Portland Cement Association

Research and Development Bulletin RDlOST

by Robert L. Day KEYWORDS: accelerated testing, alkali-aggregate reaction, alkalis, calcium aluminates, calcium hydroxide, carboaluminates, carbonation, cement, cement chemistry, chlorides, chloroaluminates, cracking, delayed formation, Duggan Test, durability, ettringite, expansion, gypsum, heat treatment, hydration, ions in solution, microcracking, monosulphoaluminate, pH, pore solution, porosity, precast concrete, precuring period, preferred orientation, secondary ettringite formation, strength, sulfates, sulfate/aluminate ratio, swelling pressure, tempera- ture, testing, thaumasite, railroad ties, transition zone.

ABSTRACT: The report comprises a review and analysis of the available literature pertaining to the causes, effects and prevention ofsecondary (delayed) ettringite in concrete. Over 300publications have been examined. Case studies of damage in concrete possibly caused by secondary ettringite formation are examined first. Fundamental research on secondary ettringite formation, its chemistry, and deposition mechanisms is then reviewed. Key investigations on the topic are analyzed in detail. Next, the potential importance of (a)method of heat-curing and (b)the chemistry of cement is outlined. In the final chapter, a rapid test for evaluation of potential secondary ettringite susceptibility (the “Duggan” test) is evaluated. The analysis indicates that there appears to be a potential for a secondary ettringite formation problem in North America; it is highly probable that secondary ettringite formation can lead to significant deterioration of heat-treated concrete. However, it is unlikely that secondary ettringite formation is, or will be, the sole mechanism responsible for premature deterioration. The critical factors that determine extent of damage due to secondary ettringite formation are(a) duration of delay period before heating the concrete;(b) severity of the heating and/or cooling regime; and (c)the S03/A120~ ratio ofthe cement. There isno evidence that non heat-treated concrete is susceptible to this phenomenon. Further research and improvements to the Duggan test may result in the development ofa useful standard test method to assess the long-term dimensional stability and durability ofconcrete.

REFERENCE: Day, R.L.,TheE~ecto~SecondwyEttringite Fonmfion on theDurabi~ityof Concrete: A Literature Analysis, Research and Development Bulletin RD108T,Portland Cement Association, Skokie, Illinois, U.S.A. 1992.

MOTS CLI%: alcali, aluminate de calcium, bi%onpr6fabriqu6, carboaluminate, carbonatation, ciment, chimie du ciment, chlorures, chloroaluminates, dormant de chemin de fer, durability, essai, essai acc416r4 essai Duggan, ettringite, expansion, fissuration, formation retardde d’ettringite, formation secondaire d’ettringite, gypse, hydratation, hydroxyde de calcium, ions en solution, microfissuration, monosulfoaluminate, orientation pr4f4rentielle, PH, pc%iodeavant cure, porositd, pression de gonflement, rapport sulfate/aluminate, rdaction alcali-granulat, rkisistance, solution de pore, sulfate, temp&ature, thaumasite, traitement h la chaleur, zone de transition.

Rl%UME: Le rapport consiste en une revue et une analyse de la littt%ature disponible en ce qui a trait aux causes, aux effets et 21la pn%ention de la formation secondaire (retard6e) d’ettringite clans le bdton. Plus de 300publications ont dtd examin~es. Des &udes de cas sur des b6tons endommagds, possiblement Acause de la formation secondaire d’ettringite, ont dtdexamindesenpremier. Larecherche fondamentalesurla formation retardde d’ettringite, sa chimie et ses mdcanismes de deposition est ensuite passde en revue. Les investigations d’importance sur le sujet sent analysdes en d&ail. Puis, l’importance potentielle de (a)la m&hode de cure iila chaleur et (b) la chimie du ciment est ddcrite. Dans le dernier chapitre, un essai accdldrdpour 6valuer le potentiel de formation retardde d’ettringite (1’essai “Duggan”) est dvalud. L’analyse indique qu’il est possible qu’il y ait des probl$mes de formation secondaire d’ettringite en Am6rique du Nerd. 11est tri% probable que la formation secondaire d’ettringite produise une d&6rioration importance du b6ton traitd Ala chaleur. Cependant, il est tr&speu probable que la formation secondaire d’ettringite, soit ou devienne, le seul mdcanisme responsible de d&6rioration pr6matur6e. Les facteurs critiques qui d&erminent l’ampleur des dommages dus ?ila formation secondaire d’ettringite sent (a) le ddlai avant chauffage du b&on, (b)la st%%itddu rdgime de chauffage ou de refroissement et (c)le rapport SO~/AlzO~du ciment. 11n’y a aucun indite ~l’effet que le b6ton non chauff6 soit susceptible ?Ice ph~nomhe. Dautres recherches et des ameliorations ~ I’essaiDugganpourraient mener iil’elaboration d’un essainormalisd utile pour determiner lastabilitd dimensionnelle Along terme et la durability du bdton.

R~FfiRENCE: Day, R.L., TheEject ofSecondary Ettringite Formation on the Durability of Concrete: A literature Analysis, Research and Development Bulletin RD108T, Portland Cement Association [L’effet de la formation secondaire d’ettringite sur la durability! du btiton une analyse bibliographique. Bulletin de Recherche et D4veloppement RD108T,Association du Ciment Portland], Skokie, Illinois, U.S.A. 1992.

Cover Illustrations: Ettringite (white needle-like crystals) in concrete exposed to phenolphthalien stain; atmospheric steam- curing cycle; and precast structure under construction.

PCA R&D Serial No. 1929 PCA Research and Development Bulletin RD108T

The Effect of Secondary Ettringite Formation on the Durability of Concrete: A Literature Analysis

by Robert L. Day

Professor of Civil Engineering Civil Engineering De artment The University oPCal a 2500 Universit Drive%x Calgary, Alberta, CanaJa T2N 1N4

ISBN 0-89312-1 69-X

42Portland Cement Association 1992 Page i

Summary of Major Conclusions

A review and analysis was performed of available literature pertaining to the causes, effects and prevention of secondary, or delayedl, ettringite in concrete. Over 300 publications were exam- ined The principal findings from this project are: Given current Cement-prduction and concrete-construction practices there appears to be a potential for a secondary ettringite formation problem in North America. It is highly probable that secondary ettringite formation can lead to significant deterioration of heat- treated cxmcrete.The extent of the potential problem cannot be predicted with accuracy, but it is likely to involve a small fraction of the pre-cast concrete cast in North America. There is no evidence that secondary ettringite formation has been a principal cause of de- terioration in non heat-treated concrete. There is some suggestion that formation of et- tringite after initial deterioration by other means can accelerate the deterioration process. Massive formations of ettringite observed in huge cracks and pores appear, for the most part, to be innocuous. The critical factors that determine extent of damage due to secondary ettringite formation (all other factors being equal) are (a) duration of delay period prior to heating; (b) sever- ity of the heating &Jor cooling r6gime; (c) the S03/A1203 ratio of the cement.

Use of cements with an WA ratio greater than about 0.7 or an ~/A ratio greater than about 2.0 may result in concrete which+under particular curing and exposure conditions, is susceptible to secondary ettringite formation and subsequent deterioration. Good quality laboratory research confirms that secondary ettringite formation can pro- duce expansion and cracking of concrete made with certain North American commercial cements. In North America it is not unusual to find Type 30 cements with S03 contents in the range 3.5 to 4.0% — which could be as high as 4.5Y0.It is not unusual to find alumina contents of 5% or more and C3A contents of 9% or more. Calculation of the ~/A or ~/A ratios for such cements places them in the maximum expansion ranges determined by He- inz and Ludwig and by Gillott. Rapid heating and/or cooling and/or an inadequate delay period can result in microcracks — predominantly at the aggregate-paste or steel-paste interface. These cracks can act as nucleation sites for the later formation of ettringite. ‘Ihe aggregate-paste and steel-paste interface is a weak zone in the concrete that is high in both calcium hydroxide and ettringite contents. The ready availability of sulphates near the cracked interface provides pessimum conditions for secondary ettringite formation. The transition zone appears to be of higher quality when the interface involves limestone aggregate. This suggests that, all other factors being equal, concrete made with limestone aggregate may be less susceptible to secondary ettringite damage. Saturated, or almost saturated, concrete or concrete subjected to frequent wetting/drying cycles is essential to the formation of secondary ettringite and concrete darnage. Early excessive heat treatments at tem~ratures above approximately 70”C results in sub- stantial amounts of sulphates being bound in an unusual form. These sulphates can be

1 see Section 1.3for an explanationof the terminology“secondaryetlringite” Page ii

slowly released back into solution at later ages. This slow release process provides a sup- ply of sulphate ions for secondary ettringite formation. The effect of secondary ettringite formation is substantially reduced by inclusion of an adequate air-entrainment void system in the concrete. Where potential expansion due to secondary ettringite formation can occur, time to start of expansion increases and rate of expansion decreases as the size of the member or speci- men increases. The 14 day sulphate expansion test does not provide an adequate prediction of cement sta- bility at later ages, Specifications for Heat-Treated Concrete by the German Committee for Reinforced Con- crete (Table 6.2) should be serious]y considered for adoption in North America, A new test method to determine optimum gypsum content of cement should be developed which considers long-term stability of the cement as well as optimum strength and setting time, Further research and improvements to the Duggan test may result in the development of a usefhl standard test method to assess the long-term dimensional stability and durability of concrete, Page ill

Table of Contents

Summary of Major Conclusions ...... i TableofContents ...... iii ListofFigures ...... vi ListofTables ...... viii Notation ...... ix

Chapter 1— Introduction 1

1,1 Objectives ...... 1 l.20utlineofReport ...... 1 l.3Terrninology— “SecondaryEttringite’’...... 2

Chapter 2 — Damage due to Secondary Ettringite Formation in Ordinary and Precast Concrete 4

2.1 CombinationsofAttackMechanisms ...... 4 2.2PelrographicObservations...... 7 2.3RoblemswithPrecastConcreteandTies ...... 7 2.4RecastConcrete ...... 9 2.4.1 EffectofHeatTreatmentonProperties...... 9 2.4.2TheCauseof InsufficientStrength Gain afterHeat’lleatment 13 2.5Comment ...... 14

Chapter 3 — Secondary Ettringite Formation: Fundamental Research 16

3.1 Overview of the Cement Hydration Process...... 16 3.2 Early Formation of Sulphoaluminates ...... 16 3.3 and Composition of Ettringite ...... 17 3.4 The Role of Ettringite in the Retardation of the Calcium Aluminates . . . . 18 3.5 The Chemical Stability of Sulphoa.luminates...... 20 3.5.1 Stability in Solution...... , 20 Page iv

3.5,2 Stability and Temperature ...... 21 3.5,3 Post-Decomposition Behaviour., ...... 22 3.6 lhe Effect of Alkalis on Hydration and Formation of Ettringite ...... 25 3.7 The Roleof C02and Carbonates . . . . ., . ., ...... 25 3.7.1 Formation of ‘i%aumasite ...... 26 3.7.2 Formation of Carboahuninates...... 26 3.8 Formation of Chloroaiuminates ...... 28 3.9 Morphology of Ettringite ...,...... 28 3,10 Mechanism of Expansion Associated with Ettringite Formation . , . . . 29 3.10.1 Crystal Growth ...... 29 3.10.2 Swelling ...... 30 3.11 Mechanism of Secondary Ettringite Formation . . , ...... , . . . . 32 3012Comment, , ...... o...... 3s

Chapter 4 — Secondary Ettringite Deposition 36

4.1 The Importance of Porosity to Ettringite Formation and Expansion. . . . . 36 4.2 The Importance of the Aggregate/Paste and Steei/Paste Interface ...... 36 403Comments ... ,,. o,.,..,,. . ., .,, , ...... 43

Cha ter 5 — Key Examinations of Secondary Ettrrngite Formation 44

5.11980 —Ghorabetal [106] .,,..,.,, , ...... 44 5.21984 — Research Institute of the Cement Industry [268] ...... 45 5.31987 —Heinzand Ludwig [132] ., .,,,.,,.,...,.,,.,,. 47 5.41988 —Sylla[297] ...... 52 5.51989 —Heinz, Ludwig &Rudiger[133] , . . , , , ...... , 54 5,61990 -Lawrence etal[177], .,... ,, ..,,..,.,,...... ,55 5.7 Comment..,,,,...... ,,. , ...... 59 Page v

Chapter 6 — Secondary Ettringite Formation and the Chemistry of Cement 60

6.1 Effect of Gypsum Content of Cement on Strength and Volume Stability . . 60 6.1.1 Test Results of A1-Rawi [6]...... 60 6.2 The Practical Significance of the S03/AIZOs Ratio ...... 62 6.3 Specifications for Heat Treatment of Concrete...... 64 6.4 Comment ...... 67

Chapter 7 —Rapid Test Method for Secondary Ettrmgite Formation 69

7.17he Duggan Test ...... 69 7.1.1 Significance of Duggan-Test exults ...... 71 7.1.2 Research of Gillott, et al ...... $, ...... 71 7.1.3 Research of Attiogbe and Wells et al [12, 317] ...... 75 7.1.4 Correlations Between Gillott’s and Attiogbe’s Results . . . . . 78 7.2 Interpretation of Duggan-Test Results ...... 79 7.2.1 The Zero Strain Datum ...... 80

Chapter 8 — General Discussion 83

Acknowledgements ...... 84

References 85

References Not Found 106

Subject Index 114 Pwzevi

List of Figures

2.1 Carbonation of Bridge Structures ...... 5 2.2 Chloride Penetration into Concrete...... 5 2.3 Relative Strengths of Heat Treated Concretes . . , ...... , . 11 2.4 Effect of Heat Curing and Drying on Relative Permeability . . . , . . . . . 12 2.5 Effect of Post Treatment Water Curing on Strength Gain . . . . , . . . . . 14 3,1 Critical Values for S03/AlzOs Ratio,...... 17 3.2 Sulphate-Ettringite Crystal Structure ...... , ...... 19 3.3 Effect of Curing Temperature on Sulphoaluminate Phases ...... 23 3.4 Variation in Ettringite Peak Intensity with Curing Temperature. . . . , . . 24 3.5 Correlation between Volume Increase and Ettringite , . , , . , . . . , . . 31 3.6 Formation of Ettringite from Sulphates Supplied by C-S-H , . , , . , , . . 32 3.7 Changes of Sulphate Ion Concentration with Heat lleatment ...... 34 3.8 Change in S04 and OH Ion Concentration During Storage ...... 34 4,1 The Paste/Aggregate Interface at 30 Minutes and 3 Days , , ...... 38 4,2 Variation in Ettringite Peak Intensity from Interface, , ...... 39 4.3 Compounds and Orientation in the Transition Zone , , ...... , . . , . 40 4.4 Orientation of Calcium Hydroxide Crystals in the Transition Zone , , . , , 41 4.5 Schematic of Changes in the Transition Zone , . , . , ...... 42 5.1 Expansion of Various Mortars after Heat Treatment. , . . , . . , , , . . , 44 5.2 Expansion of Paste Disks aftm Heat Treatment . , . . . . , ...... 46 5.3 Changes in Expansion, Weight and Resonant Frequency with Storage . . . 48 5.4 Correlation Between Expansion and Weight Gain . . , . . , , ...... 49 5.5 Correlation Between Expansion and Resonant Frequency , ...... 49 5.6 Series II Experiments, Expansion vs. Storage . . , . . . , , . . . . , . . . 50 5.7 Effect of Humidity Change on Expmsion . , , , . , , , , , , . , , , . . , 51 5.8 Effect of Water/Cement Ratio and Air Entrainment on Expansion , . . . . 52 5.9 Variation in X-Ray Peak Intensity with Curing Temperature , , . . . , . . 53 5.10 Changes in X-Ray Peak Intensities with Soaking Time , . . , , . . , . . 54 5.11 Scattergram of?/ARatio vs. Expwsion . . . . . , . . , ...... 56 5.12 Typical Expansion vs. Time Curves ...... , . . . . , ...... 57 5.13 Expansion vs Time Curves Showing Effect of Delay Period . . . , . . . , 57 Page W

6.1 Dependence of 28-Day Strengths on S03 Content ~d Cting Temp. . . ● 61 6.2 Dependence of 182-Day Strengths on S03 Content and Curing Temp. . . . 61 6.3 Effect of S03 Content on Expansion of Normal Fog-Cured Specimens. . . 62 6.4 Effect of S03 Content on Exp@on tier 48 W=k of Sotig . . . . . “ 63 6.5 Scattergram of Expansion at 14 Dais vs Expansion at 5 Years ...... 64 6.6 Scattergram of Expansion vsMgOContent ...... 65 6.7 Scattergrsm of S03/A120s Ratio vs Expansion ...... 65 7.1 Scattergram of Accelerated vs. Field Expansion ...... 70 7.2 Effect of Alkali Content and Cement Type on Expansion ...... 72 7.3 Expansion of Hardened Cement Pastes and Concretes ...... 73 7.4 Scattergramof Concrete Expansion vs. Paste Expansion ...... 74

7.5 Observations of a Pessimum SOJM203 Ratio ~ V~OM Ages . . . . ● . 76 7.6 Temperature Testing R6gimes, Duggan and Attiogbe T~* ...... 77 7.7 Expansion of Paste Cores in Attiogbe Test...... 78 7.8 Scattergram of Strsins Measured by Gillott and Attiogbe ...... 79 7.9 Comparison of 20 Day Expansions Using Three Zero Definitions . . . . . 81 Page viii

Lkt of Tables

5.1 Effect of Heat Treatment on Cracking and Phase Formation ...... 46 5.2 Expansion vs. Sulphate Content ...... 4 . S8 6.1 Compound Composition of Cfinlws used in A1-Rawi Study ...... @ 6.2 SpecillcationsforHeat-TreatedCo ncre* ...... 66 7.1 Analysis of Strsins During Duggan TeSt ...... 81 Page ix

Notation

Where appropriate, cement chemists’ notation has been used. his notation as it applies to the present nqmrt is:

C = CaO s = !N02 s= S03 A = /d@s H = H20 Chapter 1 Page 1

Chapter 1 Introduction

1.1 Objectives

At the March 15, 1991 meeting of a ‘Task Group on Secondary Ettringite Formation” — a group struck by the Canadian Standards Association Committee AS (Hydraulic Cement) Executive — the following recommendation was made: ‘Whereas the subcommittee is not aware of any documented cases of secondary ettringite failures in Canada, it is recommended that a detailed review and analysis of the literature (and especially the literature horn Europe) should be undertaken to: = attempt to identify significant factors which have caused documented cases of secondary ettringite failures in Europe ~ attempt to determine which of the factors may be relevant to North America — i.e. is there, or is there likely to be, a potential secondary ettringite formation problem in North America’?”

At the May 3, 1991 meeting of the CSA AS Committee, the Committee passed the following mo- tion in support of the Task Group’s recommendation: “The AS Hydraulic Cements Committee supports, in principle, the need for a detailed review and analysis of the literature relevant to seconduy ettringite formation”. The Portland Cement Asso ciation agreed to provide the funding for such a study; this report is the result. As well as the above objectives, the report centres upon an attempt to provide an adequate answer to a number of important questions concerning secondary ettringite formation: Is the occurrence of secondary ettringite and its effect adequatel y documented? What are the key properties of the constituents of concrete, the key properties of the con- crete itself, and the key environmental parameters that determine whether secondary et- tringite formation is likely to occur? What can be done in the cement and concrete industries in Canada to minimize the possi- bility that future construction will be prone to damage by secondary ettringite formation? Isa test available that can predict whether secondary ettringite formation for a given ce- ment or concrete is like]y to lead to durability problems? The net result is a fair]y comprehensive report. A database search indicated there were 440 refer- ences that may contain information relevant to the topic. Many of these references had obsmue origins; although the interlibrary loans office of the University tried their hardestj only 327 refer- ences were reviewed. These 327 publications are listed at the end of this document, The 113 docu- ments that could not be obtained rue also listed. After review of these references not all were deemd to be di~ctly relevant in the report proper, approximately 160 of the 327 references are cited.

1.2 Outline of Report

The report starts with an examination of various case studies relevant to the formation of secon- dary etlringite in concrete. Many of the problems that have occurred centre upon the precast con- Chapter 1 Page 2

Creteindustry; therefore, in Chapter 2 a summary of precast practices and relevant issues are given. Chapter 3 looks at the fundamental processes associated with secondary ettringite formation. The chemistry and stabiiity of sulphoaluminates are reviewed. The morphology and mechanisms of formation and expansion are studied in detail. Fhmily, chemicai issues related to the formation of secondary ettringite are noted. Chapter 4 is an examination of the physicrd processes associated with the deposition of ettringite which includes a review of the importance of the aggregate/cement-paste and steel/cement-paste interfaces. Chapter 5 is an in-depth examination of the handful of key publications that directly relate to sec- ondary ettringite formation and its consequences. In Chapkx 6 practical issues are discussed concerning the importance of gypsum and the chemis- try of cement to the type and quantity of ettringite that is deposited. Aiso in Chapter 6 the types of heat-treatment are discussed; emphasis is placed upon specifications and guidelines that are h- tended to prevent “extreme” thermai treatments. One of the principai conclusions of this report is that secondary ettringite formation could have a major effect in determining the long-term durability of some types of concrete, Accordingly, Chapter 7 concentrates upon an analysis of the Duggan test as a potentiai method to indicate the vulnerability of cements, aggregates and concretes to long-term deterioration mechanisms, In Chapta 8 the major findings of the report are summarized and discussion centres upon the steps that might be taken by the construction industry to Educe the potentiai problem of secon- dary ettringite formation.

1.3 Terminology — “Secondary Ettringite”

One of the frostquestions that arose during the initial stages of writing this report was: what to call the phenomenon by which concrete, normally at ages of a few years or more, is darnaged by the gradual formation of “ettringite” within the microstructure of the material. Lacking the imagina- tion to invent a new term, the choice for the “appropriate” terminology was made among (1) de- layed ettringite, (2) late ettringite and (3) secondary ettringite, formation. Study of the literature concerning fundamental mechanisms that cause the phenomenon leads to the conclusion that the term “delayed ettringite formation” is not accurate. lTis term implies that conditions within the microstructure might be suitable for the formation of ettringite but that et- tringite does not form; this is not true. It also implies that the delayed ettringite that forms is the same ettringite that did not form during the early hydration of the cement; this is also not the case. “Late ettringite formation” is certainly more accurate in the practicai sense, since the “problem” normally surfaces several years after the concrete is case but it is not precise. Late can have many meanings and, as study of this report will show, the process by which ettringite formation causes the problem occurs over a long time span. It is only the effect of the ettringite formation that comes to light “late”. The winner, then, is “secondary ettringite formation”. It is accurate because materials engineers and chemists normaily think of primary ettringite formation as the sulphoaluminate that forms shortly after gauging and that tends to disappear when sulphate ions in the pore solution become depleted. There can be a period for many concretes, and especially after heat treatment, where no

,,. Chapter 1 Page 3

ettringite can be detected by X-ray analysis. ‘1’bus,the gradual growth of ettringite peaks on X-ray traces at later ages can naturally be thought of as the formation of secondary ettringite. l%e term also agrees with the first two definitions of “secondary” found in the Random House Dictionary: (1) next after the fist in order, rx or tim~ (2) not primary or original, but not with the third: (3) of minor or lesser importance. Unfortunately, the decision on terminology will offend some. In geology, a “secondary mineral” is usually formed after the primary mineral has formed and then dissolved. In concrete this proc- ess may, in fact occur — in which case the terminology that has been chosen is accurate. Others would argue, however, that “the secondary formation of ettringite never did any concrete any harm”l. As review of this report will show, the precise mechanism involved in the late expansion and cracking of concrete due to ettringite formation is far from being clearly defined. The choice of the terminology “secondary ettringite” should not be taken as an implication that the author, at the outset, has taken sides concerning the fundamental mechanisms involved.

1 see, forexample,a letterfromB. Matherto C.R. DuggatIof Feb 27, 1990(aisocopiedto ASTM committeeC09.0202) chapter2 Page 4

Chapter 2 Damage due to Secondary Ettringite Formation in Ordinary and Precast Concrete

2.1 Combinations of Attack Mechanisms

In 1965 Kennerley [161] was one of the first to report possible problems with delayed ettringite formation. He examined exudations at a cold-joint in the Roxburgh Dam, Otago, New Zealand. Analysis indicated that the white deposit was ettringite. Normally,, when ettringite forms there is expansion, but he postulated that this expansion was avoided when constituents passed into solu- tion, reacted and precipitated in the voids in the mass. If the solution was low in lime the volubil- ity of both ettringite and calcium aluminates rose. Ettringite and calcium aluminates may therefore dissolve and be transported through permeable channels to areas high in lime; these compounds will then precipitate as ettringite in the lime-rich regions. Pettifer and Nixon in 1980 [252] described several case studies where secondary ettringite may have played a part in the deterioration process. = Concrete bases of some substations in the English midlands showed deterioration, Al- though there was alkali aggregate reaction, the researchers also observed pores and voids filled with ettringite and ettringite coating aggregate particles. This was surprising since there were only trace amounts of sulphates in the soil. = Similar attack was noted in other sub-stations in Western England and South Wales. The chert-containing limestone reacted, but cores also showed much ettringite co-existing with the gel reactant from the alkali aggregate reaction. = Forty year old concrete blocks were examined which exhibited severe alkali-aggregate re- action and much leaching, These blocks were manufactured with unwashed beach gravel and were gauged with sea water. Upon petrographic examination, large amounts of cal- cite, gel, ettringite and “secondary portlandite” were observed, (Presumably by “secon- dary portlandite” the researchers mean CH deposited in the pores of the microstructure), = The Pirow Street Bridge in Cape town, South Africa showed cracking only 4 years after completion, and remedial repairs were necessary after 9 years. Potentially reactive aggre- gates were used, but a low alkali cement was also used. A significant feature of the pet- rogmphic examination was a yellowish gel accompanied by moderate amounts of ettringite. Around the same time, Volkwein [314] examined 12 to 80 year old concrete bridges for carbona- tion, chloride penetration, deteriomtion and corrosion, As Figure 2.1 shows, the depth of carbona- tion was highly variable (the low carbonation in the 80 year old concrete was because it was made with Roman cement). The carbonation depth depended greatly on the moisture condition of the structure during its life. Carbonation assisted the penetration of ions into the concrete since it was found that carbonated concrete conducted vapour “about twice as well” as non-carbonated concrete. On bridges, penetration of chloride into the top of decks was negligible; most Cl pene- tmted via seeping, splashing and wind-borne Cl carried to the concrete understructure, Such proc- esses can result in the chloride ion penetmting up to 50mm in sound concrete and up to 90mm in frost &maged concrete. Figure 2.2 shows examples of the penetration depth and Cl- ion concen- trations for various ages and qualities of concrete. Volkwein found that in these deteriorated concretes, highly contaminated by Cl, “remarkable” ac- cumulations of needle-shaped crystals were found, “particularly in cracks, pores and around ag- chapter2 Page 5

CARBONATION DEllliS OF DRY CONCR13E FROM BRIDGE WRUCTURES

❑ Minimum Observed ■ Lower Range of ■ Upper Range of ❑ Maximum Obsewed Average Depth Average Depth

Figure 2.1 Carbonation of Bridge Structures [ data from ref. 314]

Chloride Penetration into Concrete Close to Road - Due to Splashing

m. ...’ ‘...., ,..’ . .. . ‘.. . J3.,. ‘“” ‘%,, ~.., “q \ ‘...... , ------35Ycaold,4SMPa,poor .....,., .\, \ .. consoli&tion __~..___ 21~~~ old,70MPa,good consolidation ~ 21yearnold,100MPa,good consolidation

...... -.-.-.-.-.,,...... ,- ‘-”-”’”-’-*-.--.------.*-.-.--.-.-. -.-.-.m

1 0 10 20 30 40 50 60 70 80 90 PENETIU4TION DEPTH (mm)

Figure 2.2 Chloride Penetration into Concrete due to Splashing Action [data from ref. 314] chapter 2 Page 6

gregates”, These crystals were analyzed and were found to be ettringite, with some thaumasite, Volkwein postulated that since the sulphate content had not changed,then the “chloridesdid cause the formationof ettringite” fromthe sulphatein the cement.Additionalreactionsoccur other than the formationof Friedel’s(monochloroaluminate)salt1dueto the penetrationof chlo- rides. Wetting and dryingalso helps to transportsolublecement compoundsto preferred loca- tions, Volkweinnoted that “it cannotbe said whetherthese ettringitecrystals had causedthe deteriorationof concrete or had grownup in alreadyopencracks”, Volkwein’sresultsare most interestingbecauseettringiteappem to have formed in a chloride rich environment.Contrastthis with the laboratoryresultsof Attiogbeet al [12] who foundthat during an accelerated exposuretest, secondaryettringite wouldnot form while concreteprisms were soaked in a sodium-chloridesatwated solution,Otherresearchconfiis that ettringite de- composes in sodiumchloride solution [243-245]. Jones & Poole in 1987[154]lookedat the effects of alkali-silicareaction on 3 structuresin the United Kingdom.They examined20mmthick disks from concretecores taken from the stmc- tures, The disks were stored in sealedcontainemat constanttemperaturesof 1, 15,20,25 and 38°C and relative humiditiesof 65,75,85 and 100%,Expansionswere measured;since the con- cretes were taken from structures,all expansionsnoted were relativeto the unknowndeformation of the structures. These researchersfoundthat initial expansionproceeded fater at elevatedtemperaturebut the Me declined more rapidly,The final expansionswere roughlyinverselyproportionalto storage temperature,Petrogmphicexaminationdid not revealthe classicalkali-silica (asr) reactionrims that are commonly observed,Crackswere often observedaround“aggregatemargins”whichm- diate into the paste matrix,Asr product was rarely seen inflllingthe cracks,but was most often seen at the centre of reactivepatticles, In one structure,large quantitiesof ettringitewere ob- sexved,This material wascoarselycrystallineand was seen inflllingthe microcracksand “lining voids”, Ettringite oftenappearedto have replacedpreviouslydepositedsilica gel, Jones & Poole noted the researchof ~reening [120]whoshowedthat monosulphatebecomesme- tastable in the presence of CaCOSand revertsto ettringite2,The reseamhersproposedthat the presence of limestoneaggregateprovidedthe calcite whichrenderedettringitethe stable sulphate phase, Ettringite, initially dispersedthroughoutthe microstructurerecrystallizesinto a coarsely crystalline form, The process is helpedby increasedpermeabilitycausedby other accompanying processessuch as asr cracking,The formationof ettringite in the voids is thus due to a through so- lutionprocessthroughthe pore fluid,The “preferentialrecrystallization”of ettringite in “gel- filled” microcracksmay contributeto the overall deteriorationprocess, E1-Sayed[96] did a post-mortemon 7 deteriomtedreinforcedconcrete structuresin a marine en- vironment in Egypt, Ettringite was observedin 3 of the 7 concretes,He concludedthat a combina- tion of factorsultimately contributeto premature deterioration, Most experimental researchnecessarilyattemptsto isolateone or two variablesand examines the effects of controlledvariationson behaviour,As E1-Sayednotes,the real situation is far differ- ent, It is hard to imaginea practical situationwhere,over the course of severalyears, only one mechanism is responsiblefor concrete deteriomtion,

I See Section3.8 for more informationon the role of chloroaluminates 2 The role of carbonatesin the formationand effectof secondaryettringiteis givenin Section3,7,2 chapter 2 Page 7

2.2 Petrographic Observations

One of the common observations of damaged structures is the occurrence of localized formations of ettringite. For example, Chandra and Bemtsson [48] examined damage in the back-lining of swimming pools in the south of Sweden. ‘Ihe concrete lining was 15 years old. They observed both translucent and opaque fibres and “white particles” on surfaces; in spots there was a white precipitate. Analysis confiied asr, but ettringite was also presen~ along with calcium hydroxide, calcite and gypsum which had been leached to the surface, The diagnosis was that deterioration was “caused by a combination of interacting factors” including alkali silica reaction, aggressive m- reactionof sulphates, and alteration of the fekispar aggregate [55], Mackod et al [245] examined damage of a concrete bridge in Strathclyde, Scotland They ob- seaved the presence of “white spheres” in the microstructure, ranging in diameter of 1 to 4mm, which had grown preferential y within the spherical voids of the concrete (air-entrained voids). X- ray analysis identified this material as ettringite, with a small amount of gypsum and “an unidenti- fied chlorine-rich phase”. This material was clearly a secondary alteration product.

Ozol [249] performed petrographic examinations of concrete attacked by alkali-silica reaction which revealed “desiccated gel balls” up to 7mm in diameter adjacent to aggregate particles. Fine crystals of ettringite were incorporated into these balls. Ettringite was also found to be present in cracks and in coarse-aggregate sockets; sometimes it was found by itself and other times it was found “in association with” dried asr gel-reaction product. Neck [237] noted a common feature of damaged concrete that has been extensively heat pre- treated and exposed to weathting is the “secondary phases” that seem to form at the contact zone betweenh matrix and the aggregate. ‘I%esephases appear to be ettringite or thaumasite or mixed C7ystals. Sylla [2971 performed microscopic examinations of samples of cracked concrete. He found the cracks to be almost completely filled with needle-shaped reaction products. The needles were all orientated vertically to either the aggregate or crack surface. It appeared from the observations that the crystals grew in a uniform front into the crack (suggesting a topochemical ~action mecha- nism)l. Once the crack is full rhen further transport of ions and growth may result in an increase in crystallization pressure and further damage. Sylla postulates that the initial cracks are “pre-ex- isting” as a result of cracking during thermal treatment.

23 Problems with Precast Concrete and Ties

In a Research Institute document [268], problems prior to 1984 are reported to have occurred in Germany with heat-treated prefabricated concrete elements, and especially railway ties and clad- ding panels. The problem initiates as crack formation at corners and edges which then tend to ex- tend into the interior. More severe separation of aggregate from the paste matrix then follows. Petrographic examination of the cracks almost always shows intilling with thaumasite or thau- masitdettringite mixed crystals. It appears that the thaumasitdettringite mixture forms in already existing cracks, The report notes that heat-treatment has two important effects to initiate these problems:

I See Section3.10 fora descriptionof reactionand expansionmechanismsof ettringite chapter2 Paga8

= an inadequatepre-treatmentallows internaldamagethrough debondingbetweenaggre- gate particlesand the paste matrix,and throughcracking, = heat treatment interruptsthe normal formationof ettringite;this formationcontinues later in the hardenedconcrete, withundesirableresults, It was noted that the solutionto the problem is to providean adequateperiod for precuring, Tepponen [303]reportedthe Scandinavianexperiencewith problemsof productionof railway sleepers(ties) that have occurredsincethe 1960s,Ties were manufacturedin the 60s and 70s with high early strengthcement (approximately4% S03 and 8% C3A). The fineness of cements was in the range 460-480 m2/kg, The cement content of the concrete mix was approximately 400 kg/m3and the water cement ratio varied from0,36 to 0,39,Ties were manufacturedwith either a 1or 2 hour delayperiod] and a maximumtemperaturein the mnge 75-80”C,The soakingperiod was2,5 or 4 hours. One-daystrengthof the 1965ties was 56 MP~ and was43 MPa in the 1971 ties. Tepponen notedthat the ties showedvisible damageafter 15years,Thin-sectionanalysis showed partial filling of the microcmcksin the ties that showedcracks,The more deterioratedties had all the pores and some of the largercracks filled with “microcrystalline ettringite” which“had accu- mulated in the free space”,In the more severelydamagedconcrete the bond betweenpaste and aggregatewas poor, Further studiesin 1970confirmedthat the ~ reason for deteriorationwas, however,the poor ilost resistanceof the concrete,To solvethis problem air content was in- creased and the maximumheat-treatmenttemperaturewas reducedto 60°C. New studiesperformed in 1980revealedthat “microcracks,,,arethe primarycause of deteriora- tion” [303],These cracks occur due to host, load,prematureheating, improperheat treatment, too severe or rapid heat treatment,etc, If temperatureis also greater than 70°C duringcuring, then “metastablemonosulphate”is producedwhich formsettringite when (a) the temperature dropsback to normaland (b) sufficientwater h supplied, New productionmethodswere adopted in the 1980swhich employednew Germanguidelines for heat treatment.These consist of a prestomgeperiod of 3 hours,a mte of temperaturerise be- tween 10=15°C/hrand a maximumtemperaturebetween6$70°C, Little changes in the cement compositionwere made and, in fact, the newcement that was used had a higher aluminateand thus C3Acontent (12Y0in 1980VS,8?40in 1965), One possibly significant change was the mduc= tion in the maximum size aggregate from 32 to 16mm, The new concrete strengthat 1day was 68MPa, Tepponen reportsthat the ties have been in service for 5 years and no expansionhas been observedeventhough microcracksdue to stressare present [303], Heinz and Ludwigs’sexperiments[132],discussedin detaii in Sections5,3 and 5,5, were promptedby practicalprobiemsthat surfacedwithprecastunits — specifhxily claddingpaneis and precast units, These high-strengthstructuralelementswere heat-treatedduringproduction, Damage surfhcedafter severalyears’ exposureto “open-airweathering”wherethere was lle- quent saturation.Crockswere first observedaroundedges,foliowedby furthercracking and loss of bond betweenthe paste matrixand coarseaggregateparticles, In another publication,Heinz and Ludwig [133] noted occurrencesof damage in Europe of high- strengthprecast units which used Type 55 high-early-strengthPortlandcement, These members were heat-treatedin production.Damagealways occurredon units exposedto the weatherand

1 also called“pre-storage”or “preset”— the time betweencastingand the startof thermaltreatment 2 See Section6,3 for a discussionof precastspecifications chapter 2 Page 9

subjected to frequent moisture satumtion. The damage was attributed to the reformation of et- tringite following heat-treatment. Testing at the British Cement Association[177]to examine the importance of secondary et- tringite formation was prompted by problems with site concretes in Germanyand Finland,and particularly with deterioration of rail ties. In the U.K. there have been a few cases of damage of particular prestressed and reinforced concrete bridge beams that have cracked at a late age due to “ettringite crystallization”. In particular, 15 prestressed and reinforced concrete beams were examined that were cast be- tween 1963 and 1969 [177]. These beams cracked due to “unexplained causes”. All were cast us- ing high early strength cement, and all were probably steam-cured. The cement content was greater than 450 kg/m3 in all concretes. The cracks took as long as 15 years to appear. The prob- lem was initially thought to be due to alkali-silica nxwtion, but examination of thin sections showed dense bands of ettringite, 25 pm thick, around coarse aggregate grains. In untracked sec- tions there were thinner layers of ettringite around aggregate particles. Ettringite has been observed in normallyast structures which, in most cases, have been dam- aged by a combination of fhctors. In such structures ettringite appears capable of being a contribu- tory cause of destruction, but unless an ample source of external sulphates is provided does not appear to be able to instigate substantial destruction. On the other hand, case studies which have examined deteriomtion of precast concrete, and many examinations to be described in following sections, strongly indicate that secondary ettringite formation can be ti main cause of destruc- tion once the heat-treatment process has established the nucleation sites for ettringite to form. Ac- cording y, it is appropriate at this point to examine the precast concrete process.

