The Effect of Secondary Ettringite Formation on the Durability of

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 ettringite
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, temperature, 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 of secondary (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 of the cement. There is no evidence that non heat-treated concrete
is susceptible to this phenomenon. Further research and improvements to the Duggan test may result in the
development of a useful standard test method to assess the long-term dimensional stability and durability of concrete.
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 300 publications ont
dtd examin~es. Des &udes de cas sur des b6tons endommagds, possiblement Acause de la formation secondaire
d’ettringite, ont dtdexamindesenpremier. La recherche 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 la stabilitd dimensionnelle
Along terme et la durability du bdton.
R~FfiRENCE: Day, R.L., TheEject of Secondary 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 steamcuring 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, Cana J a 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 examined 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 heattreated 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 deterioration in non heat-treated concrete. There is some suggestion that formation of ettringite 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) severity 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 produce 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 Heinz 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 substantial 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 supply 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 specimen increases.
The 14 day sulphate expansion test does not provide an adequate prediction of cement stability at later ages,
Specifications for Heat-Treated Concrete by the German Committee for Reinforced Concrete (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
TableofContents . . . . . . . .
ListofFigures . . . . . . . . . .
ListofTables . . . . . . . . . .
Notation . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. ..i
. ..iii
. ..vi
..viii
. ..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
2.1 CombinationsofAttackMechanisms
. . . . . . . . . . . . . . . . . . . . .
2.2PelrographicObservations..
. . . . . . . . . . . . . . . . . . . . . . ...7
2.3RoblemswithPrecastConcreteandTies
. . . . . . . . . . . . . . . . . . .
2.4RecastConcrete
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...9
2.4.1 EffectofHeatTreatmentonProperties.
..................
2.4.2TheCauseof InsufficientStrength Gain afterHeat’lleatment
2.5Comment
4
4
7
9
13
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..14
Chapter 3 — Secondary Ettringite Formation:
Fundamental Research
3.1 Overview of the Cement Hydration Process. . . . . . . . . . . . . . . . . .
3.2 Early Formation of Sulphoaluminates . . . . . . . . . . . . . . . . . . . .
3.3 Crystal Structure and Composition of Ettringite . . . . . . . . . . . . . . .
3.4 The Role of Ettringite in the Retardation of the Calcium Aluminates . . . .
3.5 The Chemical Stability of Sulphoa.luminates. . . . . . . . . . . . . . . . .
3.5.1 Stability in Solution. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ,
16
16
16
17
18
20
20
Page iv
3.5,2 Stability and Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.5,3 Post-Decomposition Behaviour., . . . . . . . . . . . . . . . . . . . . . .
3.6 lhe Effect of Alkalis on Hydration and Formation of Ettringite . . . . . .
3.7 The Roleof C02and Carbonates . . . . ., . ., . . . . . . . . . . . ...25
3.7.1 Formation of ‘i%aumasite . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.7.2 Formation of Carboahuninates.. . . . . . . . . . . . . . . . . . . . . . .
3.8 Formation of Chloroaiuminates . . . . . . . . . . . . . . . . . . . . ...28
3.9 Morphology of Ettringite ...,.....
. . . . . . . . . . . . . . . ...28
3,10 Mechanism of Expansion Associated with Ettringite Formation . , . . .
3.10.1 Crystal Growth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.10.2 Swelling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.11 Mechanism of Secondary Ettringite Formation . . , . . . . . . , .
3012Comment,
, . . . . . . .. o...
. . . . . . . . . . . . . . . .
21
22
25
26
26
29
29
30
. . . 32
. . . 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 . . . . . .
403Comments ...
,,.
o,.,..,,.
. ., .,,
, . . . . . . . . . . ...43
Cha ter 5 — Key Examinations of Secondary
Ettr rngite Formation
5.11980 —Ghorabetal [106] .,,..,.,,
, . . . . . . .
5.21984 — Research Institute of the Cement Industry [268] .
5.31987 —Heinzand Ludwig [132] ., .,,,.,,.,...,.,,.,,.
5.41988 —Sylla[297] . . . . . . . . . . . . . . . . . . . . .
5.51989 —Heinz, Ludwig &Rudiger[133]
, . . , , , . . .
5,61990 -Lawrence etal[177],
.,...
,, ..,,..,.,,......,55
5.7 Comment..,,,,......,,.
, . . . . . . . . . .
36
44
. . . . . . ...44
. . . . . . . . . 45
47
. . . . . . ...52
. . . . . . . . , 54
. . . . . . ...59
Page v
Chapter 6 — Secondary Ettringite Formation
and the Chemistry of Cement
6.1 Effect of Gypsum Content of Cement on Strength and Volume Stability . .
6.1.1 Test Results of A1-Rawi [6]. . . . . . . . . . . . . . . . . . . . . . . . . . .
6.2 The Practical Significance of the S03/AIZOs Ratio . . . . . . . . . . . . .
6.3 Specifications for Heat Treatment of Concrete. . . . . . . . . . . . . . . .
6.4 Comment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...67
Chapter 7 —Rapid Test Method for Secondary
Ettrmgite Formation
7.17he Duggan Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...69
7.1.1 Significance of Duggan-Test exults . . . . . . . . . . . . . . . . . . . .
7.1.2 Research of Gillott, et al . . . . . . . . . . . . . . . . . . . . .$, . . . . . .
7.1.3 Research of Attiogbe and Wells et al [12, 317] . . . . . . . . . . .
7.1.4 Correlations Between Gillott’s and Attiogbe’s Results . . . . .
7.2 Interpretation of Duggan-Test Results . . . . . . . . . . . . . . . . . . . .
7.2.1 The Zero Strain Datum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Chapter 8 — General Discussion
60
60
60
62
64
69
71
71
75
78
79
80
83
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
References
85
References Not Found
106
Subject Index
114
Pwzevi
List of Figures
2.1 Carbonation of Bridge Structures . . . . . . . . . . . . . .
2.2 Chloride Penetration into Concrete. . . . . . . . . . . . .
2.3 Relative Strengths of Heat Treated Concretes . . , . . . .
2.4 Effect of Heat Curing and Drying on Relative Permeability
2.5 Effect of Post Treatment Water Curing on Strength Gain .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
,
,
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
...5
. . .
, .
. .
. .
5
11
12
14
3,1 Critical Values for S03/AlzOs Ratio,.
. . . . . . . . . . . . . . . .
3.2 Sulphate-Ettringite Crystal Structure . . . . . . . . . . . . , . . . . .
3.3 Effect of Curing Temperature on Sulphoaluminate Phases . . . . . .
3.4 Variation in Ettringite Peak Intensity with Curing Temperature. . . .
3.5 Correlation between Volume Increase and Ettringite , . , , . , . . .
3.6 Formation of Ettringite from Sulphates Supplied by C-S-H , . , , . ,
3.7 Changes of Sulphate Ion Concentration with Heat lleatment . . . . .
3.8 Change in S04 and OH Ion Concentration During Storage . . . . . .
4,1 The Paste/Aggregate Interface at 30 Minutes and 3 Days , , . . . . .
4,2 Variation in Ettringite Peak Intensity from Interface, , . . . . . . . .
4.3 Compounds and Orientation in the Transition Zone , , . . . . . . , .
4.4 Orientation of Calcium Hydroxide Crystals in the Transition Zone , ,
4.5 Schematic of Changes in the Transition Zone , . , . , . . . . . . . .
5.1 Expansion of Various Mortars after Heat Treatment. , . . , . . , , ,
5.2 Expansion of Paste Disks aftm Heat Treatment . , . . . . , . . . . .
5.3 Changes in Expansion, Weight and Resonant Frequency with Storage
5.4 Correlation Between Expansion and Weight Gain . . , . . , , . . . .
5.5 Correlation Between Expansion and Resonant Frequency , . . . . . .
5.6 Series II Experiments, Expansion vs. Storage . . , . . . , , . . . . ,
5.7 Effect of Humidity Change on Expmsion . , , , . , , , , , , . , , ,
5.8 Effect of Water/Cement Ratio and Air Entrainment on Expansion , .
5.9 Variation in X-Ray Peak Intensity with Curing Temperature , , . . .
5.10 Changes in X-Ray Peak Intensities with Soaking Time , . . , , . .
5.11 Scattergram of?/ARatio
vs. Expwsion . . . . . , . . , . . . . . .
5.12 Typical Expansion vs. Time Curves . . . . . . , . . . . , . . . . .
5.13 Expansion vs Time Curves Showing Effect of Delay Period . . . , .
.
.
.
,
,
,
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
,
,
.
.
.
.
.
.
.
.
.
.
.
.
.
,
,
.
.
.
.
.
.
.
.
.
.
.
.
.
.
17
19
23
24
31
32
34
34
38
39
40
41
42
44
46
48
49
49
50
51
52
53
54
56
57
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. . . .
6.3 Effect of S03 Content on Expansion of Normal Fog-Cured Specimens. . .
61
6.4 Effect of S03 Content on Exp@on
63
64
65
65
70
72
tier 48 W=k of Sotig
6.5 Scattergram of Expansion at 14 Dais vs Expansion at 5 Years
6.6 Scattergram of Expansion vsMgOContent
. . . . . . . . . .
6.7 Scattergrsm of S03/A120s Ratio vs Expansion . . . . . . . .
7.1 Scattergram of Accelerated vs. Field Expansion
. . . . . . .
7.2 Effect of Alkali Content and Cement Type on Expansion . . .
7.3 Expansion of Hardened Cement Pastes and Concretes . . . .
7.4 Scattergramof Concrete Expansion vs. Paste Expansion . . .
.
.
.
.
.
.
.
7.5 Observations of a Pessimum SOJM203 Ratio ~ V~OM Ages
7.6 Temperature Testing R6gimes, Duggan and Attiogbe T~* . . .
7.7 Expansion of Paste Cores in Attiogbe Test. . . . . . . . . . . .
7.8 Scattergram of Strsins Measured by Gillott and Attiogbe . . . .
7.9 Comparison of 20 Day Expansions Using Three Zero Definitions
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
●
.
.
.
.
“
.
.
.
.
.
.
.
.
.
.
.
.
62
73
74
76
77
78
79
81
Page viii
Lkt of Tables
5.1 Effect of Heat Treatment on Cracking and Phase Formation
. . . . . . . . 46
5.2 Expansion vs. Sulphate Content . . . . . .
6.1 Compound Composition of Cfinlws used in
6.2 SpecillcationsforHeat-TreatedCo ncre*
.
7.1 Analysis of Strsins During Duggan TeSt . .
.
.
.
.
. . . . . . . . .
A1-Rawi Study
. . . . . . . . .
. . . . . . . . .
.
.
.
.
.
.
.
.
