, and ir on oxides. The burning of this material takes place at around 3000F in a kiln. The product that is produced is called clinker and consists of C3S, C2S, C4AF, C3A and various other minor constituents such as MgO, and alkali containing compounds. This clinke r then finely ground and combined with calcium sulfate to produce Portland cement. Although the process seems simple, it

PAGE 16

3 requires constant attention to detail sin ce the constituent proportions, grinding size, thoroughness of mixing, and kiln temperatures must be monitored closely as they have large effects on the end product. A typical kiln is located near the sour ce of the bulk raw material used in manufacturing. In Florida, the material excav ated consists of sand, silt, and clay along with limestone. This is finely ground and thoroughly mixed before sampling. Any deficiencies are corrected by adding the needed iron, silica and calcium to the finely ground mixture. These are added by using limes tone, sand, or iron pellets. The batch is then run through a rotary kiln and the temperature is closely maintained throughout the process. The clinker that is produced is cooled at a fixed ra te and ground in a ball mill to a fine powder. This powder is then mixe d with a predetermined proportion of calcium sulfate to produce cement. The calcium su lfate can take the fo rm of Anhydrite, Hemihydrite, or Gypsum. 1.3 Chemistry of Portland Cement Clinker is chiefly composed of four compounds. These are tricalcium silicate (3CaO•SiO2), dicalcium silicate (2CaO•SiO2), tricalcium aluminate (3CaO•Al2O3), and tetra-calcium aluminoferrite (4CaO•Al2O3Fe2O3). In this paper they will be abbreviates using shorthand notation as C3S, C2S, C3A, and C4AF. Also present in clinker are small amounts of free lime, alkali sulfates, alka li oxides, and magnesium oxides. Most of the strength of the cement is due to the reaction of water with C3S and C2S. This reaction with water results in the formation of calcium silicate hydrate (average composition 3CaO•2SiO2•3H2O) and calcium hydroxide (Ca [OH] 2) the

PAGE 17

4 shorthand notations for these two products ar e C-S-H and CH respectively. The chemical reaction that results in the formation of these products is written below. 2C3S + 6H2O C-S-H + 3CH 2C2S + 4H2O C-S-H + CH It is evident from the above reactions that although the pr oducts are the same, they vary in proportion. C-S-H is an amorphous gel that actua lly has a large variation in chemistry. An example of this would be the Ca/Si ratio which can vary between unity and 2:1. CH has a defined structure as compar ed to the amorphous structure of C-S-H. Although most of the strength of cement is from CH and C-S-H, the aluminates compounds have the largest effect on durabi lity. Internally or externally generated sulfates combine with the aluminates and l ead to the formation of sulfoaluminates. Because the formation of some of the sulfoaluminates is expansive, the reaction can lead to tensile stresses that result in expansion and finally the deteriorati on of concrete due to cracking. Portland cement is subdivided into five types by the American Society of Testing and Materials depending upon both chemistry a nd fineness. These cement types meet the chemical and physical requirements needed for specific purposes.

PAGE 18

5 Table 1: ASTM Portland cement types and uses taken from FHWA website Cement Type Use I General purpose cement II Use for moderate sulfate attack resistance III Use when high early strength is required IV Use when low heat of hydration is required for massive structures V Use when high sulfate resistance is required Generally the C3A content is limited to 15% for type III, 8% for type II, 7% for type IV and 5% for type V. The C3A levels are used by ASTM to determine sulfate resistance. In addition type IV also has a limit of 40% C2S and 35% C3S to help control the hydration temperature. Type I and III have the same chemistry requirements but differ in how fine they are ground. Type III achieves higher early strength by the greater rate of hydration due to it s smaller particle size. ASTM also limits the maximum sulfate content to 2.3% for type IV and V. Type I is limited to 3.0% if C3A is less than 8% and 3.5% when the C3A content is greater than 8%. Type II is limited to 3% maximum sulfate content. Type III is similar to type I with the sulfate content limited to 3.5% if C3A is less than 8% and 4.5% when the C3A content is greater than 8%. The larges t source of sulfate is from the added calcium sulfate, but sulfur can also be present in alkali sulfates as well as a trace element in all four of the basic constituents of Portland cement clinker.

PAGE 19

6 1.4 Hydration of the Major Components of Cement The hydration of cement changes the majo r components such as calcium silicates and calcium aluminates/ferrites into a se ries of calcium silicate hydrates, calcium hydroxides, and calcium aluminate/ferrite hydr ates. One of the minor components is calcium sulfate. This reacts with the calcium aluminates to form ettringite. The early formed ettringite helps cont rol the hydration process by fo rming a coating on the calcium silicates thereby restricting access to water and slowing thei r hydration. As long as there is not too much calcium sulfate, and the earl y formed ettringite is stable, the hardened cement is not susceptible to cracking caused by delayed ettringite formation. Unfortunately this early ettri ngite is partially destroyed dur ing the heat curing process. This destruction allows the bound sulfates to be temporarily adsorbed by the C-S-H gel and also to concentrate in the pore water. Calcium silicates react with water to fo rm calcium hydroxide and calcium silicate hydrate gel. The latter is an amorphous solid that can have varying chemistry. It is usually written as CxSHn with the x falling between 0.8 and 2.5. The hydration of calcium aluminates yields C3AH6 which then reacts with the sulfate present to form monosulfate or ettrin gite depending upon the su lfate content of the cement. The sulfate also is adsorbed to a le sser extent by the C-S-H ge l. The stability of ettringite depends upon the temperature at wh ich the cement is exposed, the pH of the pore solution, and the sulfate concen tration of the pore water. Calcium alumino-ferrites react with water to form a solid solution series with a chemistry that falls between C3AH6 and C3FH6 These also react with sulfate to form a more iron rich monosulfate and ettringite. The literature refers to these compounds as

PAGE 20

7 mono-sulfo-ferrite and tri-sulfo-ferrite respectively. These behave similarly to their aluminum rich counterparts with the exception of the iron rich monosulfate being slightly less stable. L.O. Hoglund ran a series of experiments to find out the regions of stability for ettringite in regards to temper ature and sulfate concentration of the pore water. Ettringite begins to decompose at temperatures above 25C if the sulfate leve l in the pore waters falls below 0.25mg/l. This fact directly influe nces the result of the experiments used in this paper. It would be expected that in th e low sulfate mortars the early formed ettringite would have fully decomposed during heat cu re. The mixes with five percent sulfate would not share a similar fate. Some of the early formed ettringite would have survived in these mixes, this would account for the higher levels of expansion experienced by these mixes at early ages. 1.5 Recent Developments In recent years there has been a shake-up of the industry due to both fuels used in the kiln, and making the process more effici ent by re-circulating the exhaust gasses to pre-heat the raw materials. In earlier times the fuel used was locally produced natural gas, coal or oil. This has changed in recen t years because of higher energy costs. Much of the domestic industry has been purcha sed by a growing Mexican company. This company generated much of the money needed to expand by making the process more efficient and using petroleum coke, a waste pr oduct of oil refining, as a fuel. The 1970’s saw the use of hazardous wastes as both a fuel and to increase profitability by the revenue generated by its high temperature destruction. Re-circulating the exhaust gasses has lead

PAGE 21

8 to higher alkali contents. As in all processe s, changing any of the inputs results in a product that is also changed. In the 1970’s durability became an issue. A series of failures of precast cement elements were initially attrib uted to alkali aggregate react ion. Upon further inspection, expansion due to internal sulfate attack was blam ed. Internal sulfate attack is the result of the destruction of ettringite that was formed during initial hydrati on. The sulfates that were then released are adsorbed into the C-S-H gel or held in the pore water. The formation of Ettringite after initial hydrati on can be expansive and lead to damage in concrete. The initial studies showed a strong as sociation with the rate of internal sulfate attack to the levels of C3A, alkalis, sulfates and high heat curing temperatures. In response both AASHTO and ASTM set limits on cement chemistry. Several states and countries have set limits on the maximum temperature that the concrete member experiences. Texas through MNM 116 and 117 allows up to 150F (66C), or 82C for truly dry service conditions. Germany and Cana da limit the temperature to 60C. In the United States the NPCA recommends a maxi mum of 65C unless safeguards are taken to prevent DEF. Even with the safeguards, th e NPCA sets the temperature limit to 70C. 1.6 Review of Previous Investigations Several studies have noticed a correla tion between the occu rrence of internal sulfate attack and high C3S content. The effect of high C3S levels is chiefly in its effect on the hydration products and the heat of hydration. Rasheeduzzafar in his 1992 paper states the many reasons why the C3S/C2S ratio affects the resistance of a past e to sulfate attack. Since C3S produces 2.2 times as much CH as an equivalent amount of C2S the largest effect of high levels of the former lies

PAGE 22

9 chiefly in its hydration product. Calcium hydroxi de acts as a buffer and is more easily dissolved from the paste than C-S-H. This dissolution can provide the calcium needed for other products, and can lead to higher porosity. Higher levels of calcium hydroxide result in increased gypsum formation. This leads to a corresp onding decrease in the strength of the paste, allowing the tensile st resses due to ettringite formation to overcome the tensile strength of concre te leading to cracking and spalling. In addition, the solubility of expansive hydrated calcium aluminates is significantly lowered in a saturated lime environment, the sulfate re action becomes topochemical and expansive by nature. Mehta hypothesized that the form that the et tringite crystals take is related to the levels of calcium hydroxide present. In ri ch environments, the ettringite formed is colloidal. The ability of ettringite to absorb water is greater than it would be if lath-like crystals were formed. The adsorption of water leads to expansion and ultimately destruction of the paste. Divet and Randriambololona found in their 1998 paper that a high C3S/C2S ratio results in a two fold effect on the formation of the C-S-H gel. Firs t, high levels of CH increase the lime/silica ratio (> 1.5) resulting in a weaker C-SH gel. This weaker paste will be damaged at lower stress levels. The stress can be caused by the reaction of sulfates with C4AF, or monosulfate. The second effect is due to the higher pH in pore solutions in hydrating cement. At higher pH le vels more sulfur tends to be adsorbed by the C-S-H gel. The result of less sulfate being available is lower amounts of primary ettringite being formed. The formation of prim ary ettringite helps slow the hydration rate. The solution used in the field to this problem is to increase the level of calcium sulfate.