2.4 Precast Concrete

2.4.1 Effect of Heat Treatment on Properties

It is clear flom the literature that the main reason accelerated-curing r6gimes were devised for precast concrete was to accelerate the strength sufficiently that production of structural elements could be performed on a one day cycle. Economic pressures provided encouragement to obtain a strength sufficient to allow release of prestress; this value varied for different applications, but in general the 18 hour to 1 day target strengths were in the range 27-31 MPa [129, 184,218]. Until the 1980’s, strength development and the result of accelerated curing on both early and later-age strengths were the prime objectives of the research. In 1951, Saul [278] noted that from 1921 to the time of his writing considerable controversy and conflicting research had occurred concerning steam-curing. He noted that there had been little agreement on optimum curing temperature, duration of treatment, delay period, etc.

Saul found with his own research that the most important factor influencing the ultimate strength of the concrete was the rate of temperature rise. He reported that good results could be achieved if the concrete temperature did not reach 50°C until 2 hours after casting, and 10O°Cuntil 6 hours after casting. Saul criticized some commercial operations because they used too rapid tempera- ture rises, but noted that this is satisfactory if a suitable delay (preset) is employed prior to the start of heating. Saul noted that very rapid temperature rise can be used if very high early strengths are needed, but 7 and 28 day strengths may suffer by as much as 1/3 (when compared to a normally cured control); later, others came to similar conclusions [e.g. 321]. The effect of heat-treatment, and especially 100”C treatment, on durability of concrete is not a theme that is considered by Saul — or by any of the early research on precast concrete. There is no indication Clwpter2 Page 10

that concretes manufactured with the aggregates and cements available at the time showed any long-term distress. Chamberlain[47] was another early researcher of precast concrete. He observed that steam-curing between temperatures of 85 and 93°C caused severe damage to the concrete when delay periods were too short. Furthermore, subsequent strength gain was small. On the other hand, concretes treated at 74°C were as strong as the control concretes at 28 days. Based on these results, it was concluded that a delay period of “several hours” is required after casting, and a maximum tem- perature rise of less than 22°C/hr is necessary if it is desired to prevent damage to concretes steam-cured at temperatures greater than 74°C. Plowman [256] confirmed that the low strengths of heat-cured concrete at later age was a well-es- tablished phenomenon; ukimate strength is influenced by the rate of early hydration. It only takes a high temperature for the first few hours of hydration to reduce ultimate strengths significantly. The work of Klieger in the late 50’s and early 60’s [164-166] fmly established the pros and cons of the heat-treatment procedure. Fairly moderate curing temperatures from 4.4 to 49°C were used by Klieger in his research. The strength at 1 day of the 49°C cured concrete was 178°/0that of the control concrete cured at 23°C, Later-age strengths were somewhat reduced but “no retro- gression in strength occurs with age”, Klieger suggested that an acceptable heat-curing cycle con- sisted of 3-6 hours delay period, a 16 hour heating period and then gmdual cooling over 3 hours to avoid excessive drying, Klieger also found that elevated curing temperatures affected flexural strength in a similar man- ner to compressive strength; although initial strengths were higher the concretes that were tested exhibited considerably lower strengths at 3 months and 1 year [165]. In one of the few examinations of durability, Klieger observed that curing at elevated tempera- ture of concrete made with high-early-strength cement does not hinder freeze-thaw durability if the concrete is dried prior to exposure, Klieger suggested that a few days of drying in the yard is enough to accomplish sufficient drying. He also found that concretes cured at elevated tempera- ture in a 24 hour cycle show lessdryingcreep and shrinkageunder sustainedstressthan those cured at normal temperatures[166]. Hanson [128] also noted that accelerated curing causes significant reductions in creep and shrink- age, This has important influences on the magnitude of prestress loss which can be reduced by as much as 40% if heat-treatment is used, The author suggestedthat a heat-treatmentrbgimeconsist- ing of 4-6 hours delay period followedby a 17°C/hour temperature rise to 66°C and a soaking pe- riod of 13 hours is about optimum with respect to strength development, In a subsequent paper, Hanson examined the properties of lightweight concrete [127], The opti- mum conditions for steam-curing lightweight concrete were not substantially different than those for normal-weight concrete, this was confirmed by other research [185]. It was also found that lightweight concrete is less susceptible to damage if severe heat-treatments are used (too high tempetuture, too rapid tempemture increase, insufficient delay period). This is probably due to the high elasticity of lightweight concrete enabling it to absorb better the internal strain that oc- curs upon heating. Merritt and Johnson [218] determined the strength of various heat-treated concretes. Soaking time and soaking temperature were the main variables. A summary of the results is given in Fig- ure 2,3, For all conditions heat-treatment resulted in a strength immediately after treatment that was greater than the fog-cured concrete. The 7 and 28-day strengths were always less than con- trol. Note, however, that the delay period for all casts was very short (0-0,5 hours), At this short preset the optimum curing r6gime, both mechanically and financially, appears to be a soaking time of 18 hours within a temperature range of 52-66°C. chapter 2 Page 11

% of Strength Compared to Fog-Cured Control ❑ 180 I I I After Heat Treatment DelayPeriodforAllCasts=0-0.5hrs. ~-’------❑ 7 Days 160 L------‘1 ------~[Q28D.ys 140 ::: 120 --;------.II::::~ ------100

80

60

40

20

0 Sosklime (hrs) 18 42 66 18 42 66 18 42 66 18 42 66 18 42 66 ~ 38°C 52°C 66°C 79°C 93°C

Figure 2.3 Relative Strengths of Heat-treated Concretes: Dependence on Soaking Time and Temperature [data from ref. 218]

Australian experience indicated that a premature application of steam can reduce the compressive strength at later ages by up to 200A.An optimum delay period between 3-5 hours is appropriate for temperature gradients from 10-40°C/hour. If an adequate delay period is used, then sub- sequent strength development is directly related to the maturity. However, with a proper delay and moderate temperature rise the 28 day strengths of heat-cured concretes can exceed those of standard-cured concretes [184]. Farslq’ noted that after a plastic strength of 1 to 1.5 MPa is achieved in the concrete the speed of heating does not influence strength. The period of delay prior to heating is, however, a decisive influence on later strength development. He proposed that a reasonable period of delay would be the time to initial set of the cement [100]. Some researchers have suggested that tailoring the cement chemistry or the use of chemical ad- mixtures can help to reduce the problem with low ultimate strengths. Saul [278] suggested that the adverse effect on strength due to a rapid temperature rise during pretreatment can be compen- sated for by additions of calcium sulphate or alkali to the mix. It was also proposed that the use of a superplasticizer can permit as much as a 25?40reduction in the water content of precast, heat-treated concrete. As a result, the normally observed reduction in long-term strengths can be avoided. Concrete cast with w/c=O.3through the use of a supeqhs- ticizer resulted in a 43% greater strength at 28 days than a normally-cured, superplasticizer free, equivalent Type I concrete [254]. Chupter2 Page 12

Pfeifer and Marusin [255] observed that the compressive strength of heatared concrete was greatly affected by the chemical composition of the cement. With respect to optimizing 1 day strength, they proposed that the best chemical composition is one with: (a) C3S:C2S = 3: 1; (b) C3A:C4AF = 2: 1; (c) C3A = 10-15%; (d) S03 = 5-7%. This is similar to a standard Type 30 ce- ment except for the higher S03 content. For optimum high early strength “the S03 content should be as high as allowed”; current ASTM specifications allow a maximum of4.5% S03, but “early-age strengths can be dramatically increased during heat curing by using somewhat higher S03 contents than currently allowed by ASTM C150”l. It is important to note that although strength has been the primary property investigated, long- term strength is not the only property that suffers by inadequate heat-curing. For example, a short steam-cure at 54°C followed by drying resulted in a 4000% increase in permeability when com- pared to a water-cured control specimen [135]. This is shown in Figure 2.4. All steam-cured con- crete resulted in an increase in permeability, Fog-curing afier heat-treatment is beneficial since it results in a dectease of permeability meas~ed at 28 days,

6 HR.STEAM@54°C, 28 DAYDRY

24 HR.STEAM@54°C, 28 DAYDRY

24HR,STEAM@64°C, 7 DAYFOG,21 DAYDRY

1 1 SHR.STEAM@71°C, 1 I I 1

28 DAYDRY 1 I 1

1 1 1 8 HR,STEAM@71°C, 7 DAYFOQ,21 DAYDRY

24 HR,STEAM@71°C, 28 DAYDRY

24 HR,STEAM@71°C, 7 DAYFOG,21 DAYDRY 1 I I t 1 1 I 1 t 1 I o 5 10 15 20 25 go 35 40 45 PERMEABILITY OF SPECIMEN/ PERMEABILITY OF 28 DAY CONTROL SPECIMEN

Figure 2.4 Effect of Heat-Curing and Drying on Relative Permeability [ data from ref. 135]

1 Note that there is no mentionof long-termpropertiesin this document,The proposalby the researchersto use as much S03 as possible is, in this author’sopinion,indicativeof the acceptanceof strengthas the universalqualityindicatorthat predominatedin the industryuntil recently,Fortunately we arenow learningvery quickly that adherenceto that philosophyoften leadsto infkriorconcretein the longterm-with resultantcostlyrepairs. chapter 2 Page 13

2,4.2 The Cause of Insufficient Strength Gain after Heat Treatment

Hansom in 1963, noted that microstructuml damage is caused during a short pre-steaming perio~ especially when the heating rate is high. Cracks that occur are due to the action of tensile stresses caused by a rapid heating rate and resultant stress distributions that occur across the specimen or structural member. The greatest cause of these stresses is probably the much larger expansion co- efficient of free water when compared to the solid ingredients of the concrete. To avoid this dam- age, Hanson suggests an optimum curing period consists of 5 hours delay, a temperature rise at 22°C/hr, and a maximum constant temperature of 66°C held until steam shutoff at 18 hours; this allows 6 hours for form stripping and readying for the next run [129]. Alexanderson [7] proposed other techniques that could be used to minimiie the effects of dam- age associated with expansions and cracking: ~ nxisting pore pressures-through the use of closed forms, or by letting the concrete at- tain a minimum strength before the stati of heating; ~ eliminating pore pressure-by heating the concrete ingredients before casting or by elimi- nating air voids; ~ letting the concrete crack-and repairing it, perhaps by mwibration after heating. Others have also attributed damage to the principal mechanism of differential thermal expan- sions. Sylla [297] stated that differential thermal expansion appears to be the primary cause of cracking during thermal treatment. Water is a key component in this regard because it has an ex- pansion coefllcient about 10 times greater than any of the solids. Sylla hypothesized that prema- ture heating causes large expansions in the thin layer of water that coats the aggregate particles at casting. If heating is delayed somewhat then some of the water in this layer is removed due to hy- dration; also the concrete has some stability so that the overall damage is less. Soroka et al [292] listed three reasons to explain why heat-curing causes a reduction in later-age strengths when compared to a normally-cured control. The first hypothesis was proposed by Verbeck and Helmuth [3 12]. Due to heat-treat- ment, the paste microstructure is more heterogeneous because of the very rapid rate of in- itial hydration. Ample time is not allowed for hydration products to dlfl%seaway and precipitate in the space between cement grains. A dense hydmte forms near the cement grains and a weaker gel and more interstitial space occur away from the grains. Cement hydration at later ages is also retarded because of the denser layer of hydrate surrounding the unhydrated grains Temperature appears to change the morphology of C-S-H gel particles. Heat treatment re- duces the relative amount of long C-S-H particles to short particles. It is not entirely clear why this should have a profound eff=t on strength development. Rapid heating results in differential thermal expansions and cracking —thus, damage is produced. To examine this reduction in strength, concrete was steam cured after a 30 minute delay period. The treatment temperatures were either 60 or 80°C and were held for periods ranging from 2 hours to 4 hours:40 minutes (Figure 2.5). For some specimens (Cure 3) no fiuther moisture was provided (65’Mor.h.) after heat-treatment until specimens were tested at 28 days. In Cure 2, speci- mens were water cured after heat-treatment until 7 days and then placed at 65% r.h. until 28 days. Control specimens (Cure 1) were treated like Cure 2 specimens except there was no heat-treat- ment. The results given in Figure 2.5 indicate that lower strengths may not be due to damage dur- ing heat treatment, but may be due to inadequate curing after the heat treatment. Chapter 2 Page 14

Cure 1 = Water cure at 20 “C for 7 days, then at 65% rh until test at 28 days Cure 2 = Heat cure at given T atler 30 minutes delay, followed by water cure at 20”C until 7 days, 65% rh to 28 days , 1 Cure 3 = Heat Cure of Cure 2, no water cure, ❑ CURE 2 ❑ CURE 3

Strength as Y. of Standard Cure Y. OF CURE 1 ‘YoOF CURE 1 1

100------..------

.80------

60------

40------

20------

ol- + 80-2.00 80-2.45 80-3.35 60-2.30 60-3.40 60-4.40 MAXIMUM TEMPERATURE ‘C -- TIME AT MAXIMUM (HRS)

Figure 2,5 Effect of Post-Treatment Water-Curing on Strength Gain [data from ref. 292]

Most mxearchers blame the presence of water and its large thermal expansion coefficient as the mason why damage occurs. However, Pfeifer & Marusin [255] propose that one must be espe- cially careful when air-entrainment is present. Moist air expands in the pores during heating, so even if the air content is as low as lYo,cracking can develop under certain conditions. Above 43°C a delay period is not necessary as long as temperature rise is gradual. At 60”C, the delay pe- riod should be 4.5 hours, and at a 90°C heat-treatment temperature, the delay period should be ap- proximately 6 hours. me delay period can be reduced if the initial temperature of the raw materials is increased, thus increasing early tensile strength gain and reducing the volume expan- sion of the various components.

2.5 Comment

The heat-treatment figimes for the precast industry appear to have evolved with one over-riding consideration — to obtain the highest strength possible in the shortest time. When engineers de- vised rt$gimesto do this they quickly tealized that later age strengths could be substantially less than that of a normal-cured concrete. If later age strength was not importan~ then a certain amount of cracking during the heat-treatment process was accepted. There appears to have been little concern or, perhaps, appreciation that long-term durability of the cracked material could be an issue. The chemistry and physical properties of cement in the 50’s and 60’s were significantly Chapter 2 Page 1S

different than now; it maybe that during this period strength MS a much stronger indicator of overall quality than it is at present. Most concrete manufacturers appear to have searched for a compromise to develop a heat-treat- ment r6gime that would still produce high early strengths but which would not result in large com- parative strength reductions at later ages. To do this it was realized that the rate of temperature rise had to be low and/or there had to be a significant delay period between the time of casting and the start of heathg. As will be shown in the following chapters, there is a third important fac- tor that must be considered if both acceptable early- and late-strengths and long-term durability are to be ensured — the maximum curing temperature. The deterioration of concrete in the field will almost always be a combination of influences, in- cluding freezdthaw, corrosion, alkali-aggregate reaction, stress cracking, sulphate attack carbona- tion, secondary ettringite formation and a number of other possible processes. The mechanisms by which the overall deterioration process advances is extremely complex and probably beyond any model that could be tested in the laboratory or on a computer. Nevertheless, particular processes can be recognized and we can ensure that the influence of those processes in the overall deterioration scheme are minimized. Corrosion and sulphate attack m two examples where simple steps have been taken to ensure, at least on paper, that their effects will be minimized. Recently, several researchers in Europe (see Chapter 5 as well as this chapter) have provided strong evidence that under the right conditions secondary ettringite formation maybe a significant factor in influencing the long-term durability of some types of concrete. Like corrosion and sul- phate attack steps can be taken to help ensure that secondary ettringite formation is either an in- significant or only a minor influence; some of these steps can be discerned from reading the material to follow. At the moment, secondary ettringite formation is different in one key aspect when compared to corrosion and sulphate attack it is by no means as widespread in occurrence. Nevertheless, it may become a very important consideration if its potential is ignored. It should, however, be kept in mind that at present the proportion of damaged concrete where damage has been directly attrib- uted to secondary ettringite formation is a very small proportion of all precast concrete cast in Europe and North America. chapter 3 Page 16

Chapter 3 SECONDARY ETTRINGITE FORMATION: FUNDAMENTAL RESEARCH

3.1 Overview of the Cement Hydration Process.

This report assumes that the reader has a working knowledge of the chemistry of cement and ce- ment hydration; therefore, no attempt is made to perform a detailed review. The reader is referred to the excellent reference book by Taylor [302] for comprehensive information on all aspects of the properties of cement. The formation and consequences of formation of ettringite are the principal topics of this report. In a normal modern cement, ettringite forms due to the reactionbetween gypsumand calciumalu- minate. X-raypeaksthat are associatedwith ettringiteare detectablewithin a few hoursand the quantity increasesduringthe first day. Afterthis time the ettringitepeaks normallyweaken,but, dependingupon the chemistryof the cementand the environmentalconditions, ettringitemay persist indefinitely,It is a general belief that if all of the gypsum is consumedthroughthe et- tringite reaction,ettringiteconvertsto monosulphoahuninate.In an average cement this conver- sion processstarts at about 1day and can be monitoredas a reductionin the size of the ettringite X-raypeaks. Much of the monosulphoaluminatephase that forms instead is poorly crystalline [302].

3.2 Early Formation of Sulphoalumlnates

The commonlyheld chemical formulafor ettringiteis: 3Ca0,AlzOy3CaSOq.32Hz0 or in chemists’ notation is: C6AS3Hn and for monosulphateis:

3Ca0.Alz03,CaS04, 12Hz0 0!’: c4A~12 The reactionthat occurs between C3A and gypsum can be found innumerous texts: C3A + sC~Hz + HZ6+ C6A~3H32 Except for the unusualcombinationof very low C3Acontentsand high gypsumcontents,the gyp- sum will be used up by this reactionbeforethe C3A,Whenthis occurs,the remainingC3Awill take part in a reaction wherebyettringitedecomposesto monosulphoaluminate: 2C3A + c(jA~3H3z+ H4+ 3C4ASHU Thus, in the net reaction,3 moles of CSAand 3 molesof gypsumwill produce 3 moles of mono- sulphoaluminate.Use of the Bogue equationand performingstoichiometriccalculationsfkomthe formulaeabove, one can cdcuiate the S03/Al@ massor molar mtios (TVAratios)requiredto consumeall of the C3Aand convert it to monosulphoaluminate,Figure 3,1 showsthe depend- ence of this mtio on the C3Acontent of the cement (givenan assumedFe20s content of 3%), If the WAmtio is greaterthan the value shown,C3Awill still be available after all of the ettringite has been consumet the remainingC3Awill react withcalciumhydroxideand water to produce C4AH13[302],The above analysisis highly idealized.It is importantto recognize,however,that the conditionsunder whichettringite forms,and the stmcture of ettringite itselfare highly com- plex, chapter3 Page 17

S03/A1203 Ratio Required to React with AU Tricalcium Aluminate to Form Monosulphoaluminate 0.7 I I -/ /n- 7 0.6 ------..- ---- . Molar Ratio /0-” - . cl-- /“ 0.5 / ------&------A /. u“ Mass Ratio S031A1203 0.4 ------. -~ ’------‘ ------RATIO # (mass or /’ molar) 03 T~------‘ ------

0.2 /’-”7‘ ------”------t 0.1 1------“------

OLH—H—HJ 2 3 4 5 6 7 8 9 10 C3A PERCENTAGE

Figure 3.1 Critical Values for S03/AlZ03 Ratio

3.3 Crystal Structure and Composition of Ettringite

The use of the term “ettringite” without qualification normally refers to “sulphate ettringite”, which has the formula C6A~3H32.However, it should be noted that “e~ingite” is) in f~ts a gen- eral term used to denote a number of minerals, all with very similar crystal structures. McConnell & Murdoch [202, 203], for example, noted in 1962 that ettringite should be represented by the general formula 16[A4(X04)34(H20)12] where “A” represents six-fold coordinated atoms (Ca, N% Al), and “X” are four-fold coordinated atoms (S, Si, I-Uand also some Al), The formula is written to make clear the interchangeability of the various ions. Moore & Taylor [231] wrote the formula for ettnngite in a different way, again in an attempt to clari~ all of the various substitution of ions that can occur, but also to illustrate the crystal struc- ture of this class of material:

@[ Al(oH)6]2.(so4) 3”26H20 chapter3 Page 18

The researchers noted that the crystal structure is based on columns with composition

[ti3[Al(OH)(i]- 12HzO]3+ which run parallel to the c, or needle, axis. The sulphate and xemaining water molecules lie be- tween the columns. Taylor [300] provided a very clear picture of the ettringhe-group crystal structure which is char- acterized by four positively charged columns in a trigonal structure, between which occur chan- nels which contain anions and sometimes water molecules. Figure 3.2 shows the sulphate-ettringite structure. The a and b lattice dimensions are 11.23A, while the repeat dimen- sion in the c direction is 10.7~. Two other minerals which closely parallel the ettringite structure, and are of interest to concrete researchers are thaumasite and jouravskite. The parallel structures can be clearly seen in the chemical formulae: ~ ettringite: {Ca~Al(OH)6~24H20}(SOA)s2H@

= thaumasite: {Ca~Si(0H)6]224H20}@C)4)y(C03]2

@ jouravskite: {Ca~Mn(0H)6~24H@}@C)A)y(C03)2

In thaumasite, silicon teplaces ahuninium in the columns (Figure 3.2), while in jouravskite man- ganese replaces aluminhum For both thaumasite and jouravskite water molecules contained be- tween the columns of the ettringite structure are replaced by carbonate, In fact, Taylor [300] notes that partial or complete replacement of sulphate ions can occur with C032-, Cr042-, Cl-, and 103-”The aluminium, on the other hand can be replaced by Ti, Cr, Mn, Fe, or Ga, The calcium can be replaced by strontium, Other researchers [196, 326] have con- firmed the wide variety of ions that can replace aluminium and/or sulphate, More recent reseamh [130, 171] has also shown that oxyanionssuch as arsenate,borate,chromate,molybdate,selenate and vanadate may substitutefor sulphate in the ettringitestructure, Taylor [300],expandingon earlier work re-definedthe compositionof the columnper half unit cell in sulphate-ettringiteas {Ca~Al(OH)G~24H20~+,whilethe material in the ettringitechan- 6- IRh iS (SOA)3,2H20 to make the entire structure electrically neutral, He noted that each col- [ umn is of nearly cylindrical1 shape and its surfhce is comprised entirely of water molecules, “or more accurately the H atoms that these contain”. It has also been shown that various ions can replace sulphate in monosulphoaluminate. The wide variety of chemical compositions of both compound types has prompted the use of the terms (a) “AFt” for Ahuninate-Ferrite-trisubstituted — the “needle-like hexagonal hydration products which crystallize with 3 molecules of a calcium salt per molecule of C3A”; and (b) “AFm” nota- tion to stand for ahminate-ferrite-monosubstituted compounds, which are hexagonal or pseudo- hexagonal plates which contain one molecule of a calcium salt per molecule of C3A [163].

3.4 The Role of Ettringite in the Retardation of the Caicium Acuminates

Gypsum is added to Portland cement clinker to @ard the rapid hydration of the calcium ahuni- nates. Lerch [180] outlined the mechanism of retardation of aluminates by sulphates. When water first contacts cement there is an initial rapid dissolution of anhydrous aluminates and rapid crys- tallization of hydrated calcium aluminates. This occurs before the solution becomes saturated Page 19 chapter3

Blank Circles = 10.7 Water Molecules

i

Structure of Columns in Ettringite Crystal Part of Single Column Projection I Blank Circles are Water Molecules I

Plan View, Showing Columns and Channels

l\ \/ Lattice a dimension= /s\ x / 11~ A S04> and Water Molecules Located in Channels Between Columns

o 2A 1

. Figure 3.2 Sulphate-Ettrlngite Crystal Structure [extracted from Taylor [300] and annotated] chapter 3 Page 20

with lime and gypsum and corresponds to the fust peak in the heat of hydration curve. Since lime and gypsum in solution decrease aluminate volubility, the rate of formation of aluminate de- creases as more of these materials are dissolved (lime alone is not sufficient to retard the hydra- tion of aluminates), Once a saturated Iirnehulphate solution is achieved, aluminates continue to hydrate more slowly. Calcium silicates dissolve and form C-S-H. Sulphoaiuminates start to form. Because of the formation of sulphoaluminates, the gypsum eventually becomes depleted and the concentration of sulphates in solution starts to drop. The volubility of alumina again rises and a rapid aluminate reaction begins again, Another theory attributes retardation to the direct formation of a coating of ettringite on the sur- face of calcium aluminate grains [69]. The retardation mechanism of gypsum is more effective in the presence of CH since ettringite crystals are more fine-grained and can therefore coat the sur- face better. Some later research, however, has questioned whether a sufficient amount of etlringite can deposit in a short time to have the pronounced retardation effect that is observed [206]. Although there has been no convincing explanation of the retardation mechanism, fairly convinc- ing evidence does exist that ettringite forms by a through-solution mechanism. There has been much controversy about this over the years, and various schools of thought have formed about whether formation is through solution [e.g. 134] or topo-chemicai, There is also some suggestion that formation may be due to a combination of mechanisms: the first is a through-solution mecha- nism. The second is a process by which some ettringite crystals sinter together by a solid-solid re- action where crystals grow together and sinter [149]. The most convincing march also relates dkctl y to the secondary ettringite issue. It has been ob- served that the transition zone at the aggregate and steel interface in concrete has a higher propor- tion of ettringite than elsewhere in the bulk matrix [225]. ‘l’hisfinding supports the through solution mechanism of formation, since constituents must dissolve and dtffuse towards the steei/aggregate surface where ettringite is precipitated, In another series of experiments Monteiro & Mehta [228] showed a distinct gradient in ettringite concentration in the interracial zone be- tween aggregate and paste, Much larger quantities of ettringite are present immediately next to the aggregate surface. They concluded that this could only be explained by a “through solution” mechanism and not by a “topochemicai” or “solid-state” theory whereby compounds do not dis- solve, The importance of the “transition” zone to secondary ettringite formation is discussed in de- tdl in Section 4,2,

3s The Chemical Stabiiity of Suiphoaiuminates

3.5,1 Stability in Solution

The earliest studies (1929, [181]) on the stability of ettringite showed that it is more stable in solu- tions of calcium sulphate and calcium hydroxide than in distilled water. On the other hand, mono- sulphate in solutions of calcium sulphate, sodium sulphate, calcium chloride and sodium chloride tend to change to the high sulphate (ettringite) form. Lerch concluded that only the high sulphate form of sulphoaluminate can be stable under the conditions expected in concrete. The instability of the monosulphate form was noted by IQdousek [156]. The AFm phase is a me- tastable compound this is also confirmed by research that shows monosulphoahuninate is metas- table below 70°C [204 reported in 132]. Once monosulphate forms it is metastable but may persist for indefinite periods under the right conditions; its metastability dictates that it will finally convert tQthe trisulphate form (ettringite) — this is confirmed by other, more recent research [163]. llie formation of monosulphate requires nearly saturated solutions of both calcium hydroxide and gypsum. If excess sulphate is not present the monosulphate may persist for weeks but will finally convert to the stable trisulphate. The stability conditions in solution for monosulphate involve high CaO, low CaSOAand high temperature[1521. Mehta found that monosulph~ hydr~ forms in cement pastes along with ettringite “but persists only in contact with solutions of much lower sulphate concentration than required for stabilization of ettringite” [207]. At the nornul temperatures experienced by concrete, Kelly found that ettringite is more stable tin monosulphate or C3AHG.Monosulphate readily forms at higher temperatures, however, which is due to its “greater crystallizability” as opposed to its higher stability [160]. The stability of ettringite in relation to other ettrhwite 1.O*1040 compounds can be explained by the “rule of monosulphoaluminate 1.7*10-28 volubility product” [326]. The compound with carboaluminate 1.4*10-30 the lowest volubility product is the most sta- chloroaluminate 1.O*10-30 ble. The table on the left shows the volubility calcite 8.7*10-W products for four major aluminates that am known to form in concrete. Calcite, a soluble compouncLis included for comparison. Based upon this table, the following transitions can occw calcium aluminate + calcium hydroxide+ gypsum+ ettringite

monosulphoaluminate + gypsum + etlringite

ettringite + NaCl + chloroaluminate

monosulphate + CaC12+ ettringite + chloroaluminate

monosulphate + CaCOs + ettringite + carboaluminate This analysis may bean oversimplification of the issue. Hampson & Bailey [124] noted that et- tringite does not exhibit a constant volubility product-it varies 5 orders of magnitude as the prod- uct of hydroxide and sulphate activity increases. This leads them to postulate that ettringite may lose its characteristic fibrous morphology at high pH and forma coherent coating over other parti- cles. They note two possibly distinct forms of ettringite, depending upon the pH of the pore solu- tion: = at pH= 11.5-11.8 — crystalline ettringite precipitates from solution ~ atpH=12.5-12.8 — the result is a “topotactic” formation of a non-crystalline material thal has a similar composition to ettringite. The one special situation that the volubility product method does not predict is the formation of monosulphate from ettringite. Monosulphate has a volubility product at 25°C of 1.7*10-28, whereas ettringite has a solubilit y product of 1.1*1040. Normally, the most insoluble sak et- tringite, will prefemntiall y be formed. However, when the sulphate in the pore solution is de- pleted, ettringite will convert to monosulphate. The extent of conversion depends entirely upon the amount of excess C3A present after the initial formation of ettringite

3.5.2 Stability and Temperature

For curing at 25°C ettringiteis detectableby X-ray analysisaf~ a few hours [302].When hydra- tion occurs at 60”C,however, ettringite X-ray peaks begin to form at 15 minutes and there is al- ready appreciable ettringite at 6 hours. The ettringite crystals formed at the higher temperature are appreciably larger than those developed at 25°C [1341. Chupter3 Page 22

At temperatures above 60”C, a different scenario develops, At approximately 70”C ettringite loses much of its water. For example, Daerr et al [79] note that the water content drops from 32 to 10 molecules at 70.5°C. Also, when water is lost between temperature of 70-85°C all ettringite compounds lose the molecular water in the channels and become amorphous [260]. Skoblinskaya et al agree; as ettringite loses water its structure remains intact only during the first-stage redu- ctionfrom 30 to 18 water molecules. However, when water drops from 18 to 6 molecules during the second stage of desiccation the crystals become “roentgenoamorphous”. Withdrawal of the re- maining 6 water molecules causes transverse rupture and the crystals disintegrate [290,291]. There is a considerable diversity of research that quotes various tempemtures at which ettringite decomposes. Abo-El-Enein et al [1] found that ettringite, when hydrated at 25”C, loses about 20 water molecules at 58°C in a dry atmosphere. If it is hydrated at 60”C, however, it loses 18 water molecules at 58°C; thus, formation at elevated temperature makes ettringite more thermally sta- ble. Similarly, pressure also tends to stabilize ettringite against thermal decomposition-et- tringite pellets were manufactured at two different pressures; the pellets made at the higher pressure were more stable when heated [2]. Some research has concentmted upon examination of ettringite decomposition under very severe conditions. Although in pure water ettnngite is stable at 60”C at a pH of 11.2, at 100”C 2.25 moles of sulphate are split from the ettringite crystal, and monosulphate forms. If temperature re- mains at 100”C the monosulphate decomposes to gypsum after 11 days of boiling. At lower tem- peratures (30”C), monosulphate is unstable and decomposes partially to ettringite after 6 hours. At intermediate temperature (60°C) monosulphate enters into a solid solution with calcium alumi- nate hydrate in 6 hours, which decomposes to ettringite after a day [104]. Ghorab et al [106] re- port that the structure of ettringite is destroyed (a) at 18°C in a vacuum of 10-6torq (b) at 74°C under normal atmospheric pressure; and (c) at 82.5°C under wetted N2. Re-storage at 90V0r.h. leads to dehydration [106]. It has been postulated that up to approximately 70°C ettringite is the stable compound in cement and concrete, while above this temperature it is the monosulphate that is stable [ref. 143 notes the work of Mchedlov-Petroysan [204]). There is some research, however that shows that ettringite can exist in equilibrium with its aqueous solution up to 90-93°C [277]. The apparent instability of ettringite at higher temperatures may be related to its higher volubility [106, 107], Mehta reports that the short-prism type of ettringite found in the paste-matrix of concrete is sta- ble at 65°C in a dry atmosphere, but decomposes partially at 93°C. In a moist environment, how- ever, ettringite shows no significant decomposition at 93”C. However, if ettringite is exposed to saturated steam at 149°C it’will decompose to the monosulphate. The fme-gmined type of et- tringite found in concretes and pastes is more thermally stable than the long slender needles of et- tringite that form from dilute solution [214].

3.5.3 Post-Decomposition Behaviour

The combination of the S03 content of the cement and the heat-treatment r6gime determines the compounds that are present in a cement paste. Kalousek [157] tested pastes with 0.4 to 4’%S03 which were cured at 24, 54 and 82”C. The pastes were put through a 19 hour heat-curing cycle, dried and analyzed. At 54°C curing, Kalousek noted that the most notable change over 24°C was the decrease in the amount of ettringite and increase in monosulphate. For curing’at 82°C et- tringite was absent except for specimens where S03 content was greater than 3V0.There is an amount of “missing S03” which is larger in the 82°C cured specimen than in the 54°C specimen. Kalousek proposed that heat curing results in an amount of S03 that is neither bound in the mono- sulphate nor in the ettringite. It was suggested that this missing ettringite becomes a “lattice sub- chapter3 Page 23

stituent” in tobermorite (C-S-H) gel; the S042- ions occupy sites of the Si044- tetmhedrons. Other substitutions of ions such as A~ A120s, FezOS, MgO and Cl- ions are well known and therefore there is no reason why S04 - cannot substitute also. Figure 3.3 shows the distribution of phases at the various treatment tempemtures. As treatment temperature rises the amount of S03 missing (i.e. in “Phase X“) increases. The presence of ettringite shifts to higher sulphate con- tents. 4.51 . 4.0 24 “CTREATMENT 3.5 3.0 1’ ‘/ AMOUNT S03 IN 2.5

DIFFERENT PHASES, 0/0 2.0 1.5 1.0 0.5 0.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 TOTAL S03 IN CEMENT, %

4.5 ~ 4.0- - 54 “C TREATMENT 3.5- - 3.0 “- AMOUNT S03 IN 2.5- - DIFFERENT PHASES, % Y ~ 2.0 1.5 !.0 0.5 O.q Unknown hydrated 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4,5 calcium aluminate TOTAL S03 IN CEMENT, %

4.5 ~ 4.0- - 82 “C TREATMENT 3.5- - 3.0- - AMOUNT S03 IN 2.5- - DIFFERENT f’HASES, ~0 2.0-- 1.5- - 1.o- - 0.5-“ ETTRINQITE 1 I 5 2.0 2.5 3.0 3.5 4.0 4.5 calcium aluminata TOTAL S03 IN CEMENT, % Figure 3.3 Effect of Curing Temperature on Sulphoaluminate Phases [data from ref. 157] Type 30 cement. DTA analysis performed at age of 19 hours In 1965 Richards [269] produced some interesting xesults that relate to the stability of ettringite. He soaked mortars in various sulphate solutiow, the temperatures of the sulphate solutions ranged fkom 20 to 80°C. He found that soaking at temperatures greater than 40°C resulted in a dramatic reduction in expansions. Measurements were takenup to 24 months.Richardsconcludedthat the expansionprocessthat is normallyvery destructivedoes not take place above 20°C. It is unfortu- nate that he did not continueto measureexpansionsafter specimenswere cooled. Odler [242] examined the hydration reactions of pastes containing cement with Oto 20% C3A, 3% S03, and a Blaine fineness of 3000 cm2/g, Curing temperatures ranged from 5 to 95”C. The nxearchers found that at 5°C gypsum was consumed within 3 days and the quantity of ettringite was large. At 25 and 50”C the conversion process from gypsum to ettringite was accelerated; both ‘ AFt and AFm phases were present at 28 days. At 75°C curing AFt was only present during the first few hours, while AFm was observed at all times up to 28 days. At 95°C no AFt was detected at all; AFm formed instead but disappeared after 3 days. Similar results were obtained for all of the cements examined. Odler concluded that as temperature rises there is an enhanced incorpora- tion of Al-, Few and S042- into the C-S-H lattice. ‘l%edisappearance of ettringite with elevated temperature was well documented by Chino & Kawamura [56]. Hardened cement pastes (w/c=O.5) were cured for 14 days at temperatures rang- ing from 20 to 90°C. Ettringite was determined by X-ray diffraction. As Figure 3.4 shows, the quantity of ettringite in the paste dramatically decreases at curing temperatures above approxi- mately 70°C. Stadelman[294]useddifferentialthermalanalysisto analyzethe stabilityof ett.ringitecontentin hardenedcementpastes when treated at temperatures ranging from 20 to 100”C. When ettringite

1.20 Cement Paste Cured for 14 Days

❑ at Given Temperature ❑ 1.00 -- ●

0.80 --

RELATIVE INTENSITY 0.60 -- X-RAY PEAK + ~ mean 0.40 -- ❑ high

● low 0.20 --

0.00 o 20 40 60 80 100 CURING TEMPERATURE (C)

Figure 3.4 Variation in Ettringite Peak Intensity with Curing Temperature [data from ref. 56J chapter 3 Page 25

was treated at temperatures above 60”C it no longer appeared on DTA tmces; instea~ monosul- phate appeared. During storage afler treatment, ettringite peaks reformed which Stadelman con- clude~ “must be seen as a cause of concrete damage”.