...4
. S8
. . . . . @
. . . . . 66
. . . . . 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
Page 1
Chapter 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 motion 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 concrete itself, and the key environmental parameters that determine whether secondary ettringite formation is likely to occur?
What can be done in the cement and concrete industries in Canada to minimize the possibility 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 cement 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 references 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 references were reviewed. These 327 publications are listed at the end of this document, The 113 documents 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 secondary etlringite in concrete. Many of the problems that have occurred centre upon the precast con-
Chapter 1
Page 2
Crete industry; 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 secondary ettringite formation and its consequences.
In Chapkx 6 practical issues are discussed concerning the importance of gypsum and the chemistry 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 htended 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 secondary ettringite formation.
1.3
Terminology — “Secondary Ettringite”
One of the frost questions 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 imagination to invent a new term, the choice for the “appropriate” terminology was made among (1) delayed 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 ettringite 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 normail y 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 process 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, for example,a letter fromB. Matherto C.R. DuggatIof Feb 27, 1990(aiso copiedto ASTM
committeeC09.0202)
chapter
2
Page 4
Chapter 2
Damage due to Secondary Ettringite Formation
in Ordinary and Precast Concrete
2.1
Combinations of Attack Mechanisms
In 1965 Kennerley [16 1] 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 solution, reacted and precipitated in the voids in the mass. If the solution was low in lime the volubility 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, Although 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 reaction and much leaching, These blocks were manufactured with unwashed beach gravel
and were gauged with sea water. Upon petrographic examination, large amounts of calcite, gel, ettringite and “secondary portlandite” were observed, (Presumably by “secondary 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 aggregates were used, but a low alkali cement was also used. A significant feature of the petrogmphic 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 carbonation, chloride penetration, deteriomtion and corrosion, As Figure 2.1 shows, the depth of carbonation 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 penetmted via seeping, splashing and wind-borne Cl carried to the concrete understructure, Such processes 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 concentrations for various ages and qualities of concrete.
Volkwein found that in these deteriorated concretes, highly contaminated by Cl, “remarkable” accumulations of needle-shaped crystals were found, “particularly in cracks, pores and around ag-
chapter
2
Page 5
CARBONATION
DEllliS
OF DRY CONCR13E
FROM
BRIDGE
WRUCTURES
❑
Minimum Observed
■
■
Lower Range of
Average Depth
Upper Range of
Average Depth
❑
Maximum Obsewed
Figure 2.1
of Bridge Structures [ data from ref. 314]
Carbonation
Chloride Penetration into Concrete Close to Road - Due to Splashing
,..’
J3.,.
~..,
\
...’ m.‘...., . .. .
‘“”
‘.. .
‘%,,
‘. ...
“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
PENETIU4TION DEPTH (mm)
Chloride Penetration
Figure 2.2
into Concrete due to Splashing Action [data from ref. 314]
90
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 “chlorides did
cause the formation of ettringite” from the sulphate in the cement. Additional reactions occur
other than the formation of Friedel’s (monochloroaluminate)salt1due to the penetration of chlorides. Wetting and drying also helps to transport soluble cement compoundsto preferred locations, Volkweinnoted that “it cannot be said whetherthese ettringite crystals had caused the
deterioration of concrete or had grownup in already open cracks”,
Volkwein’sresults are most interestingbecause ettringiteappem to have formed in a chloride
rich environment.Contrastthis with the laboratoryresults of Attiogbe et al [12] who found that
during an accelerated exposuretest, secondaryettringite would not form while concrete prisms
were soaked in a sodium-chloridesatwated solution,Other research confiis that ettringite decomposes in sodium chloride solution [243-245].
Jones & Poole in 1987 [154] looked at the effects of alkali-silica reaction on 3 structures in the
United Kingdom.They examined20mm thick disks from concretecores taken from the stmctures, The disks were stored in sealed containemat constanttemperaturesof 1, 15,20,25 and
38°C and relative humidities of 65,75,85 and 100%,Expansionswere measured; since the concretes were taken from structures,all expansionsnoted were relativeto the unknown deformation
of the structures.
These researchers found that initial expansionproceeded fater at elevatedtemperaturebut the
Me declined more rapidly,The final expansionswere roughly inverselyproportional to storage
temperature, Petrogmphic examinationdid not reveal the classic alkali-silica (asr) reaction rims
that are commonly observed,Cracks were often observedaround “aggregate margins” which mdiate into the paste matrix, Asr product was rarely seen inflllingthe cracks, but was most often
seen at the centre of reactive patticles, In one structure, large quantities of ettringite were obsexved,This material was coarsely crystallineand was seen inflllingthe microcracksand “lining
voids”, Ettringite often appeared to have replacedpreviously depositedsilica gel,
Jones & Poole noted the research of ~reening [120] who showedthat monosulphatebecomes metastable in the presence of CaCOSand reverts to ettringite2,The reseamhersproposed that the
presence of limestoneaggregateprovided the calcite which rendered ettringitethe stable sulphate
phase, Ettringite, initially dispersedthroughoutthe microstructurerecrystallizes into a coarsely
crystalline form, The process is helped by increasedpermeabilitycaused by other accompanying
processes such as asr cracking, The formation of ettringite in the voids is thus due to a through solution process through the pore fluid, The “preferentialrecrystallization”of ettringite in “gelfilled” microcracksmay contribute to the overall deteriorationprocess,
E1-Sayed[96] did a post-mortemon 7 deteriomtedreinforcedconcrete structures in a marine environment in Egypt, Ettringite was observed in 3 of the 7 concretes,He concludedthat a combination of factors ultimately contribute to premature deterioration,
Most experimental research necessarilyattempts to isolate one or two variables and examines
the effects of controlled variations on behaviour,As E1-Sayednotes, the real situation is far different, It is hard to imaginea practical situation where, over the course of several years, 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 given in Section3,7,2
chapter
2.2
2
Page 7
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 mreactionof 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). Xray analysis identified this material as ettringite, with a small amount of gypsum and “an unidentified 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 pretreated and exposed to weathting is the “secondary phases” that seem to form at the contact zone
between h 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 mechanism)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-existing” 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 cladding panels. The problem initiates as crack formation at corners and edges which then tend to extend 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 thaumasitdettringite 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 for a descriptionof reactionand expansionmechanismsof ettringite
chapter
2
Paga8
= an inadequatepre-treatmentallows internaldamage through debondingbetween aggregate particlesand the paste matrix, and through cracking,
=
heat treatment interruptsthe normal formation of ettringite; this formation continues later
in the hardenedconcrete, with undesirableresults,
It was noted that the solution to the problem is to provide an adequate period for precuring,
Tepponen [303] reportedthe Scandinavianexperiencewith problems of production of railway
sleepers (ties) that have occurred since the 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 from 0,36 to 0,39, Ties were manufactured with either a
1 or 2 hour delay period] and a maximumtemperature in the mnge 75-80”C,The soakingperiod
was 2,5 or 4 hours. One-day strength of the 1965ties was 56 MP~ and was 43 MPa in the 1971
ties.
Tepponen noted that 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 larger cracks filled with “microcrystalline ettringite” which “had accumulated in the free space”, In the more severely damagedconcrete the bond between paste and
aggregate was poor, Further studies in 1970confirmedthat the ~
reason for deteriorationwas,
however,the poor ilost resistance of the concrete, To solve this problem air content was increased and the maximumheat-treatmenttemperaturewas reduced to 60°C.
New studies performed in 1980revealedthat “microcracks,,,arethe primary cause of deterioration” [303],These cracks occur due to host, load, premature heating, improperheat treatment,
too severe or rapid heat treatment, etc, If temperature is also greater than 70°C during curing,
then “metastable monosulphate”is produced which forms ettringite when (a) the temperature
drops back to normal and (b) sufficient water h supplied,
New production methods were adopted in the 1980swhich employed new German guidelines
for heat treatment. These consist of a prestomgeperiod of 3 hours, a mte of temperature rise between 10=15°C/hr and a maximumtemperaturebetween 6$70°C, Little changes in the cement
composition were made and, in fact, the new cement that was used had a higher aluminate and
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 strength at 1 day was
68MPa, Tepponen reportsthat the ties have been in service for 5 years and no expansionhas
been observed even though microcracksdue to stress are present [303],
Heinz and Ludwigs’sexperiments[132], discussedin detaii in Sections 5,3 and 5,5, were
prompted by practical probiemsthat surfaced with precast units — specifhxily cladding paneis
and precast units, These high-strengthstructural elementswere heat-treated during production,
Damage surfhcedafter several years’ exposureto “open-air weathering”where there was llequent saturation. Crockswere first observedaround edges, foliowedby further cracking and loss
of bond between the paste matrix and coarse aggregateparticles,
In another publication, Heinz and Ludwig [133] noted occurrencesof damage in Europe of highstrength precast units which used Type 55 high-early-strengthPortland cement, These members
were heat-treated in production.Damage always occurredon units exposedto the weatherand
1
also called“pre-storage”or “preset”— the time betweencastingand the start of thermaltreatment
2
See Section6,3 for a discussionof precastspecifications
Page 9
chapter 2
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 ettringite formation was prompted by problems with site concretes in Germany and 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 between 1963 and 1969 [177]. These beams cracked due to “unexplained causes”. All were cast using 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 problem 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 sections there were thinner layers of ettringite around aggregate particles.
Ettringite has been observed in normallyast structures which, in most cases, have been damaged by a combination of fhctors. In such structures ettringite appears capable of being a contributory 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 destruction once the heat-treatment process has established the nucleation sites for ettringite to form. According 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 temperature 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
Clwpter
2
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 temperature 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-established 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 retrogression in strength occurs with age”, Klieger suggested that an acceptable heat-curing cycle consisted 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 manner 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 temperature 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 temperature in a 24 hour cycle show less drying creep and shrinkageunder sustainedstress than those
cured at normal temperatures [166].
Hanson [128] also noted that accelerated curing causes significant reductions in creep and shrinkage, 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-treatmentrbgimeconsisting of 4-6 hours delay period followed by a 17°C/hour temperature rise to 66°C and a soaking period of 13 hours is about optimum with respect to strength development,
In a subsequent paper, Hanson examined the properties of lightweight concrete [127], The optimum 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 occurs 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 Figure 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 control. 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
160
I
L
I
DelayPeriodforAll Casts=0-0.5 hrs.
~-’--------------------
---------
‘1
----
140
----
120
❑
After Heat Treatment
❑
7 Days
~[Q28D.ys
:::
--;-------.
II:~:: --------
100
80
60
40
20
0
Sosklime (hrs)
~
18
42
38°C
66
18 42
52°C
66
18 42
66°C
66
18 42
66
18 42
79°C
66
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 subsequent 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 admixtures 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 compensated 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.3 through the use of a supeqhsticizer resulted in a 43% greater strength at 28 days than a normally-cured, superplasticizer free,
equivalent Type I concrete [254].