PAGE 23

10 Thus, more gypsum is needed in the unhydrat ed cement to achieve the same level of deceleration. This greater level of sulfates co ntributes to a greater level of internal sulfate attack. The sulfates that are adsorbed by th e C-S-H gel are desorbed at a later time into the pore solutions providing a source of su lfate. Divet and Randriambololona found the rate of both adsorption and desorption are de termined by both the temperature and pH of the pore solutions. The pH influences both th e maximum quantity absorbed and the rate of absorption. In their 1998 paper, Divet and Randriam bololona also found that high levels of C3S result in higher heats of hydr ation. The occurrence of dela yed ettringite formation is linked to the maximum temperature a cement experiences. The heat can come from external sources such as a heat curing or from the internally generated heat from hydration in thick structures. In addition, the adsorption of sulfate by C-S-H gel is also temperature dependant. Both the solubili ty and amount of sulfates adsorbed are influenced by temperature. As the solubi lity decreases and th e amount of adsorbed sulfate increases, so does the likelihood of internally generated sulfate attack. The substitution of aluminate ions in C-S-H gel is also increased at highe r temperatures. Both sulfate and aluminate are essential for ettringi te formation. Higher temperatures also help limit the amount of primary ettringite formed. The sulfate not consumed in the formation of ettringite is held in the C-S-H gel wher e it can leach out over time providing a source of sulfate and aluminates for delayed ettringite formation. The alkali content of the cement also ha s a bearing on the amount of ettringite formed at a later date. F.P. Glasser in his 1996 paper proved that the higher the alkali content, the higher the level of sulfate ion that can be present in the pore solutions. This

PAGE 24

11 is due to the increase in solubility of the su lfate ion and instability of ettringite at high alkali levels. At temperatur es above 50C the solubility of the sulfate ion increases rapidly leading to the destruction of previous ly formed primary ettringite in the cement matrix. Much of this sulfate tends to be adsorbed by C-S-H gel due to its ability to adsorb a greater amount of sulf ate as the temperature increases This previously adsorbed sulfate becomes available over time for the formation of ettringite as it is slowly desorbed. The adsorption capacity of the CS-H gel is quite high and was measured in experiments by Diver and Randriambololona. They found the level of adsorption was dependant on pH, alkali content and temperature of exposure. Increases in any or all of these allow much greater levels of adsorption. The rate of desorption was also studied in the above paper. It was proven that all of th e adsorbed sulfate ion will be desorbed over time as the capacity of the C-S-H gel to hold the sulfate decreases due to lower temperatures and pH. The rate of desorption is much slower than the rate of initial adsorption. This desorption allows for the sl ow formation of ettringite over time. Early age thermal cracking is also increas ed at higher temperatures. The cracking is caused by the differential between the surface and core temperatures. These cracks provide a place for the ettringite crystals to deposit. Researchers are varied in their opinion whether these crystals can cause damage. The first study to isolate the effect of C3S upon expansion was done by Rasheeduzzafar in 1974. He noticed a correlation between the C2S/C3S ratio and expansion for external sulfate attack. Sin ce that time many other st udies have confirmed that this is true for both internal and extern al sulfate attack. In doing a review of these studies, it is evident that other factors could be causing the ex pansion that was credited to

PAGE 25

12 the C2S/C3S ratio. This study was conducted to elim inate most of the other variables by careful cement selection and equalizing cement chemistry.

PAGE 26

13 CHAPTER 2 EXPERIMENTAL PROCEDURES 2.1 Composition Determination The compositions of the cements in this study were determined chemically or by mineralogy. Chemical composition or as it is better known as oxide chemical composition determines the percentage of each oxide present. In this method the cement sample is fused at 1000C with Li2B4O7 and analyzed by x-ray fluorescence spectrometry. The results of this method are listed in table 3. The calculation of phases using the Bogue method follows a procedure outlined in ASTM C-150. Each compound has a different formula in which the oxide composition is entered. The results are dependant upon ratios the oxides to each other and give only a fairly accurate analysis of the compounds present. The results of this method are presented in table 2. The internal standard method is better known as the curves method. In a study done by Natalya Shanahan calibration curves were prepared for C3S and cubic and orthorhombic C3A according to ASTM C-1365-98. She collected a series of X-ray scans on the Phillips X’Pert PW3040 located in our lab. The samples had fixed compositions and curves that related the percentage of C3S to the results were prepared. Table 4 shows the results using this method. Because of the use of internal standards, this method was probably the most accurate.

PAGE 27

14 The final method used to determine composition was through the use of X-ray powder diffractometry. In this method Reitv eld analysis is used to determine the composition. An internal standard of titaniu m oxide is added and constitutes 10% of the sample by weight. Each compound emits a cer tain wavelength under x-ray diffraction. The peaks formed are compared to the peak emitted by the titanium oxide. Since the percentage of titanium oxide is known, the composition of the cement sample can be inferred by comparing the peak heights. 2.2 Materials This study was done using three cements. In addition, graded sand, distilled water, KOH and Terra Alba gypsum were al so used. The cement chemistry was modified by the use of Terra Alba Gypsum and KOH. The cement was mixed in the following conditions: As received alkali and su lfate, as received sulfate with alkali adjusted to 1.5% (and 2% for cement MH-3), sulfate levels of 3.6% or 5% with alkali levels remaining as received, or sulfate levels adjusted to 3.6% or 5% with alkali levels adjusted to 1.5%. Potassium hydroxide was used to adjust the alkali level. Cement MH3 had as received sulfate levels of 3.1% and was tested at this level ra ther than 3.6%. Table 2: Bogue calculations, fineness and C3S/C2S ratio Compound Cement E (%) Cement MH-3 (%) Cement C (%) C3S 57 67 66 C2S 18 7 14 C3A 6 6 7 C4AF 11 8 11 C3S/C2S 3.17 9.57 4.29 Blaine fineness (square meters per kilogram) 380 395 384

PAGE 28

15 Table 3: Chemical analysis of the two cements used in the study Cement E MH-3 C Compound Wt % WT% Wt % SiO2 21.15 20.20 20.52 Al2O3 4.78 4.02 4.92 Fe2O3 3.76 2.78 3.70 CaO 64.41 64.02 64.31 MgO 0.95 2.47 1.71 SO3 2.58 3.09 2.81 Na2O 0.18 0.21 <0.01 K2O 0.34 1.10 0.41 TiO2 0.33 0.22 0.27 P2O5 0.07 0.15 0.03 Mn2O3 0.03 0.06 0.04 SrO 0.12 0.04 0.04 Cr2O3 <0.01 <0.01 <0.01 ZnO 0.02 0.04 <0.01 Na2Oeq 0.40 0.93 0.27 Table 4: Analysis of cements using the inte rnal standard method a nd Reitveld analysis Compound Cement E (%) Cement MH-3 (%) Cement C (%) C3S 58 67 70 C3A 4.0 3.3 3.0 2.3 Procedure 2.3.1 Mix Design Mix proportioning was done by weight usi ng the spreadsheet shown in Table 5. The oxide chemistry of the cement in the as received condition was entered in the spreadsheet. The target SO3 level as well as the KOH weight in grams is entered as variables. The spreadsheet calculates the gypsum needed as well as the Na2Oeq in the spreadsheet itself. Under the spreadsheet th e weight of Terra Al ba Gypsum, KOH, and cement are automatically calculated from the data in the spreadsheet. These weights are

PAGE 29

16 used along with 243 ml of distilled water a nd 1375 g of oven dried sand to make the mortar mix. The sand was oven dried at 100 pl us or minus five degrees centigrade. Table 5: Spreadsheet used in mix design Increase Cement MH-3 to 5.0% Using Gypsum, K eep the Alkali Level in the As Received Condition Wt(g) MH-3 100 Amounts As is After KOH Target Gypsum Needed Gypsum (g) KOH (g) Remaining Added Net SO3 % 3.09 3.09 5 0.044 4.40 0 Al2O3 4.02 3.84 Na2O % 0.21 0.201 0.201 K2O % 1.1 1.052 0.000 1.052 Na2Oeq 0.93 0.89 Na2Oe= 0.89 Cement 478.00 KOH 0.00 Gypsum 22.0 2.3.2 Procedure for Mixing Mortars The mortar was mixed in compliance with ASTM C-305-99 “Standard Practice for Mechanical Mixing of Hydraulic Cement Pastes and Mortars”. Before mixing, the cement, sand, water, and needed KOH and/or Terra Alba Gypsum were weighed out. When KOH was needed this was premixed with the water using a magnetic stirrer. The gypsum if needed was also premixed with th e water after this point using the mixer for fifteen seconds. Cement was then added to the mixing water and mixed at slow speed for thirty seconds. The sand was then added to the mixture over the next thirty seconds. The mixer was then turned off to allow the change of speed to medium speed. This mixture was then mixed at medium speed for a period of thirty seconds. The mixer was then again stopped and the sides of the bowl and pa ddle scraped for fifteen seconds. The bowl was then covered for an additional minute with plastic to prevent the escape of moisture.