3.6 The Effect of Alkalis on Hydration and Formation of Ettringite

Alkalis are normally present in the clinker as the neutral sulphates Na2S04, KXS04 or the mixed salt (N~K)2SOd [71, 252]. These compounds are highly soluble. Afier contact with water the pore solution contains almost entirely N% K and OH ions with “very low” concentrations of Q S04 and Cl. The pH of the pore solution is in the range 13-14, depending upon the alkali level. If no alludi is present then pH=12.5 [71]. Alkalis accelerate the hydration process through the alteration of the CqA-C~Hz-CH system [105]. The main action of the alkalis is to increase ettringite volubility and decrease lime volubil- ity [79]. This is confirmed by Wang et al [315] for example: The molarity of pore solution when low-alkali cements are used is approximately 0.2M with respect to alkalis; normal cements pro- duce a molarity of greater than 0.7M. The pH in the pore solution might nmch 12.9 within min- utes and 13.7 or higher after 28 days. Gypsum does not affect this pH value. The OH concentmtion due to alkali is on the order of 15 times greater than the concentration of saturated calcium hydroxide solution. Therefore, the presence of alkalis decreases the solubililty of CH and accelerates the formation of ettringite; these findings are supported by similar conclusions of Daerr [79]. Experiments also show that ettringite forms even at a pH of 13.3. An alternative mechanism for ettringite formation in alkali solution is given by Chino& Kawa- mura [56]. They hypothesize that in the presence ofNaOH, ettringite formation is suppressed be- cause S042- is more stable in NaX30d than in ettringite. The retardation of ettringite formation twults in an acceleration of early hydmtion of C3A. Through the reaction of NaOH, ettringite is decomposed as: 3Ca0.Al@[email protected] lHzO + 6NaOH + 3NazSOq + 4ca(OH)z + 2A1(OH)3 Way and Shayan [316] looked at the hydration of Portland cement in 0.5 and 1.OMsodium hy- droxide solution. They found that the nature and appearance of all hydrates were the same as in water. The alldi did accelerate the time for depletion of sulphate ions by the formation of et- tringite. In alkaline solution at 30”C (0.08M NaOH), Wang et al [315] found that the stability of ettringite is the same as in water. However, at higher temperatures (60”C, 0.08M NaOH) ettringite paxtially decomposes to the low sulphate form. Both phases can exist for long exposure times under these conditions. Higher alkali concentrations (0.2 to 1.OM)decompose the ettringite very quickly, while the stability of monosulphate hydrate rises in the more concentrated alkaline solutions [104].

3.7 The Roie of C02 and Carbonates

Grounds et al [122] reported research where samples of lab-manufiwtured ettringite were stored in sealed containers at 25-95°C and 100% relative humidity (over water). At all tempemtures et- tringite showed complete decomposition to other phases. The time of decomposition was 55 days for 25 and 50”C storage, and 15-20 days for 75 and 95°C storage. The decomposition was attrib- uted to the action of vey small quantities of C02 in the system. The decomposition products chapter3 Page 26

were identified as (a) gypsum, (b) calcite, (c) aluminium hydroxide, and (d) water according to the equation:

[email protected]~04.32H20+3 ~04.2H20 + Scacos + 2A1(OH)3+ 23H20 Ghomb et al [106] found that in real pastes and in the presence of C02, ettringite decomposes to gypsum, bayerite and aragonite. Carbonation also produces a much more porous microstructure [106].

3.7.1 Formation of Thaumasite

Despite the fact that Grounds tested an “ideal” system, his research indicates the potential for carbon dioxide to alter ettringite. In practice, carbon dioxide has a role in the formation of “thau- masite” from ettringite. Both ettringite and thaumasite formation have been found to be the cause of damage to plaster [179]. It is postulated that the formation of thaumasite involves expansion similar to that due to the for- mation of ettringite [142, 173]. The formation of ettringite usually precedes that of thaumasite, but both compounds can exist together [68]. Both ettringite and thaumasite are produced more rapidly at cold ambient tempemtures (O-lO°C). Their chemical formulae are [74]: ~ Ettringite: {Ca6[Al(OH)6]z.24Hz0}[(SOA)s.2H20]

e Thaumasite: {Ca~[Si(OH)@24HzO)[(SO@][(COs)z] The similarity of the two chemical formulae reflects the fact that both thaumasite and ettringite are AFt phases with similar crystal structures (see Section 3.3) [203]. Hunter noted that thau- masite forms a solid-solution series with ettringite. At temperatures greater than 15°C ettringite probably forms first and then later converts to thaumasite [142]. Taylor [302] stated that the combination of sulphate attack and carbonation (which occurs often in practice) can result in the formation of thaumasite; this can cause severe softening or cracking. “It can easily be misidentified as ettringite”. Thaumasite’s formation requires a constant high rela- tive humidity in a cold environment (4”C), an adequate supply of S042- and C032- and the pres- ence of reactive aluminate. Thaumasite does not form directly, but requires ettringite to form first which acts as a nucleating agent. Whereas the quantity of ettringite formed is limited by the avail- able Alz03, the amount of thaumasite formed is only limited by the lime and silica contents, since alumina is not part of the compound. A continuous source of sulphate ions would allow thaumasite formation to effixtively destroy a mortar or concrete through a softening reaction. Ludwig &Mehr[193] examined deterioration of historical buildings. They found that at tempera- tures between 2 to 40”C, damage was due to ettringite formation. However, at later ages and at tempemtures less than 20°C, ettringite decomposed to thaumasite in the presence of unreacted gypsum and calcite. No evidence was gathered that the formation of thaumasite caused expan- sion. Ettringite and thaumasite form a mixture, but do not participate in a solid solution. Sylla [297] noted that thaumasite may have a role in influencing expansion after heat treatment. He stated that when conditions are right then both thaumasite and ettringite can form with time af- ter heat treatment. For example: a paste sample was heat-treated at 80°C for 5 hours, then stored at 5°C with C02 present; analysis showed a mixture of ettringite and thaumasite present with a high proportion of thaumasite. Sylla believes that both thaumasite and ettringite formation after heat treatment can cause damage; the sulphate is supplied from where it is weakly bound to the calcium silicate hydrate during heat treatment. Chapk?r3 Page 27

3.7.2 Formation of Carboaluminates

The addition of limestone to cement, or the presence of limestone aggregate in concrete, can re- sult in the formation of carboaluminates as early as 7 days after the start of hydration. In a system with Portland cement clinker, 2°/0gypsum and 6°/0limestone calcium carbonate reacts with C3A to form calcium carboaluminate hydrate. There appears to be both a high and low form of car- boahuninate, similar to a high and low form of sulphoaluminate. Although carboaluminates form, ettringite formation proceeds normally in the presence of limestone. Klemm and Adams [163] outlined three fiumtions that limestone fhlfills when it is present in con- crete: (1) it acts as a micro-filler, thus improving strength; it may also act as a nucleation agent to accelerate cement hydration; (2) carbonates can accelerate C3S hydmtion and the carbonate can be incorporated into the C-S-H phase; (3) With fme limestone the principal intemction is to react with aluminates and ferntes to form calcium carboahuninate (femite) hydmtes. This interaction can contribute as much as 10% to the compressive strength between 2 and 28 days. The carboaluminates also form AFt and AFm phases. These are: ~ AFt: [email protected]+ needles

= AFm: CsA.CaCOs. 11H20+ hexagonal plates The AFt phase is much less stable than the Al?m phase, and thus is unlikely to form in cement paste. When carbonate is present it slowly dissolves and will react with any monosulphate pre- sent to form the carboaluminate which is the more stable phase because it has a lower volubility product than monosulphate. The chemical formula is: 3C3ACaS04. 12Hz0 + 2CaCOJ + 18Hz0 + 2C3A.@C03. 11H20 + CsA.3Cas04#32Hz0 In other words, monosulphate decomposes in the presence of carbonate and water to form both a carboaluminate iiwl ettringite.Both the carboaluminate and the ettringite are very stable and can co-exist. This is a slow process; the presence of limestone (either in the aggregate or due to car- bonation) can ~sult in a later-age ettringite formation. The tests that were performed by Klemm and Adams with 5’XOlimestone addition showed the start of carboahuninate formation at an age of 91 days. Note that the sulphate ion can also react with monocarboaluminate, as well as mono- sulphoaluminate, to form ettringite. The carbonate ion will not react directly with ettringite be- cause the result would be more soluble (less stable) products. Kuzel andStrohbauch[172] note that expansive damage can occur due to ettringite formation as a result of “carbonation mctions of the calcium aluminate hydrates crystallizing in the laminar mode”. Also, if C02 is present in an environment where pH is greater than 12, the C4A$JOs.aq groups in the interlayer of the monosulphate are displaced by C032- ions. This reaction leads to the formation of a hemi-carbonate (C4A.~zC@@ and gypsum. The gypsum then reacts with the unchanged fmction of monosulphate to form ettringite. In a strong alkaline medium the hemi- carbonate reacts as follows: C4A.lACOz.aq -+ monocarbonate + CaCOJ + A1(OH)3. Although this process of reaction is explained differently, the formation of ettringite from the re- action between monosulphate and carbonate is essentially the same as proposed by Klemm and Adams [163], Calcite has one other fhnction. The reaction between C3A and gypsum to form ettringite, and the conversion of ettringite to monosulphate are accelerated by the addition of CaC03. The calcite re- acts with aluminate to produce carboahuninates. Under certain conditions all of ettringite, mono- sulphate and carboaluminate can co-exist. As well as awboaluminate, a solid solution of chapter3 Page 28

hexagonal calcium aluminate hydrate, monocarboaluminate and sulphoaluminate hydrates may also exist [265].

3.8 Formation of Chioroaiuminates

Richartz [270] stated that chloride can combine with all phases in a cement clinker. In solutions where the chloride concentrations are up to 10g Cl- litre, C3A and aluminoferrite can form Friedel’s salt, which has the formula 3Ca0,A1203CaC12, 10H2O (a monochloroalurninate), At higher Cl- concentrations, a tri-chloroahuninate phase that resembles ettringite can form; the for- mula is: 3Ca0.A120s.CaC12 .32H20, If gypsum is also present, Richartz notes that at normal temperatures ettringite always forms fimt; the chloride containing compounds form afler the gypsum is consumed. At elevated temperature (40-80”C) Friedel’s salt forms instead of either ettringite or the chloroaluminate phase, even when gypsum is present. Friedel’s salt is shown to be stable in water and in saturated CH up to 90°c. This analysis [270] suggests that heat treatment may be beneficial for reinforced concrete where substantial amounts of chloride are present. Friedel’s salt is preferred over ettringite formation and therefore the chloride is tied up and unavailable for the corrosion reaction. Note, however, that if Friedel’s salt is present at 20”C, sulphate acts on it to form ettringite while the chloride goes into solution; this would be a slow process that could result in later expansions, Richartz’ analysis is supported by Worthington et al [324]. They found that chloroah.uninates do not form at the expense of ettringite, but do form preferentially over monosulphate. Ettringite may reappear due to the breakdown of monosuiphate to form chloroaluminate, Since the amount of chloroaluminate formed is related to the alumina content, blended cements can form more than normal cements. Chloroaluminates appear to be stable; the trichloroaluminate is the more stable form at lower temperatures [283], If Cl- ions are present, chloroaluminates can form from the fmt stages of hydration. Chlorides promote the formation of a porous C-S-H, This effect subsequently promotes the penetration of magnesium ions (if present) into the hydration product and “conversion of C-S-H to the non-hy- draulic M-S-H” may occur [102], When exposedto seawater,Type I cements showboth chloroaluminateand ettringiteformation. At early ages these products deposit in the voidsand exert little expansion [158],If temperatures are low the presence ofNaCl causesettringiteto decomposeto form Friedel’ssalt. At lowertem- peratures in a chloride environmentthe Cl ion can be incorporatedinto the ettringite structure; this results in a “less-expansivephase” [243-245]

3.9 Morphology of Ettringite

The controversy over whether ettringite forms by through-solution mechanism [220] or by topo- chemical means [222] has also resulted in two hypotheses concerning the morphology of et- tringite in the hardened cement-paste matrix, Mehta [209] noted that ettringite has two types of morphology, depending upon the conditions under which it forms: = Type I: are large lath-like c~stals 10-100pm long and several pm thick. These form in solution under condhions of low hydroxyl ion concentmtion. These crystals are not ex- pansive. These are the type of crystals that have been produced in numerous lab-based re- Chapk?r3 Page 29

search projects and of which thousands of scanning electron microgmphs have been taken. = Type II: are small rods l-2~m long and 0.1-0.2km thick. These form when the hydroxyl ion concentration is high (i.e. the normal situation for concrete). Ettringite formed with this structure can be a source of strength, or can cause disruption when, in a confined space, agglomemtions of these crystals absorb water and expand. It is also possible to prepare this type of ettringite in the laboratory, but special techniques are required [216]. Mehta [207] notes that in concretes the calcium-sulphate hydrates form in a supersaturated envi- ronment. Under these circumstances the cystal structure of both monosulphate and ettringite are extremely fine-gmined. The ettringite crystals are short prisms with a thickness:length ratio of 1:3. This does not mean that the larger crystals do not form, because there is ample evidence of their occurrence in concrete. However, it is conjectured that Type I morphology only appears when sufficient space is available, such as in high w/c pastes, in large cavities, or during early hydra- tion. In more restricted spaces, Type II ettringite occurs as “short prismatic crystals”; under these conditions ettringite appears to be able to form “anywhere and everywhere” in the system [213]. As will be shown below, the presence of the smaller “colloidal” size ettringite provides the cor- nerstone of a hypothesis concerning the expansion mechanism. Accordingly, in support of the Type II morphology, Mehta [205] noted that the presence of lime in Portland cement pastes and concretes affects the formation of ettringite such that it has a colloidal morphology, and does not form as long lath-like crystals. The colloidal nature of the ettringite results in the attraction of large numbem of water molecules which cause interpaxticle repulsion and expansion [213].

3.10 Mechanism of Expansion Associated with Ettringite Formation

As mentioned above, there is controvemy surrounding the fundamental mechanisms associated with the formation of ettringite and subsequent expansion. Cohen [60] has categorized the hy- potheses into two schools — crystal growth and swelling.

3.10.1 Crystal Growth

In the “crystal growth” school expansion is caused by the formation of ettringite at the surface of reactant grains. The growth of this inner layer pushes other particles out and thus causes expan- sion.The mode of formation also appears to depend upon whether CH is present or not — if it is not present then c~stal growth is topochemical, if CH is absent then ettringite forms from solu- tion and does not produce expansion. Research on sulpho-aluminate based expansive cements (Type K) [61] shows that expansion is caused by growth of ettringite crystals on the surface of expanding particles; a topochemical IWW- tion between the particle and the surrounding solution is responsible. The magnitude of expan- sion does not depend upon the quantity of ettringite formed but on the size of crystals produced. A few long crystals at a small number of sites can produce large expansions and microcracks. Al- ternatively, formation of small crystals at a large number of sites can produce a network of smaller microcracks Chatterji and Jeffery [52] have hypothesizedthat expansiondue to the presence of sulphates is caused by the solid state conversion of C4AHI3 to monosulphate hydrate in the presence of CH. This hypothesis has been researched in detail by others. Meh@ for example, was not able to re- produce Chatterji’s results and could not detect the presence of C4AH13 [208]. It is now com- chapter 3 Page 30

monly held that expansion is due to the formation of ettringite and not monosulphate. Even if Chatterji’s theory about monosulphate formation was correct, Mehta [207] has shown that mono- sulphate formation in a confined space should not result in expansion because the plate-like monosulphate crystals can orient themselves like “leaves of a book”. On the other hand, ettringite crystals do not re-orient and therefore are more likely to be able to exert pressure. The hypothesis that pressure can be exerted from individual crystals is contrary to the findings of a number of researchers. Dron et al [88] noted that ettringite’s expansive forces can only develop in a confined space. A uniaxial force of only 1 MPa is sufficient to inhibit expansion in that direc- tion; ettringite will then proceed to grow in another direction. The destructiveness of ettringite formation depends upon where in the matrix it forms. If it forms near an aggregate particle then destruction is maximum. If it forms near a cavity, then the cavity can act as an expansion cham- ber to relieve potential stress build-up. Very recently, Ping & Beaudoin [256,257] postulated that formation of ettringite is by two proc- esses, nucleation and crystal growth, The production of “crystallization pressures” requires a con- fined space in which the cxystals grow, The lack of a confined space explains why ettringite growth does not cause pressures during early stages of hydration, Most of the ettringite may form before expansion occurs; it is only the formation in conflmed space that produces expansion. In hot, Beaudoin & Ping note that any solid product will produce crystallization pressure ifl = crystalgrowth is in a confined space = the volubilityproduct ratio is greater than 1,0(i.e. the productsare less soluble than the reactants)

3.10.2 Swelling

Cohen [62] summarizesthe “swellingschool”of thought: ettringiteformsby a through-solution mechanism,In a saturated CH environmentettringitecrystalsare gel-like and colloidal in size, The high surfhcearea results in adsorptionof significantquantitiesof waterand strongswelling pressuresdevelop, The theory was f~st proposedby Mehta [215,217]; the expansionmechanismfor ettringite is similar to that of some clays which expandbecauseof “polar orientedelectrostaticattraction of water molecules”.Small ettringitecrystals of colloidal size, formed in an alkaline environment possess a negative surfwe charge,The high surfhcearea combinedwiththe negativecharge re- sult in an attraction for large quantitiesof water.Tests on disksof pure ettringiteconfirmthe hy- pothesis;the disksadsorba large quantityof waterand large swellingstrainsoccur, A check of the XRD pattern of the ettringitebeforeand after swellingshowsno substantialchange. Early fundamental research by Seligmann & Greening [285], as well as other research [295] tend to support Mehta. Seligman & Greening found that when C3A pastes hydrate they undergo large volume increases due to imbibation of large quantities of water. As an example, pastes were pre- pared with WA mass ratios from 1/8 to 1.0. The water/cement ratio was 0.4 to provide just suffi- cient water for ettringite formation. Nevertheless, during soaking for 180 hours after casting, much larger quantities of water were taken up and the fresh pastes showed as much as 57°/0ex- pansion. When sodium or potassium hydroxide was added, all reactions accelemted; there were no observable changes in the nature of the reaction products. Sagrera’s work [273] is also supportive, He mixed paste cylinders containing either calcium sul- phate or sodium sulphate additions. Volume change was measured and the quantity of ettringite in the pastes was determined by X-ray diffraction. Figure 3.5 clearly shows the positive correla- tion that exists between ettringite content and 0/0 volume increase. Chapter 3 Page 31

180 ■

160 +

VOLUME 100 ❑ %n INCREASE ❑ % 80 9 I ■ 60 •1 ❑ ❑ ■ ■ ■ 40 u / 20 t ■ 1 1 1 1 I 1 1 1 o 1 0 50 100 150 200 250 300 350 400 ETTRINGITE CONTENT (arbitrary units)

Figure 3.5 Correlation Between Volume Increase and Ettringite Content [data from ref. 273]

However, also note that in Figure 3.5a significant amount of ettringite forms before there is any volume hwrease-Le. there is an X-axis intercept that is greater than zero. Sagrega’s results could aeeanmodate both the swelling theory and the “crystal growth in contlned space” theory as de- seribed by Ping & Beaudoin [256,257]. Like many other observed processes that occur within concrete, there is undoubtedly no single meehardsm that determines behaviour. Both mechanisms — swelling and crystal-growth — are active; environmental conditions at any particular time determine which, if any, is predominant. Mather [199], in tis discussion of Cohen’s work [62], summarizes the issue from an engineering viewpoint. He argues that the details of the formation of ettringite are of less importance than the fundamental meehanism involved in the process. The fact that the morphology of ettringite is crys- talline m not is “of minor importance”. The important factor is that the products attempt to oc- cupy a huger volume than is available for their formation. Mather quotes from ACI Committee 223 that “thermodynamics shows that there exists a very large amount of free energy, about 55 kcal/mol (formed from C3A) which equals about 169,000 ft-lb of work”. Thus, ettringite forma- tion in a eonllned space k capable of producing very high forces. Clearly, these forces cannot de- velop unless there is a rigid framework to resist the forces. Mather continues: the mechanism of development of expansive forces involves the alteration, in situ, of a quantity of aluminate reaetant to form ettringite “by a process that delivers water and other xe.actantsin solution to the naction she, and leaves elsewhere in the structure some space formerly oecupkd by solution that R no longer filled with solution”. From a review of crystal growth theory, Mather concludes that “when you put more than one unit volume of anything in a restrained three-dimensional space having, hitiall y, a size of one unit volume something has to chapter3 Page 32

give; either the quantity added must reduce itself in amount (some leaks out) or in volume (it is compressed) or the space enlarges either elastically or inelastically without rupture or it crocks”. The most important fhct is that the volume that ettringite forms per unit volume of C3A used is 8:1; “the f=t that the ettringite is crystalline or colloidal is irrelevant”,

3,11 Mechanism of Secondary Ettringite Formation

Ettringite needs sulphateto be in solutionto form. Its formationduring sulphateattack is easily explained,since sulphateionspenetratethe concretemicrostructurefrom outside.However,in

I IYI’A TRACES I

CSS with 24% gypsum belore hydration

As IlbOV~but wtth a further 1,S% RYPSUmadded whereCSAadded

100 m TEMPERATU~ “C J

Figure 3.6 Formation of Ettringite from Sulphates Supplied by C-S-H DTA results taken directly from Odler [240] and annotated

, ,, chapter3 Page 33

this repoti we are considering mechanisms by which ettringite forms at later ages without assis- tance from outside sources. The key to explaining this mechanism lies in the research of Lerch [180, 181] who found that dur- ing hydmtion some sulphate goes “missing” and form “Phase X“ (Fig. 3.3). By 1980, these ideas had been clarified by Odler [240]: Odler examined the bonding of sulphate within C3S pastes in order to confkm the initial findings of Copeland et al [72] that C-S-H can bind and then release sulphates. Odler’s DTA results are summarized in Figure 3.6. An alite cement with 2.5% gypsum was hydrated for 28 days, ground to cement freeness and then blended with C3A in a molar mtio C3A/CaS04=l :3. The new cement was mixed with additional water to form a paste and allowed to hydrate. DTA and X-ray diffraction analyses were performed at various ages after drying at 35”C. The experiment was performed for alite cements with various sulphate additions and simi- lar results to those shown in Figure 3.6 were obtained. X-ray diffraction analysis confirmed the DTA results. The formation of the ettringite peak during the second hydmtion period and the absence of a gyp- sum peak after the start of hydration confirm that the sulphate is bound within the C-S-H struc- ture. The “gypsum is bound by the C3S hydration products in a way that makes it undetectable by the [DTA and X-ray] methods employed”. The maximum amount of bound gypsum corresponds to 9.8 g S03 for 100 g of CSS. Another test showed that the sulphate in the C-S-His readily extracted by satumted calcium hy- droxide solution. Thus, when a source of aluminate is added to the ground paste containing C-S- H with bound gypsum, the ettringite peak quickly forms and reaches its final intensity within 3 hours after hydration. Lachowski’s findings are in support of Odler [174]. He examined the composition of C-S-H for pastes hydrated up to 900 days and concluded that overtime S042- ions leave the C-S-Has the pastes age. Most of this loss, however, occurs during the first 28 days. . In Neck’s research [237], concrete was manufactured with 400 kg/mj of PZ55 high-early- strength German cement, at a w/c=O.36. Strength tests at 7 days gave 60-70 MPa and 70-80 MPa at 28 days. Behaviour after 2 heat treatments was compared: In treatment “A” there was no pre- cure, a temperature rise to 80”C at 60°C/hour and 80”C was held for 4 hours; in treatment “B”, 3 hours pre-cure was used followed by a 20°C/hr rise to 65”C, which was held for 2 hours. For treatment A Neck postulated that some of the calcium sulphate remains mobile and is available for secondary phase formation when moisture is introduced. Free sulphate is a requirement for the expansion process, while moisture must be present to dissolve and transport the sulphate. Also, “crevices and disturbances” must exist for pathways for water and transport of ions. These also provide space for crystallization of the secondmy phases. Wicker etal[319,320] summarize the entire process nicely. During heat treatment the thermal de- composition of “primary” ettringite results in an increase of sulphate ion in solution and a de- crease in OH- ion concentration. Figure 3.7 shows the sulphate ion concentrations in the pore solution immediately after heat treatment for pastes with various WA ratios and alkali contents. On the other hand, after heat treatment the pore solution changes with time. Figure 3.8 shows that the S042- concentration reduces steadily over 90 days for both a 60”C treated and 90°C treated paste. Wicker et al propose that during heat treatment the ettringite formed during early stages of hydra- tion is decomposed. The process is accelerated by akalis. In the temperature mnge 50-80”C the re- versible formula is: chp16Jr 3 Page 34

■ 20 ❑ 50 ❑ 60 ■ 70 ❑ 80 ❑ 90 ❑ 100 I 1 500 ITREATMENT TEMPERATURE I

400

CWCEN- 300 TRATlON

200

100

SIA4,6 SIA=O.68 SfA=O.63 SIA=O.66 SIA=O.86 SIABI.06 SIA=O.92 ALK=O.62 ALK-O.68 ALK-O.87 ALK=O.66 ALK=l.03 ALK=l.13 ALK=l.24 Figure 3.7 Changes of Sulphate Ion Concentration with Heat Treatment S/A= S03/A1203 Ratio; ALK=% equlv. alkali (data from ref. 319)

900 60 C TREATMENT 800 .—------—--- 700 90 C TREATMENT / /“ .- 600 .“ /- ““””G/“ CONCEN- m TRATION 1 In?noln 400

300

200

100 4 60 C TREATMENT

i~ o I I

19h 26d 90CI STORAGE TIME AT 20C AFrER HEAT TREATMENT Figure 3.8 Change in S04 and OH Ion Concentration During Storage (data from ref. 320) Ckpk!Y3 Puge3s

At temperatures greater than 70*C monosulphate also decomposes according to the mwersible for- muhx

3Ca0.AlZ03.CaS04. 12Hz0 + 2NaOH 7k~~O~+ 3CaO”AlzOs”6Hz0+ NazSOq + Ca(OH)2

3.12 Comment

‘l’heAFm phase in Concrete is very complex both chemically and morphologically. One of its astonishing features is the wide variety of ions that can substitute in ettringite for Al and sulphate ions, while the basic crystal morphology remains the same. ‘llw complexity makes ettringite very useful for a wide variety of industrial applications, but it also causes problems when one wishes to determine the influence of the trisulphate phase on the engineering properties of concrete. From an engineering viewpoint one important research finding is that ettringite does not appear to be stable in concrete at temperatures above approximately 60-70”C. When pastes, mortars or con- cretes are cured at elevated temperature ettringite disappears and some of the sulphate goes “miss- ing”. Various ideas have been put forward but the most accepted view is that at elevated temperature sulphate can easily be incorporated into the C-S-H gel structure. Various solid solu- tions may form. Further transformation of ettringite to monosulphate may also play a key role in thiS regard

Whatever the exact mechanism, it appears that a sufficiently high heat treatment nxults in the sul- phate being unusually bound. ‘l%ebond is such as to allow a later slow release of sulphate ion into the pore solution and combination with aluminates to produce ettringite. It also appears that other compounds such as carbonates, chlorides and carboaluminates may influence the formation of ex- pansive compounds at later ages. The mechanism by which ettringite produces stms and strain is also not precisely knowIL Mather’s viewpoint is therefore preferred; from the practical viewpoint we are interested in know- ing under what conditions compounds form which try to occupy more space than is available within the pore structure of the material. chapter4 Page 36

Chapter 4 Secondary Ettringite Deposition

4.1 The Importance of Porosity to Ettringite Formation and Expansion

Post-mortem studies of deteriorated concretes tell us a considerable amount about how ettringite and other crystalline materials form. Many studies show crystalline deposits within the capillary and air-entrained pore structure. For example, Gillott andRitchie[112] examined cement pastes that were 50-70 years old. They observed various crystals forming in the small voids in the pastes. These were identified as: (1) thin plates, frequently extending from one side of the void to the other, which were identified as CH crystals; (2) radiating clusters of needles or spherulites; these were identified as ettringite crystals. Murat [233] examined the surface of asbestos fibre-reinforced cement paste used to manufacture hot-pressed pasteboard. He discovered a number of large voids that were filled with “flower-like” crystallization product. Mumt noted that calcium sulphate becomes concentmted in voids; “ettringite grows in flowers from a seed”. Specimens that had undergone freeze/thaw attack show crystals of portlandite and ettringite in the air voids [236]. It was hypothesized that a reduction in the lime concentmtion in solution (which can be brought about by a number of different effects) results in an increase in the volubil- ity of hydrated calcium aluminates. The fkezing action in the pore structure results in pore solu- tion being expelled into the voids where conditions are right for ettringite to form. Formation of ettringite in the pores does not result in expansive pressures. St. John [293] noted that ettringite crystals are often found in the pores and crocks of concrete that has deteriorated. These crystals are deposited in the voids as a result of movement of fluids through the pore structure. The presence of ettringite in this form does not appear to be harmfhl to the concrete. However, there are “sub-microscopic particles of sulphoaluminates dispersed through the concrete” that may be “remobilized into solution” as the concrete deteriorates. Re- crystallization can then occur “as visible needles at the nearest interfile”. This phenomenon has been observed even in young concretes that have cracked due to overstress and have fluid flow- ing through the pore structure. Observations of ettringite in the larger voids and cracks, including air-entraining voids, appear commonplace. At the same time, it seems that such formations are not directly responsible for “secon(kuy ettringite” attack. It maybe just the reverse: ettringite within these spaces maybe in- dicative of a process that might have led to expansion and cracking, but which was relieved through the transport of the reactants to “free space” where the products form in a stress-flee en- vironment. Conversely, micro-porosity, concentration of key compounds, microcracks and a weakened microstructure around the paste/aggregate and paste/steel interface appear to be nuclea- tion sites where secondary ettringite can grow and produce damage; this is the subject of the next section.

4.2 The Importance of the Aggregate/Paste and Steel/Paste Interface

At the heart of the issue with respect to ~ of secondary ettringite formation is the aggre- gate/paste-matrix interface, normally called the “transition zone”. Most petrographic examina- Chapk?r4 Page 37

tions have shown that if one wishes to find ettringite in a deteriorated concrete, the first place one should look is at the transition zone. This section will explain the reasons for this. A macroscopic indication of the existence of the transition zone in some concretes can be ob- tained simply by visual observation. For example, Hoshino [140] observed that the porosity of paste near the lower boundary of coarse aggregate is higher than elsewhere in the matrix. At the lower layer the boundary was “compamtively whitish” and contained a large number of CH crys- tals. At the upper boundary the amount of CH and the crystal size was much smaller. Using a more sophisticated procedure, Kayyali [159] noted that if one compares the porosity of concrete, aggregate and cement paste by mercury porosimetry, one finds a discrepancy in the porosities. This discrepancy indicates the existence of a higher porosity, or “interfiwial layef’ at the surface “ of the aggregates. The early fimdamental work of Hadley [123] and, later, Barnes [17] produced a clear picture of the transition zone which has since been substantiated by a large volume of research. In his PhD thesis, Hadley outlined the steps in the formation of the paste/aggregate interface after consolidation of the concrete there exists a “considerable volume” of water-filled space immediately around the aggregate Very quickly a film of lime, l/4pm thick deposits directly onto the aggregate surfaces. This lime is “perfectly oriented”, with the c-axes normal to tr,e aggregate surface a layer of C-S-H deposits on top of the lime film after 4-8 hours “large and randomly-oriented” lime crystals form in the “interracial void”. These crystals grow to encapsulate hydrating cement grains. after 1-3 days “ettringite needles” and “booklets” of “large platy hexagonal crystals” also deposit in the interfhcial region, The “booklets” were confirmed not to be lime or mono- sulphate; instead, Hadley postulates that they are “a complex solid solution of C% Si02, Alz03 and sulphur. Observation of fracture surfaces shows that fracture occurs mainly through the interlayer void space. As the concrete ages, fracture occurs through the lime portion of the interfuial film, or through aggregate particles if weak planes are evident. Fracture never occurs at the true paste-aggregate interfme. Hadley noted that some degree of orientation of CH occurs in the re- gion up to 150-200wn from the interface. Figures 4. la and 4. lb show Hadley’s schematic of the interracial region at 30 minutes and 3 days after the start of hydration. The research of Monteiro and co-workers [225, 226,229, 84] has provided valuable information about the transition zone. The aggregate/paste-matrix interfhce is a weak zone in the concrete with higher porosity than the bulk matrix [226]. Calcium hydroxide is deposited with preferred orientation in the region within 30~m from the surfiwe. Both coarse aggregate and sand groins show a calcium hydroxide coating with the c-axis perpendicular to the aggregate surface. The et- tnngite content in the transition zone is approximately 2.5 times that in the bulk matrix at lpm from the surfhce and only reaches the bulk matrix levels 20Lm from the surface. The thickness of the transition zone is larger for larger aggregates. Monteiro et al [225] have examined the transition zone at the steel-paste interhce. This zone is very similar to that occurring at the paste-aggregate interface. The CH that deposits in the transi- tion zone has preferred orientation; further tests show that pronouncedorientationstill existsafter 420 days in a 50pm region near the interface.The transitionzone has high porosity and the crys- tals are largerthan in the bulk matrix.There is a higherconcentrationof ettringitecrystals in the transitionzone, but the ettringitecrystalshave no preferentialorientation(Figure4.2). The domi- nant effect as aging occurs is the densification of sttucture behind the CH film. This densification chapter4 Page 38

I CEMENT GRAIN ON JNTERFACE AT 30 MIN. 1

/ Smail ‘%picule#’ I

I Interface I

INTERFACE AT 3 DAYS C-S-H

RI

Randomly \ Oriented Y Massive CH “~.. .3 * crysta@ ■ ● ‘ +4 nEncapsulating “ S \

,.- ,. ● r .0 . ● . : . . . . “.. . \~~ . . .“ m!liw

I I

Figure 4.1 The Paste/Aggregate Interface at 30 Minutes and 3 Days (drawings scanned from Hadiey [123] and annotated) chapter4 Page 39

180 ~ Alldata for Plain, 5% Silica Fume and 16% Silica Fume Pastes

180 t . 1 DAY i ❑ 7 DAYS

— A 30 DAYS .A I X-RAY 100 ~ ¤~A ~ INTENSITY t c

o Ul, ❑ ❑ o . OWOA @ ❑ Aw m 20 ❑B A A t I 1 1 1 1 1 1 I o 1 1 0 5 10 15 20 25 30 35 40 DISTANCE FROM THE STEEL INTERFACE, pm

Figure 4.2 Variation in Ettringite Peak Intensity from Interface (data from ref. 225)

process is assisted by the presence of silica lime. The steel/paste transition zone is not signifi- cantly different than the aggregate/paste transition zone. In 1988 the issue of whether CH was truly orientated in the transition zone was clarified. Det- wiler and Monteiro [84] addressed the criticism that the standard X-ray diffraction technique to determine CH orientation near the aggregate interfhce was inaccurate; the argument was that dur- ing analysis crystals which are not oriented with either a 0001 or T(IT1plane am missed. Instead, the researchers used a pole-figure goniometer which eliminates this presumed bias. They found essentially the same results as standard X-ray analysis— namely, that CH is oriented near aggre- gate surfizes. The type of aggregate at the interface, however, does have an important role to play. The transi- tion zone near carbonate aggregate is different than that for other aggregates, because a “basic cal- cium carbonate hydrate” forms which substitutes for the large and highly oriented CH crystals [229]. The carbonate-hydrate has smaller crystals and therefore there is a strengthening of the transition zone, The formation of this material occured with both opc and C3S pastes; carboalumi- nates in the transition zone do not appear, therefore, to be the reason for the increase in mechani- cal strength. The superiority of limestone aggregate in this regard is confirmed by Saito and Kawamura [275]. They examined the transition zone for ordinary cement pastes cast against two types of polished aggregate-limestone and granite. Fly ash and slag paste/aggregate interfaces were also exam- ined. The researchers found that both ash and slag substantially Educe the formation of CH in the interracial zone, but their analysis of plain pastes at an age of 36 days also showed some inter- chapter4 Page 40

LIMESTONE AGGREGATE/PASTE INTERFACE, AGE=36 DAYS

45 3,5

40 3 35

X-RAY 30 ‘\ 2.5 PEAK -\ HEIGHT 25 ORIENTATION “ FOR p INDEXFOR ETTRINGITE 20 CALCIUM AND “’...... - HYDROXIDE CARBOALUMINATE15 ‘\ 1,5 10 \ Y ETTRINQITE 1 5 ‘k mCARBOALUMINATE 0 Laf 0.5 1 10 100 DISTANCE FROM INTERFACE (yin) (log 8oale)