Page 12
Chupter2
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 cement 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, longterm 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 compared to a water-cured control specimen [135]. This is shown in Figure 2.4. All steam-cured concrete 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
24 HR,STEAM@64°C,
7 DAYFOG,21 DAYDRY
1
S HR.STEAM@71°C,
28 DAYDRY
1
1
I
1
I
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
1
I
1
t
I
o
5
10
15
20
25
go
35
40
45
PERMEABILITY OF SPECIMEN/ PERMEABILITY OF 28 DAY CONTROL SPECIMEN
Effect of Heat-Curing
1
Figure 2.4
and Drying on Relative Permeability
[ data from ref. 135]
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 are now learningvery quickly that adherenceto that philosophyoften leadsto infkriorconcretein
the long term-with resultantcostlyrepairs.
chapter
2,4.2
Page 13
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 coefficient of free water when compared to the solid ingredients of the concrete. To avoid this damage, 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 damage associated with expansions and cracking:
~
nxisting pore pressures-through the use of closed forms, or by letting the concrete attain a minimum strength before the stati of heating;
~
eliminating pore pressure-by
nating air voids;
~
letting the concrete crack-and
heating the concrete ingredients before casting or by elimirepairing it, perhaps by mwibration after heating.
Others have also attributed damage to the principal mechanism of differential thermal expansions. 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 expansion coefllcient about 10 times greater than any of the solids. Sylla hypothesized that premature 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 hydration; 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-treatment, the paste microstructure is more heterogeneous because of the very rapid rate of initial 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 reduces 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’Mo
r.h.) after heat-treatment until specimens were tested at 28 days. In Cure 2, specimens 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-treatment. The results given in Figure 2.5 indicate that lower strengths may not be due to damage during heat treatment, but may be due to inadequate curing after the heat treatment.
Chapter 2
Page 14
Cure
Cure
20”C
Cure
Strength as
Y.
1 = Water cure at 20 “C for 7 days, then at 65% rh until test at 28 days
2 = Heat cure at given T atler 30 minutes delay, followed by water cure at
until 7 days, 65% rh to 28 days
1
,
3 = Heat Cure of Cure 2, no water cure,
❑ CURE 3
❑ CURE 2
Y. OF CURE 1
‘YoOF CURE 1
of Standard Cure
1
----
100- -
----
..-
---
.80- -
----
----
60- -
----
----
40-
----
----
20- -
----
----
----
----
---
+
ol80-2.45
80-2.00
80-3.35
60-2.30
60-3.40
60-4.40
MAXIMUM TEMPERATURE ‘C -- TIME AT MAXIMUM (HRS)
Effect of Post-Treatment
Figure 2,5
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 especially 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 period should be 4.5 hours, and at a 90°C heat-treatment temperature, the delay period should be approximately 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 expansion 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 devised rt$gimes to 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-treatment r6gime that would still produce high early strengths but which would not result in large comparative 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 factor 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, including freezdthaw, corrosion, alkali-aggregate reaction, stress cracking, sulphate attack carbonation, 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 sulphate attack steps can be taken to help ensure that secondary ettringite formation is either an insignificant 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 attributed to secondary ettringite formation is a very small proportion of all precast concrete cast in
Europe and North America.
Page 16
chapter 3
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 cement 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 reaction between gypsum and calcium alu-
minate. X-ray peaks that are associatedwith ettringiteare detectablewithin a few hours and the
quantity increasesduring the first day. After this time the ettringite peaks normally weaken, but,
dependingupon the chemistry of the cement and the environmentalconditions, ettringite may
persist indefinitely,It is a general belief that if all of the gypsum is consumed through the ettringite reaction, ettringiteconvertsto monosulphoahuninate.In an average cement this conversion process starts at about 1 day and can be monitoredas a reduction in the size of the ettringite
X-ray peaks. Much of the monosulphoaluminatephase that forms instead is poorly crystalline
[302].
3.2
Early Formation of Sulphoalumlnates
The commonly held chemical formula for ettringite is:
3Ca0,AlzOy3CaSOq.32Hz0
or in chemists’ notation is: C6AS3Hn
and for monosulphateis:
3Ca0.Alz03,CaS04, 12Hz0 0!’:
c4A~12
The reaction that occurs between C3A and gypsum can be found innumerous texts:
C3A + sC~Hz + HZ6+ C6A~3H32
Except for the unusual combinationof very low C3Acontents and high gypsum contents, the gypsum will be used up by this reaction before the C3A,When this occurs, the remaining C3Awill
take part in a reaction wherebyettringite decomposesto monosulphoaluminate:
2C3A + c(jA~3H3z+ H4+ 3C4ASHU
Thus, in the net reaction,3 moles of CSAand 3 moles of gypsum will produce 3 moles of monosulphoaluminate.Use of the Bogue equation and performingstoichiometriccalculations fkomthe
formulae above, one can cdcuiate the S03/Al@ mass or molar mtios (TVAratios) required to
consume all of the C3Aand convert it to monosulphoaluminate,Figure 3,1 showsthe dependence of this mtio on the C3Acontent of the cement (given an assumed Fe20s content of 3%), If
the WA mtio is greater than the value shown, C3Awill still be available after all of the ettringite
has been consumet the remaining C3Awill react with calcium hydroxide and water to produce
C4AH13[302],The above analysis is highly idealized.It is importantto recognize, however,that
the conditions under which ettringite forms,and the stmcture of ettringite itself are highly complex,
Page 17
chapter
3
S03/A1203 Ratio Required to React with AU Tricalcium
Aluminate to Form Monosulphoaluminate
0.7 I
I
-/
/n-
0.6
..-
------------------
----
.
-
.
7
Molar Ratio
/0-”
cl--
0.5
/
/“
&------------
--------------
------ A
/.
Mass Ratio
/’-”7
u“
S031A1203
RATIO
0.4 - -------.
-~ ’-------
‘ ----------------------
#
(mass or
/’
molar) 03 T~------
0.2
-----------
‘ - - - - - - - - - - - - - -”-------
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
3.3
Ratio
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 general 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 structure of this class of material:
@[
Al(oH)6]2.(so4) 3”26H20
chapter
3
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 between the columns.
Taylor [300] provided a very clear picture of the ettringhe-group crystal structure which is characterized by four positively charged columns in a trigonal structure, between which occur channels 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 dimension 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
{Ca~Mn(0H)6~24H@}@C)A)y(C03)2
@ jouravskite:
In thaumasite, silicon teplaces ahuninium in the columns (Figure 3.2), while in jouravskite manganese replaces aluminhum For both thaumasite and jouravskite water molecules contained between 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 confirmed the wide variety of ions that can replace aluminium and/or sulphate, More recent reseamh
[130, 171] has also shown that oxyanions such as arsenate, borate, chromate,molybdate, selenate
and vanadate may substitute for sulphate in the ettringite structure,
Taylor [300], expanding on earlier work re-definedthe compositionof the column per half unit
cell in sulphate-ettringiteas {Ca~Al(OH)G~24H20~+,while the material in the ettringite chan6-
1
iS (SOA)3,2H20 to make the entire structure electrically neutral, He noted that each col[
umn is of nearly cylindrical shape and its surfhce is comprised entirely of water molecules, “or
more accurately the H atoms that these contain”.
IRh
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” notation to stand for ahminate-ferrite-monosubstituted compounds, which are hexagonal or pseudohexagonal 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 ahuninates. 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 crystallization of hydrated calcium aluminates. This occurs before the solution becomes saturated
Page 19
chapter
3
Blank Circles =
Water Molecules
10.7
i
I
Structure of Columns in Ettringite Crystal
Part of Single Column Projection
Blank Circles are Water Molecules
I
Plan View, Showing Columns and Channels
l\
Lattice a dimension=
11~ A
/s\
\/
/
x
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]
Page 20
chapter 3
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 decreases as more of these materials are dissolved (lime alone is not sufficient to retard the hydration 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 surface 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 surface 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 convincing 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 mechanism. The second is a process by which some ettringite crystals sinter together by a solid-solid reaction where crystals grow together and sinter [149].
The most convincing march also relates dkctl y to the secondary ettringite issue. It has been observed that the transition zone at the aggregate and steel interface in concrete has a higher proportion 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 between 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 dissolve, The importance of the “transition” zone to secondary ettringite formation is discussed in detdl in Section 4,2,
3s
The
3.5,1
Chemical Stabiiity of Suiphoaiuminates
Stability in Solution
The earliest studies (1929, [181]) on the stability of ettringite showed that it is more stable in solutions of calcium sulphate and calcium hydroxide than in distilled water. On the other hand, monosulphate 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 metastable compound this is also confirmed by research that shows monosulphoahuninate is metastable 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].
ettrhwite 1.O*1040
monosulphoaluminate
carboaluminate
chloroaluminate
calcite
1.7* 10-28
1.4* 10-30
1.O*10-30
8.7*10-W
compouncL is included for comparison.
The stability of ettringite in relation to other
compounds can be explained by the “rule of
volubility product” [326]. The compound with
the lowest volubility product is the most stable. The table on the left shows the volubility
products for four major aluminates that am
known to form in concrete. Calcite, a soluble
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 ettringite does not exhibit a constant volubility product-it varies 5 orders of magnitude as the product 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 particles. They note two possibly distinct forms of ettringite, depending upon the pH of the pore solution:
=
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 ettringite, will prefemntiall y be formed. However, when the sulphate in the pore solution is depleted, 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 hydration occurs at 60”C, however, ettringite X-ray peaks begin to form at 15 minutes and there is already appreciable ettringite at 6 hours. The ettringite crystals formed at the higher temperature are
appreciably larger than those developed at 25°C [1341.
Chupter
3
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 reductionfrom 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 remaining 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 stable. Similarly, pressure also tends to stabilize ettringite against thermal decomposition-ettringite 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 remains at 100”C the monosulphate decomposes to gypsum after 11 days of boiling. At lower temperatures (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 aluminate hydrate in 6 hours, which decomposes to ettringite after a day [104]. Ghorab et al [106] report 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 stable at 65°C in a dry atmosphere, but decomposes partially at 93°C. In a moist environment, however, 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 ettringite found in concretes and pastes is more thermally stable than the long slender needles of ettringite 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 ettringite 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 monosulphate 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 contents.
4.51
4.0
.