PAGE 30

17 After this period the mixer was again turned on at medium speed for a period of one minute. 2.3.3 Slump Test The slump or workability test was performed on all mixes. The procedure followed a modified form of ASTM C143 “Standard Test Method for Slump of Hydraulic-Cement Concrete”. After mixing the mortar was put into the metal slump mold in two layers. The first layer was a bout one inch high and was tamped using the appropriate rod twenty times working from the outside inward in a ci rcular pattern. This was followed by a second layer of mortar that was tamped twenty times also. The same procedure was followed except the level of ta mping was done just to under the level of the first layer. After the final layer was tamped, the mold was struck off by rolling the tamping rod horizontally. The table was then mechanically lifted a nd dropped a total of twenty-five times and the resulting spread of the mortar measured. 2.4 Casting of Bars The molding of bars followed ASTM C-157 “Standard Test Method for Length Change of Hardened Hydraulic-Cement Mortar and Concrete”. Four bars were cast from each batch. The bars were stored in their molds for one hour at 100% relative humidity. After this time, the bars while still in their molds were subjected to the heat curing cycle specified. 2.5 Heat Curing Cycle Four different cycles were used in this study. Each is discussed separately giving the time allowed for each part of the cycle si nce they vary with the temperature. The room temperature cured bars were left in the humid cabinet for a total of twenty-four

PAGE 31

18 hours. They were then demolded, marked a nd measured using the comparator. The bars were then stored for a period of one hour in distilled water and again measured. The second set of measurements was used as the zero point in the study. The initial steps in high temperature heat curing are the same for the 60C, 80C, and 90C heat cures. The bars were put in the humid cabinet for a period of one hour after molding. They were then placed in seal ed plastic bags and put into a furnace that was at room temperature. The furnace was turned on and the set point adjusted to the heat curing temperature of 60C, 80C or 90 C. The bars were allowed one hour and fifteen minutes to come to temperature; this is shown in Figure 1. The heat cure cycle was verified by thermocouples imbedded with in a single bar along with thermocouples placed in the furnace at various locations. A twelve hour cycle was run at temperature and then the bars cooled over a period of f our hours in the furnace. Upon removal from the furnace, the bars were demolded and appropr iately marked for identity. At this time the process differs depending upon temperature of heat curing. The bars done at 90C were allowed an additional forty-five minut es to cool to room temperature before measurement. The bars done at 80C were a llowed thirty-five minutes, and the bars done at 60C were allowed thirty minutes. After the initial measurement the bars were soaked in distilled water for one hour and re-measure d. This measurement is considered the zero point in the study. The bars were then soak ed in a saturated lime water solution and measured at proscribed intervals. The one hour pre-curing cycle was ba sed upon the 1997 paper by Fu, Ding and Beaudoin that measured the e xpansion rate as a function of pre-curing time. The short pre-cure was done to enhance the ex pansion rate of the mortar bars.

PAGE 32

19 The storage of the bars in a saturated lime water solution was done to reduce the leaching of CH and alkalis. Several studies focused on the stability of ettringite. The average range of stability was between a pH of 10.6 to 12.5. The use of a lime water bath was also to keep the pH within this range A saturated lime solution has a pH of 12.4 according to the Handbook of Chemistry and Physics. Heat Curing Cycle 0 20 40 60 80 100 05101520 Time (hours)Temperature (degrees C) 90 C Heat Cure 80 C Heat Cure 60C heat Cure Figure 1: Heat curing cycle adopted in this study 2.6 Length Change on Mortar Bars 2.6.1 Measurement Cycle Measurements were taken daily every tw enty-four hours plus or minus one half hour for the first seven days after being put in the saturated lime water solution. They were again measured at thirteen and four teen days. After this time the cycle was increased to weekly until sixty-three days had elapsed since putting the bars into the lime water solution. Thirty day cycles plus or minus one day were followed after this time, and continued until the end of the test.

PAGE 33

20 2.6.2 Measurement Procedure The length change measurements followed ASTM C-490 “Standard Practice for Use of Apparatus for the Dete rmination of Length Change of Hardened Cement Paste, Mortar, and Concrete”. The device used was a Humbolt H 3250. This device measured to 0.0001 inches. The measurement of a referenc e bar that allowed the machine to be set to a zero point preceded each set of measurements. The bars were placed in the comparator with the arrow pointing upward. This arrow was drawn on the bars when they were identified prior to their first water expo sure. The bars were then spun in a clockwise direction and the minimum gauge reading was recorded. Figure 2 shows a typical mortar bar. The bar pictured s hows the typical bending seen on bars with expansion levels greater than 1%. Figure 2: Mortar bar showing bending

PAGE 34

21 CHAPTER 3 RESULTS AND DISCUSSION Assessment of the effects of tricalcium silicate, sulfate, alkali content, and temperature on the ISA phenomenon in this study was done primarily using length measurements. Phase transformation accom pyining expansion was studied using X-ray diffraction techniques. Semi-quantification and identification of phase transformation was done using Reitveld analysis. In the follo wing pages, the findings of this study are presented and discussed. 3.1 Length Measurement Results The experimental procedure was designe d to speed up the process of secondary ettringite formation in the following ways; the one hour room temperature cure before high temperature curing helped to limit the am ount of primary ettringite formed, the high temperature heat cure cycle destroyed this primary ettringite, and storage in limewater assured the constant pH required to keep the secondary ettringite st able. The short time allowed for pre-curing was based u pon work by Fu and Beaudoin in 1996. Brown and Bothe findings in 1993 indi cate high curing temperatures and increased alkali content provide a mechan ism by which high levels of sulfate and aluminate were incorporated in the C-S-H gel. It was suggested th at this adsorption and later desorption provide the chemical basis fo r secondary ettringite formation. The heat

PAGE 35

22 cure temperatures used here were selected to determine a critical temperature for internal sulfate attack in the cements used in the study. The dominant factors in the expansive behavior of cement due to ISA can be summarized as the sulfate level, alkali level, C3A level, C3S level, C3S/C2S ratio and cement fineness. It is relevant that C3A and cement fineness were relatively constant and can be eliminated as a variable. The as received alkali and C3S content show some variation in the three cements Table 3 shows the oxide composition and Table 4 shows the composition from XRD work and internal standard method. The sulfate content varies from 2.58 weight percent in cement E to 3.09 weight percent in cement MH-3. Alkali content as measured in Na2O equivalent ranges from 0.27% in cement C to 0.93% in cement MH-3. The C3S/C2S ratio shows the same trend with cement E having a ratio of 2.32 to 5.36 for cement MH-3. The rate of expansion also mirrors the chemistry as shown in the figures below. Figure 3: All three cements as r eceived and heat cured at 90C As Received Alkali and Sulfate 90 C Heat Cure 0 0.02 0.04 0.06 0.08 0.1 0306090120150180 Age (Days)% Expansion Cement E Cement MH-3 Cement C

PAGE 36

23 Figure 4: All three cements as r eceived and heat cured at 80C Figure 5: All three cements as r eceived and heat cured at 60C As Received Alkali and Sulfate 60 C Heat Cure 0 0.02 0.04 0.06 0.08 0.1 0306090120150180 Age (Days)% Expansion Cement E Cement MH-3 Cement C As Received Alkali and Sulfate 80 C Heat Cure 0 0.02 0.04 0.06 0.08 0.1 0306090120150180 Age (Days)% Expansion Cement E Cement MH-3 Cement C

PAGE 37

24 Figure 6: All three cements as r eceived and heat cured at 23C The effect of temperature can also be read from these graphs by comparing the expansion exhibited by the cements (Figures 7 thru 9). In all three heat cures, MH-3 expanded the most followed by cement C then cement E. This behavior when compared to the as received chemistry shows the domin ant factors in expansi on to be the sulfate level and C3S/C2S ratio. As Received Alkali and Sulfate 23 C Heat Cure 0.00000 0.02000 0.04000 0.06000 0.08000 0.10000 0306090120150180 Age (Days)% Expansion Cement E Cement MH-3 Cement C

PAGE 38

25 Figure 7: Cement E showing the effect of temperature on as received chemistry Figure 8: Cement MH-3 showing the effect of temperature on as received chemistry E As Received Chemistry Temperature Effect 0 0.025 0.05 0.075 0.1 0306090120150180 Age (Days)% Expansion 90C 80C 60C 23 C MH-3 As Received Chemistry Temperature Effect 0 0.025 0.05 0.075 0.1 0306090120150180 Age (Days)% Expansion 90C 80C 60C 23 C

PAGE 39

26 Figure 9: Cement C showing the effect of temperature on as received chemistry The next logical step is to isolate each of the three variables to determine the impact of each on the rate of expansion. Ceme nt E shows little expans ion at either alkali level. The as received sulfate content is 2.58% for this cement. Figures 10 thru 12 show the effect of alkali content on this cement. Figure 10: The effect of alkali content after 90C heat cure cycle for cement E C As Received Chemistry Temperature Effect 0 0.025 0.05 0.075 0.1 0306090120150180 Age (Days)% Expansion 90C 80C 60C 23 C E: As Received SO3, Variable Alkali 90C Heat Cure 0 0.05 0.1 0.15 0.2 0306090120150180 Age (Days)% Expansion 1.5% AR

PAGE 40

27 Figure 11: The effect of alkali content after 80C heat cure cycle for cement E Figure 12: The effect of alkali content after 60C heat cure cycle for cement E E: As Received SO3, Variable Alkali 80C Heat Cure 0 0.05 0.1 0.15 0.2 0306090120150180 Age (Days)% Expansion 1.5% AR E: As Received SO3, Variable Alkali 60C Heat Cure 0 0.05 0.1 0.15 0.2 050100150 Age (Days)% Expansion 1.5% AR

PAGE 41

28 Figure 13: The effect of alkali content after 23C heat cure cycle for cement E To study the effect of increasing the alkali level, three different alkali levels were selected for cement MH-3. The alkali leve ls used were as received, 1.5% and 2.0%. Figures 14-17 show both the effect of alkali levels and curing temperatures for cement MH-3. At both 80 and 90C, hi gher alkali levels are associated with increased levels of expansion. This is due to the destruction of the primary ettringite during the high heat curing temperature and the ability of pore waters in to keep more sulfates in solution as the alkalinity increases. As the pore wate r alkalinity decreases due to leaching, the sulfates become available for incorporati on into ettringite. At 60C and 23C, the primary ettringite is not destr oyed to the same degree and the effect of high alkali content was lessened. E: As Received SO3, Variable Alkali 23C Heat Cure 0 0.05 0.1 0.15 0.2 0306090120150180 Age (Days)% Expansion 1.5% AR