GRANITE AGGREGATE PASTE INTERFACE, AGE=36 DAYS

45 3,5

40 T + 3 35 t X-RAYPEAK30 2.5 HEIGHTFOR ORIENTATION ETTRINGITE25 INDEXFOR 2 CALCIUM Ca/f%TIO 20 HYDROXIDE 15 1.5

10 1 5 MONOSULPHOALUMINATE II 1 L , I 0 1 \l ! 0.5 1 10 100 DISTANCE FROM INTERFACE (pm)(logscale)

Figure 4.3 Compounds and Orientation in the Transition Zone — Limestone and Granite Aggregate (data from ref. 275)

,,, chapter4 Page 41

esting results (Figures 4.3). The intetiace normally extends from 50 to 100pm into the matrix. Et- tringite and/or monosulphate exist in high proportions very close to the aggregate surface, but the quantity drops off rapidly and no X-ray peaks for either compound were observed i%rtherthan 5- 10p,mfrom the surface. Both types of aggregate/paste interfaces showed similar results except for the significant finding that carboaluminates appear also to preferentially form near the lime- stone/paste interfhce (see Figure 4.3). In other tests it was found that the presence of fly ash and slag resulted in an @rease of carboaluminate in the interracial zone. La.rbi& Bijen [176] report recent results which fhrther help to establish that calcium hydroxide orientation occurs. The orientation was revealed through an Orientation index which, if crystals are not oriented should be unity. Figure 4.4 shows test results from X-ray analysis near a paste- polypropylene interfue; polypropylene was used to simulate a flat aggregate surface. The orien- tation index, and thus orientation, is maximum at the interfme; some degree of orientation occurs up to 80pm ffom the aggregate surface. The introduction of fly ash has a pronounced effect in re- ducing the thickness of the transition zone from 60#m to 15~ at one month. As shown in Figure 4.4, the degree of orientation of calcium hydroxide is also significantly affected. Figure 4.4 also illustrates the beneficial effects of fly ash; CH orientation produces weak planes: if orientation can be reduced then the transition zone is strengthened. Saito & Kawamura (above) noted similar improvements [275]. The use of pozzolanas is one method to improve the transition zone. Goldman & Bentur[113], for example, illustrated the beneficial effects of silica fume. The transition zone at the aggregate/matrix interface is a highly porous ~gion. The solids contained in this zone have a heterogeneous microstructure. A “rim” of calcium hydroxide is frequently ob- served in contact with the aggregate surfme. The “porous pockets” that are interspersed through- out the region contain a needle-like material that is “probably ettringite”. The interface region is shown in the first two drawings of Figure 4.5. When silica fbme is present the transition zone is

ORIENTATIONOFCALCIUMHYDROXIDECRYSTALS,AGE=28DAYS 6

5 —m— ORDINARYPORTLAND CEMENT(OPC) 4 . OPCwith QUARTZFLOUR

ORIEN ATION OPCwith FLYASH#1 INDEX 3 l=RANDOM OPCwith FLYASH#2

2

1

I 1 1 1 1 I o 1

0 20 40 60 80 100 120 DISTANCEFROMINTERFACE(pm) Figure 4.4 Orientation of Calcium Hydroxide Crystals in the Transition Zone (data from ref. 176) chapter4 Page 42

PC=Portland cemenfi SF=sUica fume; CH=Ca(OH)2; CSH=calcium silicate hvdrate ett=ettrtiite –=– “ I

I

Fresh Concrete Transition zone in Fresh Concrete Filling of transition with fine silica fume mature fume without Fume mne with Ca(OH)2, particles concrete C-S-H, and I ettringlte. Note ~~ porous pockets, some filled with ett

I I

Figure 4.5 Schematic of Changes in the Transition Zone With and Without Silica Fume (drawing scanned from ref. 113 and annotated)

entirely altered (see third and fourth dmwings of Figure 4,5). The interface region is lower in po- rosity, is more homogeneous, and there are no porous pockets or CH rim around the aggregate. Zimbelmann [327] suggested a method by which the aggregate/paste and steel/paste interface can be improved directly. In one series of tests, aggregate is coated with a normal commercial wash- ing agent, The effect is to make the transition zone thinner. Zimbelmann found that the contact zone reduced horn its normal 2-3pm to 0.5- l~m. This resulted in an increase in bond strength of 100-180Y0.In another series of experiments aggregate surfaces were coated with a suspension of pumice, detergent and water glass (sodium silicate); the objective was to replace the contact layer of CH with a layer of calcium silicates. Theoretically there should be a high “physical affinity” between the silicate contact layer and the aggregate and between the silicate contact layer and the , hydrates. Tests showed that this method increased the bond strength by 200-270% at 100 days, and in some cases the bond strength exceeded the tensile strength of the hardened cement paste. chapter4 Page 43

4.3 Comment

There is little doubt that the transition zone in normally-cast concrete is the weak link. It has higher porosity than the bulk paste, it has oriented CH crystals, it contains other crystals of un- known origin, and it is a region where sulphates are concentrated. It is also a region where materials of distinctly different properties meet. The most important dif- ference is the thermal expansion coefficient between aggregate and hardened cement paste. One intuitively expects, therefore, that if heat-treatment of concrete is going to produce damage, the damage is going to occur at the transition zone — behveen aggregate and paste (including both coame and fine aggregate) and between steel and paste, Given the properties of the transition zone, it is therefore not at all surprising that Heinz & Lud- wig [132] conclude that for mortars showing substantial expansion due to secondary ettringite formatio~ SEM indicates that ettringite forms at the contact zone between paste and aggregate and that this is the “origin of the degradation”. Chauter5 Page 44

Chapter 5 Key Examinations of Secondary Ettringite Formation

5.1 1980 — Ghorab et al [106]

Expansion tests were performed on mortars manufactured with a typical German high-early- strength cement (with 4,0% S03) and a sulphate resistant cement. The heat-curing temperature varied fkom 60- 100”C with a duration from 12 to 72 hours. Expansion tests were performed in ac- cordance with DIN 1164 and strains were measured to an age of 300 days, Mortar bars wem 40x40x 160mm bans made with a water/cement=O.5 and cement/sand=O.33, The experiments indicated no significant late expansions occured for curing temperatures less than 80”C, regardless of the period of heating. For treatments above 80”C expansions started af- ter soaking in water for 70-80 days, Expansion was usually complete after 130 days soaking, ex- cept for the mortar cured at 90”C for 18 hours; this specimen showed continuously increasing expansion over the period 70 to 300 days. In a second series of experiments specimens were subjected to a 1 day delay period followed by 1,5 days at 80”C and then a 6 hour temperature rise to 100”C; cooling followed immediately the specimens reached the maximumtemperature,Figure 5,1 showsthe expansionof bars manufac- tured with various blendsof sulphate-resistant(SR) and high-early-strength(HES) cements, Blends which contained more than 25% HES cement showed substantial expansions, again start- ing after about 70 days after the start of soaking, In all cases, expansions were accompanied by large reductions in resonant frequencies that were measured concomitant with deformation. In a third series of tests the effect of repeated heat treatments was examined using the susceptible PZ55 cement. Set I specimens were heat treated at an early age while Set II specimens were not, Set I showed accelerating expansions after 50 days of soa.ldng; this expansion (approx, 1,2%’0)

CEMENT TYP~ 4 SR= SULPHATERESISTANT ...... r..=z.-.r.*z.=...* H13s= HIGHEARLYSTRENGTH

‘------l~z sR ‘------75%sR + 25%H~ ‘––+--- so%sR + 50XH=

~ 25%SR+ 75%HlN ---- 100ZH= -,-u-.-.-u ,-,-...-,a.-u”-’-’-~-’-””-”-’-”-’-~-”-”-” ..-,,...-..-..-..-..-..r...... —. o 50 100 150 200 250 300 STORAGE PERIOD (DAYS)

Figure 5,1 Expansion of Various Mortars after Heat Treatment, DIN1 164 Test Method (data from ref. 106) chapter5 Page 45

was finished after about 330 days. During this time, Set II specimens showed no appreciable ex- pansion. At 330 days Set I specimens were subjected to a second heat-treatment and Set II speci- mens were given their first heat-treatment. Approximately 40 to 50 days afier the treatment at 330 days Set I specimens started expanding even more and reached a total expansion of 2.5% at about 500 days. Set II specimens also started expanding about the same time (380 days) and reached a maximum expansion of about 1.7’Yo,200 days after their fwst heat-treatment occurred at an age of 330 days. The heat treatment appears to be a crucial component of subsequent expan- sion. Furthermore, it appears that the time at which the heat-treatment is applied is not very im- potint; if anything, expansions are larger when the heat-treatment is delayed. A fourth series of experiments examined the effect of pozzolans on the expansion mechanism. Various cement/pozzolan blends (containing the same amount of S03 asthe PZ55 cement) were used to prepare mortars; these mortars were tested as above. It was found that the pozzokm sub- stantially reduced expansions: e 0°/0pozzolan: 100°/0PZ55 HES cement Expansion = 1.00%

I= 15°/0trass: 85°/0 HES cement Expansion = 0.70!40

o= 15°/0 ash: 85°/0 HES cement Expansion = O.15% @ 30°/0 trass: 70°/0 HES cement Expansion =0.05’% e 30°/0 ash: 70°/0HES cement Expansion = 0.02% The authors suggest that the expansions are due to the formation of stable ettringite with time, de- rived from the monosulphate and solid solutions which contain tetracalcium aluminate hydrate. X-ray analysis clearly shows the development with time of the ettringite X-ray peak in the heat- cured sample, but not in the sample continuously cured at 20°C. Where calcium carbonate is pre- sent or carbonation is significant, Ghorab proposes that these can increase the damage caused by the gradual transformation to ettringite.

5.2 1984 — Research Institute of the Cement Industry [268]

Test disks of cement paste with dimensions 50 mm diameter x 10 mm thick were manufactured with various cements. Normal consistency was used. After various delay periods, specimens were heated over water at 100”C for 2 hours. The cements used had a C3A content of either 7.0 or 13.1’XO,3% S03, and 3500 cm2/g Blaine. In some cements C3A activity was enhanced by add- ing K20. Furthermore, the sulphate type that was used was varied-either gypsum or 1:1 anhy- drite:hemihydrate. Figure 5.2 shows the expansions of the disks measured at a constant time after the start of expo- sure, Clearly, expansion increases as the delay period decreases. Expansion was the largest at the lower C3A content of 7.0%. The addition of potassium oxide did not have a substantial effect on the results. In a parallel series of tests which used a 3 hour pre-storage, low expansions were also observed and were independent of the other parametem. It appears that the extent of prestorage is the critical factor which governs whether expansion will take place. In a second series of tests the behaviour of concrete made ftom cements with higher S03 con- tents was examined.Concreteprisms,4x4x16cm, w/c=O.38,were manufacturedwith one of two clinkers in which sulphatewasadded to give eithera 4’%o or an 8’%S03 content. After a delay pe- riod of either Oor 3 hours, specimens were heat treated at 40 or 80”C and a relative humidity of 60 or 95% for a period of 2 hours. After cooling, specimens were stored in water for 3 years at either 5 or 40”C. chapter5 Page 46

45

40 ■ O HR. DELAY """""""`""""""""""""""`""""""""""""`.n...... --“’””””””r❑ 35 .. ~HR DELAY ,, ., ...... <30 .. .,,,,,-,-...... ,,,,..,.-,,,,-,,,,...,.,,.. -,,.,,,,,.,.,,,...... ”,,,.,..,,,,.-,,,.,,,,,,- ■ 3 HR DELAY E - E 25 ...... ,.,,.,,,,---,,,,”,-,,,,,,,, -.,,,,..,,,,, c ~L.___J “$ 20 z ~ 15 10 5 0

Figure 5.2 Expansion of Paste Disks After Heat Treatment (data from ref. 268) Note that 1Pexpansion was >12 mm/m (1,2%) cracks were observed

Table 5,1 Effect of Heat Treatment on Cracking and Phase Formation (data from ref. 268)

Heat Treatment Temperature 8 9 40 c 80 C S03 Cllnke ‘r& Storage % by Relative Humidity Durtng Heat Treatment o o o (hrs) mass 60% 95 /6 60 /’s 95 /’s I Subsequent Storage Temperature for 3 Years

Dark Shading - designatesotrongcraokformationaooompanladby obviouonew “phaseformation” LightShading - designatesslightcfaok formationwith “insignlfioant’now phase formation No Shading- Indisatssno eubstanflalvisualohanges to prismsduring3 YOWOstorage

The qualitative results are presented in Table 5,1. Heat treatment at 40”C did not result in signifi- cant cracking for any combination of parametem, except for the very high sulphate cements which were not allowed a prestorage period. In most cases, heat treatment at 80”C led to cmcking Chaph5 Page 47

and new phase formation. The 3 hour delay period had an important effect on subsequent behav- iour.

5.3 1987 — Heinz and Ludwig [132]

This paper reports on continuation of experiments started by Ghorab et al [106]. Again, experi- ments were performed on mortar prisms in accordance with German specification DIN 1164. The high early strength cement had 12% C3A and 3.8% S03 (denoted PZ55 cement). In series I experiments mortar bars were heat-treated in the moulds after a delay of 2 hours. The treatment temperature varied from 20- 100°C. The duration of heating was controlled in order that all samples had the equivalent treatment of 72 hours at 20°C. Bars were heat-treated in a water bath. After cooling, specimens were stored at 20”C in water during which time, length, weight, and resonant frequency measurements were taken. Measurements were taken to 1000 days of soaking. Figure 5.3 shows the expansion, increase in weight and reduction in resonant frequency that oc- curred in the prisms that were heat-treated at different temperatures. There is clear indication that large expansions can occur if specimens are treated at temperatures above 75°C (all specimens treated at temperatures below 75°C [not shown] showed no significant changes with time of soak- ing). There is some indication that the 75°C cured prisms could expand significantly at ages greater than 1000 days (note that the rate of expansion starts to increase after 1 year) — also note the abrupt change in resonant frequency and increase in weight of the 75°C samples at 800 days. Figure 5.4 shows that a positive correlation exists between weight gain and expansion afier 28 days. This correlation is consistent with Mehta’s hypothesis [205] that expansion due to ettringite formation is the result of imbibation of water and swelling of ettringite clusters. Figure 5.5 is a scattergyam of expansion vs. change in resonant frequency; the correlation is not as good; the changes in resonant frequent y, however, could be related to microcracking within the mortar as a result of expansive stresses. X-ray analysis of specimensimmediatelyafler heat-treatmentindicatedthat the amount of et- tnngite decreasedas treatmenttemperatureincreased.Immediatelyafter a 70”Ctreatment X-ray peaks correspondingto the monosulphate(AFm) phase were observed and, as treatment tempera- ture increased, became the only sulphoaluminate present. Heinz and Ludwig postulate that as treatment temperature increases the aluminate and sulphate form hydrates that have decreased definition. Most significantly, both the aluminates and sul- phates are better incorpomted into the C-S-H stmcture. Thus, at high temperatures ettringite de- composes, not necessarily to monosulphate, but so that the sulphates and aluminates are incorporated into the C-S-H structure. The sulphate contained in the C-S-H apparently becomes mobile with time, because after 3 years storage the samples treated at 80- 100”C contain “practically nothing but ettringite as calcium alu- minate sulphate hydrate”. In comparison, a sample stored continuously at 20”C shows both AFt and AFm phases. In Series 11experiments Heinz and Ludwig examined the influence of the WA molar ratio on per- formance during soaking. In these tests anhydrite was added to the PZ55 cement to produce S03 contents within the range 2-8.60/o;thus, the WA ratio varied between 0.44 and 2.1. All bars were heat treated at 90”C for 12 hours in order to examine behaviour after a pretreatment period that is known to produce damage; a 2 hour delay period preceded the heat treatment. chapter 5 Page 48

RESONANT FREQUENCY (kHz) 13.0- , I 1 ! ~ji ! I I ! 12.5- i~!~ l-~~11~ _..&._..~,.,,.,, ;,,.,;, !- ~- ,_...__+ .T- ...... n.~.-.-~ ...... n-.F.-”-”-~-’~~ ”-”-~ -’-F- ,*4 - - ...... --- i i7’++ + ‘ ------0----+ ‘--- \ --~.-Q-~_-. ‘o_+& - ‘“n.,., I ~ 1 II 11,5.,.,...,_._.&...&.-.....- .“-..-..-.....-..-+.-..-..-.. ..-..+-. . . . “b.... < : -$ L.------Q-. .-$s-: ~ ~ + = !@& . , 11.o- “ I ‘..., : ; jl *,. %,., 10.5- “

WEIGHT GAIN, % 4.0 ‘ 3.5 ‘“ ) 3.0 ““ 2.5 “‘ / R ,....4----4P 2.0. ,/ A.....’-,#...- - -..”’ 1.5 “‘ / -0’ A * . /J~ 1,0 J 4 0,5- - 0,01 10 100 1000

EXPANSION, mtim 12

10 I / T . ------, 8 ,/ ~.- o’ 0“ ,// 6 ,’ II

4 / 2

,. 0 :

100 1000 STORAGE PERIOD (days)

~ 75C,16hr ---0--- fjoC,15hr ‘------85C, 13hr ‘------WC, 12 hr ~ 100 C,llhr

Figure 5.3 Changes in Expansion, Weight & Resonant Frequency with Storage (data from ref. 132) chapter5 Page 49

3.00

2.50

2.00 ■ s ● ?“ 4 WEIGHT GAIN h %, AFTER 28d 1“50 9: b’ ■

1.00

0.50

1

I 0.00 1 1 1 1 1

0.00 2.00 4.00 6.00 8.00 10.00 12.00 EXPANSION mm/m, AFTER 28 DAYS

Figure 5.4 Correlation Between Expansion & Weight Gain (data from ref. 132)

0.40

0.20

~ 0.00 i 1 b 1 1 ( d.oo 6.00 8,00 10.00 1; CHANGEIN 0t 2“00 RESONANT ‘0”20 ● FREQUENCY m #-* qrzzad -0.40 ■ m

■ -0,60

-0.80

-1.00

EXPANSION mtim, AFTER 28 DAYS

Figure 5.5 Correlation Between Expansion and Resonant Frequency (data from ref. 132) chapter5 Page 50

EXPANSION (mm/m) .- —

—+——— &A=,44-,66 10 -- ..----0“-- ~lA=.67 -----+-- ~/A=,79 8------+ --- 51A=.88

6- - ~ ~/A=l.5 ‘--.& --- &A=2,j

4- -

2- - s TTT3

--l

10 100 1000 STORAGE PERIOD (days)

Figure 5,6 Series II Experiments, Expansion vs. Storage (data from ref. 132)

The expansion results are summarized in Figure 5,6, The figure clearly shows that if WA is less than 0,66 there is no expansion and no indication that expansion might occur afler 1000 days, At an WA =0.67 there is a slight indication that expansion has started after about 800 days. Above WA=O.67,however,large expansionsof bars occur,Furthermore,at intermediatevalues of WA (values that are entirely practical for Canadiancements)the rate of expansioncan be dormantfor as much as 1year and then “suddenly”accelerate.At the very high WAmtios (1.5 and 2.1)the mte of expansionis appreciable fromthe start of storageand reaches0,6% at 500 days. It is importantto note that times to the stati of expansionare dependentupon the size of speci- men, In a parallel series of experiments,Heinz and Ludwignoted that a change in specimenge- ometry from 40x40x160mmprisms to 10x40x1601nxnprismsproduced expansionsmuch earlier, In Series III experimentsthe effect of relativehumidity of storage was examined,The researchers found that below 95% rh no expansionor reductionin resonantfrequencyoccurred,However,if specimenswere stored at 60% relative humidityfor an extendedperiod (120days)and then placed in a saturatedatmosphere,the subsequentdamagingnmtion and expansionwas intensi- fied (Figure 5.7). Series IV experimentslookedat the effect of water/cementratio and air-entrainmenton the ex- pansion process (Figure 5.8), A finding significantto the Canadianenvironmentwas that air-en- trainment suppressesexpansion,At the same time other experimentsshowedthat freeze/thaw cycles magnifi the problem; microcracksoccur duringf/t cycling whichprovide sites for forma- tion of ettringite. WEIGHT GAII % . 4 ! i m 3

2 -, ... J-= .+. /-.”’ ( 1 ,’ ,,R. Q ❑ ...... I .,..-.., - j...... -Q. -.-c?” t 0 — — ,/ 1 ...... - /1()0 ,/ Do -1 ,/ / -2 L f -3 - .-..<., -4 -+.. ---..,.+,_ ! I -...... _..,4 ...... -i.+ ----~ b — — -5

Humidity Change 60% to 100%

XPANSION ntim — 14 — —

12 /

f+--- 10

8 r

6

4

/ I ]60% to 100?4 ,/.’ 2 /.”’

/ 0 10 100 STORAGE PERIOD (days) I --.+...-.. 604 oo~eR.H. ~ WATER SOAK ‘--n--- 100% R.H, I

Figure 5.7 Effect of Humidity Change on Expansion (data from ref. 132) 2 hr. prestorage, 12 hr. treatment at 90”C, PZ55 HES mortar bars Chapter S Page 52

Expansion (mm/m) 10 3 . I 9- - T 8- - ~ w/c=o,4 & w/@0,5 with air 7- - entrainment / 6- - ---cl----- w/*o,5 9’ 5 ‘- ----+--- w/~o,45 / 1’ 4 ‘- ---*--- WIC=O.7 1’ / / ‘“ 3 -- /’ f / 2 “- / , k ,6 1 $“”” I TTT A d 4 I 100 1000 STORAGE PERIOD (days)

Figure S$ Effect of Water/Cement Ratio and Air=Entrainment on Expansion (data from ref 132) delay perlod=2 hrs,, treatment for 12 hoursat 90*C,mork barswith HEScement -

5.4 1988- Sylla [297]

Sylla reporteduponEuropeanexperiencewith secondaryettringiteformation,As temperature rlsa in concrete the reactivity and speedof dissolutionof C3Aincreases,while at the same time the availablesulphatein solutiondecreases,Therefore,there is an increasedtendencyto form monosulphaterather than ettringite,l%e importanceof the treatmenttemperatureand heat-treat- ment period on the quantityof ettringitepresentare clearly shownin Figure S,9,Resultsare re- ported from pastes made with two cements:bothcementshad approximatelythe same sulphate content,but cement B had 12,1%C3A.CementA had 8.7%C3A.The extra aluminatein cement B resultsin monosulphateformingin some pasteswhentreated at 60 and 80°C,The high C3Ace- ment also showsa morerapidreductionin ettrh@e content, Sylla notes that the formationof monosulphatealonedoes not account for all the sulphatere- leased whenettringitedisappears,He postulatesthat the predominantpotion of sulphatejs “in- creasinglyattached”to the C-S-Hat high temperatures.In fact this hypothesisis confirmedby parallel “rnicrofluorescence” and thermal analysis studies, l%e results given in Figure 5.9 are for pastes that had no delay period, Sylla notes that for tests performed with a delay of 3 hours the et- tringite formed during the delay period remained, even @r 5 hours of treatment at 80°C, Sylla also performedexperimentsto determinewhethernew phasesform during storage after heat lreatment.Paste specimenswere storedat 20°C and 100%relative humidityafter 5 hours heat- treatment at 40,60 and 80°C.There was no delay time. Resultsare given in Figure 5,10. For 40 and 60°Ctreatmentthere are no large changesin ettingite content as sorddngtime pro- ceeds; In fact for cement B ettringitecontentdecreaseswith soakingtime. After en 80”Ctreat- ment however,there is a verylarge increasein ettringiteduringthe first 28 days of soaldngand a chapter5 Page 53

X-RAY PEAK INTENSITY 20C 20 –

18- -

16- - 40C Ettringite and mono- sulphate peak measured 12 immediately after heat- treatment 10 E 8- -

6- -

4- -

2- - y 0 t 1 m, 1 2 3 4 5 \ Cement A: HEAT-TREATMENT PERIOD (hrs) 3.55% S03, 8.7% C3A

Cement B: / 3.36% S03, 12.1% C3A X-RAY PEAK INTENSITY 30 – Treated without pre- =: storage at given 25- - ~ temperature for periods Zoc from 1 to 5 hours.

15- -

10: L w

5- - . . -- - O---;’----*------1- & -- --- . --- . . 0+ * wA 1 I 1 2 3 4 5 HEAT TREATMENT PERIOD (hrs)

Figure 5.9 Variation in X-Ray Peak Intensity with Curing Temperature (data from ref. 297) chapter 5 Page 54

~’1 Note Iarga Inmass in eltrkrgks with tires for 80C curing

WC ~ ,,,, EITRINGITE PEAKS \::::,,;,,,, CEMENT A Y .: :.:

40c w ,:::,,,,:::.:/..::{ 60C ::.,:;:;! MONOSULPHATE PEAKS M C “’. CEMENT B\ 28 SOAKING o TIME, days ? AT 20C.

Figure 5.10 Changes in X-Ray Peak Intensities with Soaking Time (data from ref. 297)

I

! small increase thereafter. Afler 56 days soaking, all ettringite peak heights are the same value for a given cement. Note also that for cement B the monosulphate content also increased with time for all specimens. No cmcking was observed in the pastes and Sylla hypothesizes that the newly formed ettringite is finely distributed over the entire sample. If microcracks were present, then transport processes may result in large local accumulations of ettringite, with subsequent destruction of a longer pe- riod.

5.5 1989 — Heinz, Ludwig & Rudiger [1331

In this paper the authors summarize the findings of a number of previous publications [125, 157, 159, 160, 161]. Chqxkr 5 Page 55

Mortars or concrete prisms were cast and heat-treated in water at temperatures between 50 and 100”C. Various heat-treatment temperatures and r6gimes were used, with all r6gimes designed to give the same 3-day strength for all specimens. A wide vaxiety of cements were tested, including those with slag, fly ash, fume and limestone. Storage after treatment was normally in water at 20”C, but other storage media were also tried. In general, it was found that to avoiddarnagethe maximumtreatmenttemperatureshouldbe kept to less than 70°C.There is no clear indicationof the extentof delay periodrequired;prestorage (delay)for 1day to 1year prior to heat treatmentmay lead to more damagebecausethe structural matrixis stiffer and thus is less able to withstandthe build-upof pressures.When the delay period is withinthe first 24 hoursthen a longer durationresults in less damage. The cement type and sulphate content of the cement is important. For the same treatment, the high-early-strength cement is damaged, while the sulphate resistant cement is not (even at l(N)”C maximum temperature and when 490 additional sulphate is added to the sulphate-resistant ce- ment). Use of 8% ground limestone to HES cement neither diminishes nor increases the extent of the problem. On the other hard trass, ash and slag when used to partially replace some cement (30% to 50%) substantially improved performance; no darnage was observed after 10 years storage. Five to ten percent mass replacement by fume also resulted in a clear reduction in expansion. . Heinz and Ludwig postulate that the composition of cement is very important in determining whether expansions will occur. In particular, the main influence derives fmm thesO~~203 ~~o of the cement. However, the sulphate proportion appxirs to have a higher weight in determining behaviouq therefore, the authors suggest that the ratio (S03)2/A1203 is a parameter that shows the strongest correlation to subsequent effects of secondary ettringite formation. In this ratio the alu- mina content is that contained in C3A only — i.e. the so-called “active” alumina.

Figure 5.11 is a scattergram of expansionvs ~/A ratio. All data for various cements are plotted. me authors suggest a “Safe Ratio” of 2.0 — cements with a smailer ratio than 2.0 are not suscep- tible to secondary ettringite attack. Note that at higher @/A ratios expansion is again lower. In other words, there appears to be a pessimum ratio where expansion and darnage are a maximum. One other important findingfrom these authorsis that inclusionof 14%air-entrainmentinto the mortiu “drasticallyy“ reduces expansion; ettringite formation occurs, but takes place in the air-void system rather than within the confhmi space of the pore structure. On the other hand, if a poten- tial exists for secondary ettringite formation, a material that undergoes fxeezdthaw cycles is likely to be damaged to a greater extent. After such damage Heinz and Ludwig observed AFt crystals on aggregate particles. Cracks also occur in the matrix; in these cracks ettringite crystallizes at right angles to the crack edges.

5.6 1990 — Lawrence et al [177]

‘l%etest programme at the British Cement Association was performed primarily in an attempt to contlrm German observations of secondaryettringiteformation;U.K. cements were used Prisms, 40x40x 160mm, were cast in accordance with RILEM standard. The mortar composition was 1:3:0.5cement/sanct/water. Twenty U.K. cements, mostly commercial rapid-hardening ce- ments were employed. A pre-cuxing period of 2 hours in the moist room was used. Specimens were then placed in a water bath and the temperature was raised to 100”C over the course of two hours. ‘he boiling environment was maintained for another 3 hours and specimens were then al- chapter 5 Page S6

EXPANSION, mtim Expanslon=31 mm/m 16 4

A ‘A Suggested “Safe Ratio” of 14 I 1 2.0. Cements with ratio less , ❑ ■ than 2,0 not susceptible to I seoondary ettringite attack. 12 1 ■ M o . , ● 0 10 0 I ■ I

+

# a 8 I 1

, :0 ■ 6 o

1 1 ■ ■ 1 I o 4 t I i 2 1 O%**

14 ❑ @*& , J3&ia A-W+ II I I I I 0 r

0 2 4 8 10 12 14 16 la @/A RATIO

Illgure 5.11 Scattergramof ?/A ratiovs. Expansion(datafrom ref. 133)

lowedtocool naturallyuntil 16 hours whenthey were storedin water at 20°C,The aevemtreat- ment was chosen to reveal damagein the shortestamountof tlmq the regime chosen is more “ex= treme” than is likely in practice, Examplesof expansiveand non-exansiveresults are shownMFigure 5,12; other results are given in Figure5.13, Lawrenceet al were ableto classifythe types of expansivebehavlourinto 7 cate- gories: ~ (a) zero expansion— e.g. cementES3579,Figure5.12 * (b) sm~l continuousexpansion ~ (c)small acceleratingexpansion = (d) l$rge=elerating expansion— e.g. cementES3594,Figwe 5,13 ~ (e) @el~ating or limitingexpansion = (f) large S-shapedexpansionIesdhg to a definitemaximum— e.g. cementsES3595 (Fig.5.12) and ES3572 (Fig.5,13), chapter5 Page 57

EXPANSION, Yo 1,00

0.90 ------t 0.80 ------t- 0.70 ------t — 0.60 ------Exampleof=pansive CementES3595 0.50 ------

0.40 ------,-#_#”f- -

------0.30 ---fl ------Exampleof Nonexpansive 0.20 ------J ------Cement ES3579 w=’ 0.10 ------

0,00 1 0 5 10 15 20 25 SQUAREROOTOF TIME(days)

Figure5.12 Typical Expansion vs. Time Curves (data from ref. 177)

EXPANSION, % 1,00

0.90 ~ES3572,3HRS

0.80 — Cl— ES3572,16HR 0.70 — x- –ES35943HRS 0.60 — +— ES3594,16HR 0.50 ~ ES3740,3HRS 0.40 — * – ES374016HR 0.30

0.20 ------*------: X2X’------’

0.10

0.00 0 5 10 15 20 25 SQUARE ROOT OF TIME (days)

Figure 5.13 Expansion vs Time Curves Showing Effect of Delay Period (data from ref. 177) chapter5 Page 58

Table 5,2 The addition of 1’%0S03 to cement Expansion vs. Sulphate Content (ref. 177) ES3578 changed its behaviourflom a class (e) to a class (b) category— the addition of sulphate removed the tendency to reach a limiting expan- sion and resulted in a more gradual expansion process, The addition of sodiumsulphateto some cements did not change the 4.00 1.00 0.70 time at which expansionstarted,but 4.64 1.64 1.10 the magnitudeof the maximumex- 5.27 2.27 1,50 pansion increasedas sodiumsul- phate content increased.Table 5,2 showsthis effect. Figure 5.13 illustrates the influence of duration of heat treatment on subsequent expansion. Speci- mens that were heat-treated for 16 hours, mther than 3 hours showeda reduction in the time to the start of expansion,The effect is pronouncedfor cementES3740;the cementthat showedvery little expansionwhen treated for 3 hours showedthe maximumexpansionwhen treated for 16 hours. Petrography indicated that for specimens which crocked,the crockswere filled with ettringite,Et- tringite also formed ridges around surfacecracks.Expansiondid not necessarilymean that the specimenscracked, however;some bars showedas much as 1.2%expansionwithoutcracking, Analysisof thin sectionsshowedrims of ettringitearound sand grains.An “acicular” ettringite morphologyexisted withinthese rims, Lawrenceputs forwarda number of points to explainsecondaryettringite formation: Any ettringitethat formsduringthe delay period decomposesduring heating at 75°C or above Water storageat roomtemperatureafter heat treatment results in the m-formationof et- tringite withoutsignificantchanges in the quantity of monosulphate.For ettringiteto form in preference to monosulphatethe pore solution must be high in sulphatecontent, Lawrenceproposesthat this is the result of modificationof the calcium aluminate hy- drates— hydrogarnetlike materials— duringheat treatment;the modifiedaluminate compoundscannot rapidlyabsorb sulphatesand allow free sulphate ions, such as those slowlyreleasedby C-S-Hto persist in solution, The naturalprocess of slow releaseof aluminateswith time then results in ettringitecrys- tallization from solution, A less severetemperatureregime doesnot modi~ all of the aluminates. Somecan mp- idly absorb S04 immediatelyon cooling.Monosulphatethen forms insteadof ettringite because there is insufficientsulphate in solution, Ettringite formationoccursat nucleationsites: on the surfhceof voids, in microcmcks and at the cement-aggregateinterfhce, The air-voidsor flaws must become filled in order for internal stressand concretedisrup- tion to occur, chapter5 Page 59

5.7 Comment

The first important works directly related to seconda~ ettringite formation were performed by Ghorab et al in 1980. They found delayed expansions when specimens were cured at tempera- tures greater than 80”C. Heat treatment was found to be an essential component if expansion dur- ing soaking was to occur, but the time at which the treatment occurred was found not to be very important. In many respects the work done at the Cement Industry Research Institute [268] sup- ported these initial findings. The work of Heinz& Ludwig, reported in 1987 and later in 1989 firmly established the phenome- non of secondary ettringite formation. They found several crucial factors influencing the poten- tial formation of secondary ettringite: the critical heat-treatment temperature above which damage could potentially occur is in the range 70-75°C the longer the delay period within the first 24 hours, the lower is subsequent expansion, However, delay periods of several days or weeks, followed by heat treatment, could also produce sufilcient cracking to provide the nucleation sites necessary for secondary et- tringite effects the cement type and particularly the ratio of sulphates to aluminates are important for de- termining the potential for secondary ettringite formation. In 1987 Heinz& Ludwig es- tablished that an WAratio of 0.67 appeared to be a “critical” ratio above which substantial expansion and cracking had the potential to occur. In 1989 Heinz et al estab- lished another ratio, ~/A (where A is the “active” alumina contained in C3A); the safe value for this ratio was determined to be 2.0. the time at which significant expansion starts and the rate of expansion is a function of specimen size. Although the research was fairly limited in this respect, it appears that the larger the specimen (or member), the longer is the time to start of expansion, and the slower is the mte of expansion. If Heinz and Ludwig’s interpretation is correct, it ex- plains why expansions and cracking in the field have been observed after several years, while laboratory experiments, on smaller samples, show large expansions after the first two or three years of soaking. air-entminment of the mortar phase dmstically reduces the expansions that are observed when comparison is made to a non-entrained mortar. This phenomenon has been ob- served by several researchers and indicates that one simple method of deterring damage due to potential secondary ettringite formation maybe to ensure that adequately en- trained concrete is cast. Chapter 6 Page 60

Chapter 6 Secondary Ettringite Formation and the Chemistry of Cement

6.1 Effect of Gypsum Content of Cement on Strength and Volume Stablllty

Retardation of set and early strengh requirements are both important practical considerations that call for the production of cement with an “optimum” gypsurn content. The fineness of the ce- ment plays an important role in this respect and affects both retar&tion and early strength gain; the amount of gypsum required to produce proper retardation is higher the higher the freeness [180]. In 1917 ASTM C19-17 limited S03 content to 2%; considering that current Type 30 (and occasionally Type 10) cements are ground above 600 m2/kg Blaine, the 2% S03 limit is clearly not acceptable. Current CSA standards allow 4.5°/0S03 or more, A perusal of the Chapter 5, and some fbrthertest resultsgiven below, indicatethat what is “opti- mum” for strengthand retardation is not necessarily(and perhaps not normally)optimum with re- spect to long-termstability of concrete.