24 “CTREATMENT
3.5
3.0
AMOUNT S03 IN
DIFFERENT PHASES, 0/0
‘/
1’
2.5
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
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
DIFFERENT PHASES, %
2.5- -
Unknown hydrated
calcium aluminate
0.0
~
Y
2.0
1.5
! .0
0.5
O.q
0.5
TOTAL S03 IN CEMENT, %
4.5 ~
4.0- -
82 “C TREATMENT
3.5- -
3.0- AMOUNT S03 IN
DIFFERENT f’HASES,
2.5- ~0
2.0- 1.5- 1.o- -
0.5- “
ETTRINQITE
I
5
calcium aluminata
2.0
2.5
3.0
3.5
4.0
1
4.5
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 taken up to 24 months. Richardsconcludedthat the
expansionprocess that is normallyvery destructivedoes not take place above 20°C. It is unfortunate that he did not continue to measure expansionsafter 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 incorporation 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 ranging 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 approximately 70°C.
Stadelman [294] used differentialthermal analysisto analyzethe stabilityof ett.ringitecontentin
hardened cement pastes 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
~
0.40 --
+
mean
❑
high
●
low
0.20 --
0.00
o
20
40
60
80
CURING TEMPERATURE (C)
Figure 3.4
Variation in Ettringite Peak Intensity with Curing Temperature
[data from ref. 56J
100
‘
Page 25
chapter 3
was treated at temperatures above 60”C it no longer appeared on DTA tmces; instea~ monosulphate appeared. During storage afler treatment, ettringite peaks reformed which Stadelman conclude~ “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 volubility [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 produce a molarity of greater than 0.7M. The pH in the pore solution might nmch 12.9 within minutes 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& Kawamura [56]. They hypothesize that in the presence ofNaOH, ettringite formation is suppressed because 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 [3 16] looked at the hydration of Portland cement in 0.5 and 1.OMsodium hydroxide 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 ettringite.
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 ettringite 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 attributed 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
+ 23H20
+ 2A1(OH)3
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 “thaumasite” 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 formation 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 thaumasite 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 relative humidity in a cold environment (4”C), an adequate supply of S042- and C032- and the presence 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 available 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 temperatures 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 expansion. 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 after 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?r
3
3.7.2
Page 27
Formation of Carboaluminates
The addition of limestone to cement, or the presence of limestone aggregate in concrete, can result 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 carboahuninate, 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 concrete: (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 present 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 carbonation) can ~sult in a later-age ettringite formation. The tests that were performed by Klemm
and Adams with 5’XO
limestone 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 monosulphoaluminate, to form ettringite. The carbonate ion will not react directly with ettringite because 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 hemicarbonate 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 reaction 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 reacts with aluminate to produce carboahuninates. Under certain conditions all of ettringite, monosulphate and carboaluminate can co-exist. As well as awboaluminate, a solid solution of
chapter
3
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 formula 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-hydraulic M-S-H” may occur [102],
When exposedto seawater,Type I cements show both chloroaluminateand ettringite formation.
At early ages these products deposit in the voids and exert little expansion [158],If temperatures
are low the presence of NaCl causes ettringiteto decomposeto form Friedel’s salt. At lower temperatures 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 topochemical means [222] has also resulted in two hypotheses concerning the morphology of ettringite 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 expansive. These are the type of crystals that have been produced in numerous lab-based re-
Chapk?r
3
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 environment. 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 hydration. 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 cornerstone 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 hypotheses 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 expansion.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 solution and does not produce expansion.
Research on sulpho-aluminate based expansive cements (Type K) [6 1] shows that expansion is
caused by growth of ettringite crystals on the surface of expanding particles; a topochemical IWWtion between the particle and the surrounding solution is responsible. The magnitude of expansion 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. Alternatively, formation of small crystals at a large number of sites can produce a network of
smaller microcracks
Chatterji and Jeffery [52] have hypothesizedthat expansion due 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 reproduce 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 monosulphate 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 direction; 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 chamber to relieve potential stress build-up.
Very recently, Ping & Beaudoin [256,257] postulated that formation of ettringite is by two processes, nucleation and crystal growth, The production of “crystallization pressures” requires a confined 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
= crystal growth is in a confined space
= the volubilityproduct ratio is greater than 1,0(i.e. the products are less soluble than the
reactants)
3.10.2
Swelling
Cohen [62] summarizesthe “swelling school” of thought: ettringite forms by a through-solution
mechanism, In a saturated CH environmentettringitecrystals are gel-like and colloidal in size,
The high surfhcearea results in adsorption of significant quantitiesof water and strong swelling
pressures develop,
The theory was f~st proposed by Mehta [215,217]; the expansionmechanismfor ettringite is
similar to that of some clays which expand because of “polar oriented electrostaticattraction of
water molecules”.Small ettringite crystals of colloidal size, formed in an alkaline environment
possess a negative surfwe charge, The high surfhcearea combined with the negative charge result in an attraction for large quantitiesof water. Tests on disks of pure ettringiteconfirm the hypothesis; the disks adsorb a large quantity of water and large swelling strains occur, A check of
the XRD pattern of the ettringite before and after swelling shows no 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 prepared with WA mass ratios from 1/8 to 1.0. The water/cement ratio was 0.4 to provide just sufficient 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°/0expansion. 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 sulphate 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 correlation that exists between ettringite content and 0/0 volume increase.
Chapter 3
Page 31
180
■
160 +
❑
VOLUME 100
INCREASE
%
80
%n
❑
9
■
I
60
•1
❑
❑
■
■
40
/
■
u
20
t
■
o
1
1
1
1
I
1
1
1
50
100
150
200
250
300
350
400
1
0
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 deseribed 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 crystalline m not is “of minor importance”. The important factor is that the products attempt to occupy 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 formation in a eonllned space k capable of producing very high forces. Clearly, these forces cannot develop 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
Page 32
chapter
3
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 sulphate to be in solution to form. Its formation during sulphate attack is easily
explained, since sulphate ions penetrate the concretemicrostructurefrom outside. However, in
I
IYI’A
I
TRACES
CSS with 24% gypsum belore hydration
As IlbOV~but wtth a further 1,S%
RYPSUm
added whereCSAadded
m
100
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
,
,,
chapter
3
Page 33
this repoti we are considering mechanisms by which ettringite forms at later ages without assistance from outside sources.
The key to explaining this mechanism lies in the research of Lerch [180, 181] who found that during 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 similar 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 gypsum peak after the start of hydration confirm that the sulphate is bound within the C-S-H structure. 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 hydroxide solution. Thus, when a source of aluminate is added to the ground paste containing C-SH 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-earlystrength 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 precure, 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 decomposition of “primary” ettringite results in an increase of sulphate ion in solution and a decrease 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 hydration is decomposed. The process is accelerated by akalis. In the temperature mnge 50-80”C the reversible formula is:
chp16Jr 3
Page 34
■
500
20
I
❑
50
❑
60
■
❑
70
❑
80
90
❑
100
1
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
ALK=O.62 ALK-O.68 ALK-O.87 ALK=O.66 ALK=l.03
ALK=l.13
SIA=O.92
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
1““””G
90 C TREATMENT
600
/-
.“
.-
/“
/
/“
CONCEN- m
TRATION
In?noln
400
300
200
4
100
60 C TREATMENT
i~
o I
I
19h
26d
STORAGE
TIME AT 20C AFrER
90CI
HEAT TREATMENT
Figure 3.8
Change in S04 and OH Ion Concentration
(data from ref. 320)
During Storage
Puge3s
Ckpk!Y3
At temperatures greater than 70*C monosulphate also decomposes according to the mwersible formuhx
3Ca0.AlZ03.CaS04.
3.12
12Hz0 + 2NaOH 7k~~O~+
3CaO”AlzOs”6Hz0 + NazSOq + Ca(OH)2
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 concretes are cured at elevated temperature ettringite disappears and some of the sulphate goes “missing”. 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 solutions 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 sulphate 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 expansive 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 knowing under what conditions compounds form which try to occupy more space than is available
within the pore structure of the material.
chapter
4
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 fracture 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 volubility of hydrated calcium aluminates. The fkezing action in the pore structure results in pore solution 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. Recrystallization 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 flowing 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 indicative 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 environment. Conversely, micro-porosity, concentration of key compounds, microcracks and a
weakened microstructure around the paste/aggregate and paste/steel interface appear to be nucleation 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 aggregate/paste-matrix interface, normally called the “transition zone”. Most petrographic examina-
Chapk?r
4
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 obtained 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 crystals. 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 monosulphate; 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 cleavage planes are evident. Fracture never occurs at the true
paste-aggregate interfme. Hadley noted that some degree of orientation of CH occurs in the region 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 ettnngite 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 transition zone has preferred orientation; further tests show that pronounced orientation still exists after
420 days in a 50pm region near the interface.The transition zone has high porosity and the crystals are larger than in the bulk matrix.There is a higher concentrationof ettringite crystals in the
transition zone, but the ettringitecrystals have no preferentialorientation (Figure 4.2). The dominant effect as aging occurs is the densification of sttucture behind the CH film. This densification
“
chapter
4
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
n
Randomly
Oriented
Massive CH
crysta@
Encapsulating
\
“~..
■
“
●
Y
.3
‘
S
*
+4
\
,.- ,.
r
.
.
. \~~
.
.
.
●
.0
●
. .
.
“..
:
.“
m!liw
I
Figure 4.1
The Paste/Aggregate Interface at 30 Minutes and 3 Days
(drawings scanned from Hadiey [123] and annotated)
I
chapter
4
Page 39
180 ~
All data for Plain,
5%
Silica Fume and 16% Silica Fume Pastes
180
. 1 DAY
t
i
❑
I
—
.A
X-RAY
100 ~
INTENSITY
t
¤~A
7 DAYS
A 30 DAYS
c
~
o
Ul,
@
OWOA
❑
m
Aw
20
A
t
o I
0
❑.
o
❑
1
A
❑B
1
1
1
1
1
1
I
1
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 significantly different than the aggregate/paste transition zone.
In 1988 the issue of whether CH was truly orientated in the transition zone was clarified. Detwiler 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 during 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 aggregate surfizes.
The type of aggregate at the interface, however, does have an important role to play. The transition zone near carbonate aggregate is different than that for other aggregates, because a “basic calcium 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; carboaluminates in the transition zone do not appear, therefore, to be the reason for the increase in mechanical 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 examined. 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-
chapter
4
Page 40
LIMESTONE
45
40
35
X-RAY
30
PEAK
HEIGHT
25
FOR
ETTRINGITE 20
AND
CARBOALUMINATE
15
10
5
AGGREGATE/PASTE
INTERFACE,
m
‘\
DAYS
AGE=36
3,5
3
2.5
-\
“’......-
p
‘\
\
Laf
ORIENTATION
INDEXFOR
CALCIUM
HYDROXIDE
1,5
Y
ETTRINQITE
1
‘k
CARBOALUMINATE
0
0.5
10
1
100
DISTANCE FROM INTERFACE (yin) (log 8oale)
GRANITE
AGGREGATE
PASTE INTERFACE,
AGE=36
DAYS
45
3,5
40
T
+
3
35
t
2.5
X-RAYPEAK30
HEIGHTFOR
ETTRINGITE25
2
Ca/f%TIO
20
15
10
5
0
ORIENTATION
INDEXFOR
CALCIUM
HYDROXIDE
1.5
1 \l
L
1
1
1
MONOSULPHOALUMINATE
II
!