PAGE 42

29 .Figure 14: The effect of alkali content af ter 90C heat cure cycle for cement MH-3 Figure 15: The effect of alkali content af ter 80C heat cure cycle for cement MH-3 MH-3: As Received SO3, Variable Alkali 80C Heat Cure 0 0.025 0.05 0.075 0.1 0306090120150180 Age (Days)% Expansion 2.0% 1.5% AR MH-3: As Received SO3, Variable Alkali 90C Heat Cure 0 0.05 0.1 0.15 0.2 050100150 Age (Days)% Expansion 2.0% 1.5% AR

PAGE 43

30 Figure 16: The effect of alkali content af ter 60C heat cure cycle for cement MH-3 Figure 17: The effect of alkali content af ter 23C heat cure cycle for cement MH-3 The same four charts were made for cement C. Unlike MH-3, there is only slightly more expansion at an alkali level of 1.5% than the as r eceived alkali level when MH-3: As Received SO3, Variable Alkali 23C Heat Cure 0 0.025 0.05 0.075 0.1 0306090120150180 Age (Days)% Expansion 1.5% AR MH-3: As Received SO3, Variable Alkali 60C Heat Cure 0 0.025 0.05 0.075 0.1 0306090120150180 Age (Days)% Expansion 2.0% 1.5% AR

PAGE 44

31 the sulfate is maintained in the as receiv ed condition of 2.81. These four charts are shown in Figures 18 thru 21. Figure 18: The effect of alkali content after 90C heat cure cycle for cement C Figure 19: The effect of alkali content after 80C heat cure cycle for cement C C: As Received SO3, Variable Alkali 90C Heat Cure 0 0.05 0.1 0.15 0.2 0306090120150180 Age (Days)% Expansion 1.5% AR C: As Received SO3, Variable Alkali 80C Heat Cure 0 0.05 0.1 0.15 0.2 0306090120150180 Age (Days)% Expansion 1.5% AR

PAGE 45

32 Figure 20: The effect of alkali content after 60C heat cure cycle for cement C Figure 21: The effect of alkali content after 23C heat cure cycle for cement C The sulfate level present ha s a large effect on expansion. The experiment was setup to generate results graphing sulfate levels against both al kali and temperature. At a given temperature, an increase in the sulfate level will generate increased rates of C: As Received SO3, Variable Alkali 60C Heat Cure 0 0.05 0.1 0.15 0.2 0306090120150180 Age (Days)% Expansion 1.5% AR C: As Received SO3, Variable Alkali 23C Heat Cure 0 0.05 0.1 0.15 0.2 0306090120150180 Age (Days)% Expansion 1.5% AR

PAGE 46

33 expansion. In figures 22 thru 27 the effect of the sulfate leve l is graphed against an alkali level of 1.5% or 2.0% when heat cured at 90, 80 and 60C. Only cement MH-3 was mixed and measured at an alkali level of 2.0%. A comparison of the corresponding figures for each curing temperature show th e difference that the additional alkali makes to the level of expansion expe rienced by the bars. The higher alkali level of 2.0% results in both earlier and increased le vels of expansion for MH-3. Figure 22: Expansion of cement E with 1.5% alkali content after 90C heat cure E: 1.5% Alkali, Variable SO3 90C Heat Cure 0 0.25 0.5 0.75 1 0306090120150180 Age (Days)% Expansion 5% 3.6% As Rec'd

PAGE 47

34 Figure 23: Expansion of cement MH-3 with 2.0 % alkali content after 90C heat cure Figure 24: Expansion of cement MH-3 with 1.5% alkali content after 90C heat cure MH-3: 2.0% Alkali, Variable SO3 90C Heat Cure 0 0.25 0.5 0.75 1 0306090120150180 Age (Days)% Expansion 5% 3.1% MH-3: 1.5% Alkali, Variable SO3 90C Heat Cure 0 0.25 0.5 0.75 1 0306090120150180 Age (Days)% Expansion 5% 3.1%

PAGE 48

35 Figure 25: Expansion of cement C with 1.5% alkali content after 90C heat cure A comparison of expansion data from the previous four figures shows that cement C has the highest rates of expansion of all three cements at an alkali level of 1.5%. Cement MH-3 has an as received sulfate le vel of 3.09, this compares favorably with the other two cements at 3.6% sulfate. Figure 26: Expansion of cement E with 1.5% alkali content after 80C heat cure C: 1.5% Alkali, Variable SO3 90C Heat Cure 0 0.25 0.5 0.75 1 0306090120150180 Age (Days)% Expansion 5% 3.6% As Rec'd E: 1.5% Alkali, Variable SO3 80C Heat Cure 0 0.25 0.5 0.75 1 0306090120150180 Age (Days)% Expansion 5% 3.6% As Rec'd

PAGE 49

36 Figure 27: Expansion of cement MH-3 with 2.0% alkali content after 80C heat cure The additional alkali makes a large differe nce in the expansion levels of the 5.0% sulfate samples in MH-3. This is due to the alkali content and its effect on the capacity of the C-S-H gel and pore waters to ho ld higher levels of sulfate. Figure 28: Expansion of cement MH-3 with 1.5% alkali content after 80C heat cure MH-3: 2.0% Alkali, Variable SO3 80C Heat Cure 0 0.25 0.5 0.75 1 0306090120150180 Age (Days)% Expansion 5% 3.1% MH-3: 1.5% Alkali, Variable SO3 80C Heat Cure 0 0.25 0.5 0.75 1 0306090120150180 Age (Days)% Expansion 5% 3.1%

PAGE 50

37 Figure 29: Expansion of cement C with 1.5% alkali content after 80C heat cure Figures 26 thru 29 show the expansion da ta for cements with 1.5% or 2.0% alkali and variable sulfate content af ter the 80C heat curing cycle. As in the 90C data the sulfate level determines the total expansion e xperienced by the mortar bar. The rate of expansion is lower for the bars heat cure d at 80C than for those done at 90C. Figure 30: Expansion of cement E with 1.5% alkali content after 60C heat cure C: 1.5% Alkali, Variable SO3 80C Heat Cure 0 0.25 0.5 0.75 1 0306090120150180 Age (Days)% Expansion 5% 3.6% As Rec'd E: 1.5% Alkali, Variable SO3 60C Heat Cure 0 0.1 0.2 0.3 0.4 0.5 0306090120150180 Age (Days)% Expansion 5% 3.6% As Rec'd

PAGE 51

38 Figure 31: Expansion of cement MH-3 with 2.0% alkali content after 60C heat cure Figure 32: Expansion of cement MH-3 with 1.5% alkali content after 60C heat cure MH-3: 2.0% Alkali, Variable SO3 60C Heat Cure 0 0.1 0.2 0.3 0.4 0.5 0306090120150180 Age (Days)% Expansion 5% 3.1% MH-3: 1.5% Alkali, Variable SO3 60C Heat Cure 0 0.1 0.2 0.3 0.4 0.5 0306090120150180 Age (Days)% Expansion 5% 3.1%

PAGE 52

39 Figure 33: Expansion of cement C with 1.5% alkali content after 60C heat cure Figures 30 thru 33 show the same effects for the 60C heat cure. The temperature has an effect on the amount of expansion experienced by the bar, but in all cases expansion increases with increa sing sulfate and alkali content. In all three cements the expansion is dependant upon the sulfate content. Equalizing the sulfate content, temperature, and alkali content between all three cements yields the data shown in figures 34 thru 37. In Figures 34 and 35 representing the 90 and 80C heat cure, cement C exhibits the most e xpansion followed by MH-3 then lastly E. This is what was to be expected by the C3S content using the calibration curve method. Cement C has the greatest value at 70% followe d by MH-3 at 67% and Cement E at 58%. In figures 36 and 37, the 60C and room temper ature heat curing cycles, the expansion is greatest for cement MH-3. At these temper atures there is not much difference in expansion rates. C: 1.5% Alkali, Variable SO3 60C Heat Cure 0 0.1 0.2 0.3 0.4 0.5 0306090120150180 Age (Days)% Expansion 5% 3.6% As Rec'd

PAGE 53

40 Figure 34: The effect of cement composition on expansion at constant alkali and sulfate Levels after 90C heat cure Figure 35: The effect of cement compos ition on expansion at c onstant alkali and sulfate Levels after 80C heat cure 1.5% Alkali and 5% SO3 90C Heat Cure 0 0.5 1 1.5 2 0306090120150180 Age (Days)% Expansion E MH-3 C 1.5% Alkali and 5% SO3 80C Heat Cure 0 0.5 1 1.5 2 0306090120150180 Age (Days)% Expansion E MH-3 C

PAGE 54

41 Figure 36: The effect of curing temperature on cement with 1.5% alkali and 5% sulfate after 60C heat cure Figure 37: The effect of curing temperature on cement with 1.5% alkali and 5% sulfate after 23C heat cure 1.5% Alkali and 5% SO3 60C Heat Cure 0 0.25 0.5 0.75 1 0306090120150180 Age (Days)% Expansion E MH-3 C 1.5% Alkali and 5% SO3 23C Heat Cure 0 0.25 0.5 0.75 1 0306090120150180 Age (Days)% Expansion E MH-3 C