6.1.1 Test Results of Ai-Rawl [6]

The most significant work in this area was performed by AMtawi in 1977 [6], The rate of forma- tion of sulphoahuninates is accelerated by an increase in temperature, Therefore, one might ex- pect that the optimum S03 content for maximum strength to increase as temperature increases. Brown [34], for example, noted that maximum strength was achieved in a heat-cured concrete with 7.1’%S03. However, such concretes exhibited high expansions and he noted that it maybe necessary to limit S03 to 5-6% to keep expansions to a tolerable level, A1-Rawi examined the strength and expansion of concretes made with two different clinkers, Both clinkers (Table 6.1) had high C3A and C3S contents which are believed to be beneficial to steam-curing. The clinkers were blendedwith gypsumto give S03 contents of2,0, 3,8, 5.7 and 7,5%. Concrete cubes and prisms were cast with w/c=O,5,and a/c=5. Specimenswere subjectedto an 18hour cur- ing cycle consistingof 4 hoursdelay,a 12°C/hourheating rate to either 50, 70 or 90”C,followed by 6 hours of curing, Strengthawere determinedimmediatelyafter cooling, Other specimens were fog cured for ages up to 182days.Deformationreadingson prisms were taken to an age of 48 weeks, Strengthsat ages of 28 and 182daysare shown in Figures6,1 and 6,2 (the notation C1-70A means concrete made with cement 1heat-curedat 70°C),Both figures indicatethat the cements should have between 3.8 and 5.7% S03 contentto produce optimumstrengthsof heat-cured Table 6.1 Compound Composltlon of Clinkers used in A1.Rawl Study [6]

.. Chapter 6 Page 61

65.0

60.0 — C1-20A 55.0 50.0 ~ C1-50A STRENGTH 450 ~ C1-70A MPa “ 40.0 ~ C1-90A

35,0 - - C2-20A 30.0 ~ C2-70A 25.0 20.04 I , , r , r , , T , { 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5 8 S03 CONTENT%

Figure 6.1 Dependence of 28-day Strengths on S03 Content and Curing Temp. (data from ref. 6)

— C1-20A

~ C1-50A

~ C1-70A

~ C1-90A

Figure 6.2 Dependence of 182-day Strengths on S03 Content and Curing Temp. (data from ref. 6) Chapter 6 Page 62

specimens.On the other handroom-curedconcretesshowa clear regressionin strengthas S03 content hicreases above 2%. Generalconclusions,however,shouldnot be made upon the basisof only two high C3Scontent cements. optimum S03 contents for strengthare not necessarilycompatiblewith optimumcontentsfor vol- ume stability.Figure 6,3 showsthat S03 contentsaboveabout4% result in large expansions within the first 20 weeks of water-curing.On the other hand,results shownin Figure 6.4 indicate that an 18hour heat-treatmentcan reduce expansionduring48 weeksof water-soaking, me results given in Figure 6.4 are contraryto the commonly-heldviewthat heat-treatmentpro- duces an increase in expansiondue to secondaryettringiteformation.Damagedue to SEFhas been connectedto the productionof microcracksduringheat-treatment.Suchcracks are reduced, or eliminated,by provkiingan adequatepre-curingperiod.Note that A1-Rawiemployeda 4 hour pre-cure prior to heat-curingto as high as 90°C.

6.2 The Practical Significance of the S03/A1203 Ratio

FaI Type 30, high early stnmgth cemenc CSA A5-M88 allows a maximum of 3,5% S03 when the C3A content is less than or equal to 7.5%, and a maximum of 4,5% if C3A content is greater than 7.5%. The American Society for Testing and Materials Standard C150-89 also allows as much as 4.5% S03, but the C3A limit above which this is allowed is 8%, rather than the 7.5% found in the CSA standard.ManyotherCountriesin the worldfollowthe Americanstandar~ 1 while others set a single S03 limit varyingfrom 3.()to 4.5%. [Cembureau,Cement Standardsof ! the World,46].

0.14

0.12

0.1

0.08 Expansion (%) o ~ .

0.04

0.02

0 0 5 10 15 20 25 30 35 40 45 50 Curing Time, Weeks

Figure 6.3 Effect of S03 Content on Expansion of Normal Fog-Cured Specimens (data from ref.6) Chapter 6 Page 63

0.2 0.18 0.16 0,14

0,06 0,04 0,02 0 2 3.8 5.7 7.5 S03 Content (%)

Figure 6.4 Effect of S03 Content on Expansion after 48 Weeks of Soaking (data from ref. 6)

Canadz Colombiaand Hungaryaxethe oniy three countriesnotedin the Cembureaureport that al- low 4.5% S03 in cements whichcontain as little as 7.5%C3A(Colombiaand Hungaryset a 4.5% S03 maximumwhichis independentof C3Acontent).Canada and thosecountriesthat use ASTM C150-89 also permit S03 contentsgreaterthan 4.5%if it can be shownthat mortarbars prepared in a standard manner and soaked in water for 14 days expand less than 0.02%. Thus, a low expan- sion value in a 14 day testis used to assess the long-term volume stability of cements with high sulphate contents. The question that arises is: is a 14-day test an adequate method to predict long- term stability? Lerch and Ford in 1948[183]testedprismsmade from 27 differentcementsthat were moist- cured for 1day and then storedcontinuouslyin water at 21“Cfor 5 years.The S03 contentsof the cements were fairlylow, but the results,shownin Figure 6.5 point out the lack of correlationbe- tween expansionsat 14days and expansionsat 5 years.Severalcementsshowless than 0.02%ex- pansion at 14days, but showappreciableexpansionsat 5 years.The test resultshelp to accentuate that predictionof long-termbehaviourfrom short-termtests is seldomreliable. In the case of Lerch’s results,the reliabilityis low. Aithough the WA ratio appears to be the most reiiable “indicator” of whether secondary ettringite formation could be importan~notethat other factors also ~ntribute to the ~~~~tion of whethera given cement is like]y to be stable in concrete.Wells et al, for exampleproposedthat it is not just the WAratio that determineswhetherexpansionoccurs after heat treatment;both a low ratio and a low alkali content are necessaryto avoidlarge expansions[3171. Chen and Grattan-Bellew[55]in their studyof alkali-aggregatenmctionaddedvariousquantities of KzSOAand NaX30qto a clinker from a singlecementplant. A suite of 17cements was manu- Chapter6 Page 64

0,25-

0,2- “ ?“

0,15 EXPANSION AT5 YRS,% 0,1 I O“O:b=zaLL- 0 0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04 EXPANSIONAT14DAYS,%

Figure 6,5 Scattergram OFExpansion at 14 Days vs Expansion at 5 Years (data from ref. 183)

fhcturedwhich containedvariousalkali contents.The cements were intergroundwith an opti- mum quantity of gypsurnto a finenessof 370 m2/kg.Expansionwasthen monitoredwith time of storage in a moist environment.A good correlation(r=0,91) was foundbetween acid-solubleal- kali content and expansion.However,fbrtheranalysis of the data for the present report also showeda positive correlationbetween expansionand MgOcontent (Figure6.6) and between rate of expansionand SO#Al*Osratio (Figure 6,7); both comelationcoefficients(0.73 and 0.80)are significantat the 5°Alevel. Chen & GnMan-Bellew also considered the potential effect of sulphates and looked at the hy- pothesis that “ettringite in cracks and air-voids might be contributing to the late stages of expan- sion”, They foundthat there was no correlationbetween mte of expansion of the high-expansion bars (those withNa20 equivalent>1.0%)and the amount of ettringitoin the moctarat 6 months.

6.3 Speclficatlons for Heat Treatment of Concrete

Concern overpossibleproblemswith secondaryettringiteformationhas prompted some m- researchersto recommendheat-treatmentguidelines,some authoritiesto change their specifica- tions, and some authoritiesto considerchangingtheir specifications,Properheat-treatment methods wouldtheoreticallyeliminate,or substantiallyreduce,the potential for problems even if the cement and/or aggregate were foundto be inferior, Neck [237]proposedthree types of heat treatment,dependingupon the proposed durabilityexpo- sures, withthe objectiveof avoiding damagedue to secondaryettringite formation: Type 1:1 hour pre-cure with a maximumsoakingtemperatureof 80”C— for interiorex- posureswith no moisture Type II: 2 hour pre-curewith a maximumsoakingtemperatureof 70”C— for exterior ex- posurewith no contact with the earth Chapk?r6 Page 65

CORRELATION BETWEEN EXPANSION AND MgO CONTENT 17 POINTS, CORR. COEFF. = 0.732 EXPANSION = .651 + 5.974 * MgO

30

25

20

5

0

5

0

30

5

0

.5 .6 .7 .8

S03/AL203 RATlO

Figure 6.7 Scattergram of SO#Ai@3 Ratio vs Expansion (data frOm ref. 55) ~ hapter 6 Page 66

= Type III: 3 hour pre-cure with a matimum soaldng temperature of 60°C — for exterior exposure wh.hcontact with the earth The European Committee for Standardization in 1989 has responded to the issue by setting more slringent specifications for the heat treatment of concrete [98], These are: ~ The concrete temperatureduringthe first 3 hoursof curingcannotexceed 30°Cand can- not exceed40°C duringthe first 4 hours

~ The rate of temperaturerise cannot be greater than 20 OC/hr = The average maximumtemperaturecannotexceed 60°C;individualtemperaturereadings cannotexceed 65°C ~ The codng rate aftexheat-treatmentcannotexceed 10OC/hr ~ Concrete must be protectedagainstmoisture10ss Alsoin 1989,The GermanCommitteefor ReinforcedConcrete [103]recognizedthat intensive heat treatment can affect durabilityof structuralmembers~e, Faultsin the concrete due to too rapid warmingor sharplydiffering expansion are susceptible to “sulphate bonding [that] can occur with a high level of moisture in the concrete”. Accordingly, speciflca- tiona for damp conditions are different for dry concrete, as shownin Table 6,2,The 60”CmaxL mum temperaturefor concretethat will be damp in serviceis clearly a change madeto address potentialproblems with [email protected], The Britishhave been slowerto respond,perhapsbecausethey have been waMngfor detailedre- search results from the BritishCementAssociation(e,g, Lawrenceet al [177]),It 1sinteresting that the Britishcode of practicein 1972CP110:1972notedthat damagedue to excessiveheat treatment wouldnot occur if ~ temperaturerise duringthe first 3 hours1snot greater than 15°C!/hr ~ thereafter,the rate of increaseor decreaseis not greater than 35°C/hr = the maximumtemperaturereached by the concreteis not greaterthan 80°C, However,the 1985Britishcode of practice BS8110:1985ondtted the abovethres guidelines, Other Brltlsh authoritieshave taken stepsto 1111the gap, The BritishDepc of TransportSpecifica- tions for HighwayWorks (1986),Part 5, Clause 1709:5(Ii)requh’es(1) the concrete must be lefl for 4 hours withoutaddkionalheating;(2) the concretetemperaturecannotbe raked at a rate

Table 6.2 Specifications for Heat=Treated Concrete German Committee for Reinforced Concrete, 1989[103]

I Moisture Category Dry Damp Minimumholding (delay)period (hrs) 3 4 + or + Maximumconcretetemperatureduring holding, “C 30 40 Maximum heating rate, OC/hr <20* <20* Maximumconcretetemperature**,“C 80 60 heating rate should be reducedto 10OC/hrfor lightweightconcrete ** individual values may be up to 5°Chigher Chapter 6 Page 67

greater than 20°C/hr;(3) the maximumconcretetemperaturecannotexceed 70°C;(4) the rate of coolingcannot exceed the rate of heating [177]. [n North Americameasuresto controlexcessiveheat treatmentshave been takenon a locaJscale. The results of Merritt& Johnson[218]on concretestrengthdevelopmentafter variousheat treat- ments were incorporatedinto the Iowa State HighwayCommisionspecificationsfor precastcon-

‘1’leinitial concrete temperature should not be greater than 38°C for a minimum of 2 hours after casting The rate of temperature increase after a 2 hour period should not be greater than 14°C ‘I%emaximum temperature attained should not be greater than 66°C The maximum temperature should beheld for a period sufficient to develop the nx@red strength ‘he rate of temperature decrease should not be less than 11°C/hr. Units should be kept covered for at least 24 hours after casting. This is all good, practical, engineering sense; it was not intended to address directly the issue of prevention of secondary ettringite formation. There is some concern, however, that such steps should be taken immediately, and by national authorities.

Hill [136], points out in a letter to ACI committee 517.2R,the issues that ntA to be addressed concerning guidelines for precast concrete manufacture. He notes that the present guidelines do not mention the current potential problems related to secondary ettringite formation of precast concrete. l%e three conditions that appear necessary for this problem to surface need to be in- cluded in a warning statement. These conditions are: = a curing (heat-treatment) temperature greater than 60-71“C ~ concrete that will be submerged or exposed to 100Zorelative humidity - concrete made with a mment with a high WA ratio and high alkali contents.

Hill was concerned that under the guidelines noted in517.2R a user could face a situationwhere the abovethree conditions occur. Hill recommended that the current document be withdrawn. It appears that in Canada a revision of CAN3-A23.4-M78 on precast concrete methods is about to take place. The 1978 standard specifies that the concrete should attain initial set before heat is ap- plied (this is normally 2 to 4 hours but could be as low as 45 minutes). During this time the con- crete must be maintained at at least 10”C. Temperature increase should occur at approximately 20°C/hr to an optimum temperature range of approximately 65-70”C, but in no case should the curing temperature exceed 80”C. Temperature decrease should not occur at a rate exceeding 30°C/hr. Proposed changes for the new version of A23.4 include reducing the maximum tempera- ture tlom 80°Cto 70°C.

6.4 Comment

Determination of “optimum” gypsum content involves the determination of the proportion of gyp- sum that will produce optimum strength and a satisfactory setting time. Research performed in Europe and in Canada (see Chapter 7) indicates that the sulphatehlumina ratio of the cement may play an importantrole in determininglong-termstabilityof concrete.Thus, “optimum”gypsum Chap&r 6 Page 68 needsto be redefined;long-termdurabilityof productsmanufacturedwith Portlandcement needs furtherconsideration, Clearly,cement is just one ingredientof concrete.Otheringredientscan affect durabilityas can the mix-design,casting and curing procedures.However,all other factorsbeing satisfactory,it is necessaryto be able to evaluatecementsfor long-termdurability.This requiresthat a satisfactory test methodis avtdlable, It is unrealisticto expect a good correlationbetweenthe expansionof mortar bars soakedh water for 14days and expansionador crackingof concreteat 10or 15years,The scattergramof Fig- ure 6.5 showedone exampleof the poor correlationthat exists between short and long-termex- pansions;the cementswere manufacturedat rrdd-century,It is Mghlyunlikely,for a numberof masons,that the correlationfor moderncementsISbetter, The viabilityof the 14-daysoakingtest shouldbe *examined in the light of the chemistryand physicalpropertiesof modem cements.‘Ms is especiallyimportantgiven (a) secondaryettrlngite formationmay bean importantdurabilityissue, and (b) both ASTMand CSA standardsallow more than 4,5% S03 if a cement can be shownto pass the 14-daysulphateexpansiontest. In CanadL A23.4 specificationsfor precastconcreteam presentlyunder revIsIomIt is suggested that careful considerationshouldbe paid to the potentialsecondaryettringlteissue andto the Ger- man specifications(Table6.2) that have addressedtMspotentialproblem,It is especiallytrouble- some that the proposedA23,4still relies onthesetdng time to define the minimumdelay period prior to treatment this settingtime could be as short as 45 minutesand still satls~ both A23,4 and ASstandards, chapter7 Page 69

Chapter 7 Rapid Test Method for Secondary Ettringite Formation

Researchers have been looking for rapid test methods to “predict” long-term behaviour of con- crete for many years. For example, in 1949 Scholer [282] reported on results of a revised acceler- ated performance test which was first developed by Gibson in 1932; Gibson wished to look at the map-cracking of concrete pavements [108]. In Scholer’s test method, saturated concrete specimens (3’’x4”X16”) are dried at 54°C for 8 hours, then immersed in water at 21-27°C for 16 hours. This process is repeated for 6 days a week while on the seventh day they rested — immersed in water. This accelerated test can be performed for as long as the researcher wishes, with expansions being measured at regular intervals. To substantiate his test method, Scholer compared field performance of various concretes to ac- celerated performance. Concretes were made with 2 types of aggregate (one good performance, one poor performance) and a wide variety of cements. Prisms (3’’x4”X16“ prisms — the same size as those used in the accelerated test) were placed on the ground and exposed to 5 years of natuml freeze/thaw cycles. The accelerated exposure W* runOver285 ~YS. Figure 7.1 shows that a Dositive correl~tion between field and a~elerated performance occ~red. ‘Failure of the specimens occurs by excessive expansion, cracking and loss of strength. It is interesting to note that as early as 1949 there is some inference that something other than the known reactiotdexpansion processes (such as aar) may be at work. For example, Scholer recog- nized the importance of the bond between the paste and the aggregate and that after this bond fhils, a “jacking action” tends to develop to increase expansion. Although Scholer achieved good correlations between laboratory and field performance, and his work was heralded as landmark research (by Bryant Mather, for example [see discussion to pa- per]), the “accelemted” exposure lasting 285 days would hardly be considered “accelemted” to- day. In recent times, many other researchers, including the present author, have attempted to develop laboratory tests that reflect long-term field performance. The primary difficulty in this development is that it takes a very longtime to determine whether an accelemted test is accurate because it takes a very long time to obtain ~levant field data. An adequate test must ensure that it “measures” or predicts potential expansion in an unbiased fashion. Figure 5,13 (Chapter 5) illustrates the influence of duration of heat treatment on sub- sequent expansion. Specimens that were heat-treated for 16 hours, rather than 3 hours showed a teduction in the time to the start of expansion. The effixt is especially pronounced for cement ES3740; the cement that showed very little expansion when treated for 3 hours showed the maxi- mum expansion when treated for 16 hours. The development of a rapid test method must take this into account [177]. Observations by Heinz et al that freeze/thaw cycling exacerbates any potential for secondary et- tringite formation and damage, leads them to the suggestion that freeze/thaw cycling can be used to accelerate potential deterioration of concrete that has a dense microstructure [133]. However, they did not attempt to develop an accelerated test method based on this principle.

7.1 The Duggan Test

One accelerated test method that has been extensively developed will be the subject of the re- mainder of this chapter, because it is purported to “measure” the effect of secondary ettringite for- mation of various concretes. This is the Duggan test, whose development, results, and chapter 7 Page 70

DATA FROH SCHOLER, 1949 48 eamasl CORR, COEFF. ■ ,707 FIELD EXP, = -,017 + 1.382 * ACCEL. EXP.

0.25

0.2

0,.. . ..”! ! 0 . ..’” 0.15 !1

0,1 ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,

0.06

0

!

I I -1 0,02 0.04 0.00 0,08 0,1 0,12

ACCCLCRA7E0 CXPANSXON, X

Figure7.1 Scattergram ofAccelerated vs. Field Expansions (data from ref.282)

significanceofmsults havebeenreported indetailinDuggan &Scott [91-93]; Scott&Duggan [284]; GillOttetal[110, 111];Jones & Gillott [155];Attiogbeet al [12];and Wells et al [317], As originally devisedthe Duggantest involvesthe followingprocedure [284]: Concrete cores (5 minimum)are taken from.a structureor from laboratory-castprismsor cylinders,The coresare 25mm in diameterand at least 65 mm long and are cut to 50mm lengths.The ends are groundsmoothand parallel, Initial length measurementsare taken with a comparatorjust before the start of heat treat- ment. In latertests by Duggan& Scottand in tests by Gillott et al [110,111]this measure- ment wasused as the zero point from whichother strain ~dings were taken. Coresare soaked for 3 days in distilledwater @21°Cin a closedcontainer Coresare then placed in a @-air ovenat 82°C for one day Cores are removedfrom the oven,allowedto cool for 1hour, then placed back in dis- tilled water for one day chapter7 Page 71

~ A second one-day heating, one-day soaking cycle is performed = A third cycle is performed, but this time cores are left in the oven at 82°C for 3 days @ At the end of the third heating cycle, cores are removed from the oven, allowed to cool for 1 hour and then length measurements are taken relative to a steel standard. In the early tests by Duggan and Scott this measurement was used as the zero point for further strain readings. @ Cores are placed in distilled waterat21 “C = Length measurements are taken relative to the steel standard at intervals of 3-5 days.

7.1.1 Significance of Duggan-Test results

The test was initially thought to be an accelerated means by which potential alkali-aggregate reac- tivity of various cement/aggregate combinations could be assessed. Scott & Duggan set about es- tablishing that the Duggan test successfidly predicted aar of various lab-cast concretes. By taking cores from both sound and deteriomted site concretes, they also established a rough correlation between observed behaviour and results of the Duggan test. Presumably due to time/fmncial limitations, no attempt was made to establish a formal correlation between Duggan results on young concretes and the long-term performance of those concretes. Scott& Duggan [284] noted that the test could be used both to classi@ lab trial mixes, and to evaluate existing structures. Observations that (a) concrete known to be susceptible to aar and concrete especially prepared with reactive aggregate showed large expansions in the Duggan test and (b) concrete with small expansions in the Duggan test did not contain alkali-reactive aggregate, led the researchers to con- clude that the test could be useful for evaluating potential alkali-aggregate reactivity. Later research (Duggan & Scott, 93) showed that both the type of aggregate, the type of cement and the alkali content of the cement play a role in determining the 20-day expansion value (which was defined by the researchers as the critical time at which pass/fail criteria could be es- tablished), Figure 7.2 shows some of the results. Duggan & Scott concluded that only concrete should be ranked by the Duggan test — it is the cement/aggregate combination, and not just the cement alkali content or the extent of reactivity of the aggregate that determines potential expan- sion behaviour, They advocated perforrnace testing, exemplified by the Duggan test, such that “the testing emphasis currently placed upon acceptance or rejection of aggregates should be shifled to acceptance or rejection of concrete”. In this regard, the researchers should be com- mended for espousing a philosophy that is starting to gain wide acceptance; a significant number of researchers, the present author included, believe that it is a dangerous and potentially costly practice to acceptor reject the individual material components of concrete, such as aggregate, ce- ment and pozzolan in isolation.

7.1.2 Research of Gillott, et al

The research of Gillott et al [110, 111] and Jones et al [155] provides valuable insight into the sig- nificance of the Duggan test. In the first series of test, [111], various concretes were manuf~tured with 4 commercial cements with different alkali contents, 2 alkali active aggregates and 1 inert aggregate. Specimens were subjected to both the Duggan test and to the standard concrete prism test for Potential Expansiv- ity of Cement/Aggregate Combinations (CSA A23.2-14A). chapter7 Page 72

I ❑ AGG. A (INERT) ❑ AGG. B (REACTIVE) ❑ AGG, C (REACTIVE) I

0.8 I m 0.7

0,6

0.5 EXPANSION, 0.4 ‘/0 0,3

0.2

0.1

0 0.95 0.95 1.00 1.00 1,07 1,10 1.17 TIO T 30 T 30 T 30 TIO T 30 T 10 CEMENT ALKALI CONTENT, ?’0 AND CEMENT TYPE

Figure 7.2 Effect of Alkali Content and Cement Type on Expansion (data from ref. 93)

Throughcarefhl experimentationand comparisonsof the responseof the variousconcretesto the two types of test, Gillott was able to make the two primary conclusionsthat: - al~li.aggregate rwtion is not the major cause of expansion measured during the first21 days of the Duggan test = expansion in the Dugga.n-test at ages up to 90 days “correlates with the relative propor- tion of ettringite and frequency of microcracking in the concrete”, Other more tentative conclusionswerethat (a) Duggan-testexpansiondependsprimarily on the properties of the cement, and (b) the form of the sulphatethat is present (either gypsum or anhy- drite) may affect the extent of microcrackingcausedby the Duggantest, and thus may affect sub- sequent expansions. In a second set of experiments, reported in detail in Gillott et al [110] both cement pastes and con- cretes were manufwtured from various cements and aggregates. Six different Type 30 Portland cements were chosen on the basis of variations in S03 content. Two non alkali-expansive aggre- gates were used in the preparation of the concrete. Cement pastes were prepared for each of the six cements, with a w/c ratio of 0.4. Ten concrete mixes were pre ared with different ~ment/ag- gregate combinations; the nominal cement content was 475 kg/m!l, and the water/cement ratio was approximately 0.4 for all concretes. All specimen sizes were 3“x3”X14” and, in an attempt to simulate precast concrete, specimens were subjected to an accelerated curing regime OR2 hours pre-cure, 2 hours temperature increase to 85°C, 4 hours at 85”C, and slow cooling overnight, The 24 hour strengths of the concretes were in the range 27-44 MPa, while the 28 day strengths were in the range 56-64 MPa. chapter 7 Page 73

1,2 EXPANSION OF CONCRETE I

~ CEMENT A 1.0 ..-. ------—a-— CEMENT B /’ 0.8 – ‘“– CEMENT C ------/ ------

M . ~ CEMENT D / A ------!!0.6 —-&— CEMENT E B

0.4 ------./ -. / >4 ------

0.2

0.0 0 10 20 30 40 50 60 70 80 90 SOAKING TIME, DAYS

EXPANSION OF CEMENT PASTE

, 1 ~ CEMENT A .------— U- – CEMENT B

—. ● .- — CEMENT C ------

~ CEMENT D -- — -A- — CEMENTE ‘ ------‘ ------”------_—- -

0.4 ---- . . . . ------.F ------

~/ “F 0.2 ---- -. -R- ---- T.-Y----:----.:-7 ::: ; :: :?:: t=

0.0 o 10 20 30 40 50 60 70 80 SOAKINGTIME,DAYS

Figure7.3 Expansion of Hardened CementPastes and Concretes Made with theSame Cements (data extracted from ref. 110) chapter 7 Page 74

Gillott found that all cement pastes had significant expansion aller being subjected to the Duggan r6gime. All pastes (Figure 7.3) showed a monotonically decreasing expansion over the course of 90 &ys of soaking, The initial and final expansion values and the mte of expansion over various time periods was different for the different cements, All ten concretes that were tested also showed significant expansions during the Duggan test. It is understandable that after Gillott’s findings were released those in the cement and concrete indus- try became rather concerned about the Duggan test; Five of the six concretes made with the Nel- son aggregate (a dolomitic limestone) fbiled the proposed Duggan limit of 0.05%at21 days, while two of the four concretes made with Exshaw aggregate (a calcitic limestone) tiiled. The results presented by Gillott provided an opportunity to compare the expansion behaviour of pastes and concretes, Figure 7.3 shows the expansion of cement pastes and of concretes manufac- tured with the same cements at approximately the same water/cement ratio. The aggregate used to make the concretes was the Nelson dolomitic limestone. There is a fair correlation between the expansion of pastes and the expansion of “equivalent con- cretes” at soaking times up to approximately 40 days. Beyond 40 &ys, however, behaviour be- tween the two “types” of material is different. The expansion curves for pastes become concave downwards — the expansion process is starting to exhaust — while the curves for concrete re- main straight or, in the case of the concrete with Cement D, become concave upwards, The con- crete strains at 90 days for two of the concretes are much greater than those of the corresponding cement pastes, The lack of correlation at higher expansion levels is clearly shown in the scatter-

1,20

■ 22 DAYS 1al ==%””” /“ ❑ 48 DAYS 0’ la’ /“ 0,80 ● 90 DAYS /“ i? / k , /“ 5& 0.60 b ● /’ ❑ ❑ . ‘n ’ [ 0.40 / k /’ bt ■ , ❑m ‘/ “ 0,20 + /’ •1 /’ ‘/’. / O,cm 1 t I I t Oocm 0.20 0.40 0.60 0.80 1.(x) 1,20 % EXPANSION OF CONCRETE AT GIVEN TIME

Figure 7.4 Scattergram of Concrete Expansion vs. Paste Expansion (data from ref. 110) chapter 7 Page 75

gram given in Figure 7.4; this is a plot of expansion of concrete vs. expansion of the equivalent cement paste at a given time. Based upon an examination of the research noted in this report, it is suggested that the reasons for the disparity between paste expansion and concrete expansion are that with time ettringite be- comes more concentrated in localized regions such as in microcracks and at cracks at the aggre- gate/paste interface. The total ettringite content in the concrete matrix is the same as in the paste, but in the concrete ettringite is localized and is therefore more efllcient in producing expansions. By using a suite of cements with various sulphate and aluminate contents Gillott was, in essence, able to confirm the findings of Heinz et al [133] (see Figure 5.11, Chapter 5) who noted a pes- simum value for ?/A ratio and that cements with high @/A ratios showed smaller expansions than those with smaller mtios. Figure 7.5 plots expansions vs WA ratios for both cement pastes and concretes (made with Nelson aggregate). There appears to be a pessimum range for WA mo- lar ratio centred about a value of 1.0. One important aspect of the test progmmme was that the expansion tests that were performed on concrete in the Univesity of Calgary laboratories were repeated in the CN laboratories. It is sig- nificant that when one compares both the qualitative behaviour of expansion vs. time and the sizes of expansions observed at given time, there is excellent agreement between the results from the two laboratories; the Duggan test on concrete appears to have excellent reproducibility. Gillott’s mearch resulted in a number of practical conclusions: Delayed ettringite formation is the major cause of expansion of cements and concretes exposed to the Duggan test The cement is the main component causing expansion; expansion is highest, all other tic- tors equal, when the WA molar ratio is approximately 1.0 (Note that the present author has attempted to correlate oxide compositions and Blaine surfhce areas of cements used in Gillott’s work and in Attiogbe’s work with expansion. Various correlation methods have met with failure; there does not appear to be single parameter or group of parame- ters which show significant correlation — other than the 13/Aratio). The severe heat treatment of the Duggan test causes microcracks in the concrete. Et- tringite can form fwter due to easier penetration of water and an increase in the number of potential nucleation sites. The formation of ettringite at nucleation sites effects fbrther microcracking and affects expansion mtes.

7.1.3 Research of Attiogbe and Wells et al [12,3171

Attiogbe also examined the use of a severe heat treatment followed by soaking in water and meas- urement of expansion to evaluate the cements and concretes with potential durability problems. The researchers prepared 75x75x380mm prisms of concrete and 150mm cubes of both concrete and cement paste. The water/cement ratio for the pastes was 0.50. Specimens were stored at 23°C in a moist room for the first 24 hours after casting, then in a fog room until 7 days. Unlike Gillott’s test progmmme, a simulated precast heat-treatment regime was not used prior to the start of test. The “Duggan test” performed by Attiogbe normally consisted of two cycles of (a) heating to 80”C for 3 days, followed by (b) 1 hour cooling to room temperature, and then (c) soaking in distilled water for 2 days. The Duggan test and the Attiogbe test are compared in Fig- ure 7.6. Note that Attiogbe defined “zero strain” two days after the end of the last heating period (see Figure 7.6). In other respects the Attiogbe test method was similar to the Duggan test. chapter 7 Page 76

t

1!20 ‘

1,00------

0.80- . - - ” ------” - - - ”----”-”- t ii! 0.60. . - - - ” - - ” - - ” ------”” - J= B K OAO------

------21 OR22 v DAYS 000 + C3 0.8 0,9 1 1,1 1,2 1.3 1.4 MOlARRATt0303/A1203

1,20 ‘

lSKI .------

0,80. . ------” - -”------” m i! 0.60. . - ” ------h w 0.40. ----- ...... 4!5OR 48 0DAYS 0s)0 - 008 049 1 1,1 1)2 1,3 1.4 MOlARRAT10303/A1203

1.20 ‘

i!! 8

I

Osxl ~ 0.8 0.9 1 1.1 1.2 103 1,4 MOLARRAT10S03/A1203

Figure 7.5 Observatlonsofa PessimumS03/A1203 Ratio atVarious Ages (data from ref. 110) Ptwe 77

DATUM LENGTH MEASUREMENT (A) SATURATED SPECIMEN AlllOGBE et al ATIIOGBE REGIME m

distilled water at roomtemperature

(b) + / mm DUGGAN rREGIME LLH_Hl”

3 45678 9 10 11 12 TIME (CiEty$) /

o,l’% I I (c) DEFORMATION VSTIME STRAIN El A= . 1ST &2~JE-SAT

FIRST DRY* ● C o.1’% 2ND DRY 3RD DRY ., . b...... z DUGGAN ZERO STRAIN, EARLY E)(PTS cmtraction STRAINSAT END OF EACH DUGGAN CYCLE

Figure 7.6 Temperature Testing R6gimes, Data and Examples Strains in Duggan and Attiogbe Tests chapter 7 Page 78

2,5

~— CEMENT A

~ CEMENT B 2.0 ~ CEMENT C

~ CEMENT D 1.5 ● EXPANSION ‘Ye 1.0 /’ 0.5

0.0 o 5 10 15 20 25 30 35 TIME (days)

Figure 7.7 Expansion of Paste Cores in Attiogbe Test (data from ref. 12) w/c=O.5, cores taken from 150mm cubes

Example expansion results for cement pastes are shown in Figure 7,7, Note the extremely large expansions that were observed with the pastes made with cement B and cement C, - - The pamllel tests performed by Attiogbe are perhaps more interesting than the main series, The researchers looked at expansion behaviour of pastes and concretes stored in saturated NaCl solu- tion rather than in water, Companion specimens stored in water expanded, but specimens stored in NaCl did not expand and even contracted in some cases. In the presence of chlorides the volu- bility of ettringite increases; therefore, no ettringite is observed in NaCl soaked specimens and no expansion occurs, This is strong evidence that the expansions in water are due to the formation of ettringite, Fundamental examinations confirm that pastes and concretes that exhibit large expansions con- tain “massive formations of ettringite” and those with small expansions have “little or no visible amounts of ettringite”, The results suggest that “late formation of ettringite” is the cause of expan- sion, Attiogbe’s research is in agreement with Gillott’s in this respect; the Duggan test, or tests similar to Duggan’s test, indicates expansion due to ettringite formation and not alkali-aggregate mction.

7.1.4 Correlations Between Glllott’s and Attiogbe’s Results

Both research programmed have come to similar conclusions concerning the Duggan test, The test is a measure of secondary ettringite formation and not alkali aggregate reaction, at least in the fust 20 days of soaking. Ftuthermore, the findings in both studies do not contradict the chapter 7 Page 79

0.25

CEMENT B ❑ 0.2 t

CEMENT D 0.15 J ■ CEMENTB ~ CEMENT C

0,1 1n CEMENTA t I @ CONCRETE , CEMENTC 0.05 I CEMENT A PETE_ 0 1 o 0.5 1 1,5 2 2.5 STRAINSMEASUREDBY ATTIOGBE AT 20 DAYS,%

Figure 7.8 Scattergram of Strains Measured by Gillott [110] and those Measured by Attiogbe [12] Data are Iabelled according to origin of Cement

mechanism of secondary ettringite formation first proposed by European researchers (see Chap- ter 5). Attiogbe and Gillott used some of the same types of cement in their studies. Thus, it is possible to examine whether any comelation exists in the expansions obtained by the two studies. Figwe 7.8 shows a scattergram that was obtained from the available results. The very large stmins shown in the Attiogbe study and confiied by Wells et al [317] are some- what surprising. Apparently the Attiogbe heat-treatment r6gime is much more severe than that of the Duggan r~gime. This is also somewhat surprising, since Gillott subjected his specimens to a simulated precast r6gime prior to the Duggan test. One aspect that could explain the very high strains of the Attiogbe paste specimens is the higher w/c ratio used (0.5 vs 0.4 for Gillott). In any event, there is little correlation between these two studies, despite the fact that the same types of cement were used. Although it has been shown that the Duggan test is reproducible among laboratories, it is clearly important that a single test procedure is used for all studies.