I
10
100
,
0.5
DISTANCE FROM INTERFACE (pm)(logscale)
Compounds
and Orientation
Figure 4.3
in the Transition Zone — Limestone and Granite Aggregate
(data from ref. 275)
,,,
“
chapter
4
Page 41
esting results (Figures 4.3). The intetiace normally extends from 50 to 100pm into the matrix. Ettringite 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 510p,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 limestone/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 pastepolypropylene interfue; polypropylene was used to simulate a flat aggregate surface. The orientation 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 reducing 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 observed in contact with the aggregate surfme. The “porous pockets” that are interspersed throughout 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
ORIENTATION
OFCALCIUMHYDROXIDE
CRYSTALS,
AGE= 28DAYS
6
5
—m—
4
.
ORIEN ATION
INDEX
3
l=RANDOM
ORDINARYPORTLAND
CEMENT(OPC)
OPCwith QUARTZFLOUR
OPCwith FLY ASH #1
OPCwith FLY ASH #2
2
1
o
I
1
1
1
1
1
I
0
20
40
60
80
100
DISTANCEFROMINTERFACE(pm)
Orientation
Figure 4.4
of Calcium Hydroxide Crystals in the Transition Zone (data from ref. 176)
120
chapter4
Page 42
PC=Portland cemenfi SF=sUica fume; CH=Ca(OH)2; CSH=calcium silicate hvdrate ett=ettrtiite
–=–
I
“
I
Fresh Concrete
without Fume
Fresh Concrete
with fine silica fume
particles
Filling of transition
mne with Ca(OH)2,
C-S-H, and
ettringlte. Note
porous pockets,
some filled with ett
I
I
Transition zone in
mature fume
concrete
~~
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 porosity, 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 washing 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.
chapter
4
4.3
Page 43
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 unknown 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 difference 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 & Ludwig [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”.
Chauter
5
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-earlystrength 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 accordance 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 after soaking in water for 70-80 days, Expansion was usually complete after 130 days soaking, except 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 expansion of bars manufactured with various blends of sulphate-resistant(SR) and high-early-strength(HES) cements,
Blends which contained more than 25% HES cement showed substantial expansions, again starting 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)
4
............r..=z.-.r.*z.=...*
CEMENT TYP~
SR= SULPHATE
RESISTANT
H13s= HIGHEARLY
STRENGTH
‘--------- 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
STORAGE PERIOD (DAYS)
Figure 5,1
Expansion of Various Mortars after Heat Treatment,
(data from ref. 106)
DIN1 164 Test Method
300
chapter5
Page 45
was finished after about 330 days. During this time, Set II specimens showed no appreciable expansion. At 330 days Set I specimens were subjected to a second heat-treatment and Set II specimens 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 expansion. Furthermore, it appears that the time at which the heat-treatment is applied is not very impotint; 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 as the PZ55 cement) were
used to prepare mortars; these mortars were tested as above. It was found that the pozzokm substantially reduced expansions:
e
0°/0pozzolan: 100°/0PZ55 HES cement
Expansion = 1.00%
I=
15°/0trass:
Expansion = 0.70!40
o=
15°/0
ash:
85°/0
85°/0
HES cement
HES cement
@ 30°/0 trass: 70°/0 HES cement
e
30°/0 ash: 70°/0
HES cement
Expansion = O.15%
Expansion =0.05’%
Expansion = 0.02%
The authors suggest that the expansions are due to the formation of stable ettringite with time, derived 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 heatcured sample, but not in the sample continuously cured at 20°C. Where calcium carbonate is present 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 adding K20. Furthermore, the sulphate type that was used was varied-either gypsum or 1:1 anhydrite:hemihydrate.
Figure 5.2 shows the expansions of the disks measured at a constant time after the start of exposure, 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 manufactured with one of two
clinkers in which sulphate was added to give either a 4’%o or an 8’%S03 content. After a delay period 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.
chapter
5
45
40
.~
~L.___J
-
35
■
-“’””””””r
❑
O HR. DELAY
E
E
25
c
“$ 20
"""""""`""""""""""""""`""""""""""""`.n............................
HR DELAY
..
<30
Page 46
■
,,
., ....................... ...
.,,
,,,
.,
.,,
,.........”
,,,
.,..,,,,.-,,,.,,,,,,.,,
,,,
-,-...
...,,,,..,.,,,
,-,
,,,
...
,.,
,.. -,,
..,
.,,
.,,,,--,,,
,”,-,,,,,
,,, -.,,,,..,,,,,
3 HR DELAY
..
...................................... ................. .....................
..................
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
Cllnke
I
S03
‘r&
Storage % by
(hrs)
mass
(data from ref. 268)
Heat Treatment Temperature
8
9
40 c
80 C
Relative Humidity Durtng Heat Treatment
60%
95 o/6
60 o/’s
o
95 /’s
Subsequent Storage Temperature for 3 Years
Dark Shading - designates otrongcraok formationaooompanladby obviouonew “phaseformation”
LightShading - designatesslightcfaok formationwith “insignlfioant’now phase formation
No Shading- Indisatssno eubstanflalvisual ohanges to prismsduring3 YOWO
storage
The qualitative results are presented in Table 5,1. Heat treatment at 40”C did not result in significant 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 behaviour.
5.3
1987 — Heinz and Ludwig
[132]
This paper reports on continuation of experiments started by Ghorab et al [106]. Again, experiments 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 occurred 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 soaking). 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 specimens immediatelyafler heat-treatmentindicatedthat the amount of ettnngite decreasedas treatment temperature increased.Immediatelyafter a 70”Ctreatment X-ray
peaks correspondingto the monosulphate (AFm) phase were observed and, as treatment temperature 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 sulphates are better incorpomted into the C-S-H stmcture. Thus, at high temperatures ettringite decomposes, 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 aluminate 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 performance 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-
,
!
~ji
_..&._..~,.,,.,,
,*4
.T- .. .. ...n.~.-.-~ . .. .. ..n-.F.-”-”-~-’~
~
--i
!-
;,,.,;,
”-”-~ -’-F-
1
I
i~!~
12.5- ,_...__+
I
!
I
!
l-~~11~
~-
i7’++
+ ‘
--~.-Q-~_-. ‘o_+&
------0----+ ‘--\
- ‘“n.,., I
~
II
...
..-..+-.
.“-..-.......-..-+.
-..-..-..
11,5., .,...,_._.&...&.
-.....“b....
< : -$ L.------Q-.
.-$s-:
~
~
+ = !@ & . ,
11.o- “
I
‘...,
*,.
: ; jl
%,.,<r-.+
------- .-*”------+-------.-0-...’---’ .& .+..+.,,
10.5- “
I
‘
II
!1
II
10.0 ~
)
10
100
1000
-
. . . .. .
1
WEIGHT GAIN, %
4.0 ‘
3.5 ‘“
)
3.0 ““
2.5 “‘
/ R
2.0.
1.5 “‘
/
P
,#...
- -,....4----4
A.....’-..”’
,/
-0’
. /J~
4
A *
1,0 J
0,5- 0,01
10
100
1000
EXPANSION, mtim
12
10
.
T
8
~.- o’
0“
,//
6
,/------,
I
I
/
I
,’
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)
chapter
5
Page 49
3.00
2.50
2.00
■ s
●
4
WEIGHT GAIN
%, AFTER 28d 1“50
h
b’
■
?“
9:
1.00
■
0.50
1
I
0.00
1
0.00
1
2.00
4.00
1
6.00
1
1
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
~
i
0.00
(
CHANGEIN
RESONANT ‘0”20
FREQUENCY
qrzzad
1
0t 2“00
d.oo
b
6.00
1
1
10.00
8,00
●
m
■
-0.40
#-*
m
■
-0,60
-0.80
-1.00
EXPANSION mtim, AFTER 28 DAYS
Correlation
Figure 5.5
Between Expansion and Resonant Frequency
(data from ref. 132)
1;
chapter
5
.-
Page 50
EXPANSION (mm/m)
—
—+——— &A=,44-,66
10 --
..----
0“-- ~lA=.67
-----+-8- -
6- -
4- -
2- -
--- -+
~/A=,79
--- 51A=.88
~
~/A=l.5
‘--.& ---
&A=2,j
TTT
s
3
--l
10
100
1000
STORAGE PERIOD (days)
Series II Experiments,
Figure 5,6
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 Canadian cements) the rate of expansioncan be dormant for
as much as 1 year and then “suddenly”accelerate. At the very high WA mtios (1.5 and 2.1) the
mte of expansion is appreciable from the start of storage and reaches 0,6% 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 Ludwig noted that a change in specimengeometry from 40x40x160mmprisms to 10x40x1601nxnprisms produced expansionsmuch earlier,
In Series III experimentsthe effect of relative humidity of storage was examined,The researchers
found that below 95% rh no expansionor reduction in resonant frequencyoccurred,However, if
specimens were stored at 60% relative humidity for an extendedperiod (120 days) and then
placed in a saturated atmosphere,the subsequentdamagingnmtion and expansionwas intensified (Figure 5.7).
Series IV experimentslooked at the effect of water/cementratio and air-entrainmenton the expansion process (Figure 5.8), A finding significantto the Canadianenvironmentwas that air-entrainment suppressesexpansion, At the same time other experimentsshowedthat freeze/thaw
cycles magnifi the problem; microcracksoccur during f/t cycling which provide sites for formation of ettringite.
4
3
2
1
0
%
WEIGHT GAII
.L..-
.
m
.,..-..,
-2
...
.+. -,
J-=
/-.”’
,,R.
. .. ..
- j. ❑ .. ...-Q. -.-c?”
(
Q
t
—
,’
..
—
I
,/,/
)0
/1(
1
-1
i
!
Do
,/
/
f
-3
-4
- .-..<.,
-+..---..,.+,_
I
-........
_..
,4
!
....... ..
.-i.+
----~
-5
b
—
—
Humidity Change 60% to 100%
14
XPANSION
ntim
—
— —
12
/
r
10
8
f+---
6
4
/
I
]60% to 100?4
2
,/.’
/.”’
/
0
100
10
STORAGE PERIOD (days)
I
~
WATER SOAK
‘--n---
100% R.H,
--.+...-.. 604 oo~eR.H.