PAGE 55

42 The same comparisons are duplicated in ch arts 38 thru 41 with the sulfate content at 3.6%. As in the previous charts, cement C shows the greatest e xpansion after both the 90 and 80C heat curing cycles. Figure 38: The effect of curing temperature on cement with 1.5% alkali and 3.6 % sulfate after 90C heat cure Figure 39: The effect of curing temperat ure on cement with 1.5% alkali and 3.6 % sulfate after 80C heat cure 1.5% Alkali content and 3.6% SO3 90C Heat Cure 0 0.5 1 1.5 2 0306090120150180 Age (Days)% Expansion E MH-3 (3.1%) C 1.5% Alkali content and 3.6% SO3 80C Heat Cure 0 0.5 1 1.5 2 0306090120150180 Age (Days)% Expansion E MH-3 (3.1%) C

PAGE 56

43 Figure 40: The effect of curing temperature on cement with 1.5% alkali and 3.6 % sulfate after 60C heat cure Figure 41: The effect of curing temperature on cement with 1.5% alkali and 3.6 % sulfate after 23C heat cure In Figures 42 thru 44 the rate of expans ion for all three cements are presented at 1.5% Alkali and 5% Sulfate conten t. These graphs clearly show the effect of heat curing temperature on expansion. Higher the temperatures result in greater expansion over time. 1.5% Alkali content and 3.6% SO3 60C Heat Cure 0 0.25 0.5 0.75 1 0306090120150180 Age (Days)% Expansion E MH-3 (3.1%) C 1.5% Alkali content and 3.6% SO3 23C Heat Cure 0 0.25 0.5 0.75 1 0306090120150180 Age (Days)% Expansion E MH-3 (3.1%) C

PAGE 57

44 Figure 42: The effect of heat cure te mperature on cement E with alkali = 1.5% and SO3 = 5.0% Figure 43: The effect of heat cure temp erature on cement MH-3 with alkali = 1.5% and SO3 = 5.0% MH-3: 1.5% Alkali, 5% SO3 0 0.5 1 1.5 2 0306090120150180 Age (Days)% Expansion 90C 80C 60C 23 E: 1.5% Alkali, 5% SO3 0 0.5 1 1.5 2 0306090120150180 Age (Days)% Expansion 90C 80C 60C 23

PAGE 58

45 Figure 44: The effect of heat cure te mperature on cement C with alkali = 1.5% and SO3 = 5.0% Before the results are discussed, let us c onsider the rate of expansion as measured in % expansion per day for all three cement s. Figures 45 thru 48 show the rate of expansion for all three cements at an equalized chemistry of 5.0% sulfate and 1.5% alkali. At both 80C and 90C cement C has th e greatest rate of expansion. After the 60C and 23C heat cure cement MH-3 expands at the greatest rate. This is another example showing that cement C has the grea test expansion after the high temperature heat cures. C: 1.5% Alkali, 5% SO3 0 0.5 1 1.5 2 0306090120150180 Age (Days)% Expansion 90C 80C 60C 23

PAGE 59

46 Figure 45: The rate of expansion for cements with SO3 = 5% and alkali =1.5% cured at 90C. Note: results on cement C terminated at 91 days due to bending Figure 46: The rate of expansion for cements with SO3 = 5% and alkali =1.5% cured at 80C. Note: resu lts on cement C terminated at 105 days due to bending The same trend is shown at a sulfate level of 3.6%. Figures 49 thru 52 are graphs of cements at 3.6% sulfate and 1.5% alkali. Rate of expansion: 1.5% Alkali, 5.0% SO3 90 C 0 0.005 0.01 0.015 0.02 0306090120150180 Time (days)Rate of expansion (% per day) E MH-3 C Rate of expansion: 1.5% Alkali, 5.0% SO3 80 C 0 0.005 0.01 0.015 0.02 0306090120150180 Time (days)Rate of expansion (% per day) E MH-3 C

PAGE 60

47 Figure 47: The rate of expansion for cements with SO3 = 5% and alkali =1.5% cured at 60C Figure 48: The rate of expansion for cements with SO3 = 5% and alkali =1.5% cured at 23C At a sulfate level of 5% and alkali level of 1.5%, cement and 60C heat cure, cement MH-3 is the only one of the three to have any meaningful level of expansion. Rate of expansion: 1.5% Alkali, 5.0% SO3 60 C 0 0.005 0.01 0.015 0.02 0306090120150180 Time (days)Rate of expansion (% per day) E MH-3 C Rate of expansion: 1.5% Alkali, 5.0% SO3 23 C 0 0.005 0.01 0.015 0.02 0306090120150180 Time (days)Rate of expansion (% per day) E MH-3 C

PAGE 61

48 Although the level of expansion is far from that required to produce physical damage, the heat curing had some effect on this cement. There is little expansion for any of the cements when cured at 23C. Figure 49: The rate of expansion for cements with SO3 = 3.6% and alkali = 1.5% cured at 90 C Figure 50: The rate of expansion for cements with SO3 = 3.6% and alkali = 1.5% cured at 80C Rate of expansion: 1.5% Alkali, 3.6% SO3 90 C0 0.0025 0.005 0.0075 0.01 0306090120150180 Time (days)Rate of expansion (% per day) E MH-3 (3.1%) C Rate of expansion: 1.5% Alkali, 3.6% SO3 80 C0 0.0025 0.005 0.0075 0.01 0306090120150180 Time (days)Rate of expansion (% per day) E MH-3 (3.1%) C

PAGE 62

49 Figure 51: The rate of expansion for cements with SO3 = 3.6% and alkali = 1.5% cured at 60C Figure 52: The rate of expansion for cements with SO3 = 3.6% and alkali = 1.5% cured at 23C Charts 53 thru 56 show the expansion rate for cements with as received sulfate and 1.5% alkali content. Charts 57 thru 60 s how the expansion rate for cements with as Rate of expansion: 1.5% Alkali, 3.6% SO3 60 C0 0.0025 0.005 0.0075 0.01 0306090120150180 Time (days)Rate of expansion (% per day) E MH-3 (3.1%) C Rate of expansion: 1.5% Alkali, 3.6% SO3 23 C0 0.0025 0.005 0.0075 0.01 0306090120150180 Time (days)Rate of expansion (% per day) E MH-3 (3.1%) C

PAGE 63

50 received sulfate and as received Alkali content. At all three temperatures cement MH-3 shows the greatest expansion followed by cement C then cement E. These results correlate with the sulfate content. Figure 53: The rate of expansion for cements with as received SO3 and alkali = 1.5% cured at 90C Figure 54: The rate of expansion for cements with as received SO3 and alkali = 1.5% cured at 80C Rate of Expansion: 1.5% Alkali, As Received SO3 90 C 0 0.001 0.002 0.003 0.004 0.005 0306090120150180 Time (days)Rate of expansion (% per day) E MH-3 C Rate of Expansion: 1.5% Alkali, As Received SO3 80 C 0 0.001 0.002 0.003 0.004 0.005 0306090120150180 Time (days)Rate of expansion (% per day) E MH-3 C

PAGE 64

51 Figure 55: The rate of expansion for cements with as received SO3 and alkali = 1.5% cured at 60C Figure 56: The rate of expansion for cements with as received SO3 and alkali = 1.5% cured at 23C Rate of Expansion: 1.5% Alkali, As Received SO3 60 C 0 0.001 0.002 0.003 0.004 0.005 0306090120150180 Time (days)Rate of expansion (% per day) E MH-3 C Rate of Expansion: 1.5% Alkali, As Received SO3 23 C 0 0.001 0.002 0.003 0.004 0.005 0306090120150180 Time (days)Rate of expansion (% per day) E MH-3 C

PAGE 65

52 Figure 57: The expansion rate for cements with as received SO3 and alkali cured at 90C Figure 58: The expansion rate for cements with as received SO3 and alkali cured at 80 C Rate of expansion: Alkali and SO3 As Received 90 C 0 0.001 0.002 0.003 0.004 0.005 0306090120150180 Time (days)Rate of expansion (% per day) E MH-3 C Rate of expansion: Alkali and SO3 As Received 80 C 0 0.001 0.002 0.003 0.004 0.005 0306090120150180 Time (days)Rate of expansion (% per day) E MH-3 C

PAGE 66

53 Figure 59: The expansion rate for cements with as received SO3 and alkali cured at 60C Figure 60: The expansion rate for cements with as received SO3 and alkali cured at 23C A series of charts were made plotting the level of expansion (%) against the heat curing temperature for all three cements. All ce ments were plotted separately to show the Rate of expansion: Alkali and SO3 As Received 23 C 0 0.001 0.002 0.003 0.004 0.005 0306090120150180 Time (days)Rate of expansion (% per day) E MH-3 C Rate of expansion: Alkali and SO3 As Received 60 C 0 0.001 0.002 0.003 0.004 0.005 0306090120150180 Time (days)Rate of expansion (% per day) E MH-3 C

PAGE 67

54 chemistry variations; the plots were done at spec ific ages. These charts tend to highlight the effect of heat curing temperatures upon th e variation in chemistr y. Figures 61 thru 63 are for cement E, and represent the time periods of 120, 180, and 270 days. Because the level of expansion for cement E was so small, the scale was reduced from 2% to 0.5%. Temperature, sulfate, and alkali levels had very little effect on the expansion levels for this cement. At 270 days the 5% sulfate a nd 1.5% alkali mortar bars that were cured at 90C only expanded slightly more than the other mi xes. It is to be remembered that this cement had the lowest C3S and C3S/C2S ratio of all three cements. Figure 61: Expansion exhibited by cemen t E at one-hundred and twenty days Cement E @ 120 days 0 0.1 0.2 0.3 0.4 0.5 2030405060708090100 TemperatureExpansion % 5.0-1.5 3.6-1.5 AR-1.5 AR-AR