7.2 Interpretation of Duggan-Test Resuits

Duggan & Scott initially proposed a limit of acceptable behaviour for concrete as 0.l% expan- sion after 20 days of soaking. The proposal of this limit for the Duggan test was criticized. Lawrence et al [177], for example, had several concerns about the test which they say prevent sat- isftiory interpretation of results: chapter 7 Page 80

@ 22mm diameter cores are not representative of the concrete structure. The testing of small concrete cores could magni~ the intensity of the effect of secondary ettringite for- mation, Note that Lawrence tested mortar bars which am also not representative of the concrete = Drilling cores could itself produce &mage and thus create sites for additional expansion of ettringite @ Site concretes may not have been heat-treated and therefore may not be susceptible to secondary ettringite formation. They become susceptible when the core is heat-treated in the Duggan test = Expansions of the cores during testing maybe the mult of a number of different effects EIJ the heating cycle proposed by Duggan does not represent practical heating regimes. The heating programme is very severe and may nxult in the rejection of cements that may perform adequately in pmctice, = Duggan’s test results have not been correlated with field concrete,

7.2.1 The Zero Strain Datum

A primary criticism centres about the definition of zero strain on a dry datum, Figure 7.6b shows a schematic of the heating/cooling regime followed in the Duggan test, Also shown (Figure 7,6c) are example values of observed strains measured at the end of each component of the Duggan test and during soaking (typical values extracted from Duggan & Scott [92]). Figure 7.6a shows the temperature regime followed by Attiogbe et al [12], In this example, first drying results in a shrinkage strain of about 0,07?40(bottom left of Figure 7.6c). Most, but not all, of this deformation is recoverable on first resaturation, Second drying produces less relative shrinkage than in the 1st cycle, but the total shrinkage is slightly greater, Second resaturation is not much different than 1st resaturation and the magnitude of 3rd drying is approximately the same as second drying, The length of the specimen at the third drying point is the definition of zero strain fust used by Duggan. Thus, when bars are placed back in water for swelling vs. time measurements, a substan- tial component of that swelling deformation is due to the uptake of water and not due to swelling as a result of secondary ettringite formation, Under these conditions, then, the definition of O,1°/0 expansion at 20 days as “unacceptable behaviouf’ is shaky, at best, because the swelling compo- nent due to water ingress will vary from concrete to concrete independent of whether secondary ettringite forms, Afier criticism by ASTM committee C09.02.02, Duggan & Scott [92] redefined the zero datum as the length of the saturated core hefizcethe start of the test regime (see Figure 7,6), This new zero was also used by Gillott et al [11O, 111] in their experiments. This is a better approach be- cause it attempts to remove the component of swelling vs time that is due to water uptake only. The rationale behind this definition is that the swelling due to water uptake atler 3rd drying is equal to the total shrinkage that occurs between point A and point B (Figure 7,6c). Thk new defi- nition led Duggan & Scott to define a new Pass/Fail limit of 0.05°/0at 20 days, rather than O.10/o. This is, however, still not correct because during the Duggan r6gime the cores undergo a signifi- cant amount of permanent deformation. In the example shown in Figure 7.6 this amounts to 0.01 to 0,02% (compare point A with the 1st and 2nd resatwation points). It is well known that most permanent (irrecoverable) deformation in concrete that is subjected to shrinkage/swelling r4gi- mes occurs during the first cycle; shrinkage and swelling deformation during subsequent cycles chapter 7 Page 81

Table 7.1 Analysis of Strains During Duggan Test (data from ref. 92)

Expans. Old New R&ised Type of corvxete tested Zero Zero Zero Z1 Z2 Z3

A Non-reactive agg., low alk. cemen -0.067 0.046 -0.021 -0.030 0.041 4.010 -0.045 0.062 0.027 0.041 B: Non-reactive agg., silicafum~ -0.065 0.055 -0.010 -0.054 0.074 0.010 -0.062 0.066 0.016 -ma C Non-reactive agg., silica fume, repaa -0.051 0.045 -0.006 -0.052 0.062 0.024 -0S)62 0.056 0.02 -0.024 D Non-reactive agg., low alk. cemen -0.021 0.015 -0.006 -0.050 0.044 -0.012 -0.044 0.096 0.04 0.052 E: Expensive oonc., failed in servkx -0.064 0.060 -0.004 4.079 0.067 0.004 -0.076 0.322 0.246 0.235 F: Non-reactive agg., high elk. cemen -0.050 0.050 0.000 -0.072 0.062 -0.010 -0.060 0.304 0.234 0.242 Q: Reactive agg. #l, low-alkali cemen -0.020 0.016 -0.004 0.010 0.000 0.006 -0.027 0.119 0S)98 0.119 H: Reactive agg$2, low-alkali cemen S3-0.056 0.062 0.006 %-0.066 0.050 -0.030 -0.050 0.096 0.016 0.046

0.35 ❑ FIRST DUGGAN ZERO - Z1 ------0.3 ■ SECOND DUGGAN ZERO - Z2 — ---- 0.25 ❑ REVISED ZERO - Z3

- ---- 0.2 STRAIN r AT ------20 DAYS 0.15 % 0.1 “

o.05-

0-

-o.05- I I I I I I I r’ A B c D E F G H CONCRETEDESIGNATION

Figure 7.9 Comparison of 20 day Expansions Using Three Zero Definitions chapter 7 Page 82

are then approximately constant. Therefore, a more correct estimate of water-swelling after the 3rd drying cycle is the swelling that occurs between the 2nd chying point (point C) and the 2nd re- saturation point (point D). Results reported by Duggan & Scott [92] were used to compare the effect of the three definitions of the zero point on the deformation at 20 days. The extracted data and analysis am shown in Ta- ble 7.1 and Figure 7.9. The zero data are defined as: = Z1: Initial Duggan Zero, point B, Figure 7.6 ~ Z2: New Duggan Zero = Z1 - estimate of water swelling (pt A - pt B) = Z3: Revised Zero = Z1 - estimate of water swelling @t D-pt C) It should be noted that Attiogbe et al [12] also attempt to compensate for the water-swelling com- ponent of deformation by defining the zero point two days after the last heat cycle (see Figure 7.6). The disadvantage of this definition is that strains over the fwst two days that are, in tit, due to internal chemical reactions and expansion will not be measured, The use of the revised zero, or any other zero, and definition of pass/fail criteria based upon abso- lute deformation values is not the best approach for two reasons: = Regardless of the zero datum that is used, there must still be some uncertainty in the meaning of a given deformation at, say, 20 days. This uncertainty occurs because we can- not precisely differentiate between deformation caused by water-swelling and deformat- ion caused by secondary expansions. ~ A single point criterion is proposed to describe a time-dependent rate process — be it aar or secondary ettringite formation. Any number of reaction mechanisms could be envis- aged which pass through the proposed window of, say; between Oand 0.05% expansion at 28 days. A single point criterion is not satisfactory, A much more mtional, and still simple, approach is one which can be developed from the work of Gillott et al [110]. Gillott calculates rates of expansion over two time periods, among others — i.e. from 3-22 days and from 22-90 days. He also defines a ratio of deformation rates= rate (3- 22) /rate (22-90). It is suggested that if the Duggan testis to be developed into a standard test method, then defini- tion of pass fail criteria along the lines of “rate of deformation” must be used, There are three dis- tinct advantages: ~ The expansion processes involved are rate processes and therefore the use of rate criteria are fimdarnentally valid = The use of rates of deformation to define pass/fail eliminates the uncertainty associated with having to differentiate between water-swelling and other types of swelling, Duggan and Scott [92] indicate that the time-dependence of wetting-swelling is effectively com- plete 8 hours after the start of soaking. Thus, mtes of deformation determined afler this eight hour period should be related to the extent and consequences of secondary et- tringite formation. ~ Comparison of tates of deformation over different time periods (e.g. 3-22 days and 22-90 days) allows a rapid determination of whether the time-dependent process is accelerating or decelerating, A single deformation-time criterion as proposed by Duggan & Scott can- not do this. Chapkr 8 Page 83

Chapter 8 General Discussion

A review of case studies indicates that deterioration of concrete is usually a result of a combina- tion of effects. Alkali-aggregate niwtion, corrosion and sulphate attack appear to be principal mechanisms. Some case studies indicate that secondary ettringite formation may also be a princi- pal mechanism for some types of concrete. The number of cases that can be directly related to sec- ondary ettringite formation is small; however, a detailed review of research in the area combined with a few simple deductions indicate that the rate of incidents that can be wholly or partially at- tributed to secondary ettringite formation will increase. The main question posed at the beginning of this report was: is there, or is there likely to be, a po- tential secondary ettringite formation problem in North America? Examination of the available lit- erature does not reveal any incident of deterioration in North America dhectly attributed to secondary ettringite formation. lherefore, one conclusion based on this review is that there is not presently a secondary ettringite problem in North America. However, good quality laboratory research confirms that secondary ettringite formation can pro- duce expansion and cracking of heat-treated concrete made with North American commercial ce- ments. In North America it is not unusual to find Type 30 cements with S03 contents in the range 3.5 to 4.0% — and could be as high as 4.5Y0.It is not unusual to find alumina contents of 570 or more and C3A contents of 9% or more. Calculation of the WA or ~/A ratios for such cements places them in the maximum expansion ranges determined by Heinz and Ludwig and by Gillott- Cement with high sulphate and alumina contents results in the potential for production of a large quantity of ettringite and thus high potential expansion. The high BMnes currently used for Type 30 cements results in a heterogeneousmicrostructure, easier ion transport to nucleation sites and m- requireseven higher sulphate contents to prevent rapid set. The presence of cement, aggregate and steel together in concrete result in a weak interracial re- gion which has high calcium hydroxide content, with orientated crystals, high in porosity and high in ettringite content. This is also the region that cracks when excessive heat treatment is applied. On the concrete production side, the author is aware of concrete treated with maximum tempera- tures in the range 60-70”C and delay periods in the range 1-3 hours. Furthermore, some produc- tion processes using high heating and cooling rates (as high as 25-30°C/hr) have been observed. A 70”C maximum temperature is within the range where ettringite starts to decompose and sul- phates integrate into the structure of calcium silicate hydrate; research indicates that sulphate in this form can slowly re-enter the pore solution at a later age, and react with aluminates to produce secondary ettringite. The potential inadequate delay periods and excessive heating and cooling rates could &mage the concrete to provide sufficient nucleation sites for secondary ettringite for- mation and destruction. In Canada we often have precast concrete that is subjected to frequent wetting and drying and freezing and thawing cycles; cycles which will exacerbate any potential secondary ettringite prob- lem. Fortunately, much of the concrete is also air-entrained, which appears to result in substantial dampening of the secondary-ettringite deterioration process. The conclusion is that given current cement production and construction practices there does ap- pear to be a potential for a secondary ettringite formation problem in North America. Possible remedies in the cement plant are to strive towards the reduction of C3A, S03 and the tW- nesses of cements used by the precast industry. Better test methods need to be developed to evalu- ate the potential long-term stability of cements in concrete. Chapter 8 Page 84

The precast concrete industry should strive towards more stringent specifications. It is suggested that the new German sgx!cificationsmaybe appropriate for consideration by North American authorities. There is substantial evidence that air-entrainment, pozzolams and lime-stone aggregate can all contribute towards Educing the potential for secondary ettringite formation. Much more research is also necessary, Much of past research has been performed on cmnmercial czments. Clearly, to determine the conditions under which secondary ettringite forms, and thus to propose more precise specifications, we need to perform careful experiments on specially pr- eparedcements — cements where cement composition, gypsum content, fineness, and other fac- tors are strictly controlled. At the same time, experiments on commercial cements should also continue in order to provide relevance to the results from the controlled experiments. The development of adequate test methods to evaluate long-term durability is also a crucial theme for fiture research, It is unfortunate that the Duggan test started out on a poor footing — lack of proper attention to detail and lack of control during early experiments severely limited the signM- cance of the early findings. More careful research by Gillo~ however, and maUzation that some improvements to the test method can still be made suggest that a future variation of the Duggan test may prove to be a useful practical test method, and one that could be developed into an accu- rate standard test method.

Acknowledgements

The author is grateful to the Portland Cement Association (Project 92-05) for funding this project. Thanks also go out to members of the Secondary Ettringite Task Group (H, Chen, S. Cumming, P, Grattan Bellew, P. Breeze, C, Rogers and J,F, Scott) for forwarding copies of papers and other information relevant to the subject, Special thanks are due to Paddy Grattan-Bellew and the Na- tional Research Council for doing the computer-database search on topics related to ettrlngite. Paddy also provided tome english translations of some important German papers. I would also like to acknowledge the painstaking work of the Interlibrary Loans Oftlce at the University of Cal- gary and Mr. CMS Huizer for doing the legwork to obtain access to the relevant literature. The contents of this report reflect the research and views of the author, who is responsible for the accuracy of the report. “fhe opinions expressed and conclusions drawn in this report am not neces- sarily those of the Portland Cement Association. References Page 8S

References

All references examined for this report are listed beloW not all are cited in the report since some had only peripheral relevance to the main topic.

1 Abo-El-Eneim S.A., Abo-Nuaim.i, KKh., Marusin, S.L., E1-Hemaly, S.A.S., Hydration ki- netics and microstructure of ettringite, Proc. 6th Conf. on Cement Microscopy, Int. Ce- ment Microscopy Assoc., Duncanville, Tx, 219-231, 1984

2 Abo-El-Eneim S.A., Hanati, S., Hekal, E.E., Thermal and physiochemical studies on et- tringite. II: dehydration and thermal stability, Cemento, V 85, No 2, 121-32,1988

3 Abo-El-Enein, S.A., Mikhail, R.S., Hekal, E.E., Zettlemoyer, A.C., Microstructure of et- tringite produced by different methods, Proc. Int. Conf. Cem. Microscopy, 4th, Ed. Gouda, G.R., 295-302, 1982

4 Abo-El-Eneiu S.A., Salem, T.M., HanafL S., Abd-El-Wahab, S. M., Marusiu S.L., Elec- trical conductivity and microstructure of me hydrata c3A-3cas04 Pas@, proc. 7~ Intnl. Conf. on Cem. Microscopy, Int. Cement Microscopy Assoc., Duncanville, Tx, 277- 287, 1985

5 Adonyi, Z., Gyarmathy, G., Kiliau J., Szekely, I., Investigations by thermogravimetry into the hydration processes in tricalcium alumimte and tricalcium aluminaegypsum mixtures, Periodica Polytechnic& Chemical Engineering, v 13, n 1-2, 131-47, 1969

6 A1-Rawi, R.S., Gypsum content of cements used in concrete cured by accelerated meth- ods, Journal of Testing and Evaluation, Vol 5, No. 3, May, p231-236, 1977

7 Alexanderson, J., Strength losses in heat-cured concrete, Swedish Cement and ConCr. Res., Ins. Proc., No. 43, Stockholm, 1972

8 Alksnis, F., Alksne, V., Role of the siliceous phase in the degradation processes of ce- ment stone in sulfate media 8th Intnl. Congress Chem. Cem., Rio de Jan.iero, V 5, 170-4, 1986

9 Alksnis, F.F., Alksne, V.I., Silicate phase effect on the degradation processes of cement blocks in sulfate containing environments, Tsemen4 n 9, Sept, 13-15, 1984

10 Ahmno-Rossetti, V., Chiocchio, G., Collepardi, M., Steam curing and resistance of port- land cements to sodium sulfate solution, Cemento, V 70, n 1,23-32, 1973

11 Anon, Technical guidelines for the use of sulphate-resisting cementi Indian Concr. Journ., v 62, n 6, June, 307-309, 1988

12 Attiogbe, E.K., Wells, J.A., Rizkall% S.H., Evaluation of Duggan concrete core expan- sion tes~ Research report, Sponsored by Canadian National Railways &Transport Instit- ute, University of Manitoba, Winnipeg, Canada, Sept. 1990

13 Ayyar, T.S. R., Ramesan, K., A study of lime columns in an expansive clay, Indian Geotechnical Conf., Andhra Univ., Visakhapatnam, India 185-189,1989

14 Bailey, J.E., Hampson, C.J., Microstructure and chemistry of tricalcium aluminate hydra- tion, Technology in the 1990s, Developments in Hydraulic Cements, Proc. Royal Sot. Discussion Mtg., London, Feb 16-17, 105-111, 1983 References Page 86

15 Bannai, H., Nakagaw& K., TlwrmaJ transformations of ettringite, Sekko To Sekkai, V 97, 253-9, 1968

16 Bannister, F.A., Hey, M.H., Bernal, J,D., Ettringite from Scawt Hill, County Antrim, Miner. Msg., 24,324-329, 1936

17 Barnes, B., Morphology of the paste-aggregate interface, PhD thesis, University of Pur- due, 1975

18 Barneyback R.D., Diamond S., Expression and analysis of pore fluids from hardened ce- ment pastes and mortars, Cem. Concr. Res,, V 11, 279-285, 1981

19 Bedard, C. and Bergeron, M., The effect of steam curing on high-early strength Portland cement containing carbonate addition, in Carbonate Additions to cemen~ ed. P. Klieger and R.D. Hooton, ASTM Special Tech. pub]. 1064,51-59, 1990

20 Bensted, J., Discussion of “the effects of several heavy metal oxides on the deformation of ettringke and the microstructure of hardened ettringite”, by Tashiro, C., Oba, J. and Akarnai K., Cem. ConCr. Res,, V 10, n 1, 119-20, 1980

21 Bensted, J., Varm& S,P., Studies of thaumasite, Part 1 and Part 2, Silic. Industr,, 38, 1973, 29-32; Silic, Industr,, 39,11-19,1974

22 Bensted, J., Varm& S,P,, Studks of ettringite and its derivathws — 4, the low sulphate form of calcium sulphorduminate (monosulphate), Cement Technology, v4, n 3, May-Jun, 112-114, 116, 1973

23 Bensted, J,, Varm& S,P,, Studies of ettringite and its derivatives, Cem. Technol, v 2, n 3, May-June, 73-6,100,1971

24 Bensted, J,, Varm& S,P., Ettringite and ita derivatives 11,Chromate substitution, Silicates Ind., V 37, n 12,315-18, 1972

25 Bentur, A,, Ish-Shalom, M,, Properties of type K expansive cement of pure components - 2, proposed mechnkm of ettringite formation and expansion in unrestrahed paste of pure expansive component, Cem. Concr, Res,, v 4, n 5, Sept,, 709-721, 1974

26 Berra, M., Baronlo, G., ‘Ihaumasite in deteriorated concretes in the presence of sulphates, Amer. Concr. Inst., Publ SP-1OO,Concrete Durability, V2, DetroiL 2073-89,1989

27 Berry, L,G., The unit cell of ettringite, An, Mineralogist, V 48, Nos. 7-8,939-940,1963

28 Bezjak A,, Jelenic, I,, Crystal structure investigation of calcium aluminum sulfate hy- drate - ettringite, Croat. Chem. Act& V 38,239-42, 1966

29 Blanks, R.F,, Kennedy, H.L., The technology of cement and concrete, Vol. 1, Concrete Materials, Wiley, New York 1955

30 Benin, A., Air aging and stiffening of cements, Int. Congr. Chem, Cem., 7th Pm,, V 2, IU209-W213, 1980

31 Bon-in,A., Cariou, B., Calcium aluminate-gypsum-lime-water system, Int. Congr. Chem. Cem., 7th Proc., V 3, Paris, V/158-V/163, 1980 References Page 87

32 Boros, M., Balazs, Gy., Strength development of 3Ca0.Alz03-CaS04.HzO systems dur- ing steam curing, Zero. Kalk Glps, V 24, n 10,473-6, 1971

33 Braga Reis, M.O., Formation of expansive calcium sulphoalurninate by the action of the sulphate ion on weathered granites in a calcium hydroxide saturated medium, Cem. Concr. Res., V 11, N 4, Jui, 541-547,1981

34 Browu H.E., An investigation of the delay period and related factors in steam-cured con- crete, Virginia Council of Highway Investigation and Research and U.S. Department of Commerce, 1965

35 Browm P,W., ‘Ile implications of phase equilibria on hydration in the tricalcium silicate water and the tricalcium aiuminate gypsum weater systems, Proc. 8th Intnl. Congress Chem. Cementi Rio deJaniero,V111,231-238, 1986

36 Browm P.W., LaCroix, P., Kinetics of ettringite formation, Cem. ConCr. Rw., V 19, No 6, NOV,879-884, 1989

37 Brown, P.W., Liberman, L.O., Frohnsdorff, G., Kinetics of the early hydration of trical- cium aluminate in solutions containing calcium sulfate, Journ. Amer. Ceramic Sot., v 67, n 12, Dee, 793-795, 1984

38 Buck A.D., Relationships between ettringite and chloroaluminate, strength, and expan- sion in paste mixtures, 8th Intni. Congress Chem. Cement, Rio de Janiero, V 3,417-24, 1986

39 Budnikov, P.P., Erschler, E. Ya., Studies of the processes of cement hardening in the course of low-pressure steam curing of concrete, Highway Research Board, S~ial Re- port 90, Washington, 1966

40 Budnikov, P.P., Kravchenko, I.V., Expansive Cements, Proc. 5th Int. Symp. on Chemis- try of Cement, Tokyo, Vol 4,319-330,1968

41 Building Research l%tablishmen~ Alkali aggregate reactions in concrete, Building Re- search Establishment Digest 258, Watford, England, 8, 1982

42 ButL Yu. M., Kolbasov, V.M., Timashev, V.V., High temperature curing of concrete un- der atmospheric pressure, 7th Intnl. Symp. Chem. Cem., Tokyo, 437-471,1968

43 Campiteli, V.C. and Florindo, M.C., The influence of limestone additions on optimum SUL fur trioxide content in portland cements, in Carbonate Additions to Cement, ed P. Kiieger and R.D. Hooton, ASTM Spech.I Tech. Publ. 1064,3040, 1990

44 Caries-Gibergues, A., et al, Contact zone between cement paste and aggregate, Proc. Intnl. Conf. on Bond in Concrete, 1982,24-33, 1982

45 Carpenter, A.B., Oriented overgrowths of thaumasite on ettringite, Am. Mineralogist V 48, NOS. 11-12, 1394-1396, 1964

46 Cembureau, Cement Standards of the World, 7th Edition, Brussels,1985

47 Chamberline, W.H., Brewer, H.W., Shideler, J.J., Compressive strength of steamared concrete, Report C-621, Bureau of Reclamation, U.S. Dept of the Interior, Denver, Au- gust, 1952 References Page 88

48 Chandra, S., Bemtsson, L., Deterioration of concrete in swhnrning pools in the south of Sweden, ACI Materials Journal, American ConCr. Inst., V 85, N 6, Nov-Dee, 489-494, 1988

49 Chatterji, S., Discussion of mechanism of sulfate attack on Portland cement concrete — another look by P.K. Mehm Cem. ConCr. Res., V 14, n 1,151-2, 1984

50 Chatterji, S., Discussion of expansion associated with ettringite formation at different tem- peratures,by A. Negro and A. Bachiomird, Cem. Concr, Res., V 13, n 4,587-8,1983

51 Chatterji, S., Discussion of paper Scanning electron micrographic studies of ettringite fm- mation by P.K Meh@ Cem. ConCr. Res., V 6, n 5,711-12, 1976

52 Chatterji, S., Jeffery, J.W., A new hypothesis of sulphate expansion, Msg. ConCr. Res 15,83-86, 1963

53 Chatterji, S., Jeffery, J.W., Further studies of the three-dimensional arrangement of the hydration products of Portland cements, Indian ConCr. J., v 45, n 4, Apr., 167-72,1971

54 Chatterji, S., Thaulow, N., Jense~ A,D., Studies of alkali-silica reaction. Part 4. Effect of different alkali salt solutions on expansion, Cem. Concr. Res., V 17, No 5,777-783,1987

55 Chew H, and Grattan-Bellew, P.E., Effect of cement composition on expansion of mortar bars due to alkali-silica reaction, 8th Intnl. Conf. on Alkali-Aggregate Reactiom Kyoto, Japan, 635-640, 1989

56 ., Chino, K, Kawamura F,, Stability evaluation for cement package containing radioactive waste, Nuclear Technology, v 81, n 1, Apr. 95-99, 1988

57 Chiocchio, G,, Collepardi, Mario, Turriziani, R., Substituted hydrated calcium silicates obtained in autoclave hydration, Joum. Amer. Ceramic SW., v 58, n 5-6, May-Jum 185- 188, 1975

58 ClarlGA.I,, Peon, C, S,, Perry, R., Rational use of cement-based stabilization techniques for the disposal of hazardous wastes, U.S. Environmental Protection Agency, Offke of Research & Development, Report EPA 600/9-85/025, Publ. by EPA, Cincinatti, OH, 229- 347,1985

59 Clastres, P., Murat M., Bachiorrini, A., Hydration of expansive cements - correlation be- hveen expansion and formation of hydrates, Cem. Concr. Res., v 14, n 2, Mar, 199-206, 1984

60 Cohen, M.D., Modeling of expansive cement, Cem. Concr. Res., Vol 13,519-528, 1983

61 Cohen, M.D., Microstructural analysis of ettringite formation: scanning electron micro- scope, Proc. Intnl. Conf, on Cement Microscopy, 4th, Publ, by Int. Cement Microscopy Assoc., Duncanville, Tx. 204-224, 1982

62 Cohen, M.D., Themies of expansion in sulfoaluminate - type K expansive cements: schools of thought, Cem. Concr. Res., v 13, n 6, NOV.809-818, 1983

63 Cohen, M.D., Reply to discussion by B. Mather of the paper Theories of expansion in sul- foaluminate-type expansive cements: schools of though~ Cem. Concr. Res., V 14, n 4, 610-612,1984

., References Page 89

64 Cohen, M.D., Discussion of paper Expansion of ettringite by water adwrption by P.K Mehta and S. Wang, Cem. ConCr. Res., V 12, n 3,409-10,1982

65 Cohen, M.D. and Mobssher, B., Physiochemical, rnicrostructural and micromechanicd investigations of sulfoaluminate -type expansive cements, Proc. 8th Intnl. Congress Chem. Cement, Rio de Janiero, V, 476-480,1986

66 Cohen, M.D., Campbell, E., Fowle, W., Kinetics and morphology of ettringite formation, Proc. 7th InL Conf. Cem Micmsc., Ed. James Bayles, Publ. by ht. CementMkfCXWPy Assoc., Duncanville, Tx, 3081,1985

67 Cohen, M.D., Hashemi, H. N., Crystal morphology under pressure, Roe. 7th Intnl. Cat on Cement Microscopy, Publ. by Int. Cement Microscopy Assoc, Duncanville, ‘I% 267- 276, 1985

68 Collepardi, M., Degradation and restoration of mssonry walls of historical buildings,Ma- terialsand Structures, v 23, n 134, Mar, 81-102, 1990

69 Collepardi, M., Baldini, G., Pauri, M., Corradi, M., Tricalcium aluminate hydration in the presence of lime, gypsum or sodium sulphate, Cem. ConCr. Res., v 8, n 5, SepL 571-580, 1978

70 Collepardi, M., Marcialis, A., Massid@ L., Turriziani, R., Paste hydration of 4caC).3A1203.S03 in the presence of Ca(0~2, C*04.2H20 ~d some org~c com- pounds, Cemento, v 72, n 2, Apr-Jun, 53-60,1975

71 Concrete Society, Alkali-silica reaction, minimizing the risk of damage to concrete: guid- ance notes and model specification clauses, Technical report no. 30, Lmdon, England Oct. 1987

72 Copeland, L.E., Bodor, E., Chang, T.N. and Weise, C.H., Reactions of tobermorite gel with aluminates, ferrites and sulphates, J. Portland Cem. Assoc. Res. Dev. Lab., V 9(l), 61-74,1967

73 Cos@ U., Massaz.za,F., Interstitial mass composition and setting inegularities of port- land cements, Cemento, v 72, n 4, Ott-Dtx, 181-194, 1975

74 Coutinho, A. de S., As~ts of sulfate attack on concrete, Cem. ConCr. Aggregate& V 1, n 1,10-12,1979

75 CrsmmoncL N.J., Quantitative X-ray diffraction analysis of ettringite, thaumasite and gyp- sum in concretes and mortars, Cem. ConCr. Res., V 15,431-441, 1985

76 Crammon& N.J., llmumasite in failed cement mortars and renders from exposed brick- work Cem. Concr. Res., v 15, n 6, Nov, 1039-1050, 1985

77 CrammonG N.J., Quantitative X-ray difllaction analysis of ettringite, thaumasite and gyp- sum in concretes and mortars, Cem. ConCr. Res., v 15, n 3, May, 431441, 1985

78 D’Ans, J., Eick+H., Das system CaO-A1203-CaS04-H20 bei 20”C, Zement-Kalk-Gips, Dec., 449459, 1954

79 Daerr, G.M., punze~ M., Ludwig, U., On the chemical and thermal stability of et.tringite, ReacL Aluminates Setting Cem., Summ. Contrib. Semin., 42-5, 1977 Refmnces Page 90

80 Dahrns, J., Beton hoher Fruhfestigkeit (concrete with high early slxength), Betonwerk + Fertigteil-Technik 40, H. 6, S. 402/408, 1974

81 Davies, G. and Oberholster, R.E., ‘Ihe effect of different outdoor exposure conditions on the expansion due to alkali silica-reaction, 8th Intnl. Conf. on Alkali-Aggregate Reaction, Kyoto, Japan, 623-628, 1989

82 Deloye, F-X, Combined action of sulphur and alkalies in cement-aggregate reactions in concrete, Bulletin de liaison des labratoires des Ponts et Chaussees, V 161, May-June, 41449,1989

83 Detwiler, R., Effect of granulated blast fiumace slag on the transition zone in concrete, Concrete Durability: Katharine & Bryant Mather International Conference, Amer. ConCr. Inst. publ. SP-1OO,DetroiL 63-72,1987

84 Detwiler, R.J., Monteiro, P.J.M., Texture of calcium hydroxide near the cement paste-ag- gregate interface, Cem. ConCr. Res., Vol 18,823-829, 1988

85. Diamond, S,, ASR - Another look at mechanisms, 8th Intnl. Conf. on Alkali-Aggregate Reaction, Kyoto, Jam 83-94,1989

86 Diamond, S., Lachowski, E.E., Investigation of the composition and morphology of indi- vidual particles of Portland cement paste 2. calcium sulfoaluminates, Cem. ConCr. Res., v 13, n 3, May, 335-340, 1983

87 Dorner, H., Rippstain, D., Effect of an aqueous sodium chloride solution on ettringite, Tiz-Fachberichte, V 109, n 9,680-3, 1985

88 Dron, R., Brivot, F., Ettringite Expansion, Bulletin de liaison des laboratoims des Ponts et Chaussee-s,English translatio~ ISSN 0458-5860,1990

89 Dron, R., Brivot, F., Contribution a l’etude du gonflement ettringitique, paper mad at VIII Intnl. Symp. on the Cemistry of Cement Rio de Jsnko, V 4.2,1-6,1986

90 Dron, R., Briv@ F., Expansion due to ettringite, Bulletin de Liaison des Laboratoires des Ponts et Chaussees, n 161, May-Jun, 25-32,1989

91 Duggaz C.R., Sco& J.F., Proposed new test for alkali-aggregate reactivity, Canadian Na- tional Railways, Technical Research Report, Montreal, Canada, April 13,1987, revised Oct. 31,1989

92 Duggau C.R. and Sco~ J.F.~Establishment of new accepancehejection limits for por- posed test method for detection of potentially deleterious expansion of concrete, Re- sented to ASTM Subcommittee C09.02.02, sep~ 1989

93 Duggu R. and Scm F., New test for deleterious expansion in concrete, 8th Intnl. Conf. on Alkali-Aggregate Reactiom Kyoto, Japan, 403408, 1989

94 El Korchi, T., Melchinger, K., Kolvites, B., Gress, D., Effect of heavy metals and organic solvents on the microstructure of cement stabilized hazardous waste, Proc. Intnl. Conf. on Cement Microscopy, 8th Publ. by Int. Cement Microscopy Assoc., Duncanville, Tx, 132- 143, 1986 Refmnces Page 91

95 E1-Didamony, H., Khalil, A.A., Investigation of the hydration products of the system CaO-A120s-Si02 in the prwence of varying amounts of Si02, ZKG. zemen~ KallL Gips, V 34, No 12,660-663, 1981

96 E1-Sayed, H.A., E1-Wahed, M.G. At@ Ali, A.H., Some aspects of the corrosion of nSn- fodng steel in concrete in marine atmospheres, Durab. of Building Materials, v5, N 1, Jul, 13-25,1987

97 Ess.ingtou M.E., Composition and volubility of ettringite precipitated Ilom combusted oil shale, Roe. 23rd Oil Shale Symp., Colorado Sch. of Mines Ress, Publications Center, Golden, Co, 16-25,1990

98 EuK)pean Committee for Standardization Concrete-performance, production, placing and compliance aiteri~ Final DrafL European IWstandsr& February, 1989

99 Farram J., Grande4 J., Maso, J.C., Variations in the concentration of portlandite and et- tringite in the portland cement mixtures in contact with pxous baked earth, Acad. Sci., C.R., Ser. D, Vol 271,4-7, 1970 lti Farsky, R., Basavarajaiah, B.S., Some studies of steam curing of cement mortar, Indian Concrete Journal, Vol. 45, No. 1, Jan., p35-41, 1971

101 Fataliev, S.A. and Imanov, A.M., Effect of reactive aggregates on the crystallization of new formations in concrete, Zh. Rikl. Khim. (Leningrad), V 45, n4, 752-5, 1972

102 Ftikos, Ch., Parissakis, G., Combined action of Mg2+ plus Cl- ions in cement pas@, Cem. ConCr. Res., v 15, n 4, Jul, 593-599,1985

103 German Committee for Reinforced Concrete, Recommendation on the Heat Treatment of Concrete, Deutscher Ausschuss fur Stahlbeton, DIN Deutsches Institut fur Normung e.b. Bundeaalle 216/218, D-1000 Berlin 15, Sept., pp13, 1989

104 Ghorab, H.Y. and Kishar, E.A., The stability of the calcium sulfoaluminate hydrates in aqueous solutions, Roe. 8th Intnl. Congress Chem. CemenL Rio de Janiero, V V, 104- 109, 1986

105 Ghorab, H.Y., Abou El Fetouh, S.H., New approach to the hydration reaction of the trical- cium aluminate with gypsum at 30°C. Part I: Effect of Lime, Zement-Kalk-Gips, v 38, n 5, May, 267-270,1985

106 Ghorab, H.Y., Heinz, D., Ludwig, U., Meskendahl, T., Welter, A., On the stability of cal- cium aluminate sulphate hydrates in pure systems and in cements, 7th Intnl. Congr. Chem. Cem., Vol IV, 496-503,1980

107 Ghorab, H.Y., Kishar, E.A., Kishar, E.A., Studies on the stability of the calcium sulfoalu- rninate hydrates: Part I: effect of temperatureon the stability of ettringite in pure water, Cem. ConCr. Res., v 15, n 1, Jan, 93-99,1985

108 Gibeson, W.E., A study of mapcracking of sand-gravel concrete pavements, Roe. High- way Research Board V 18, Pt. 1, 1938

109 Gills J.E., Petrology of dolomitic limestones, Kingston, Ontario, Camda, Bull. Geol. Sot. Amenc& Vol. 74, No. 6,759-778,1963 References Page 92

110 Gill@ J.E., Grabowsld, E., Czarnecki, B., Quinn, T., Duggan, C.R., Sco% J.F., Mecha- nism of expansion in rapid test method for alkali-aggregate reactiom Final Progress Re- po% Dept. of Civil Engineering, University of Calgary, Calgary, Canada, October, 1990

111 Gillo& J.E., Grabowski, E., Jones, T.N., Quinn, T., SW% J.F., Duggan, C.R., Mechanism of expansion in rapid test method for alkali-aggregate reaction, Regress Report #1, Dept. of Civil Engg., University of Calgary, Calgary, Cana@ May, 1989

112 Gillm J.E., Ritchie, T., A note on the occurrence of crystals of calcium hydroxide and et- tringite in voids in hardened Portland cement paste, Msg. ConCr.Res., V 17, No. 50, ~C~ 34,1965

113 Goldman, A., Bentur, A., Bond effects in high-strength silica-fime concrete& ACI Mate- rials Journal, Vol 86, No 5, Amer. Concr. Inst., Detroig 440-447, 1989

114 Goswami, G., Mohapatra, B., Pan@ J.D., Gypsum @hydration during commhmtion and its effect on cement properties, Journ, Amer. Cerm. Sot., v73, n3, Mar, 721-723, 1990

115 Goto, S., Daimon, M., Ettringite, Semento Konkuriito, n 408,36-9, 1981

116 Grabowsld, E., Gill~ J.E., Effect of replacement of silica flour with silica fume on engi- neering properties of oilwell cements at normal and elevated temperatures and pressures, Cem. ConCr. Res., V 19,333-344,1989

117 Grabowski, E., Gillo~ J.E., Modification of engineering behaviour of thermal cement blends containing silica fume and silica flour by replacing flour with silica W@ Cem. ConCr. Res., V 19,499-508,1989

118 Grabowski, E., Gillo& J.E., The effect of initial curing temperature on the performance of oilwell cements made with different types of silica Cem, Concr. Res., Vol 19,703-714, 1989

119 Grandet, J. and Ollivier, J.P., New method for the study of cement-aggregate interface, Roe. 7th InL Conf. Chem. Cement, Paris, Vol II, VII.85-VII.89, 1980

120 Greening, N.R., Landgren, R., Surface discoloration of concrete flahvor~ Journ., Port- land Cement Association Research and Development Laboratories, Vol. 2, No. 3, Septem- ber, Also PCA Research DepL Bulletin No. 203, 1960

121 Grounds, T., Midgley, H.G., Newell, D.V., Carbonation of ettringite by atmospheric carb- on dioxide, ‘Ihermochim. Acts, V 135,347-52, 1988

122 Grounds, T., Midgley, H.G., Newell, D.V., ‘Ihe use of thermal methods to estimate the state of hydration of calcium trisulfoaluminate hydrate, lhermochim. ACW V 85,215- 218, 1985

123 Hadley, D.H., l%e nature of the paste-aggregate interface, PhD thesis, University of Pur- due, 1972

124 Hampson, C.J., Bailey, J.E., On the structure of some precipitated calcium aluminosul- phate hydrates, Joum.of Materials Science, v 71, n 11, Nov, 3341-3346,1982

125 Hansen, W.C., A discussion of the paper ‘scanning electron micrographic studies of et- tringite formation by P.K. Mehta, Cem. Concr. Res., V 6, N 4,595-6, 1976 References Page 93

126 Hansen, W.C., OffutZ J.S., Roy, D.M. and Grutzek M.W., Gypsum & Anhydrite in Port- land Cemeng US Gypsum Co., 3rd editiom95pp, 1988

127 Hansom J.A., Optimum steam-curing procedures for structural lightweight concrete, Journ. Amer. ConCr. Inst., Vol 62, June, 661, 1965

128 Hansom J.A., Prestress loss as affected by type of curing, Journ. of Restressed Concr. Inst, Vol 9, NO 2, April, 1964

129 HansorL J.A., Optimum steam curing procedure in precasting plants, Journ. Amer. Concr. Inst., V 60, Jan., 1-25, 1963

130 Hassett+D,J., McCarthy, G.J., Kumarathasan, P., Pflughoeft-HassetL D., Synthesis and characterization of selenate and sulfate-selenate ettringite structure phases, Materials Res. Bulletim V 25, n 11, Materials Research Society, Pittsburgh, Nov, 1347-1354,1990

131 He, J., Scheetz, B.E., Roy, D.M., Hydration of fly ash-Portland cements, Cem. ConCr. Res., v 14, n 4, July, 505-512, 1984

132 Heinz, D., Ludwig, U., Mechanism of secondary ettringite formation in mortars and ccm- cretes subjected to heat treatment, American Concr. Inst., SP 100-105, Detroit, 2059- 2071, 1987

133 Heinz, D., Ludwig, U., Rudiger, I., Delayed ettringite formation in heat treated mortars and concretes, Concrete Precasting Plant and Technology, Issue 11/1989, 56-61, 1989

134 Hekal, E.E., Abo E1-Enein, S.A., Effect of compression on microstructrue and thermal sta- bility of ettringite, Proc. Intnl. Conf. on Cement Microscopy, 8th, Int. Cement Micros- copy Assoc., Duncanville, TX, 227-243, 1986

135 Higginson, E.C., Effect of steam curing on the important properties of concrete, Journ, Amer. ConCr. Inst., Vol 58, No. 3, Sept., 281-298, 1961

136 Hill, Eugene D., Letter to ACI Committee 517,m:ACI517.2R-87, November 7, 1990

137 Hjorth, L., !%udieopholdved Building Research Station, England-Rontgen-difWdctometri, cements hydratisering, I and II, Concrete Research Laboratory, Karktrup, Denmark 1966