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
I
Page 52
Chapter S
10
Expansion (mm/m)
T
9- ~
w/c=o,4 &
w/@0,5 with air
entrainment
6- -
---cl-----
w/*o,5
5 ‘-
----+---
w/~o,45
4 ‘-
---*---
WIC=O.7
8- 7- -
3 -2 “1
3
. I
/
9’
/
1’
1’
TTT
f
A
100
/
/’
/
k
$“””
d
/
‘“
/
,
I
,6
4
I
1000
STORAGE PERIOD (days)
Figure S$
Effect of Water/Cement Ratio and Air=Entrainment on Expansion (data from ref 132)
delay perlod=2 hrs,, treatment for 12 hours at 90*C,mork bars with HES cement -
5.4
1988-
Sylla [297]
Sylla reported upon Europeanexperiencewith secondaryettringite formation, As temperature
rlsa in concrete the reactivity and speed of dissolutionof C3Aincreases, while at the same time
the availablesulphatein solutiondecreases,Therefore, there is an increased tendencyto form
monosulphaterather than ettringite,l%e importanceof the treatment temperature and heat-treatment period on the quantity of ettringitepresent are clearly shown in Figure S,9, Results are reported from pastes made with two cements: both cementshad approximatelythe same sulphate
content, but cement B had 12,1%C3A.Cement A had 8.7% C3A.The extra aluminatein cement
B results in monosulphateformingin some pastes when treated at 60 and 80°C, The high C3Acement also shows a more rapid reductionin ettrh@e content,
Sylla notes that the formationof monosulphatealonedoes not account for all the sulphatereleased when ettringitedisappears,He postulatesthat the predominantpotion of sulphatejs “increasingly attached”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 ettringite formed during the delay period remained, even @r 5 hours of treatment at 80°C,
Sylla also performedexperimentsto determinewhether new phases form during storage after heat
lreatment.Paste specimenswere stored at 20°C and 100%relative humidity after 5 hours heattreatment at 40,60 and 80°C. There was no delay time. Results are given in Figure 5,10.
For 40 and 60°C treatmentthere are no large changesin ettingite content as sorddngtime proceeds; In fact for cement B ettringitecontent decreaseswith soaking time. After en 80”Ctreatment however,there is a very large increase in ettringiteduring the first 28 days of soaldng and a
chapter
5
X-RAY
20
PEAK
Page 53
INTENSITY
20C
–
18-
-
40C
16- -
E
Ettringite and monosulphate peak measured
immediately after heattreatment
12
10
8- 6- 4- 2- t
1
m
,
2
3
4
0
1
y
\
5
Cement A:
3.55% S03, 8.7% C3A
HEAT-TREATMENT PERIOD (hrs)
/
X-RAY PEAK INTENSITY
30 –
=:
25- -
~
Zoc
Cement B:
3.36% S03, 12.1% C3A
Treated without prestorage at given
temperature for periods
from 1 to 5 hours.
15- -
10: L
w
5- ---
0+
1
---
-*
2
-- - O---;’----*-.
..
A
w
3
. .
&
------11
I
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
EITRINGITE
PEAKS
CEMENT A
WC
Note Iarga Inmass in eltrkrgks with tires for 80C curing
~
,,,,
\::::,,;,,,,
.: :.:
Y
40c w ,:::,,,,
:::.:/..::{
::.
,:;!:;
60C
MONOSULPHATE
PEAKS
M C
“’.
28
CEMENT B
\
o
?
SOAKING
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 period.
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 avoid darnagethe maximumtreatmenttemperatureshould be kept
to less than 70°C. There is no clear indicationof the extent of delay period required; prestorage
(delay) for 1 day to 1 year prior to heat treatmentmay lead to more damage because the structural
matrix is stiffer and thus is less able to withstandthe build-upof pressures.When the delay period
is within the first 24 hours then 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 cement).
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 alumina 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 susceptible 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 finding from these authorsis that inclusion of 14%air-entrainmentinto the
mortiu “drastically
y“ 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 potential 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 1:3:0.5 cement/sanct/water. Twenty U.K. cements, mostly commercial rapid-hardening cements 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
4
16
A
‘A
14
I
1
,
12
❑
■
I
1
o
Suggested “Safe Ratio” of
2.0. Cements with ratio less
than 2,0 not susceptible to
seoondary ettringite attack.
■ M
.
,
10
0
●
0
I
I
■
+
a
#
8
I
1
,
■
:0
o
6
1
1
1
I
■
■
o
t
4
I
I
i
2
1
O%**
❑
14
@*&
0
A-W+
0
,
II
r
I
2
4
I
I
I
8
10
J3&ia
12
14
16
la
@/A RATIO
Illgure 5.11
Scattergramof ?/A ratio vs. Expansion(datafrom ref. 133)
lowedto cool naturallyuntil 16 hours when they were stored in water at 20°C, The aevemtreatment was chosen to reveal damagein the shortest amountof tlmq the regime chosen is more “ex=
treme” than is likely in practice,
Examples of expansive and non-exansiveresults are shownMFigure 5,12; other results are given
in Figure 5.13, Lawrenceet al were able to classify the types of expansive behavlourinto 7 categories:
~
(a) zero expansion— e.g. cementES3579,Figure 5.12
*
(b) sm~l continuousexpansion
~
(c)small acceleratingexpansion
=
(d) l$rge =elerating expansion— e.g. cement ES3594, Figwe 5,13
~
(e) @el~ating or limitingexpansion
=
(f) large S-shapedexpansionIesdhg to a definitemaximum— e.g. cements ES3595
(Fig. 5.12) and ES3572 (Fig. 5,13),
chapter
5
Page 57
EXPANSION,
1,00
----
0.90
t
0.80
t-
Yo
----
----
----
----
----
----
----
----
----
----
----
----
----
----
----
----
----
----
---
----
----
----
----
----
----
----
----
----
----
0.60
----
----
----
----
----
----
----
----
----
--
0.50
----
----
----
----
----
----
----
----
----
---
0.40
----
----
----
----
----
----
----
,-#_#”f
0.30
----
----
----
----
----
---
0.20
----
----
----
----
----
--- J
0.10
----
----
----
----
-
0.70
t
—
---fl
Exampleof=pansive
CementES3595
-
---------------
----
----
w=’
Exampleof Nonexpansive
Cement ES3579
0,00 1
SQUAREROOTOF
Typical Expansion
25
20
15
10
5
0
TIME(days)
Figure5.12
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
~
ES3740,3HRS
0.50
0.40
— *
– ES374016HR
0.30
----
0.20
----
----
--
*--------:
X 2X’------------
--’
0.10
0.00
0
5
10
15
20
SQUARE ROOT OF TIME (days)
Expansion
Figure 5.13
vs Time Curves Showing Effect of Delay Period (data from ref. 177)
25
chapter
5
Expansion
Page 58
Table 5,2
vs. Sulphate Content (ref. 177)
The addition of 1’%0
S03 to cement
ES3578 changed its behaviour flom
a class (e) to a class (b) category—
the addition of sulphate removed the
tendency to reach a limiting expansion and resulted in a more gradual
expansion process,
4.00
4.64
1.00
0.70
1.64
5.27
2.27
1.10
1,50
The addition of sodium sulphateto
some cements did not change the
time at which expansionstarted, but
the magnitude of the maximumexpansion increasedas sodium sulphate content increased.Table 5,2
showsthis effect.
Figure 5.13 illustrates the influence of duration of heat treatment on subsequent expansion. Specimens that were heat-treated for 16 hours, mther than 3 hours showeda reduction in the time to
the start of expansion,The effect is pronounced for cement ES3740; the cement that showedvery
little expansion when treated for 3 hours showedthe maximumexpansionwhen treated for 16
hours.
Petrography indicated that for specimens which crocked,the crockswere filled with ettringite,Ettringite also formed ridges around surface cracks. Expansion did not necessarily mean that the
specimens cracked, however; some bars showedas much as 1.2%expansion without cracking,
Analysis of thin sections showed rims of ettringitearound sand grains. An “acicular” ettringite
morphology existed within these rims,
Lawrence puts forward a number of points to explain secondaryettringite formation:
Any ettringite that forms during the delay period decomposesduring heating at 75°C or
above
Water storage at room temperatureafter heat treatment results in the m-formationof ettringite without significant changes in the quantity of monosulphate.For ettringite to
form in preference to monosulphatethe pore solution must be high in sulphate content,
Lawrence proposes that this is the result of modificationof the calcium aluminate hydrates — hydrogarnet like materials— during heat treatment; the modified aluminate
compounds cannot rapidly absorb sulphatesand allow free sulphate ions, such as those
slowly released by C-S-Hto persist in solution,
The natural process of slow release of aluminateswith time then results in ettringitecrystallization from solution,
A less severe temperature regime does not modi~ all of the aluminates. Some can mpidly absorb S04 immediatelyon cooling. Monosulphatethen forms instead of ettringite
because there is insufficientsulphate in solution,
Ettringite formation occurs at nucleationsites: on the surfhceof voids, in microcmcks
and at the cement-aggregateinterfhce,
The air-voids or flaws must become filled in order for internal stressand concrete disruption to occur,
chapter5
5.7
Page 59
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 temperatures greater than 80”C. Heat treatment was found to be an essential component if expansion during 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] supported these initial findings.
The work of Heinz& Ludwig, reported in 1987 and later in 1989 firmly established the phenomenon of secondary ettringite formation. They found several crucial factors influencing the potential 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 ettringite effects
the cement type and particularly the ratio of sulphates to aluminates are important for determining the potential for secondary ettringite formation. In 1987 Heinz& Ludwig established that an WA ratio 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 established 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 explains 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 observed by several researchers and indicates that one simple method of deterring damage
due to potential secondary ettringite formation maybe to ensure that adequately entrained concrete is cast.
Page 60
Chapter 6
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 cement 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 fbrther test results given below, indicatethat what is “optimum” for strength and retardation is not necessarily(and perhaps not normally) optimum with respect to long-term stability 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 formation of sulphoahuninates is accelerated by an increase in temperature, Therefore, one might expect 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 blended with gypsum to 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 curing cycle consisting of 4 hours delay, 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. Deformationreadings on prisms were taken to an age of
48 weeks,
Strengths at ages of 28 and 182days are shown in Figures 6,1 and 6,2 (the notation C1-70A
means concrete made with cement 1 heat-curedat 70°C), Both figures indicatethat the cements
should have between 3.8 and 5.7% S03 content to produce optimum strengths of heat-cured
Compound
Composltlon
..