PAGE 68

55 Figure 62: Expansion e xhibited by cement E at one -hundred and eighty days Figure 63: Expansion exhibited by cemen t E at two-hundred and seventy days The same graphs were generated for cem ent MH-3. The effect of temperature, sulfate and alkali content are very apparent for this cement as well as cement C. The scale was set at a 2% expansion level to cl early show the effect of temperature and chemistry upon the level of expansion experien ced by the mortar bar. The mixture with Cement E @ 180 days 0 0.1 0.2 0.3 0.4 0.5 2030405060708090100 TemperatureExpansion % 5.0-1.5 3.6-1.5 AR-1.5 AR-AR Cement E @ 270 days 0 0.1 0.2 0.3 0.4 0.5 2030405060708090100 TemperatureExpansion % 5.0-1.5 3.6-1.5 AR-1.5 AR-AR

PAGE 69

56 5% sulfate and 1.5% alkali expanded at a much greater rate than the other mixes. With this mix, the heat curing temperature determines the level of expansion. Figure 64: Expansion exhibited by cement MH-3 at one-hundred and twenty days Figure 65: Expansion exhibited by cement MH -3 at one-hundred and fifty days. The 5.0-2.0 data at 90C is a bent bar and the expans ion is higher than shown Cement MH-3 @ 120 days 0 0.5 1 1.5 2 2030405060708090100 TemperatureExpansion % 5.0-2.0 3.1-2.0 5.0-1.5 3.1-1.5 3.1-AR Cement MH-3 @ 150 days 0 0.5 1 1.5 2 2030405060708090100 TemperatureExpansion % 5.0-2.0 3.1-2.0 5.0-1.5 3.1-1.5 3.1-AR

PAGE 70

57 The trends observed for cement E carry th rough the experiment. The heat cure temperature and cement chemistry determine the levels of expansion. Cement MH-3 was mixed at three al kali levels. The expansion follows the increasing alkali content, with the 2.0% alka li mortar bars expanding at a much greater rate than those of lower alkali content. Figures 65 and 66 shows cement MH-3 at 120 and 150 days. No data is available at this time for the 180 and 270 day expansion levels. Figure 66: Expansion exhibited by cemen t C at one-hundred and twenty days Cement C @ 120 days 0 0.5 1 1.5 2 2030405060708090100 TemperatureExpansion % 5.0-1.5 3.6-1.5 AR-1.5 AR-AR

PAGE 71

58 Figure 67: Expansion e xhibited by cement C at one -hundred and eighty days Figure 68: Expansion exhibited by cemen t C at two-hundred and seventy days Cement C @ 180 days 0 0.5 1 1.5 2 2030405060708090100 TemperatureExpansion % 5.0-1.5 3.6-1.5 AR-1.5 AR-AR Cement C @ 270 days 0 0.5 1 1.5 2 2030405060708090100 TemperatureExpansion % 5.0-1.5 3.6-1.5 AR-1.5 AR-AR

PAGE 72

59 Figure 69: Comparison of expansion exhibited by all cements with a sulfate content of 5.0% and an alkali content of 1.5% at 120 days At 120 days cement C exhibits the great est expansion. The temperature effect shows for cement C above 60C as opposed to cement MH-3 which shows little expansion after the 80C heat cure. The same trends hold true at 180 days. Figures 70 and 71 show the same cements at the 3.6% sulfate level and 1.5% alkali level. Figure 70: Comparison of expansion exhibited by all cements with a sulfate content of 3.6% and an alkali content of 1.5% at 180 days All 5% Sulfate and 1.5% Alkali @ 120 days 0 0.5 1 1.5 2 20406080100 TemperatureExpansion % Cement E Cement MH-3 Cement C All 5% Sulfate and 1.5% Alkali @ 180 days 0 0.5 1 1.5 2 20406080100 TemperatureExpansion % Cement E Cement MH-3 Cement C

PAGE 73

60 Figure 71: Comparison of expansion exhibited by all cements with a sulfate content of 3.6% and an alkali content of 1.5% at 120 days The expansion levels are much lower at the 3.6% sulfate level. Cement C still shows a temperature effect above 60C and MH-3 above 80C as in the 5.0% sulfate charts. At 180 days the same trends conti nue for all cements with E having almost no expansion, while C and MH-3 continue to fo llow the same trends as the 5.0% sulfate mortar bars. All 3.6% Sulfate and 1.5% Alkali @ 120 days 0 0.5 1 1.5 2 20406080100 TemperatureExpansion % Cement E Cement MH-3 (3.1%) Cement C

PAGE 74

61 Figure 72: Comparison of expansion exhibited by all cements with a sulfate content of 3.6% and an alkali content of 1.5% at 180 days 3.2 XRD Results Using Reitveld Analysis Bars that expanded to the level that there was noticeable bending present were run on the XRD machine to determine the levels of key components that included ettringite. Tables 6 thru 14 show the results of these runs In every case, except one, ettringite was present in measurable quantities. A series of XRD runs were also ma de on cement C heat cured at 90C at two different chemistries. The cement was run in the as received condi tion and at 1.5% alkali and 5.0% sulfate level. The results of thes e runs are shown in Tables 15 thru 24. The results show how the phases present and thei r percentages change during the process and cover a time period that stretches from the in itial mixing to sixty days. There were 9 different times considered, this paper will only show the tabled results for five time periods from each. All 3.6% Sulfate and 1.5% Alkali @ 180 days 0 0.5 1 1.5 2 20406080100 TemperatureExpansion % Cement E Cement MH-3 (3.1%) Cement C

PAGE 75

62 Table 6: XRD results from cement C with 5.0% sulfate content and 1.5% alkali content heat cured at 90C after 102 days storage in a saturated lime solution Cement: C SO3 = 5.0%, Alkali = 1.5% Cured @ 90 C 102dl 1.4% Expansion Phase Diffraction angle Intensity Ratio(Iphase/ITiO2) Rutile (TiO2) 27.4 1 Calcite (CaCO3) 29.4 0.258 Silica (SiO2) 26.6 0.092 C-S-H (gel) 50.0 0.258 Portlandite (Ca(OH)2) 34.1 0.983 Ettringite (Ca6Al2S3•32H2O) 9.09 0.205 Gypsum (CaSO4 2H2O) 11.5 0 Table 7: XRD results from cement C with 5.0% sulfate content and 1.5% alkali content heat cured at 80C after 116 days storage in a saturated lime solution Cement: C SO3 = 5.0%, Alkali = 1.5% Cured @ 80 C 116dl 1.4% Expansion Phase Diffraction angle Intensity Ratio (Iphase/ITiO2) Rutile (TiO2) 27.4 1 Calcite (CaCO3) 29.4 0.221 Silica (SiO2) 26.6 0.134 C-S-H (gel) 50.0 0.214 Portlandite (Ca(OH)2) 34.1 1.004 Ettringite (Ca6Al2S3•32H2O) 9.09 0.151 Gypsum (CaSO4 2H2O) 11.5 0 At higher heat curing temperatures the bars expand at a faster rate due to ettringite formation. The time required for the bars to show a noticeable bend was 102 days for the 90C heat cure as compared to 116 days for th e 80C heat cure. The amount of ettringite present is much greater in the 90C heat cure d bars. Tables 6 and 7 represent the most expansive combination: 5.0% sulfate a nd 1.5% alkali cured at 80C and 90C.

PAGE 76

63 Table 8: XRD results from cement C with 3.6% sulfate content and 1.5% alkali content heat cured at 90C after 122 days storage in a saturated lime solution Cement: C SO3 = 3.6%, Alkali = 1.5% Cured @ 90 C 122dl 0.2% Expansion Phase Diffraction angle Intensity Ratio (Iphase/ITiO2) Rutile (TiO2) 27.4 1 Calcite (CaCO3) 29.4 0.309 Silica (SiO2) 26.6 0.35 C-S-H (gel) 50.0 0.095 Portlandite (Ca(OH)2) 34.1 1.663 Ettringite (Ca6Al2S3•32H2O) 9.09 0.225 Gypsum (CaSO4 2H2O) 11.5 0 The lower sulfate level of 3.6% show ed noticeable bending at 122 days in a saturated lime solution when heat cured at 90 C. The 80C heat cure did not exhibit this level of expansion until 195 days. Tables 9 thru 12 show cement MH-3 at 5.0% sulfate and 1.5% or 2.0% alkali. Although the ages are similar, the 2.0% al kali mixes run at both 80C and 90C show lower levels of ettringite. Table 9: XRD results from cement MH-3 w ith 5.0% sulfate conten t and 1.5% alkali content heat cured at 90 C after 123 days storage in a saturated lime solution Cement: MH-3 SO3 = 5.0%, Alkali = 1.5% Cured @ 90 C 123dl .5% Expansion Phase Diffraction angle Intensity Ratio (Iphase/ITiO2) Rutile (TiO2) 27.4 1 Calcite (CaCO3) 29.4 0.000 Silica (SiO2) 26.6 0.257 C-S-H (gel) 50.0 0.115 Portlandite (Ca(OH)2) 34.1 1.665 Ettringite (Ca6Al2S3•32H2O) 9.09 0.201 Gypsum (CaSO4 2H2O) 11.5 0