138 Hobbs, D.W., Some tests on fourteen year old concrete affected by alkali-silica reactiom Proc. 7th Intrd. Conf. on Concrete Alkali-Aggregate Reactions, ed. P.E. Grattan-Bellew, NO)WS Publ.,N.J., USA, 1987

139 Hooton, R.D., Effects of carbonate additions on heat of hydration and sulfate nxistance of Portland cements, in Carbonate Additions to CemenL ed. P. Klieger and R.D. Hooton, ASTM Special Tech. Publ. 1064,73-81, 1990

140 Hoshino, M., Difference of the WICratio, porosiy and microscopical aspect between the upper boundary paste and the lower boundary paste of aggregate in concrete, Materials and Structures, V 21, 336-340, 1988

141 Hunter, D., Lime-induced heave in sulfate-bearing clay soils, Journal of Geotechnicrd Engg., Feb, Vol 114, No. 2, 150-167, 1988

142 Hunter, R.D., The geochemistry of lime-induced heave in sulfate-bearing clay soils, PhD thesis, Univ. of Nevada, Reno, 286pp, 1989 References Page 94

143 Idorn, G.M., Hydration of Portland cement paste at high tempwiture under atmospheric pressur, Principal Paper, Session III-% 5th Intnl. Symp. Chem. Cem., Tokyo, 1968

144 Impulsora Tlaxcalteca de Industrial, S.A. de C.V., Concrete Durability Research, Labora- tory Report, October 1991

145 Inderwic~ A.F., Should acid soluble sodium equivalent alkali be the measure of potential portland cement reactivity, Pm. 7th Intnl. Conf. on Concrete Alkali-Aggregate Reac- tions, ed. P.E. Grattan-Bellew, Noyes Publ., N.J., USA, pp456-459, 1987

146 Ingram, K., Poslusny, M., Daugherty, K and Rowe, W., Carboaluminate reactions as in- fluenced by Limestone Additions, in Carbonate Additions to Cement, ed. P. Klieger and R.D. Hooton, ASTM Special Tech, Publ. 1064,14-21,1990

147 Irassar, F., Batic, O., Effects of low calcium fly ash on sulfate Esistance of OPC cement, Cem. Concr. Res., V 19, N 2, Mar, 194-202,1989

148 Ish-Shalom, M., Bentur, A., Properties of type K expansive cement of pure components, Cem. ConCr. Res., 4,1974, 519-532; Cem. Concr. Res, 4, 1974, 709-721; Cem. Concr. Res. 5,139-152,1975

149 Isozald, K, Nakagaw& K, Behaviour of ettringite in high-strength hardened cement paste, 8th Intnl, Congress Chem, Cement, Rio de Janiero, V 3,401-6,1986

150 Isozaki, K, Nakagaw& K and Watanabe, Y., Studies on the strength of high ettringit.ic cement stones steam cured at normal pressure, Rev. Gen. Meet., Tech. Sess., Cem. Assoc. Jpn, V 32,43-3,1978

151 Johnston, C.D., Alkali-silica reactivity in concrete — importance of cement content and alkali eqdvalen~ Proc. 7th Intnl. Conf. on Concrete Alkali-Aggregate Reactions, ed. P.E. Grattan-Bellew, Noyes Publ., N.J., USA, 477-482,1987

152 Jones, F.E., ‘l%equaternary system CaO-A1203-CaS04-H20 at 25 C, Trans. Faraday SW,, 35,1484,1939

153 Jones, T.N., New interpretation of alkali-silica reaction and expansion mechanisms h con- crete,Chemistryand Industry (London), n 2, Jan 18,40-44, 1988

154 Jones, T.N. and Poole, A.B., Alkali-Silica Reaction in Several U.K Concretes: The Ef- fect of Temperature and Humidity on Expansion and Significance of Ettrlngite Develop- ment Proc. 7th Intnl. Conf. on Concrete Alkali-Aggregate Reactions, d P.E. Grattan-Bellew, Noyes Publ., N.J,, USA, p@46450, 1987

155 Jones, T.N., Grabowski, E., Quhm, T., Gillott, J.E., Duggan, C.R., SCOGJ.F., Mechanism of expansion in Duggan test for alkali-aggregate reaction, Canadian Developments in Testing Concrete Aggregate for Alkali Aggregate Reactivity, MinMry of Transportation, Ontario, Engg. Materials Report 92, Mar., Toronto, 70-82,1990

156 KalouseL G.L,, Sulphoaluminates of calcium as stable and metastable phases, PhD the- sis, University of Maryland 1941

157 Kalousek, G.L., Analyzing SOs-bearing phases in hydrating cement, Mat. Res. & Std., 6, 292-304,1965 References Page 95

158 Kalousek, G.L., Benton, E.J., Mechanism of seawater attack on cement pastes, Journ, Amer. Concr. Inst., G7-9, Feb., 187-193, 1970

159 Kayyali, O.A., Porosity of concrete in relation to the nature of the paste-aggregate inter- face, Materials and Structures, Vol 20, 19-26, 1987

160 Kelly, R., Solid-liquid reactions. Part 2: Solid-liquid reactions amongst the calcium alumi- nates and sulphoaluminates, Can. Journ. Chem., Vol 38, 1209-1226, 1960

161 Kemerley, R.A., Ettringite formation in dam gallery, Journ. Amer. Concr. Inst,, Proc. V 62, No. 5, May, 559-574, 1965

162 Kim K., Makino, Y., Mura@ Y., Dehydration and dehydration of ettringite, Gypsum Lime, V 170,7-13, 1981

163 Klemm, W.A., Adams, L.D., Investigation of the formation of carboaluminates, ASTM Special Tech. pub], v STP n 1064,60-72,1990

164 Klieger, P., Early high-strength concrete for prestressing, Portland Cement Association, Research Bulletin RX91, March, 1958

165 Klieger, P., Effect of mixing and curing temperature on concrete strength, Journ. Amer. Concr. Inst., V 54, No. 12, June 1958, 1063-1082. Also Portland Cem. Assoc. Research llep~ Bulletim RX103, 1958

166 Klieger, P., Some aspects of durability and volume change of concrete for prestressing, Journal, PCA Research and Development Laboratories, Vol 2, No 3, Sept., 2-12, Also PCA Research Dept. Bulletin RX 118, 1960

167 Kodama, K, and Nashino, T., Observation around the cracked region due to alkali-aggre- gate reaction by analytical electron microscope, Proc. 7th Intnl. Conf. on Concrete AlkAi- Aggregate Reactions, ed. P.E. Grattan-Bellew, Noyes Publ., N.J., USA, 398402, 1987

168 Kudryavtsev, A.B., Kouznetsova, T.V., Pyatkova, A.V., Proton relaxation and high reso- lution solid state Al 27 NMR study of the hydration of calcium oxide sulfoaluminate, Cem. ConCr. Res., v20, no 3, May, 407-418,1990

169 Kujal& K, Nierninen, P,, On the reactions of clays stabilized with gypsum lime, Improve- ment of Ground, 8th Eur. Conf. on Soil Mech. and Fndtn, Engg., Helsinki, V 2,929-932, 1983

170 Kumarathasam P., Studies supporting the disposal of North Dakota lignite ashes: hydra- tion behaviowy synthesis, and characterization of trace metal oxyanion substituted et- tringites, Dissertation, North Dakota State Univ. Agric. Appl. Sci., Fargo, ND, 207pp, 1990

171 Kumarathasam P., McCarthy, G.J., Hassett, D.J., Pflughoeft-HassetL D,F., Oxyanion sub- stituted ettringites: synthesis and characterization; and their potential role in irnmobL lization of arsenic, borom chromium, selenium and vanadium, Mater. Res. Sot. Symp, Proc., V 178, Materials Research Society, Pittsburgh, 83-104,1990

172 Kuzd, H., -J., Strohbauch, G., Reactions associated with the action of C02 on heat treated hardentxl cement pastes, ZKG International, Vol 42, Augus, p413-418, 1989 173 Lachaud, R., Thaumasite and ettringite in building materials, Annales de l’hstitut Tech- nique du Batiment et des Travaux Publics, n 370, Mar, 1-7,1979

174 Lachowski, E.E., Mohan, K., Taylor, H.F.W., Analytical electron microscopy of cement pastes: III — pastes hydrated for long times, Journ. Amer. Ceramic Sot., 64,6:319-321, 1981

175 Lane, D.S., Long-term mortar-bar expansion tests for potential alkali-aggregate reactivity, Rw. 7th Intnl. Conf. on Concrete Alkali-Aggregate Reactions, ed. P.E. Grattan-Bellew, NoyW PubI., N.J., USA, 336-341,1987

176 Larbi, J.A. and Bijen, J.M., Evolution of lime and ndcrostructural development in fly ash- Portland cement systems, Fly Ash &Coal Conversion By-Products: characterization, Utilization and Disposal VI, Materials Res. SOC.,Pittsbugh, VO1178. @. R.L. Day & F.P. Glasser, 127-138,1990

177 Lawrence, C.D., Dalziel, J.A., Hobbs, D.W., Sulphate attack arising from delayed et- tringite formation, Interim Technical Note, 12, British Cement Association, Wexham Springs, Slough, U.K., May, 1990

178 Le Roux, A. and Toubeau, P., Importance des mineraux accessoires des sois traites, XV Congress of the International Association of Engineering Geology, Buenos Aim, 915- 921,1986

179 Leifel& G., Muenchberg, W,, Stegmaier, W., Ettringke and thaumasite as causes of ex- pansions in lime-gypsum plasters, Zement-IWk-Gips, v 23, n 4, Apr, 174-7,1970

180 Lerck W., ‘l’heinfluence of gypsum on the hydration and properties of Portland cement pastes, Research Laboratory of the Portland Cement ASSOC.,Bulletin MOW sko~e. IL March, 1946

181 Lerc~ W., Ashton, F.W., Bogue, R,H,, Sulphoaluminates of calcium, Jourm of Res. of the National Bureau of Standards, Vol 2, No. 4,715,1929

182 LercL W., Bogue, R.H,, Heat of hydration of Portland cement pastes, Journal of Re- search, National Bureau of Standards, 12, No. 5, May, 645-664, 1934

183 LercL W., Ford, C.L., Long-time study of cement performance in concrete, Chapter 3, Journ. Amer. Concr, Inst., 44, April, 743-795,1948

184 Lewis, R.K,, A summary of investigations of steam curing of concrete related to cement characteristics, Constructional Review, March, 18-25, 1968

185 Lewis, R.K, The effect of delay prior to steaming and of temperature time gradient on the steam curing of lightweight concrete, Constructional Review (Sydney), Vol 36, No. 2, Feb., 20-25,1963

186 Lieske, H., New development in the tield of prestressed concrete sleepers, Betonwerk + Fertigteil-Technik, 52, H. 6, S. 381/383, 1986

187 Lobo, C., Cohen, M.D., Pore structure development in type K expansive cement pastes, Cem. Concr. Res., V 21, N 2-3, Mary-May, 229-241,1991

188 Lecher, D,Q., Richartz, W., Sprung, S., Influence of the calcium sulphate additive, Ze- ment-Kalk-Gips, 33,271-277, 1980 References Page 97

189 Lecher, F.W., Setting and initial hardening of cement, Zement-Kalk-Gips, 26,53-62,1973

190 Lou, Z., Xu, X., Ham R., On ettringite and modification of slag cemenu Int. Congr. Chem. Cem., 7th Proc., V 2, Paris, 11/82-11/87,1980

191 Lou, Z., Xu, X., Yang, L., Sheng, Q., Dissociation of aluminium from slag glasses and formation of ettringite, 8th Intnl. Congress Chem. CemenL Rio de Janiero, V 4,30-34, 1986

192 Ludwig, U., Effects of environmental conditions on alkali-aggregate xeaction and preven- tive measures, 8th Intnl. Conf. on Alkali-Aggregate Reaction, Kyoto, Japan, 583-596, 1989

193 Ludwig, U., Mehr, S., Destruction of historical buildings by the formation of ettringite or thaumasite, 8th Intnl. Congrms Chem. CemeenL Rio de Janiero, V5, 181-8, 1986

194 Ludwig, U., Mehr, S., Moldam D., Causes of damages to historical buildings and propos- als for a conception of mortars for restoration work Baumaschine und Bautechnik v 31, n 10, Ott, 396, 398,401,404, 1984

195 Lukas, W., Substitution of silicon in the lattice of ettringite, Cem, ConCr. Res., V 6, n 2, 225-33, 1976

196 Lukas, W., Roeck R., Hagleitner, K., Energy dispersive analysis of ettringites in cement paste samples, Zero. Kalk Gips, V30, n 7,328-30, 1977

197 Macleod, G., Hall, A.J., FallicL A.E., An applied mineralogical investigation of concrete degradation in a major concrete road bridge, Mineral. Msg., 55(377), 637-44, 1990

198 Mansfield, G.A., Curing — most efficient meth(xis determined by test, Rock produc~ Vol 50, Feb., 154, 1947

199 Mather, B., Discussion of paper Theories of expansion in sulfoahminate-type expansive cements: schools of thought, by M.D. Cohen, Cem. ConCr. Res., V 14, n 4,603-9, 1984

200 Mather, B., Expansive cement concrete-present state of knowledge, Comments, J. Amer. Concr. Inst., V 68, n 4,293,1971

201 Mather, B., New concern over alkali-aggregate reaction, presented at Joint Engineering Session of the National Sand and Gravel Assoc. and National Ready Mixed ConCr. As- soc., Jan 28, New Orleans, LA., pp20, 1975

202 McConnell, D., Murdoch, J., Crystal chemistry of ettringite (ABS.), Gecd. Sot. America Spec. Paper 68,228, 1962

203 McConnell, D., Murdoch, J., Crystal Chemistry of ettringite, Mineralog. Msg., V 33, No. 256,59-64, 1962

204 Mchedlov-Petroysan, O,P. et al, cd., Thermodynamic der silikate, VEB, Verhig, BerliQ Germany, 1965 (Thermodynamics of Silicates, Berlin/Heidelberg/New Yor~okyo, Translation of 4th edition, 1985)

205 Mehta, P.K., Mechanism of expansion associated with ettringite formatiom Cem. ConCr. Res., Vo] 3, 1-6, 1973 References Page 98

206 Mehta, P.K, Scanning electron micrographic studies of ettringite formation Cem. ConCr. Res., Vol 6,169-182,1976

207 Mehta, P.K, Morphology of csJcium sulphoaluminate hydrates, Journ. American Ce- llUlliCSot., V 52, No. 9,521, 1969

208 Mehta, P.K, Expansion characteristics of calcium sulphoaluminate hydrates, Journ. Amer. Cersm. Sot., 50,204-208,1967

209 Mehta, P.K, Mechanism of sulfate attack on Portland cement concrete - another loolG Cem. ConCr. Res., v 13, n 3, May, 401-406,1983

210 Mehta, P.K, Reply to discussion by S. Chatterji of :mechanism of sulfate attack on port- land cement concrete - another look, Cem. Concr, Res,, V 14, n 1, 153, 1984

211 Meh@ P.K, Reply to discussion by S. Chatterji of scanning electron micrographic stud- ies of ettringite formation, Cem. Concr. Res., V 6, N 5,713-14,1976

212 Mehta, P.K, A reply to W.C, Hansen’s discussion of the paper scanning electron rni- crographlc studies of ettringite formatiom Cem. Concr. Res., V 6, n 4,597-8, 1976

213 Mehta, P.K, Scanning electron micrographic studies of ettringite formatlo~ Cem. ConCr. Res., V 6, n 2,169-82,1976

214 Mehta, P.K, Stability of ettringite on heating, J, Amer. Cerarn. Sot., V 55, n 1,55-6,1972

215 Mehta, P.K., Hu, F., Further evidence for expansion of ettringite by water adsorption J. Am. Ceramic Sot., V 61, n 3-4, p179-81, 1978

216 Mehta, P.K, Klein, A,, Formation of ettringite by hydration of a system containing anhy- drous calcium sulfoaluminate, Joum, Amer. Ceram. Sec., V 48, No, 8,435-436,1965

217 Mehta, P.K,, Wang, S., Expansion of ettringite by water adsorption, Cem. ConCr. Res., V 12, n 1,121-2,1982

218 MerriK R.R., Johnson, J.W., Steam curing of Portland cement concrete at atmospheric pressure, Bulletin 355, Highway Research Board, 1-26, 1962

219 Midgley, H.G., Pettifer, K, The microstructure of hydrated supersulphated cement Cem. Concr. Res., V 1,101,1971

220 Mikhail, R, Sh., Abo-El-Enein, S,A,, Hantil, S,, Good, R.J., Ettringite formation in com- pressed expansive cement pastes, Cem. Concr. Res., v 11, n 5-6, Sep-Nov, 665-673,1981

221 Moldovam V,, Butucescu, N., Expansion mechanisms of expansive cements, InL Congr. Chem, Cem., 7th Pm,, V 3, Paris, V/1-V/5, 1980

222 Moldovan, V. and Paul, F., Studies on expansion mechanism based on ettringite forma- tion, Mater. Const. (Bucharest), V 10, N 3,123-7,1980

223 Monosi,S., Collepardi, M., Effect of silica fume on the hydration products of C3A- CS(bar)H2 system, Proc. 8th Intnl. Congress Chem. Cement, Rio de Janiero, V III, 120- 124, 1986 References Page 99

224 Montanaro, L., Negro, N., Expansion of the ferrite phase in the presence of gypsum and lime: comparison with the tricalcium aluminate, Cem. Concr. Res., v 18, n 5, Sept, 812- 818,1988

225 Monteiro, P.J.M., Gjorv, O.E., MehQ P.K,, Effect of condensed silica fume on the steel- cement paste transition zone, Cem. Concr. Res., v 19, n 1, Jan, 114-123, 1989

226 Monteiro, P.J,M., Maso, J.C., Ollivier, J.P., Aggregate-mortar interface, Cem. ConCr. Res., v 15, n 6, Nov, 953-958, 1985

227 Monteiro, P.J.M., Meh@ P.K, The transition zone between aggregate and Type K expan- sive cemen~ Cem. Concr. Res., V 16, No, 1, 111-114, 1986

228 Monteiro, P.J.M,, Meh@ P.K, Ettringite formation on the aggregate--cement paste inter- fac, Cem. Concr. Res., V 15, No 2, March, 378-380, 1985

229 Monteiro, P.J.M., Meht& P.K., Interaction between carbonate rock and cement paste, Cem. ConCr. Res., V 16, No 2, March, 127-134, 1986

230 Monteiro, P.J.M., Ostertag, C.P., Analysis of the aggregate-cement paste interface using grazing incidence X-ray scattering, Cem. Concr. Res., V 19,987-988, 1989

231 Moore, A.E., Taylor, H.F.W., Crystal structure of ettringite, Acts Crystallogr., V 26, Sect. B, Part4, 386-393, 1970

232 Mortureux, B., Hornain, H., Regourd, M., Role of C3A in the attack of cements by sul- fate solutions, 7th Intrd. Congress Chem, Cem., Paris, V 4,5704,1980

233 Murat, M., Internal morphology of fiber-reinforced cement, Cem. ConCr. Res., v 4, n 2, Mar, 327-333, 1974

234 Nakagawa, K, Watanabe, Y., Nakaya, S., Ettringite under high-temperature, high-prm- sure, Rev. Gen. Meet., Tech. Sess., Cem. Assoc. Jpn., V 30, 34-6, 1976

235 Nakamura T., Sudoh, G., Akavi4 S,, Proc. 5th Int. Symp. on Chemistry of Cement TO kyo, V 4,35, 1968

236 Nasser, K, Lai, P.S.H., Effect of fly ash on the microstructure and durability of concrete, Serv. and Durability of Construction Materials, Proc. 1st Materials Engg. Congress, Part 2, Denver, Co, Aug 13-15, ASCE, 688-697, 1990

237 NeclL U., Auswirkungen der Warmbehandlung auf Festigkeit un Dauerhaftigkeit von Be- ton, Beton+Vol 38,488-489, 1988

238 Negro, A., Bacchiomini, A., Expansion associated with ettringite formation at different temperatures, Cem. Concr. Res., 12,677-684, 1982

239 Negro, A., Bachiorrini, A., Reply to a discussion by S. Chatterji if expansion associated with ettringite formation at different temperatures, Cem. Concr. Res., V 13, n 4,589-90, 1983

240 Gdler, I., 318, Interaction between gypsum and CSH-phase formed in C3S hydration, 7th Intnl. Cong. Chem. Cem., Paris, Vol IV, 493-495, 1980 References Page 100

241 Odler, I., Abdul-Maula S., Possibilities of Quantitative determination of the Aft-(et- tringite) and Afm-(monosulphate) phases in hydrated cement pastes, Cem. ConCr. Res., v 14, n 1, Jan., 133-141,1984

242 Odler, I., Abdul-Maula S., Lu, Z., Effect of hydration temperature on cement paste struc- ture, Microstructural Development During Hydration of Cements, Proc. Materials Re- search Society, L. Struble & P. Brown eds., Materials Research Society, Pittsburgh Vol 85,139-144,1987

243 Ogawa, K., Roy, D.M., Hydration of 3Ca0.3A1203.CaS04, formation of ettringite, and its expansion mechanism. I. Expansion and stability of ettringite, Onoda Kenkyu Hokoku, V 33, n 106,86-94, 1981

244 Ogawa, K, Roy, D.M., C4A3S hydration, ettringite formation, and its expansion mecha- nism, III. effect of calcium oxide, sodium hydroxide and sodium chloride; conclusions, Cem. ConCr. Res., V 12, n 2,247-56,1982

245 Ogawa, K., Roy, D.M., C4A3S hydration, ettringite formation, and its expansion mecha- nism, IL microstructural observation of expansion, Cem. Concr. Res., V 12, n 1, 101-9, 1982

246 Ogawa, K, Roy, D.M., C4A3S hydration ettringite formation, and its expansion mecha- nism: I: expansion; ettringite stability, Cem. Concr, Res., V 11, n 5-6,741-50,1981

247 Okushim% M,, Kondo, R., MugusumA H., One, Y., Proc, 5th In. Symp, on Chemistry of Cement, Tokyo, Vol 4,419,1968

248 One, K, Assessment and Repair of Damaged Concrete Structure, 8th Intnl. Conf. on Al- kali-Aggregate Reaction, Kyoto, Japan, 647-658,1989

249 Ozol, M.A., Alkali silica reaction of concrete in electrical substation piers sccderated by electric curren~ ASTM Spec. Tech. PubI,, 1061,106-120, 1990

250 Panchenko, A.I., Control of expansion and structure formation of expansive cement Cem, Concr, Res., V 20, n 4, Jul, 602-609,1990

251 Peacor, D.R., Dunn, P,J., Duggan, M., Sturmanite, a ferric iron, boron analogue of et- tringite, The Canadian Mineralogist Nov., V 21, Part 4,705-709,1983

252 Pettifer, K,, Nixon, P.J., Alkali metal sulphate — a factor common to both alkali aggre- gate mction and sulphate attack on concrete, Cem, Concr, Res., V 10,173-181,1980

253 Pettifer, K,, Nixon, P.J., A reply to a discussion by D.A. St. John of AlkaU metal sulfate — a factor common to both alkali aggregate reaction and sulfate attack on concrete, Cem. ConCr. Res, V 11, n 5-6,801-2,1981

254 Pfeifer, D.W., Landgrem J.R., Energy-efficient accelerated curing of concrete, a labora- tory study for plant-produced prestressed concrete, A Research Investigation by Wiss, Janney, Elstner & Associates Inc., Ill., for Prestressed Concrete Instititute, Chicago, De- cember, 1981

255 Pfeifer, D,W., Marusin, S,, Energy-efficient accelerated curing of concrete, a state of the art review, Research Investigation by Wiss, Janney, Elstner & Associates, 111,,for Prestressed Concrete Institute, Chicago, March, 1981 References Page 101

256 Ping, X. and Beaudoin, J.J., Mechanism of Sulphate Expansion I. Thermodynamic Princi- ple of Crystallization Pressure, private communication, 1991

257 Ping, X. and Beaudoin, J.J., Mechanism of Sulphate Expansion II. Validation of Thermo- dynamic ‘Ilmry, private communication, 1991

258 Plowmam J.M., Maturity and the strength of concrete, Msg. ConCr. Res., 8, No. 24, Nov, 175-178,1956

259 Poellman, H., Kuzel, H.J., Wenda, R., Solid solutions of ettringites, Part I: incorporation of hydroxide and carbonate ions in 3Ca0.AlzOs.3CaSO+ 32Hz0, Cem. ConCr. Res., V 20, No 6,941-7, 1990

260 Poellmann, H., Kuzel, H,J., Wend& R., Compounds with ettringite structure, Neues Jahrb. Mineral., Abh., V 160, n 2,233-58,1989

261 Polivh M., Mehta, P.K, Baker, J.A. Jr., Freeze-thaw durability of shrinkage-compensat- ing cement concrete, Dumb. of Concrete, 79-87, pub]. by Amer. ConCr. Inst., Sp-47, De- troit, 1975

262 Pommersheim, J., Chang, J., Kinetics of hydration of tricalcium aluminate in the presence of gypsum, Cem. ConCr. Res., V 18, No6,911-22, 1988

263 Poole, A.B., Thomas, A., Staining technique for the identification of sulfates in aggregaes and concretes, Mineral. Msg., V 40,n311, 315-16, 1975

264 Portill& M., Enthalpy of dehydration of Portland and pozzolanic cements, Cem. ConCr. Res., v 19, n 3, May, 319-326, 1989

265 Ramachandran, V.S., Zhang, C-M., Hydration kinetics and microstructural development in the 3Ca0.A120s.CaSOA.2Hz0 .CaCOs.HzO system, Mater. Struct., v 19, n 114, Nov- Dec, 437-444,1986

266 Reading, T.J., Physical aspects of sodium sulfate attack on concrete, Am. ConCr. Inst., Publ. SP 77, George Vertwck Symp. Sulfate Resist. Concr., Detroit, 75-81, 1982

267 Regourd-Moronville, M., Products of nxiction and petrographic examination, 8th Intnl. Conf, on Alkali Aggregate Reaction, Kyoto, Japan, 445-449,1989

268 Research Institute of the Cement Industry, Chemical Reactions in Concrete, Activity Re- pofi 1981-1984, Translated by TEX Associates Inc., Philadelphia, 1990

269 Richards, J.D., The effect of various sulphate solutions on the strength and other proper- ties of cement mortars at temperatures up to 80”C, Msg. ConCr. Res., V 17, No. 51, June, 69-76, 1965

270 Richartz, W., Combining of chloride in the hardening of cemen~ Zement-Kalk-Gips, v 22, n 10, Ott, 447-456, 1969

271 Rogers, C.A., Alkali-Aggregate Reactivity in Canada, 8th IntnL Conf. on Alkali-Aggre- gate Reaction, Kyoto, Japan, 57-63, 1989

272 Rogers, C.A., Hooton, R.D., Leaching of alkalis in alkali-aggregate mction testing, Roe. 8th Int. Conf. on Alkali Aggregate Reaction, Kyoto, Japan, July, 17-20, 1989 References Page 102

273 Sagrera, J.L., Study on the durability of sulphate application of le Chatelier-Anstett method to an ordinaryportland cemen~ Cem. Concr. Res., v 2, n 3, May, 253-260, 1972

274 Saito, M,, Tensile fatigue strength of the mortar-aggregate bond, Cem. Concr. Res., Vol 18,710-714,1988

275 Saito, M., Kawamur& M., Resistance of the cement-aggregate interracial zone to the propagation of cracks, Cem. ConCr. Res., Vol 16,653-661,1986

276 Saito, M., Kawamur& M., EiTect of fly ash and slag on the interracial zone between ce- ment and aggregate, Roe. 1989 Trondheim Conf., Amer. Concr. Inst. publication%SP- 114, Detroit, p669-688, 1989

277 Satav& V., Vepre~ O., Thermal decomposition of ettringite under hydrothermal condi- tions, J. Am. Ceram. Sot., V 58, n 7-8,357-9,1975

278 Saul, A.G.A., Rinciples underlying the steam curing of concrete at atmospheric pressure, Msg. Concr. Res., Vol. 2, No, 6, March p127-140, 1951

279 Scheetz, B.E., Roy, D.M., Duffy, C., Elevated temperature (hydrothermal) stability of ce- mentitious sealants for a deep geological ~pository in tuff, Waste Management, v 9, n 4, 253-259,1989

280 Schlorholtz, S., Demirel, T., Autoclave expansion of Portland cement-fly ash pastes, Mat. Res. Sot., Symposium Proc., V 43, Mater. Res. Society, Pittsburgh, PA, 85-93,1987

281 Schlorholtz, S., Demirel, T,, Quick lime-gpsum interaction in stabilized soil bases fa concrete highways, Cem, Concr. Res., v 14, n 4, July, 529-532,1984

282 Scholer, C.H., A wetting and drying test for predicting cement-aggregate reactions, Pre- sented at the Fifty-Second Annual MeW.ingof ASTM, June27-Julyl, ASTM Roe. V, 942- 953,1949

283 Schwiete, H.E., Ludwig, U,, AlbeclGJ., Combining of calcium chloride and calcium sul- phate in hydration of ahnn.inate-femitic clinker constituents, Zement-Kalk-Gips, v 22, n 5, May, 225-34, 1969

284 Scott, J.F., Duggan, C.R., Potential new test for alkali aggregate mctivity, Roe. 7th Intnl. Conf, on Alkali Aggregate Reactions, Ottawa Canada, @. P.E. Grattan-Bellew, Noyes pub]., N.J., USA, 319-323, 1986

285 Sel.igmann, P. and Greening, N.R., Studies of early hydration reactions of Portland ce- ment by X-ray diffraction, Highway Research Record, No. 62, 1964

286 Shayan, A., Experiments with accelerated tests for predicting alkali-aggregate reactivity, 8th Intnl. Conf, on Alkali-Aggregate Reaction, KyotQ,Japan, 321-326, 1989

287 Shayam A. and Lancucki, C.J., Alkali-aggregate reaction in the causeway bridge, Pa% Western Australia Proc. 7th Intnl, Conf. on Concrete Alkali-Aggregate Reactions, ed. P.E. Grattan-Bellew, Noyes pub]., N.J., USA, 321-326,1987

288 Shayam A., and Quick G., Microstructure and composition of aar products inconven- tional standard and new accelerated testing, 8th Intnl. Conf. on Alkali-Aggregate Reac- tion, Kyoto, Japan, 475-482, 1989 References Page 103

289 Sheiken, A.E. and Oleinikova N.I,, Effect of hydrothermal treatment and t%mess of grinding of cement on the stmcture and properties of cement stone, Preprint, RILEM Symposium, Moscow, USSR, Session I/17, 1964

290 Skoblinskaya, N.N., Krasil’nikov, K.G., Changes in crystal structure of ettringite on dehy- dration I, Cem. ConCr. Res., V 5, n 4,381-93,1975

291 Skoblinskaya, N.N., Krasil’nikov, KG,, Nikitin& L,V., Varlamov, V.P., Changes in the crystal structure of ettringite on dehydration. II, Cem, Concr. Res,, V 5, n 5,419-31, 1975

292 Soro@ I., Jaegermann, H., Bentur, A., Short-term steam-curing and concrete later-age strength, Materials and Structures, V 11, No. 62,93-96, March/April, 1978

293 St. John, D.A., A discussion of the paper ‘Alkali Metal Sulphate — a factor common to both alkali-aggregate reaction and sulphate attack on concrete’ by K Pettifer and P.J. Nixon, Cem. Concr. Res., V 11,799, 1981

294 Stadelmann, C., Hem, R., Wicker, W. and Ilon& K, Determination of ettringite in hard- ened Portland cement pastes, SillikattechniL V39, N 4, 120-2, 1988

295 Stefani& V., Abbiati, G., Negro, A., Bachiorrini, A., Mathematical model for the et- tringite expansion, Cemento, V 79, n 2, 101-6, 1982

296 Struble, L., Synthesis and characterization of ettringite and related phases, Proc. Intrd. Congress Chem. Cem., (8th), Rio de Janiero, Sept 22-27, p582-588, 1986

297 Sylla, H., M., Reaktionen im Zementsteine durch Warmebehandlung (reactions in cement stone due to heat treatment), Beton, Vol 38,449-454, 1988

298 Tash.ire, C., Ob& J., Akama, K, Reply to discussion of the effects of several heavy metal oxides on the formation of ettringite and microstructure of hardened ettringite, Cem, Concr. Res., V 10, n 1, 121, 1980

299 Tashiro, C., Ob& J., Akama, K., Effects of several heavy metal oxides on the formation of ettringite and microstructure of hardened ettringite, Cem. ConCr. Res., V 9, n 3, 303-8, 1979

300 Taylor, H.F.W., Crystal structures of some double hydroxide minerals, Mineral. Msg., V 39, No. 304,377-389,1973

301 Taylor, H.F.W., Bound water in cement pastes and its significance for pore solution com- positions, Meter. Res. Sot. Symp. Proc., V 85, Microstr. Dev. Hydration Cem., 47-54, 1987

302 Taylor, H.F.W., Cement Chemistry, Academic Press, Londom 1990

303 Tepponon, P., Erickson, B.E., Darnages in Concrete Railway Sleepers in Finland Nordic ConCr. Res. Publication #6, 199-209, 1987

304 Tezu@ Y., Djanikian, J.G., Uchikawa, H. and Uchid& S., Hydration characteristics and properties of mixtures of cement and high content of calcium sulfate, Proc. 8th Intrd, Con- gress Chem. Cement Rio de Janiero, V III, 323-329,1986 References Page 104

305 Tognon, G.P. and Cangiano, S., Strength and Interface Bond in Concrete, in Bond in Con- crete, ed. by P. Bartos, Applied Science Publishers, Londom 4-14, 1982

306 Tsuyuld, N., Hirota, N., Miyakawa, T. and Kasai, J., The physical properties and the hy- dration mechanism of C3A in the presence of Cas04.2H20 and Ca(0H)2, PCOC.8th In~. Congress Chem. Cement Rio de Janiero, V V, 400-404,1986

307 Tsuyuki, N., (h, R., Miyakawa, T., Kasai, J., Effects OfSodium gkonti on the fo~- tion of ettringite, Journ. of Ceramic Society of Japan, v98, n 4,332-338,1990

308 Uchikawa H., Uchida, S,, Analysis of ettringite in hardened cement paste, Cem. ConCr. Res., v 4, n 5, Sept, 821-834, 1974

309 Van Aardg J.H.P., Visser, S., Reaction of Ca(OH)2 and of Ca@H)2+C*@.2H@ ~ v~- ous temperatures with feldspars in aggregates used for concrete making, Cem. ConCr. Res., v 8, n 6, Nov, 677-681,1978

310 Verbeck G., Cement hydration nmctions at early ages, Portland Cement Assoc., Research Dept., Bulletin 189, 1965

311 Verbeck G., Copeland, L.E., Some physical and chemical aspects of high pressure steam curing, Amer. Cone, Inst., publication SP-32, Menzel Symposium on Wgh Pressure Steam Curing, Detroit 1972

312 Verbeck G.J., Helmuth, H.H., Structure and physical properties of cement pastes, Princi- pal Paper, Session III-1, 5th Intnl. Symp. Chem. Cement, Tokyo, pi-44, 1968

313 Verne4 C., Sequence and kinetics of hydration reactions of C3A with gypsum, lime and calcareous fillers, Pm. 8th Intnl. Congress Chem. Cemen6 Rio de Jankro, V III, 70-74, 1986

314 Volkwein, A., Springenschmid, R,, Corrosion of reinforcement in concrete bridges at dif- ferent age due to carbonation and chloride Penetration 2nd Intnl. Conf. on Durability of Building Materials and Components, Sept 14-16, Publ, by NBS, Washington D.C., 199- 209,1981

315 Wang, S., Ji, S., Liu, Y., Hu, K., Effect of alkali on expansion of sulfoaluminate Cemeng Gyisuanyan Xuebao, V 14, No 3,285-92,1986

316 Way, S.J., Shayan, A., Early hydration of a Portland cement in water and sodium hydrox- ide solutions, Composition of solutions and nature of solid phases, Cem, Concr. Res., v 19, N 5, Sept, 759-769,1989

317 Wells, J.A., Attiogbe, E.K. and Rizkall& S.H., Deleterious expansion of cement paste sub- jected to wet-dry cycles, preprint, 1990

318 Whittaker, J.B., Low temperature curing of precast prestressed concrte products, Journ. Restressed Cone. Inst., July-Aug, 40-49,1977

319 Wicker, W., Herr, R., Wirdder, A,, Connexions between the concentration of hydroxide - and sulphate ions in pore solutions of heat-treated cements and the expansion by secm- dary ettringite formation, XV Conference on Silicate Industry and Silicate Science, Buda- pes~ 1989.6.12-16,392-395, 1989 References Page 105

320 Wicker, W., Herr, Roland, Some problems concerning the chemistry of Portiand cement, Zeitschrift fur Chemie, September, pl- 12,1989

321 Wierig, H. -J., Kurzzeitwarrnbehandiung von Beton (rapid heat treatment of czmcrete), Betonwerk + Fertigteil-TechniL 41 (1975) H. 9, S 418/423 und H, 10, S. 492/495, 1975

322 Wischers, G., Einfluss der Zussamensetzung des Betons auf seine Fruhfestigkeit (infiu- ence of the composition of concrete on its early strength), beton, 13, (1963), h. 9,s, 427/432: Ebenso Betontechnische Berichte 1963, Beton-Verlag, Dusseldorf, S. 137/151, 1964

323 Wischers, G,, Memorandum concerning production of closed concrete surfaces with heat treatment, Betontechnische Berichte, 35-50, 1967