Table 6.1
of Clinkers used in A1.Rawl Study [6]
Page 61
Chapter 6
65.0
60.0
—
C1-20A
~
C1-50A
~
C1-70A
40.0
~
C1-90A
35,0
-
55.0
50.0
STRENGTH 450
MPa
“
30.0
-
~
C2-20A
C2-70A
25.0
, , T , {
, ,
r ,
r
I
20.04
1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5 8
S03 CONTENT%
Dependence
Dependence
Figure 6.1
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
of 182-day Strengths on S03 Content and Curing Temp. (data from ref. 6)
Chapter 6
Page 62
specimens. On the other hand room-curedconcretes show a clear regressionin strengthas S03
content hicreases above 2%. General conclusions,however,should not be made upon the basis of
only two high C3Scontent cements.
contents for strengthare not necessarilycompatiblewith optimum contentsfor volume stability.Figure 6,3 showsthat S03 contents above about 4% 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 weeks of water-soaking,
optimum S03
me results given in Figure 6.4 are contrary to the commonly-heldview that heat-treatmentproduces an increase in expansiondue to secondaryettringiteformation.Damage due to SEF has
been connectedto the productionof microcracksduringheat-treatment.Such cracks are reduced,
or eliminated,by provkiing an adequatepre-curingperiod. Note that A1-Rawiemployed a 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
1
!
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.Many other Countriesin the world follow the Americanstandar~
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
Curing Time, Weeks
Figure 6.3
Effect of S03 Content on Expansion of Normal Fog-Cured Specimens (data from
ref. 6)
50
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 Colombia and Hungaryaxethe oniy three countriesnoted in the Cembureaureport that allow 4.5% S03 in cements which contain as little as 7.5% C3A (Colombiaand Hungary set a 4.5%
S03 maximumwhich is independentof C3Acontent).Canada and those countries that use ASTM
C150-89 also permit S03 contents greater than 4.5% if it can be shownthat mortar bars 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 longterm stability?
Lerch and Ford in 1948 [183] tested prisms made from 27 differentcements that were moistcured for 1 day and then stored continuouslyin water at 21“C for 5 years. The S03 contents of the
cements were fairly low, but the results, shownin Figure 6.5 point out the lack of correlationbetween expansionsat 14days and expansionsat 5 years. Several cements show less than 0.02% expansion at 14days, but show appreciableexpansionsat 5 years. The test results help to accentuate
that prediction of 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
of
formation could be importan~note that other factors also ~ntribute to the ~~~~tion
whether a 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 avoid large expansions[3171.
Chen and Grattan-Bellew[55]in their study of alkali-aggregatenmctionadded various quantities
of KzSOAand NaX30qto a clinker from a single cementplant. A suite of 17 cements was manu-
Page 64
Chapter
6
0,250,2- “
0,15
EXPANSION
AT5 YRS,%
0,1
?“
I
O“O:b=zaLL0
0.005
0.01
0.015
0.02
0.025
0.03
0.035
0.04
EXPANSION
AT 14 DAYS,%
Scattergram
Figure 6,5
OFExpansion at 14 Days vs Expansion at 5 Years (data from ref. 183)
fhcturedwhich containedvarious alkali contents. The cements were intergroundwith an optimum quantity of gypsurnto a fineness of 370 m2/kg.Expansion was then monitored with time of
storage in a moist environment.A good correlation(r=0,91) was found between acid-solublealkali content and expansion.However,fbrther analysis of the data for the present report also
showed a positive correlationbetween expansionand MgO content (Figure 6.6) and between rate
of expansionand SO#Al*Os ratio (Figure 6,7); both comelationcoefficients (0.73 and 0.80) are
significant at the 5°Alevel.
Chen & GnMan-Bellew also considered the potential effect of sulphates and looked at the hypothesis that “ettringite in cracks and air-voids might be contributing to the late stages of expansion”, They found that there was no correlationbetween mte of expansion of the high-expansion
bars (those with Na20 equivalent>1.0%) and the amount of ettringitoin the moctarat 6 months.
6.3
Speclficatlons for Heat Treatment of Concrete
Concern over possible problems with secondary ettringiteformation has prompted some mresearchers
to recommendheat-treatmentguidelines,some authorities to change their specifications, and some authorities to consider changingtheir specifications,Proper heat-treatment
methods would theoretically eliminate,or substantiallyreduce, the potential for problems even if
the cement and/or aggregate were found to be inferior,
Neck [237] proposedthree types of heat treatment,dependingupon the proposed durability exposures, with the objective of avoiding damage due to secondaryettringite formation:
Type 1:1 hour pre-cure with a maximumsoakingtemperature of 80”C— for interior exposures with no moisture
Type II: 2 hour pre-cure with a maximumsoakingtemperature of 70”C — for exterior exposure with no contact with the earth
Page 65
Chapk?r
6
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
.6
.5
S03/AL203
Scattergram
of SO#Ai@3
.7
.8
RATlO
Figure 6.7
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 temperatureduring the first 3 hours of curing cannot exceed 30°C and cannot exceed40°C during the first 4 hours
~
The rate of temperaturerise cannot be greater than 20 OC/hr
= The average maximumtemperaturecannot exceed 60°C; individualtemperaturereadings
cannot exceed 65°C
~
The codng rate aftexheat-treatmentcannotexceed 10 OC/hr
~
Concrete must be protectedagainst moisture10ss
Also in 1989,The German Committeefor ReinforcedConcrete [103]recognizedthat intensive
heat treatment can affect durabilityof structuralmembers~e,
Faults in the
concrete due to too rapid warmingor sharply differing expansion are susceptible to “sulphate
bonding [that] can occur with a high level of moisture in the concrete”. Accordingly, speciflcationa for damp conditions are different for dry concrete, as shownin Table 6,2, The 60”C maxL
mum temperaturefor concrete that will be damp in service is clearly a change made to address
potential
problems with [email protected],
The British have been slowerto respond,perhaps because they have been waMngfor detailedresearch results from the British Cement Association(e,g, Lawrenceet al [177]),It 1sinteresting
that the British code of practice in 1972CP110:1972noted that damage due to excessive heat
treatment would not occur if
~
temperaturerise during the first 3 hours 1snot greater than 15°C!/hr
~
thereafter, the rate of increase or decreaseis not greater than 35°C/hr
= the maximumtemperaturereached by the concreteis not greater than 80°C,
However,the 1985British code of practice BS8110:1985ondtted the above thres guidelines,
Other Brltlsh authoritieshave taken steps to 1111
the gap, The British Depc of TransportSpecifications for HighwayWorks (1986),Part 5, Clause 1709:5(Ii)requh’es(1) the concrete must be lefl
for 4 hours without addkional heating; (2) the concretetemperaturecannot be raked at a rate
Table 6.2
Specifications for Heat=Treated Concrete
German Committee for Reinforced Concrete, 1989 [103]
I
Minimumholding (delay)period (hrs)
Moisture Category
Dry
Damp
3
+
30
Maximum concretetemperatureduring holding, “C
Maximum heating rate, OC/hr <20*
80
Maximumconcretetemperature**,“C
heating rate should be reduced to 10 OC/hrfor lightweightconcrete
** individual values may be up to 5°Chigher
or
4
+
40
<20*
60
Chapter 6
Page 67
greater than 20°C/hr; (3) the maximumconcretetemperaturecannot exceed 70°C; (4) the rate of
cooling cannot exceed the rate of heating [177].
[n North America measuresto control excessiveheat treatmentshave been taken on a locaJscale.
The results of Merritt& Johnson [218]on concrete strengthdevelopmentafter variousheat treatments were incorporatedinto the Iowa State HighwayCommisionspecificationsfor precast con‘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 included 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 above three 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 applied (this is normally 2 to 4 hours but could be as low as 45 minutes). During this time the concrete 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 temperature tlom 80°C to 70°C.
6.4
Comment
Determination of “optimum” gypsum content involves the determination of the proportion of gypsum 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 products manufacturedwith Portland cement needs
further consideration,
Clearly, cement is just one ingredientof concrete. Other ingredientscan affect durability as can
the mix-design,casting and curing procedures.However,all other factors being satisfactory,it is
necessaryto be able to evaluatecements for long-termdurability.This requires that a satisfactory
test method is avtdlable,
It is unrealisticto expect a good correlationbetween the expansionof mortar bars soaked h water
for 14days and expansionador cracking of concrete at 10or 15 years, The scattergramof Figure 6.5 showedone example of the poor correlationthat exists between short and long-termexpansions;the cements were manufacturedat rrdd-century,It is Mghlyunlikely, for a number of
masons, that the correlationfor moderncementsISbetter,
The viabilityof the 14-daysoakingtest should be *examined in the light of the chemistry and
physical propertiesof modem cements. ‘Ms is especiallyimportantgiven (a) secondaryettrlngite
formation may bean importantdurabilityissue, and (b) both ASTMand CSA standardsallow
more than 4,5% S03 if a cement can be shownto pass the 14-daysulphate expansiontest.
In CanadL A23.4 specificationsfor precast concrete am presentlyunder revIsIomIt is suggested
that careful considerationshould be paid to the potential secondaryettringlte issue and to the German specifications(Table 6.2) that have addressedtMspotentialproblem,It is especiallytroublesome that the proposed A23,4 still relies onthe setdng time to define the minimumdelay period
prior to treatment this setting time could be as short as 45 minutes and still satls~ both A23,4
and AS standards,
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 concrete for many years. For example, in 1949 Scholer [282] reported on results of a revised accelerated 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 accelerated 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 recognized 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 paper]), the “accelemted” exposure lasting 285 days would hardly be considered “accelemted” today. 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 subsequent 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 maximum 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 ettringite 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 remainder of this chapter, because it is purported to “measure” the effect of secondary ettringite formation of various concretes. This is the Duggan test, whose development, results, and
chapter 7
Page 70
48
FIELD
DATA FROH SCHOLER,
eamasl
CORR, COEFF.
EXP,
=
-,017
+
1.382
1949
■
*
,707
ACCEL.
EXP.
0.25
0.2
0.15
!1
,.. . ..”!
0
! 0
. ..’”
,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,
0,1
0.06
0
!
0,02
0.04
0.00
ACCCLCRA7E0
Scattergram
ofAccelerated
I
I
0,08
0,1
CXPANSXON,
-1
0,12
X
Figure7.1
vs. Field Expansions
(data from ref.282)
significanceofmsu
lts
havebeenreported indetailinDuggan
&Scott [91-93]; Scott&Duggan
111];Jones & Gillott [155]; Attiogbe et al [12]; and Wells et al [317],
[284]; GillOttetal[110,
As originally devised the Duggan test involvesthe followingprocedure [284]:
Concrete cores (5 minimum)are taken from.a structureor from laboratory-castprisms or
cylinders, The cores are 25mm in diameterand at least 65 mm long and are cut to 50mm
lengths. The ends are ground smooth and parallel,
Initial length measurementsare taken with a comparatorjust before the start of heat treatment. In later tests by Duggan& Scott and in tests by Gillott et al [110,111]this measurement was used as the zero point from which other strain ~dings were taken.
Cores are soaked for 3 days in distilled water @21°C in a closed container
Cores are then placed in a @-air oven at 82°C for one day
Cores are removed from the oven, allowed to cool for 1 hour, then placed back in distilled 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
=
7.1.1
Length measurements are taken relative to the steel standard at intervals of 3-5 days.