PAGE 77

64 Table 10: XRD results from cement MH-3 with 5.0% sulfate content and 1.5% alkali content heat cured at 80C after 121 days storage in a saturated lime solution Cement: MH-3 SO3 = 5.0%, Alkali = 1.5% Cured @ 80 C 121dl 0.04% Expansion Phase Diffraction angle Intensity Ratio (Iphase/ITiO2) Rutile (TiO2) 27.4 1 Calcite (CaCO3) 29.4 0.197 Silica (SiO2) 26.6 0.740 C-S-H (gel) 50.0 0.316 Portlandite (Ca(OH)2) 34.1 1.487 Ettringite (Ca6Al2S3•32H2O) 9.09 0.235 Gypsum (CaSO4 2H2O) 11.5 0 The results are similar for both the above charts. The difference in the amount of ettringite present is probably due to the survival of primary ettringite after the 80C heat cure. Table 11: XRD results from cement MH-3 w ith a 5.0% sulfate content and 2.0% alkali content heat cured at 90 C after 125 days st orage in a saturated lime solution Cement: MH-3 SO3 = 5.0%, Alkali = 2.0% Cured @ 90 C 125dl 1.0% Expansion Phase Diffraction angle Intensity Ratio (Iphase/ITiO2) Rutile (TiO2) 27.4 1 Calcite (CaCO3) 29.4 0.057 Silica (SiO2) 26.6 0.102 C-S-H (gel) 50.0 0.087 Portlandite (Ca(OH)2) 34.1 1.564 Ettringite (Ca6Al2S3•32H2O) 9.09 0.140 Gypsum (CaSO4 2H2O) 11.5 0 Table 12: XRD results from cement MH-3 w ith a 5.0% sulfate content and 2.0% alkali content heat cured at 80 C after 120 days st orage in a saturated lime solution Cement: MH-3 SO3 = 5.0%, Alkali = 2.0% Cured @ 80 C 120dl 0.96% Expansion Phase Diffraction angle Intensity Ratio (Iphase/ITiO2) Rutile (TiO2) 27.4 1 Calcite (CaCO3) 29.4 0.271 Silica (SiO2) 26.6 0.165 C-S-H (gel) 50.0 0.155 Portlandite (Ca(OH)2) 34.1 1.511 Ettringite (Ca6Al2S3•32H2O) 9.09 0.159 Gypsum (CaSO4 2H2O) 11.5 0

PAGE 78

65 Interestingly, both sets of data show th e ettringite levels higher after the 80C heat cure cycle than at the 90C cycle. This may be due to the fact that primary ettringite exists at greater levels in the bars cured at 80C than those done at 90C. Table 13: XRD results from cement E with 5.0% sulfate content a nd 1.5% alkali content heat cured at 90C after 285 days storage in a saturated lime solution Cement: E SO3 = 5.0%, Alkali = 1.5% Cured @ 90 C 285dl 0.04% Expansion Phase Diffraction angle Intensity Ratio (Iphase/ITiO2) Rutile (TiO2) 27.4 1 Calcite (CaCO3) 29.4 0.324 Silica (SiO2) 26.6 .540 C-S-H (gel) 50.0 0.146 Portlandite (Ca(OH)2) 34.1 0.915 Ettringite (Ca6Al2S3•32H2O) 9.09 0 Gypsum (CaSO4 2H2O) 11.5 0 It is interesting to note the low levels of ettringite present in cement E after 300 days in lime water. In the bars heat cure d at 90C, no ettringite was found. This is in direct contrast to the other cements in whic h higher curing temperatures lead to greater amounts of delayed ettringite formation. Ce ment E has a very low expansion rate, and did not show any signs of bending during the ex periment. Cement C was mixed in the as received condition and as doped to 1.5% alka li and 5.0% sulfate. The initial XRD runs were made directly after the paste nuggets we re placed in the humid cabinet for their one hour age. Tables 14 and 15 show the results at this time.

PAGE 79

66 Table 14: XRD results from cement C in th e as received condition 20 minutes after the initial hydration Cement: C SO3 = AR, Alkali = AR (20 minutes) Phase Diffraction angle Peak Height Rutile (TiO2) 27.4 0 Calcium Silicate (Ca3SiO2) 32.5 545 Portlandite Ca(OH)2 34.1 0 Calcite (CaCO3) 29.4 443 Anhydrite (CaSO4) 25.4 81 Calcium Silicate (Ca2SiO2) 31.0 0 Brownmillerite (Ca4Al2Fe2O10) 12.2 0 Ettringite (Ca6Al2S3•32H2O) 9.09 0 Gypsum (CaSO4•2H2O) 11.6 20 Bassanite (CaSO4 •1/2 H2O) 14.7 40 Table 15: XRD results from cement C as dope d to a 1.5% alkali and 5.0% sulfate level 30 minutes after initial hydration Cement: C SO3 = 5.0%, Alkali = 1.5% (30 minutes) Phase Diffraction angle Peak Height Rutile (TiO2) 27.4 0 Calcium Silicate (Ca3SiO2) 32.5 449 Portlandite Ca(OH)2 34.1 0 Calcite (CaCO3) 29.4 465 Anhydrite (CaSO4) 25.4 50 Calcium Silicate (Ca2SiO2) 31 0 Brownmillerite (Ca4Al2Fe2O10) 12.2 43 Ettringite (Ca6Al2S3•32H2O) 9.09 27 Gypsum (CaSO4•2H2O) 11.6 78 Bassanite (CaSO4 •1/2 H2O) 14.7 44 Rutile was not added to the paste during the initial run because the sample was run as wet paste. It is hard to accurately compare the mixes at this point due to the time difference of ten minutes and the lack of an internal standard. After the one hour age in the humid cabinet the paste nuggets were pr epared and mixed with 10% rutile which acted as an internal standard. Tables 16 a nd 17 show the results at this step. It is interesting to note in the time progression of the samples that ettringite was present before the heat curing process, and was destroyed during the heat curing process.

PAGE 80

67 Table 16: XRD results from cement C with as received sulfate levels 75 minutes after the initial hydration Cement: C SO3 = AR, Alkali = AR (75 Minutes) Phase Diffraction angle Intensity Ratio (Iphase/ITiO2) Rutile (TiO2) 27.4 1 Calcium Silicate (Ca3SiO2) 32.5 0.907 Portlandite Ca(OH)2 34.1 0.233 Calcite (CaCO3) 29.4 0.615 Anhydrite (CaSO4) 25.4 0.109 Calcium Silicate (Ca2SiO2) 31.0 0 Brownmillerite (Ca4Al2Fe2O10) 12.2 0.059 Ettringite (Ca6Al2S3•32H2O) 9.09 0.038 Bassanite (CaSO4 •1/2 H2O) 14.7 0.054 Table 17: XRD results from cement C as dope d to a 1.5% alkali and 5.0% sulfate level 75 minutes after initial hydration Cement: C SO3 = 5.0%, Alkali = 1.5% (75 Minutes) Phase Diffraction angle Intensity Ratio (Iphase/ITiO2) Rutile (TiO2) 27.4 1 Calcium Silicate (Ca3SiO2) 32.5 1.026 Portlandite Ca(OH)2 34.1 0 Calcite (CaCO3) 29.4 0.916 Anhydrite (CaSO4) 25.4 0.092 Calcium Silicate (Ca2SiO2) 31.0 0 Brownmillerite (Ca4Al2Fe2O10) 12.2 0.073 Ettringite (Ca6Al2S3•32H2O) 9.09 0.104 From a comparison of the above two tabl es it is evident that the higher sulfate levels result in higher levels of ettringite. The paste nugge ts were run at 6, 12, 16 and 17 hours. Sixteen hours from the start of the heat curing process the paste nuggets were once again at room temperature. Tables 18 and 19 show the results at this time.

PAGE 81

68 Table 18: XRD results from cement C with as received sulfate le vels 17 hours and 15 mi nutes after the in itial hydration Cement: C SO3 = AR, Alkali = AR Cured @ 90C (17.25 Hours) Phase Diffraction angle Intensity Ratio (Iphase/ITiO2) Rutile (TiO2) 27.4 1 Calcium Silicate (Ca3SiO2) 32.5 0.543 Portlandite Ca(OH)2 34.1 0.662 Calcite (CaCO3) 29.4 0.348 Anhydrite (CaSO4) 25.4 0.087 Calcium Silicate (Ca2SiO2) 31.0 0.063 Brownmillerite (Ca4Al2Fe2O10) 12.2 0.053 Ettringite (Ca6Al2S3•32H2O) 9.09 0.000 Table 19: XRD results from cement C as dope d to a 1.5% alkali and 5.0% sulfate level 17 hours and 15 minutes after initial hydration Cement: C SO3 = 5.0%, Alkali = 1.5% Cured @ 90C (17.25 Hours) Phase Diffraction angle Intensity Ratio (Iphase/ITiO2) Rutile (TiO2) 27.4 1 Calcium Silicate (Ca3SiO2) 32.5 0.413 Portlandite Ca(OH)2 34.1 1.014 Calcite (CaCO3) 29.4 0.311 Anhydrite (CaSO4) 25.4 0.063 Calcium Silicate (Ca2SiO2) 31.0 0.079 Brownmillerite (Ca4Al2Fe2O10) 12.2 0.000 Ettringite (Ca6Al2S3•32H2O) 9.09 0.000 At this point in the process, the phases that are present in each mix are at almost the same level. There is no measurable ettr ingite present in either mix. The 28 and 60 day measurements are presented in Tables 20 thru 23. Ettringite is reformed in the paste which contained 5.0% sulfate, and was at meas urable levels within twenty-eight days. The paste that had the as received sulfate leve l still had no measurable ettringite after 60 days.

PAGE 82

69 Table 20: XRD results from cement C with as received sulfate levels 28 days after the initial hydration Cement: C SO3 = AR, Alkali = AR Cured @ 90C (28 Days) Phase Diffraction angle Intensity Ratio (Iphase/ITiO2) Rutile (TiO2) 27.4 1 Calcium Silicate (Ca3SiO2) 32.5 0.046 Portlandite Ca(OH)2 34.1 1.129 Calcite (CaCO3) 29.4 0.000 Anhydrite (CaSO4) 25.4 0.061 Calcium Silicate (Ca2SiO2) 31.0 0.148 Brownmillerite (Ca4Al2Fe2O10) 12.2 0.063 Ettringite (Ca6Al2S3•32H2O) 9.09 0.000 Table 21: XRD results from cement C with as received sulfate levels 60 days after the initial hydration Cement: C SO3 = AR, Alkali = AR Cured @ 90C (60 Days) Phase Diffraction angle Intensity Ratio (Iphase/ITiO2) Rutile (TiO2) 27.4 1 Calcium Silicate (Ca3SiO2) 32.5 0.042 Portlandite Ca(OH)2 34.1 1.160 Calcite (CaCO3) 29.4 0.110 Anhydrite (CaSO4) 25.4 0.000 Calcium Silicate (Ca2SiO2) 31.0 0.160 Brownmillerite (Ca4Al2Fe2O10) 12.2 0.000 Ettringite (Ca6Al2S3•32H2O) 9.09 0.000 The hydration process was still taking place at 60 days. The levels of portlandite and calcite were increasing while anhydrite and calcium silicates were being consumed in the process. At 60 days there were no meas urable quantities of ettringite. The same process was taking place in the doped cement as shown in Tables 22 and 23.