324 Worthington J.C., Bonner, D.G., Newell, D.V., Influence of cement chemistry on chio- ride attack of concrete, Materials Science and Technology, V4, N 4, Apr, 305-313, 1988

325 Zapl’skii, A.K., Nadel, L.G., Deshko, 1.1.,Pavlova, N.A., Hydration and structure forma- tion in the system beta C2S-Krent-Gypsum-Water, Journ. Applied Chem. of the USSR (Engiish translation of Zhurnai Prikiadnoi Khirnii), V 59, N 5, pt 2, May, 1084-1087, 1986

326 Zhang, F., Zhou, Z., Lou, Z., Volubility product and stability of ettringite, 7th Intnl. Con- gress Chem. Cem., V 2, Paris, 11/88-11/93,1980

327 Zimbelman, R., A method for strengthening the bond between cement stone and aggre- gates, Cem. Concr, Res., Vol 17,651-660, 1987 References Not Found Page 106

References Not Found

1 Alesz@ J., Schnittgrund, G., 7, NTIS, An evaluation of the fracture of plain concrete, fi- brous concrete, and mortar using the scanning electron microscope, Army Construction Engg. Research Lab, Champaign Ill., Mar, 1975, 67p, Final Report on Characterization of Fracture

2 Altner, W,, Reichel, W., 11, 155, Betonschneilhartung; Grundlagen und Verfahren (rapid hardening of concrete; basics and procedures), Beton-Verlag, Dusseldorf, 1981

3 Azum& T., Mineshi@ Y., Ichimaru, K, 16, CA, Reaction retardation in ettringite manu- facture, Patent, Matsushita Electric Works Ltd., Japan Kokai Tokkyo Koho JP 77104530, Sept 2,1977

4 Bachiorrini, A., Negro, A., 17, CA, Study of the expansion of tricalcium aluminate caused by calcium sulfate dihy&ate, Atti Accad. Sci. Torino, Cl. Sci. Fis., Mat Nat., V 115, no 3-4, p 125-9,1983

5 Baykal, G. I., 23, GEO, The effect of micromorphological development on the elastic moduli of fly ash-lime stabilized bentonite, Doctoral thesis, Louisiana State Univ., Baton Rouge, 1987,283 pp

6 Bensted, J., Varm& S.P., 29, CA, Sulfoaluminate phase, Klei Keram., V 23, 1973, n 8, p 61-3

7 Berra, M., 34, CA, Shrinkage-compensating concretes and their durability, VTT Symp., 1984, v 50, Intnl. conf. durability build mater. compon., 3r& v 3, p 267-79

8 Candlo4 E.,51, 117, Proprieties des ciments et des liants hydrauliques, Bull, sec. Encour. Ind. Nat., 1890,682-685

9 Changgeng, L., 56, 155, Industrielle Hersteihmg von Betonfertigteilen in China (indus- trial production of ready-made concrete parts in China), Betonwerk + Fertigteil-Technik 52 (1986) H. 5, S. 318/322

10 Concrete Institute of Australi% 78,23, Third progress reprt of the low pressure steam- curing of concrete, 1972

11 Courtois, A., Dusausoy, Y., Laffaille, A., Protas, J., 80, GEO, Preliminary study of the crystalline structure of ettringite, Acad Sci., C,R., Ser. D, 1968, V 266, No 19, 1911-1913

12 Courtois, A., Dusausoy, Y., LafMle, A., I%otas,J., 81, CA, Preliminary studies on the crystalline structure of ettringite, C.R. Acad. Sci., Paris, Ser. D, V 266, 1968, n 19, p 1911-13

13 Deloye, F.-X., Bernard, A., PoindeferL A,, 90, 132, Approche analytique de la dedolorniti- sation des granulats clansle beton durci, VII Intnl. Symp. on Chem. Cement, Paris, 3, June-5 July 1980; Bull. liaison Labo. P. et Ch., 117 (Jan-Feb): 82-84

14 Deloye, F,X., Louarn, N., Loos, G., 91, EI, Examples of masonry analyses. Case of the Puberg tunnel, Bulletin de Liaison des Laboratoires des Ponts et Chaussees, n 163, Sep Ott 1989,17-24 References Not Found Page 107

15 Dollmayer, M., Vadasz Rusko, E., Nemeth Novak, I., Krudy, C, et al, 95, GEO, Manufac- ture of rapid-setting pozzolan concrete and mortar, Hung.TeljesHU52011 A2, 28 Jun 1990, 11 pp

16 Dron, R., Brivot, F., 100, CA, Ettringite-caused expansion, 8th Intnl. Congress Chem, Cem., Rio de Janiero, V 5, p 115-20, 1986

17 Duggan, C.R., Scow J.F., 101, 21, Concrete durability: avoidance of deleterious expan- sion, paper presented at ASTM Committee C090202 meeting, 1988

18 Du~ P.J., Peacor, D.R., Leavens, P.B,, Baum, J.L., 104, GEO, Charlesite, a new mineral of the ettringite group, from Franklin, New Jersey, American Mineralogist Ott 1983, V 68, No. 9-10, p 1033-1037

19 Erlin, B., Stark, D.C., 108, 155, Identification and occurrence of thaumasite in concrete, Highway Res. Rec., 1965, Nr. 113, S. 108/113

20 Fataliev, S.A., 114, CA, Conditions for crystallization of portlandite and ettringite in con- crete with reactive rocks, Mineral. Rudn. Mestorozhd, Editoc Kashkai, M.A., 1974, p 167-73

21 Feldman, R,F., 116, 147, DiscusSon on R, Kondo and S. Ueda’s paper on Kinetics and Mechanisms of the hydration of cements, Proc. 5th Int’1Symp. on the Chemistry of Ce- ments, Cement Association of Japan, Tokyo, 1968, p. 252

22 Filatov, L.G., Kuz’mina, L.A., Kuznetsov, S.N., 117, GEO, Investigation of the conver- sion of ettringite and portkmlite in cement during reactions involving both negative and positive temperature, in Eksperimental’nyye Issledovaniya Mineraloobrazovaniya v Suk- hikh Okisnykh I Silikatnykh Sisteman p 171-174, 1972, Moscow

23 Filatov, L.G., Kuz’mina, L.A., Kuznetsov, S.N., 118, CA, Transformation of ettringie and portlandite in concrete under the effect of negative and positive temperatures, Eksp, Issled Mineraloobrazov. Sukhikh Okisnykh Silikat. Sist., Lapin V.V., (cd), 1972, p 171-5

24 Freedman, S., 119,23, High-strength concrete, Portland Cement Association (IS 176.OIT), Concrete Information, 1971

25 Gabadadze, T.G., 121, CA, Mechanism of expansion of cements, Izv. Akad Nauk Gruz. SSR, Ser. Khim. V 4,1978, p 358-62

26 Galibina, E.A., 122, CA, Effect of free calcium oxide and ettringite on structure forma- tion of highly basic shale ashes, Stroit. Matex., n 4, 1980, p 21-2

27 Ghorab, H.Y., Ludwig, U., 127, CA, Model studies to clarify damage causes on heat- treated concrete. Part I. stability of monophases and ettringites, TIZ, V 105, n 9, 1981, pp 634,636-40

28 Gusev& V.I., Rozman, D.A., 144, CA, Interrelation between properties of stressing ce- ment and the morpholoy of the resulting ettringite, Nov. Effekt. Vidy Tsemen@ 1981,70- 5

29 Hanada, M., Nakazaw% S., Shiroi, T., 147, CA, Static demolishing agent for concrete and rock, PatenC Japan Kokai Tokkyo Koho; JP 8759687 A2, 1987, Yoshizawa Lime Indus- try Co., Ltd. References Not Found Page 108

30 Heinz, D., 157, 19, Schadigende Bildung ettringltahnlicher Phasen in warmebehandelten Morteln und Betonem Thesis, Technical high school, Aachen, Dec. 1986

31 Heinz, D., 158, NTIS, Formation of a phase similar to ettringite in heat-treated mortar and concrete, Technische Hochschule Aachen Fakultaet Fuer Bergbau, Huettenwesen und Geowissenschaftem 10 Dec 1986, 209p

32 Heinz, D., Ludwig, U., Nasr, R,, 161, 10, Model experiments for clarification of the causes of damage to heat treated pm-fabricated concrete elements: Part 2- heat treatment of mortar and late ettringite formatiom Tonind. -Ztg 106, 1982, 178-183

33 Hodawand-Khani, M., Dosch, W., 166, CA, Strontium aluminate sulfate, Fortschr. Min- eral., V 53, n 1, 1975, p 30

34 Howard, L.L., 168,23, Notes from PCI energy conservation in curing, Prestressed ConCr. Inst., March 1980

35 Isozald, K,, Nakagawz K., 177, CA, Hydrothermal KXWtionsbetween ettringite and inor- ganic salts, Rev. Gen. Meet., Tech. Sess., - Cem. Assoc. Jpn., v 31, 1977, p 45-7

36 JambOr, J., Zivica, V,, BlazelG A (ala)., 179, GEO, Investigation of the relative resistance of hydration products of cement against comosion due to aggressive carbon dioxide water, Thermal Analysis, Proc. of ICTA (NEW), 8th, 1985, 2, 605-8, Bratislava, Czech

37 Karapchansky, N., Pejev, A., 187,23, Study of some problems relating to steam curing of concretes, RILEM, Moscow, July, 1964

38 Keene, P.W., 189,23, Notes on curing concrete in steam at atmospheric pressure, Con- crete/Beton, No. 1, March, 1976

39 Kir& K., Makino, Y,, Yamamoto, H,, 192, CA, Carbonation of synthesized ettringite, Sekko To Sekkai, V 159, 1979, p 48-53

40 Klavins, Z., Alksnis, F., Kauke, A., Baumanis, O., 193, CA, Effect of active silica on the interaction of tricalcium aluminate with gypsum, Neorg. Stekl& Pokrytiya Mater., V 4, 1979, p 117-24

41 Knoefel, D., Boettger, K.G., 198, NTIS, The influence of a sulfur dioxide-enriched atmos- phere on concrete, GesellschafC FortscMtte der ,Beiheft (Deu) Aug 1984, V 62, No, 1, p 127-128

42 Kollmann, H,, 199, CA, Mineralogical studies on efflorescence and the appearance of ex- pansion in structural material in the presence of sulfate, Giessener Geol. Schr., V 18, 1978, 159pp

43 Kollmann, H., Struebel, G., 200, GEO, Etlringite-thaumasite mixed crystals in Brenk Eifel, Chemie der Erde, 1981, V 40, No. 2, p 110-120

44 Kollmann, H., Struebel, G., 201, CA, Experimental studies on the question of solid solu- tion formation between ettringite and thaumasite, Fortschr. Mineral., Beih., V 56, n 1, 1978, p 65-6

45 Kollmann, H., Struebel, G., Trosg F., 202, CA, Reaction mechanisms in the formation of expansion nuclei in lime-gypsum plasters by ettringite and thaumasite, Zero. Kalk Gips, V 30, n 5,1977, p 224-8 References Not Found Page 109

46 Kuznetsova, T.V., 208, CA, Structure formation of cement stone in hydration of alumi- nate and sulfoaluminate cements, Fiz.-Khim. Mekh., V 15, 1987, p 16-20

47 LaFraugh,R,W.,211, 23, The use of superplasticizers in the precast industry, Proc. Intnl. Symposium, CANMET/ACI, Ottawa, Canada, May, 1978

48 Larionov%Z.M.,213, CA, Stability of ettringite in cement systems, Tr. -Mezhdunar, Kongr. Khim. Tsem, 6th, Editor Boldyrev, A.S., Vol 2, n 1, 1976, p 168-70

49 Larionova, Z.M., 214, CA, Stability of ettringite in cement systems, Gidratatsiya Tverd. Tsem., V 2, Pt 1, 1976, p 168-70

50 Larionova, Z.M., NikitinA L.V., 215, CA, Transformation of ettringite to calcium mono- sulfoaluminate in cement stone, Tr. Nauch. -Issled. Inst. Betona Zhelezobeton& V 7, 1972, p 39-47

51 Larionova, Z.M., Nikitina, L.V., Lapshina, A.I., Garashin, V.R., Garashin4 E.V., 216, CA, Behaviour of ettringite during heating, Tr. NII Betona i Zhelezobetona. Gosstroi SSSR, n 17,1975, p 30-8

52 Le Roux, A., 218, 132, Traitement des SOISargileux par la chaux, Bull. liaison Labo. P. et Ch., 40, Sept-0c4 1969,59-96

53 Le Roux, A,, Cador, C., 219, 132, Importance de la petrographic clansl’approache des mechanisms des reactions alcali-granulats, Inter. Aggregate Symposium, Nice, 1984

54 Le Roux, A., Toubeau, P., 221, 132, Mise en evidence du seuil de nocivite et du mecan- isme d’action des sulfures au tours d’un traitement ala chaux, IX Conf. on Geotechnics, Bangkok, 1987

55 Lieske, H., 231, 155, Neuentwickhmgen im Bereich der Spannbetonschweilen (new de- velopment in the area of railroad ties of prestressed concrete), Betonwerk + Fertigteil- Technik 52 (1986) H. 6, S. 381/383

56 Ludwig, N.C., Pence, S.A., 237,23, Properties of Portland cement pastes cured at ele- vated temperatures and pressures, Journ. Amer. Concr. Inst., 27, 6, 1956

57 Ludwig, U., 238, 10, Concerning the setting and hardening of cements, Tonindustriezei- tung, 96, No. 4, 1972,85-92

58 Ludwig, U., Heinz, D., 239, 155, Einflusse auf die Schadreaktionen in warmebehandelten Betonen (influences on the damage reaction in heat treated concretes), in: Baustoffe 1985, Aachen. Festschrift Prof. K.H, Wesche, Vauverlag Wiesbaden-Berlin

59 Macleod, G., Hall, A.J., 244, GEO, Whisker crystals of the mineral ettringite, Mineral. Petrol., 1991, 43(3), 211-15

60 Mangotich, E., 246,23, Some tests of the compressive strength of concrete masonry units as affected by the time-temperature-maturity with curing at atmospheric pressure, Tech. Report No. 47, National Concrete Masonry Assoc., Washington, D.C., May, 1954

61 Marchi, M., 248, CA, Ettringites in Portland cement, Studi Sassar., Sez. 3, V 25, 1978, pp 237-42

62 Michaelis, W,, 277, 117, Der Zementbazillus, Tonind. Ztg., Vol 16, 105 Rejiw rices Not Found Page 110

63 Millet J., Bernard, A,, Hommey, R. Poindefem A., Voinovitch, I.A., 280, CA, ‘Ihe deter- mination of ettringite in cement and mortar pastes, Bull. Liaison Lab. Ponts Chaussees, V 109,1980, p 91-5

64 Mille~ J., Bernard, A., Hommey, R., Poindefert, A., 281, CA, Determination of ettringite in portland cement pastes and mortars, Analusis, V 9, n 7, 1981, p 311-17

65 Mironov, S.A., 282,23, Effect of temperature and admixtures on the acceleration of the hardening process in concrete, Symposium on Chemistry of Cements, State Publication on Structural Materials, Moscow, USSR, 1956

66 Mironov, S.A., Malinin& L,A., Fedorov, V.A., 283,23, me increase in strength snd lin- ear strains of concrete during steam curing, Beton i Zhelezobeton, No. 4, 1961, 170-174, Dept. of Scientific and Industrial Research, Bldg. Res. Station, Library Communication No. 1156

67 Mueller, H., Riedmueller, G,, Schwaighofer, B., 301, CA, Scanning electron microscopy of the transformation of the cement mineral ettringite, Carinthia 2, V 85, 1975, p 71-82

68 Negro, A., 309, CA, Mathematical model of the expansion of ettringite, Cim., Betons, Pla- tres, Chaux, V 756, 1985,319-27

69 Negro, A., Stafferi, L., 313, CA, Ettringites and cement expansion, Cemento, V 64, n 11- 12, 1967, p 239-42

70 Nildtin& L.V., Lapshina A.I., Krasil’nikov, K.G., 314, CA, Relation between ettringite crystallization conditions and the development of expansion deformations during the hard- ening of sulfate-containing cements, Tr. Nauch. -Issled Inst Betona Zhelezobetona, V 7, 1972, p 21-9

71 Nikitina L.V,, Larionov& Z.M., LapshinA A.I., Garashina, E.V., Garashin, V,R,, 315, CA, Phase transformations of ettringite in expanding systems, Tr. NII Betona i Zhelezobe- tons., Gosstroi SSSR, n 17,1975, p 39-55

72 Nikitin& L.V., Larionov4 Z.M., Levushkin, L,N,, Lapsina, A.I., 316, CA, Formation and transformation of ettringite in expanding systems, Tr. VNII Tsement. Rom-sti, 1983, n 77,75-8

73 Nikonova, N.S,, Matyukhina, O.N., Mityushin, V.V,, Nesterina E.M., 317, CA, Kinetics of heat release in hydration of tricalcium silicate in the presence of ettringite, Izv, Vyssh. Uchebn. Zaved., Khim. Khim. Tekhnol,, V 32, no 9,1989, p 77-80

74 Odler, J., 322, 155, Uber die Rissbildung bei der Warmbehandlung von Beton (concern- ing cracking during heat treatment of concrete), beton 35 (1985) H. 6, S. 235/237

75 Olesova, T.N., Razumovskii, B.E., 329, CA, Kinetics and mechanism of ethlngite forma- tion in hydration and hardening of calcium sulfo- and fluoroaluminates, Tsement, n 4, 1984, p 14-17

76 One, M.; Nagashima, M., Otsuka, K., Ito, T., 330, CA, Some aspects of mechanism of sea water attack on hardened cement paste, Semento Gijutsu Nenp, n 32, 1978, p 100-3 References Not Found Page 111

77 Pestunova, E.K., 336, CA, Volume changes during the hardening of expanding cements, Issled. Betonu Zhelezobeton, Konstr., Mater. Konf. Molodykh Spets., 7th, Ed. Aleksan- drovskii, S.V., 1974, p 125-32

78 Pfeifer, D.W,, 340,23, Maximum stmgth of heat-cured concrete in relation to the delay period and penetrometer strength of mortar extrac~ notes from PCI Energy Conservation in Curing, Prestressed Concrete Institute, March, 1980

79 Piksarv, E., Kikas, V., Hain, A., Nurm, V., 344, CA, Effect of alkalies on the durabiity of steam-cured Portland cement, Tr. Tallin Politekh. Inst., V 388, 1975, p 71-8

80 Ponomarev, I.F., Kholodnyi, A.G., Izmailov& R.A., Gribko, V.F., Potapova, L.M., Omel’chenko, A.F., 350, CA, Effect of the crystallization kinetics of ettringite on the properties of plugging mortars, Tsement, No 9, p 22-3, 1987

81 Preston, H.K,, 353,23, Effect of temperature drop on strand stresses in a casting bed, Prestressed Concr. Inst. Journal, June, 1959

82 Prince, W. and Perarni, R., 354, , Mechanism de cristallisation de l’ettringite clansles hants a base al kaolin active de chaux et de sulfates, C.R. Acad. Sci, Paris, t. 310, 11, 1990, p 535-540

83 Punzet, M., Ludwig, U., 355, CA, Chemical stability of ettringite, Tonind. -Ztg. Keram. Rundsch., V 98, n 8, 1974, p 181-7

84 Pushnyakova, V.A., Kotsupalo, N.P., Berger, A.S., 356, CA, Reaction of ettringite with aqueous alkali metal hydroxide solutions, Izv. Sib. Otd. Akad. Nauk SSSR, Ser. Khim. Nauk, n 6, 1972, p 108-13

85 Ramachandran, V.S., Beaudoin, J.J., 357, EI, Physico-mechanical characteristics of hy- drating tetracalcium aluminofemite system at low water:solid ratios, Journ. of Chem. Technology and Biotechnology, Chemical Technology, v 34A, n 4, May 1984, 154-160

86 Rehm, G,, Zimbelmann, R., 361, 10, Study of the factors determining the adhesion be- tween additive and cement matrix, Deutscher Ausschuss fur Stahlbeton, 1977, No. 283,5- 37

87 Reinsdorf, S., 362,23, Unter suchungen uber &n Einfluss von Portlandzementen unter- schiedlicher kilnkerabstimmung und verschiedener mineralischer poro ser Zuschlagstaffe .... Unpublished Doctor’s dissertation, The University for Architecture and Building, Wei- mar, 1957

88 Reinsdorf, S., 363,23, Improvement relating to low-pressure steam curing and heating concrete at temperatures up to 100C, Proceedings, RILEM Symposium, Moscow, RILEM, Paris, Session H/ General Report, 1964,32 pp

89 Reinsdorf, S., 364,23, Der Einfluss der Klinkerminerale des Portlsnd zementm auf die Eigenschaften Dampfbehandelfer Betone, Silikat Teckni~ May, 1959

90 Revay, M., Nagy, M., 366, CA, Stability of ettringite, Epitoanyag, V 20, n 12, 1968, p 473-5 References Not Found Page 112

91 Richards, C.W., Helmuth R.A., 367, NTIS, Expansive cement concrete-micromechanical models for free and restrained expansion, Stanford Univ., Dept. of Civil Engg., National Science Foundation, Final Report, Jan 1975, 42p

92 Schmi@ E., Shutz, R,J., 383,23, Steam curing, Journ. Restressed Concr. Inst., Sept., 1957

93 Schwander, H.W., Stern, W.B., 384, EI, Experimental production of transformation prod- ucts on hardened cement paste surfaces, 1988 (ref. not given)

94 Schwiete, H.E., Ludwig, U., Jager, P., 387, 147, Investigations in the system 3Ca0.A1203.CaS04,Ca0.H20, Symposium on Structure of Portland Cement Paste and Concrete, Highway Research Board Spcial Report 90,1966,353

95 Sheiken, A.E., Kurbatova, 1.1.,Fedorov, A.E., Shvedov, V.N., 390, CA, Effect of sulfaa containing phases on cement stone strength, Gidratatsiya Tverd. Tsem,, V 2, Pt 1, 1976, p 166-7

96 Shpynova, L.G., Kril, A.S., Goldinov& T.S., 393, CA, Two ways for the formation of et- tringite, Vestn. L’vov. Politekhn. In-m V 8, n 127, 1977, p 174-6

97 Shubin, V.I., Lashchenko, N.V., 394, CA, Strength of ettringite intergrowths in hardening of cement stone, TsemenL No 10, 1986, p 9-10

98 Skoblinskaya, N.N., Krasil’nikov, K,G., Nildtin& L.V., Varlamov, V.P., Lapshina, A.I., Garashin, V.R., 398, CA, Stability of ettringite during change in the humidity and tem- perature of the environment, Tr. NII Betona i Zhelezobetona. Gosstroi SSSR, n 17,1975, p 4-29

99 Skramtaev, B.G., 399,23, High strength rapid-hardening concrete, Proc, 4th International Symp. Chemistry of Cement Washington, D.C., Vol. 2,1960,1099-1100

100 Tan@ M., Hashizume, G,, Matsui, H., Nakagawa, S., 408, CA, Ettringite whiskers and their uses, PatenL Hyogo Prefecture, US #4255398, Mar 10,81

101 Tashiro, C., Akama, K., Ob& J., 409, CA, Effect of heavy metal oxides on formation of ettringite and on microstructure of its hardened mortar, Semento Gijutsu Nenp, n 32, 1978, p 86-7

102 Thiel, A., 415, CA, Factors affecting volume changes of expansive cements. Part I, Cem. - Wapno-Gips, n 4, 1983, p 104-9

103 Uklonskaya, N.T., 420, CA, Infrared spectroscopy of sulfate minerals, Zap. Uzb. Otd. Vses. Mineral. Obshchest., V 25, 1972, p 137-8

104 VenuaL M., 422, 19, Effect of elevated temperatures and pressures on the hydration and hardening of cement VIth Intnl, Congress on the Chemistry of Cement, Moscow, Princi- pal Paper, 1974

105 Volkwein, A., 426, CA, Etlringite-like phases in strong chloride-containing old cement stone and concrete, TIZ, Tonind. Ztg., V 103, n 9, 1979, p 530-1,534

106 Wang, S., 429, CA, Thermal stability of ettringite, Guisuanyan Tongbao, V 6, n 2,1987, p 19-24 References Not Found Page 113

107 Wend& R., Poellmann, H., 433, NTIS, Investigations of crystal chemistry and thermal sta- bility of 3Ca0.A1203.3CaS04.30-36H20, Amual Mtg. of German Mineralogical Sot., Gesellschaf$ Fortschritte der Mineralogies,Beiheft (Deu) Aug 1984, Vol 62, No 1,258- 259

108 Wierig, H. -J., 438, 155, Einige Beziehungen zwischen den Eigenschaften von,, grunen” und,, jungen” (some relationships between the characteristics of green and young concrete and those of hard concrete), Betonen und denen des Festbetons. beton 21 (1971) H. 11, S, 445/448, und H. 12, S. 487/490

109 Wischers, G., 440, 155, Anmerkungen zum Merkblatt fur die Hersteilung geschlossener Betonoberflachen bei einer Warmebehandlung (notes for production of sealed concrete surfaces with heat treatment), beton 17 (1967) H. 3, S, 101/103, und H. 4 S. 139/142: ebenso Betontechnische Berichte 1967, Beton-Verlag, Dusseldorf 1968, S 35/50

110 Xue, J., 443, CA, Expansion associated with ettringite formation, Guisuanyan Xuebao, V 12, n 2, 1984, p 251-7

111 Yanchikov, V.G., 444, CA, Thermodynamics of decomposition of ettringite during ther- moturbulent activation of cement pastes, Issled. Svoistv Tsement. i Asfal’t. Betonov, Omsk, 1984, p 51-5

112 Yang, N., Zhong, B., Dong, P., Wang, J., 445, CA, Formation and stability conditions for ettringite, Guisuanyan Xuebao, V 12, n 2, 1984, p 155-65

113 Zimonyi, G., 449, EI, Research into the hydration of portland cemen~ Period Polytech, Electr. Eng. v 16, n 2, 1972, p 177-188 Subject Index Paze 114

Subject Index

Accelerated Tests ...... 6,69-82 decomposition of...... 6,22 Aggregate deposition of ...... 4.6.7.9.39 limestone ...... 6,39 expansion mechanism ...... 29-32,44-68 andtransition zone, ...... , . ...6.7,39 formation atTectedby alkalis ..,....., ,.. .25 Airentrainment ...... 52 formation a.fkcted by temp...... 8.22.24 Alkali-aggregate reaction ...... 4,6,71 formation of ...... 16.20.36 Alkalis morphology of. , . ., ...... 28 effect onettringite formation ...... 25 precipitation of ...... 4.36 Attack mechanism ...... 4 stability of ...... 2O. 24,25,35 Blast furnace slag ...... 39 Expansion ...... 4.6.24.44.74 Calcium silicate hydrate andtype of cement, ...... 57.73 andsulphates ...... 23.33 due to ettringite formation...... 30,47-51,71 Calcium aluminates typeof ...... 56 nXardationof...... 18 Flyash ...... 39,45 Calcium hydroxide ...... ,4 Frost durability, ...... 8,36 orientation in transition zone ...... 3942 Gypsum, , ...... l9.24.33.46 Carboaluminates ...... 21,28,39 effect of percent on expansion. ,....46,58, 62 carbonates Heat curing effect onettringite formatb ...... 25 guidelines for. ,., ...... lO.64-67 Carbonation ...... 4 Humidity Cement of storage, effect on expansion ...... 51 llneneas ...... o ...... o .$.24.45 Hydration hydradonchemistry ...... 16 intransition zone, ...... 41-42 Cement chemistry Ion substitution andettringite formation ...... 60-68 andeffectoftemperature. .,...... 24 chlorides inettringite crystal structure ...... 18 diffusion of ...... ,4,6 Jouravskite .,, ...... ,.,...... 18 penetration of. .,,..,.0...... ,,, . 4 Microcracks ...... 6,33 Chloroaluminate ...... 6,28 at aggregate/paste interface ...... 7, 8 Corrosion ...... 4,21 infilling of ...... 8 Creep ...... 10 Monosulphoalumi nate Crystal growth formation of ...... 16 andpressure exerted ...... 29 stability of ...... 20,35 Duggantest, ...... 69-82 Permeability .,...... ,...,,. . ..12 Ettringite pH wdmwtitimofdudnti, ...... 18 andstability of aluminates, ...,...... 21 andthe transition zone ...... 39-42 effect ofalkalis on. , .,.,...... 25 chemistry of ...... 17 Pore solution Crystal structure of . . ..o . . . .. o...... 17 ionsin ...... o. o...... 33 Subject Index Page 115

Porestructure ...... 4,36 Porosity andthe transition zone ...... 42 Precast c.mcrete .,...... 7,8,9-14 Precipitation inpores ..,...... ,...4,36 Precuringperiod...... 8,10,33,46 Relative humidity effect on expansion...... ,., ...,, 46 Secondary ettringite mechanism of formation. 32-35,47,52-54,58 Shrinkage ...... ,. .10 Silica fume ...... ,, .42 Sleepers ...... see Ties Soaking period . , . . . , , see Precuringperiod Volubility product ...... 21 Strength andgypsum content ...... 60-62 effect of elevated temperature on . . . 10, 13, 61 of ties ...... 8 Sulphate effect oftype onexpansion ...... 46 release ofions withtime ..,.,, , ...... 47 SulphateJaluminate ratio . . . . 50,55,62-64,76 Swelling pressure...... ,,.. ., ..30 Temperature effect ofrate of rise ...... 9 effect instability ...... 21.33.54 elevated, effect of ...... 6,9-12,44-68 Testing andthe Duggantest. .,,,..,,.,., . ...69-82 Thaumasite ...... 6,7,18,26 l%ermalexpansion ..,,...... 14 Ties ...... 7.8 Transition zone andcalcium hydroxide ...... 39 andettringite content ...... , ...... 39 at aggregate/paste interface...... 36-42 atsteel/paste interface ...... 20.42 Trass ...... 45 METRIC CONVERSIONS

Following are metric conversions of the measurements used in this text. They are based In meet cases on the International System of Units (S1), 1In, = 25,40 mm 1 sq In, = 545.16 mmz 1 ft = 0,3048 m lsqff = 0.0929 ,mz 1 sq ft per gallon = 0,0245 mzlL 1 gal = 3.785 L 1 kip = 1000 Ibf = 4.448 kN 1 lb = 0.4536 kg 1 lb fwr cubic yard = 0.5933 kglm’ 1 psf = 4,682 kg/m~ 1 pal = 0.006895 MPa No, 4 sieve = 4.75 mm No. 200 sieve = 75 ~m 1 bag of cement (U,SJ = 94 lb = 42,6 kg 1 bag of cement (Canadian) = 881b = 40kg 1 bag per cubic yard (US,) w 55,8 kg/m3 deg. c = (deg. F - 32)/1.8 PALABRAS CLAVE: pruebas aceleradas, reacci6n alcali-agregado, alcalis, aluminates de calcio, hidr6xido de calcio, carbo-aluminates, carbonataci6n, cemento, quimica del cemento, cloruros, cloru-aluminates, agrietamiento, formaci6n de etringita retrasada, prueba de Duggan, durabilidad, etringita, expansi6n, cal, tratamiento de calor, hidrataci6n, iones en soluci6n, microagrietamientos, monosulfa-aluminato, pH, soluci6n para poros, concreto precolado, periodo de precurado, orientaci6n preferida, formaci6n secundaria de etringita, resistencia, sulfates, relaci6n sulfato/aluminato, presi6n de expansion, temperature, ensaye, taumasita, durmientes de concreto, zona de transici6n.

SINOPSIS: Este reporte comprende una revisi6n y an~lisis de la literature existente relacionada con las causas, 10Sefectos y la prevenci6n de la etringita secundaria (retrasada) en el concreto. Se examinaron m&sde 300 publicaciones. Se tratan primeramente Ios estudios en 10Scuales el daiio en el concreto posiblemente fue causado por la formaci6n de etringita secundaria. Posteriormente, se trata la investigaci6n fundamental relacionada con la formaci6n de la etringita secundaria, su quimica y 10Smecanismos de formaci6n. Se analizan as investigaciones mi%importances sobre el tema. Asimismo, se determina la importancia potential (a)del m&odo de curado por calentamiento (b)de la qufmica del cemento. En el tiltimo capitulo, se evalua una prueba rapida sobre la evaluaci6n de la suceptibilidad del potential de la etringita secundaria (ensaye de Duggan). El an~lisis indica que parece haber potential en Norte America para la formaci6n de etringita secundaria; es muy probable que la formaci6n de etringita secundaria pueda llevar a un deterioro significativo del concreto tratado con calor. Sin embargo, es poco probable que la formaci6n de etringita secundaria pueda ser el iinico mecanismo responsible del deterioro premature. Los factores criticos que determinant la extensi6n del dafio debido a la formaci6n de etringita secundaria, son: (a) la duraci6n del periodo de retraso antes del calentado del concreto, (b) la severidad del regimen de calentado o enfriamiento, y (c) la relaci6n SOJA1,OS del cemento. No existe evidencia de que 10Sconcretos no tratados con calentamiento son suceptibles a este fentimeno. Mayor investigaci6n y mejoras en el ensaye de Duggan puede resultar en el desarrollo de un mdtodo de prueba estandard para analizar la estabilidad y la durabilidad del concreto a largo plazo.

REFERENCIA. Day, R. L., The Eflect of Secondary Ettringite Formationon the Durability of Concrete: A LiteratureAnalysis, Research and Development Bulletin RD108T, Portland Cement Association [El efecto de la formaci6n de etringita secundaria en la durabilidad del concreto: Un an~lisis de la literature, Boletin de Investigation y Desarrollo RD108T, Asociacion de Cemento Portland], Skokie, Illinois, U.S.A., 1992.

STICHWORTER: Schnellverfahren, Alkali-Zuschlagreaktion, Alkali, Kalziumaluminate, Kalziumhydroxid, Kohlenstoffaluminate, Carbonation, Zement, Zementchemie, Chloride, Chloraluminate, Rit3bildung, verzogerte Ettringitbildung, Duggan-Test, Haltbarkeit, Ettringit, Expansion, Gips, Warmebehandlung, Hydration, geloste Ionen, Bildung von Mikrorissen, Monosulphoaluminat, pH, Porenloslichkeit, Porositat, Fertigbeton, Vorharteperiode, wiinschenswerte Ausrichtung, sekundare Ettringitbildung, Festigkeit, Sulfate, Sulfat-/Aluminatverhaltnis, Treibdruck, Temperature,Testen, Thaumasit, Bindung, ~ergangszone.

AUSZUG: Der Bericht umfat3teine Besprechung und Analyseder verfiigbarenLiteratur uber die Ursachen, Auswirkungen und Vermeidung von sekundarer (verzogerter) Ettringitbildung in Beton. Es wurden uber 300 Publikationen untersucht. Fallstudien i.iberSchaden in Beton, die wahrscheinlich durch sekundare Ettringitbildung hervorgerufen wurden, werden zu Beginn untersucht. Die Grundlagenforschung uber die sekundare Ettringitbildung, deren Chemie und die Ablagerungsmechanismen werden danach behandelt. Wichtige Punkte uber das Thema werden detailliert analysiert. Danach wird die potentielle Bedeutung (a) einer Methode der Warmebehandlung und (b) die Chemie des Zements besprochen. Im letzten Kapitel wird ein Schnellverfahren zur Bewertung moglicher Empfindlichkeit auf sekundare Ettringitbildung (“Duggan’’-Test) beschrieben. Die Analyse weist darauf bin, daf3in Nordamerika ein Potential fi-irdie Bildung sekundaren Ettringits besteht. Es ist sehr wahrscheinlich, daf?die Bildung von sekundarem Ettringit zu starker Degenerierung von warmebehandeltem Beton fuhrt. Es ist jedoch unwahrscheinlich, dafl sekundare Ettringitbildung der einzige Mechanisms ist, der zu vorzeitiger Degenerierung fuhren kann. Die kritischen Faktoren, die das Ausmafl der Schaden aufgrund von sekundarer Ettringitbildung bestimmen, sind (a) die Dauer der Verzogerungsperiode vor Beginn der Warmebehandlung des Betons, (b) das Ausmai3 des Warme-/Kuhlbereichs und (c) das SO~/Al,O~-Verhaltnis im Zement. Es gibt keine Hinweise, dafi nicht warmebehandelter Beton fur dieses Phanomen anfallig ist. Weitere Forschung und Verbesserungen des Duggan-Tests konnten zur Entwicklung eines verwendbaren Standardtests fuhren, der die langfristige dimensional Stabilitat und Haltbarkeit von Beton bewertet.

REFERENZ: Day, R.L., The Eflecf of Secondary EttringiteFormation on the Durability of Concrete: A Literature Analysis, Research and Development Bulletin RD108T, Portland Cement Association [Die Auswirkungen sekundarer Ettringitbildung auf die HaItbarkeitvonBeton: Eine Analyse der Literatur,Forschungs-und EntwicklungsbulletinRDlO8T, Portlandzementverband], Skokie, Illinois, U.S.A., 1992.

PCA R&D Serial No. 1929 This publication is intended SOLELY for use by PROFESSIONAL PERSONNELwho are competent to evaluate the significance and Iimifations of the information provided harein, and who will accept total responsibility for the application of this information. The Portland Cement Association DISCLAIMS any and all RESPONSIBILITY and LIABILIW for the accuracy of and the application of the information contained in this publication to the full extent parmitted by law.

Portland Cement Association 5420 Old Orchard Road,SMie, Wzoia 60077-1083 (708) 966-6200,FAX (708) 966-9781

An organization of cement manufacturers to improve and extend the uses of portland cement and concrete II through market developmen~ engineerin~ research, meducation, and public affairs work. Printed in U.S.A. RD108,O1T