PAGE 83

70 Table 22: XRD results from cement C as dope d to a 1.5% alkali and 5.0% sulfate level 28 days after initial hydration Cement: C SO3 = 5.0%, Alkali = 1.5% Cured @ 90C (28 Days) Phase Diffraction angle Intensity Ratio (Iphase/ITiO2) Rutile (TiO2) 27.4 1 Calcium Silicate (Ca3SiO2) 32.5 0.232 Portlandite Ca(OH)2 34.1 1.171 Calcite (CaCO3) 29.4 0.146 Anhydrite (CaSO4) 25.4 0.061 Calcium Silicate (Ca2SiO2) 31.0 0.236 Brownmillerite (Ca4Al2Fe2O10) 12.2 0.000 Ettringite (Ca6Al2S3•32H2O) 9.2 0.122 Table 23: XRD results from cement C as dope d to a 1.5% alkali and 5.0% sulfate level 60 days after initial hydration Cement: C SO3 = 5.0%, Alkali = 1.5% Cured @ 90C (60 Days) Phase Diffraction angle Intensity Ratio (Iphase/ITiO2) Rutile (TiO2) 27.4 1 Calcium Silicate (Ca3SiO2) 32.5 0.000 Portlandite Ca(OH)2 34.1 0.986 Calcite (CaCO3) 29.4 0.000 Anhydrite (CaSO4) 25.4 0.075 Calcium Silicate (Ca2SiO2) 31.0 0.207 Brownmillerite (Ca4Al2Fe2O10) 12.2 0.000 Ettringite (Ca6Al2S3•32H2O) 9.09 0.185 In the higher sulfate paste nuggets ettringi te has been found at measurable levels at 28 days. The relative amount of portlandite, calcite, and calcium silicates decreases in the 32 day time period from 28 to 60 days. Ettrin gite levels continue to increase as the aging process continues as show n in the chart below.

PAGE 84

71 Figure 73: Ettringite and portlandite formati on over time from XRD results of cement C with 1.5% alkali and 5.0% sulfate level 0 0.2 0.4 0.6 0.8 1 1.2 1.4 0.050.250.50.670.711.712860 Time (days)Intensity Ratio (Phase/Rutile) Ettringite Portlandite

PAGE 85

72 CHAPTER 4 CONCLUSIONS AND RECOMENDATIONS The study indicated the significance of se veral variables in th e ISA process. It confirmed the impact of sulfate and alkali le vels along with the heat curing temperature. Equalization of the cement chemistries and fineness illustrated the impact of tricalcium silicate levels on the ISA process. The chemistry of all three cements was shown in tables 2 thru 5. In the as received condition cement MH-3 has the highest SO3, Na2Oeq and MgO levels. This is reflected in the as received expansion results in figures 3 thru 6 and 53 thru 59. In the first four graphs and the last seve n graphs, the expansion follows the SO3 levels of the cements. Bars made with MH-3 expanded the greatest amount followed by C then E. The expansion picture changes when the cements are equalized in alkali and sulfate content. At both 80C and 90C cemen t C bars expanded at a greater rate and experienced greater expansion levels than cemen t MH-3. This is what is to be expected when the only variable is the C3S content of the cement. The effect of C3S levels on expansion is a function of the alkali and sulfate content. Research conducted by Rasheeduzzafar and others have come to the conclusion that the C3S/C2S ratio has a bearing on ettringite formation and ultimately the level of expansion experienced by a particular cem ent over time. One problem with the experiments that were used to show this point was the variations in the alkali and sulfate

PAGE 86

73 levels of the cements used in the studies. The cements used in this study were carefully chosen to represent a small ra nge in sulfate levels and a wide variation in the alkali levels. Doping was used to both equalize the cement chemistries and to explore the effect of sulfate and alkali levels on cements with variable tricalcium silicate content. The conclusions reached by use of the experimental data are outlined below. The heat curing temperature has a larg e effect on expansion experienced by the bars. In all cases the rate and level of expansion increa sed the higher the heat curing temperature. Divet and Randriambololona found that the high heat curing temperatures destroy the aluminosulfates initially formed dur ing the early hydration stage, and have an effect on the structure of the C-S-H gel. The sulfates that were a constituent of the initially formed ettringite were adsorbed by the C-S-H gel or held in the pore waters. Reformation of ettringite takes place due to sulfate desorption by the C-S-H gel or sulfates in solution in the pore waters. The present study found that 60C heat cure has little or no effect on the durability of the cement for ages up to 180 days. Even at high sulfate and alkali levels the cements exhibi ted little expansion for the duration of the experiment. The sulfate level has a large impact on the durability of the cement paste. This is recognized by both ASTM and AASHTO a nd regulated accordingly. ASTM C-150 requirements were discussed under Section 1.3 “Chemistry of Portland Cement” and will not be restated here. It is of interest to note that in every case increasing th e sulfate level lead to higher expansion rates. The alkali levels present in the unhyd rated cement have a large impact on the cements durability. According to F.P. Glasser, alkalis allow greater amounts of sulfate to

PAGE 87

74 be held both in pore waters and adsorbed by the C-S-H gel. This sulfate is readily available through desorption from the C-S-H gel providing essential oxide for secondary ettringite formation. The experiments ran two different alkali levels in cements C and E, and three levels in MH-3. A review of Figur es 10 thru 21 shows the effect of the alkali content on the overall expansion rates. Thes e figures prove that increasing the alkali level will lead to a corresponding increase in both the level and rate of expansion. Finally, the experimental data shows the effect of tricalcium silicate levels on expansion due to ettringite formation. High C3S content is detrimental to durability. This is due to a combination of fact ors such that the effect of hi gh alkalis and sulfates tend to be exaggerated at C3S contents of 60% or more. Figur e 73 shows the data for the worst case scenario tested: 5.0% sulfat e and 1.5% alkali at 180 days. Figure 74: Effect of C3S level upon expansion of bars with 5.0% sulfate and 1.5% alkali at 180 days in limewater The best approach to cement durability issues would not be a limit on any single component, but the chemistry considered as a whole. The graphs show that the 0 0.5 1 1.5 2 5560657075 C3S ContentExpansion % 90 80 60

PAGE 88

75 expansion that a cement experiences can be directly related to interplay between the alkali, sulfate, and tricalcium silicate levels. A discussion of the experimental conditions as they would compare to the field is warranted. Exposure to 90 C during curing is not out of line with the temperatures experienced in the field in thick sections. The sulfate level of 3.6% is also close to that of field concretes. The exposure of the bars to a lime water solution is far from field cements. This was done to stabilize the ettr ingite formed initially and through ISA. According to Brown and Bothe calcium has a large effect on the stability of ettringite. In the experiment, the pre-cure was limited to one hour. This is to speed up the process of expansion as proven in experiments by F u, Ding, and Beaudoin in their 1997 paper. They found a dramatic increase in the rate of expansion if the moist curing time was reduced to one hour. Further research is needed to confirm the results of this study and determine the effect of tricalcium silicate levels on the microstructure of the cement paste. It was apparent from the XRD data th at the levels of ettringite present do not always correspond to the amount of expansion experienced by th e mortar bars. Perhaps a link can be made between the cement chemistry and curing temper atures to the streng th and expansion of the paste.

PAGE 89

76 REFERENCES V. Baroghel-Bouny, P. Mounanga, A. Khelidg, A. Loukili, and N. Rafai Autogeneous Deformations of Cement Pa ste: Part II W/C Effects, Micro-Macro Correlations, and Threshold Values in Cement and Concrete Research (article in Press 2005) P.W. Brown and J. V. Bothe Jr, The Stability of Ettringite in Advances in Cement Research volume 5, 1993 X. Cong et al., Effects of Temperature and Relative Humidity on the Structure of C-S-H Gel in Cement and Concrete Research vol. 25, 1995 L. Divet and R. Randriambololona, Delayed Ettringite Form ation: The Effect of Temperature and Basicity on the Inte raction of Sulphate and C-S-H Phase in Cement and Concrete Research vol. 28, 1998 Y. Fu and J.J. Beaudoin, Expansion of Portland Cement Mort ar Due to Internal Sulfate Attack in Cement and Concre te Research volume 27, 1997 Y. Fu, J. Ding, and J.J. Beaudoin, Microcracking as a Precurso r to Delayed Ettringite Formation in Cement Systems in Cement and Concrete Research volume 26, 1996 Y. Fu, P. Xie, P. Gu, and J.J. Beaudoin, Effect of Temperature of Sulphate Adsorption/Desorption by Tricalcium Silicate Hydrates in Cement and Concrete Research volume 24, 1994 F.P. Glasser, The Role of Sulfate Mineralogy and Cure Temperature in Delayed Ettringite Formation in Cement and Concrete Composites volume 18, 1996 P.K. Mehta, Mechanisms of Expansion Associated with Ettring ite Formation in Cement and Concrete Research volume 3, 1973 Rasheeduzzafar, Influence of Cement Composition on Concrete Durability in ACI Materials Journal vol. 89 1992 N. Shanahan, Influence of C3S Content of Cement on Concrete Sulfate Durability in her Masters Thesis U.S.F. 2003

PAGE 90

77 M.D.A. Thomas, C.A. Rogers, and R.F. Bleszynski, Occurrences of Thaumasite in Laboratory and Field Concrete in Cement and Concrete Composites volume 25, 2003