Significance
of Duggan-Test
results
The test was initially thought to be an accelerated means by which potential alkali-aggregate reactivity of various cement/aggregate combinations could be assessed. Scott & Duggan set about establishing 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 conclude 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 established), 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 expansion 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 commended 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, cement 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 significance 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 Expansivity of Cement/Aggregate Combinations (CSA A23.2-14A).
Page 72
chapter
7
I
0.8
❑
❑
AGG. A (INERT)
AGG. B (REACTIVE)
❑
AGG, C (REACTIVE) I
m
I
0.7
0,6
0.5
EXPANSION,
‘/0
0.4
0,3
0.2
0.1
0
0.95
TIO
0.95
T 30
1.00
T 30
1.00
T 30
1,10
T 30
1,07
TIO
1.17
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)
Through carefhl experimentationand comparisonsof the response of the various concretes to 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 proportion of ettringite and frequency of microcracking in the concrete”,
Other more tentative conclusionswere that (a) Duggan-testexpansion depends primarily on the
properties of the cement, and (b) the form of the sulphatethat is present (either gypsum or anhydrite) may affect the extent of microcrackingcaused by the Duggan test, and thus may affect subsequent expansions.
In a second set of experiments, reported in detail in Gillott et al [110] both cement pastes and concretes 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 aggregates 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/aggregate 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.
Page 73
chapter 7
1,2
EXPANSION OF CONCRETE
1.0
0.8
M
~
CEMENT A
—a-—
CEMENT B
–
CEMENT C
‘“–
~
.
!!
0.6
..-.
----
I
----------------------/’
-----
-----
-----
-----
/
-----
----
CEMENT D
—-&—
----
CEMENT E
/
--
---
A
--
B
-----
0.4
-----
-----
-----
- --./
30
40
-. / >4
------------------
--
0.2
0.0
10
0
20
50
70
60
80
90
SOAKING TIME, DAYS
EXPANSION OF CEMENT PASTE
,
.-
1
~
CEMENT A
----
----
----
----
----
----
----
----
-
— U- – CEMENT B
.-
—.
●—
~
--
CEMENT C
-----------------------
----------
CEMENT D
— -A- — CEMENTE
0.4
----
. . . . ----
0.2
----
-.
----
‘ ---------------
---
‘ - - - - - - ---”----_—----
.F
----
----
----
-
----
--
~/ “F
-R-
---- T.-Y----:----.:-7
:::
; ::
:?::
t=
0.0
o
10
20
30
40
50
60
70
80
SOAKINGTIME,DAYS
Expansion
Figure7.3
of Hardened CementPastes and Concretes Made with theSame
(data extracted from ref. 110)
Cements
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 industry became rather concerned about the Duggan test; Five of the six concretes made with the Nelson 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 manufactured 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 concretes” at soaking times up to approximately 40 days. Beyond 40 &ys, however, behaviour between 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 remain straight or, in the case of the concrete with Cement D, become concave upwards, The concrete 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
❑
48 DAYS
●
90 DAYS
1al
==%”””
/“
0’
la’
i?
k
0,80
/“
/“
/
, /“
5
& 0.60
b
●
❑ . ‘n ’
❑
[ 0.40
k
bt
/’
/
■
+
0,20
•1
O,cm
/
/’
,
❑m
‘/ “
/’
/’
‘/’
.
1
Oocm
0.20
t
I
I
t
0.40
0.60
0.80
1.(x)
% EXPANSION OF CONCRETE AT GIVEN TIME
Scattergram
Figure 7.4
of Concrete Expansion vs. Paste Expansion (data from ref. 110)
1,20
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 becomes more concentrated in localized regions such as in microcracks and at cracks at the aggregate/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 pessimum 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 molar 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 significant 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 tictors 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 parameters which show significant correlation — other than the 13/Aratio).
The severe heat treatment of the Duggan test causes microcracks in the concrete. Ettringite 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 measurement 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 Figure 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.
Page 76
chapter 7
t
1!20 ‘
1,00- -----------------------
-
0.80- . - - ” - - - - - - - ” - - - ”----”-”t
ii! 0.60. . - - - ” - - ” - - ” - - - -------””
J=
B
K
OAO-- -----------------
-
-------
-------
21 OR22
DAYS
v
000 +
0.8
C3
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. -----
. . . . . . . . .
OR 48
DAYS
0s)0 -
008
049
1
4!5
0
1,1
1)2
1,3
1.4
103
1,4
MOlARRAT10303/A1203
1.20 ‘
i!!
8
I
Osxl ~
0.8
0.9
1
1.1
1.2
MOLARRAT10S03/A1203
Observatlonsofa
Figure 7.5
PessimumS03/A1203
Ratio atVarious
Ages (data from ref. 110)
Ptwe 77
DATUM LENGTH MEASUREMENT
SATURATED SPECIMEN
AlllOGBE et al
(A)
ATIIOGBE
REGIME
m
distilled water
at roomtemperature
(b)
+
r
/
mm
LLH_Hl”
DUGGAN
REGIME
3
/
9
45678
10 11
12
TIME (CiEty$)
o,l’%
I
I
(c)
DEFORMATION VSTIME
El
STRAIN
A=
.
1ST &2~JE-SAT
●
o.1’%
cmtraction
FIRST DRY*
C
3RD DRY
2ND DRY
,.
..
b. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..z
STRAINS AT END OF EACH DUGGAN CYCLE
Temperature
DUGGAN ZERO STRAIN, EARLY E)(PTS
Figure 7.6
Testing R6gimes, Data and Examples Strains in Duggan and Attiogbe Tests
chapter 7
Page 78
2,5
2.0
~—
CEMENT A
~
CEMENT B
~
CEMENT C
~
CEMENT D
1.5
●
EXPANSION
‘Ye
/’
1.0
0.5
0.0
o
5
10
15
20
25
30
35
TIME (days)
Expansion
Figure 7.7
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 solution 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 volubility 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 contain “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 expansion, 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
1
CEMENT D
0.15
0,1
J
■
CEMENTB
~ CEMENT C
n CEMENTA
t
I
0
I
PETE_
@ CONCRETE
1
,
0.05
CEMENTC
CEMENT A
o
1
0.5
1,5
2
2.5
STRAINS MEASUREDBY ATTIOGBE
AT 20 DAYS,%
Scattergram
Figure 7.8
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 Chapter 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 [3 17] are somewhat 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% expansion 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 satisftiory interpretation of results:
Page 80
chapter 7
@ 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 formation, 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,
=
7.2.1
Duggan’s test results have not been correlated with field concrete,
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 substantial 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 component 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 because 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 definition 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 significant 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 r4gimes 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)
Type of corvxete tested
Old
Zero
Z1
Expans.
New
Zero
Z2
R&ised
Zero
Z3
B: Non-reactive agg., silicafum~ -0.065
C Non-reactive agg., silica fume, repaa -0.051
0.055
-0.010
-0.054
0.074
0.010
-0.062
0.066
0.016
0.041
-ma
0.045
-0.006
-0.052
0.062
0.024
-0S)62
0.056
0.02
-0.024
-0.021
0.015
-0.006
-0.050
0.044
-0.012
-0.044
0.096
0.04
0.052
A Non-reactive agg., low alk. cemen
D Non-reactive agg., low alk. cemen
E: Expensive oonc., failed in servkx
F: Non-reactive agg., high elk. cemen
Q: Reactive agg. #l, low-alkali cemen
H: Reactive agg $2, low-alkali cemen
-0.067
0.046
-0.021
-0.030
0.041
S3
0.25
0.027
0.060
-0.004
4.079
0.067
0.004
-0.076
0.322
0.246
0.235
0.050
0.000
-0.072
0.062
-0.010
-0.060
0.304
0.234
0.242
-0.020
0.016
-0.004
0.010
0.000
0.006
-0.027
0.119
0S)98
0.119
-0.030
-0.050
0.096
0.016
0.046
-0.056
-0.066
0.006 %
0.062
❑
FIRST DUGGAN ZERO - Z1
■
SECOND DUGGAN ZERO - Z2
❑
REVISED ZERO - Z3
0.050
----
----
----
----
----
----
—
r
----
-
----
0.062
-0.064
0.2
STRAIN
AT
0.15
20 DAYS
%
-0.045
-0.050
0.35
0.3
4.010
--
----
----
----
---
0.1 “
o.050-o.05-
I
I
A
Comparison
B
I
c
I
I
D
E
F
CONCRETEDESIGNATION
I
I
G
Figure 7.9
of 20 day Expansions Using Three Zero Definitions
r’
H
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 resaturation 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 Table 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 component 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 absolute 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 cannot precisely differentiate between deformation caused by water-swelling and deformation 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 envisaged 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 (322) /rate (22-90).
It is suggested that if the Duggan testis to be developed into a standard test method, then definition of pass fail criteria along the lines of “rate of deformation” must be used, There are three distinct 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 complete 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 ettringite 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 cannot 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 combination 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 principal mechanism for some types of concrete. The number of cases that can be directly related to secondary 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 attributed 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 potential secondary ettringite formation problem in North America? Examination of the available literature 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 produce expansion and cracking of heat-treated concrete made with 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% — 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 GillottCement 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 mrequireseven higher sulphate contents to prevent rapid set.
The presence of cement, aggregate and steel together in concrete result in a weak interracial region 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 temperatures in the range 60-70”C and delay periods in the range 1-3 hours. Furthermore, some production 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 sulphates 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 formation 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 problem. 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 appear 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 tWnesses of cements used by the precast industry. Better test methods need to be developed to evaluate 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 preparedcements — cements where cement composition, gypsum content, fineness, and other factors 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 signMcance 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 accurate 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 National 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 Calgary 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 necessarily those of the Portland Cement Association.
Page 8S
References
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 kinetics and microstructure of ettringite, Proc. 6th Conf. on Cement Microscopy, Int. Cement Microscopy Assoc., Duncanville, Tx, 219-231, 1984
2
Abo-El-Eneim S.A., Hanati, S., Hekal, E.E., Thermal and physiochemical studies on ettringite. 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 ettringite 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., ElecPas@, proc. 7~
trical conductivity and microstructure of me hydrata c3A-3cas04
Intnl. Conf. on Cem. Microscopy, Int. Cement Microscopy Assoc., Duncanville, Tx, 277287, 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 methods, 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 cement 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 portland 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 expansion tes~ Research report, Sponsored by Canadian National Railways &Transport Institute, 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 hydration, 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 Purdue, 1975
18
Barneyback R.D., Diamond S., Expression and analysis of pore fluids from hardened cement 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 hydrate - 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
ing steam curing, Zero. Kalk Glps, V 24, n 10,473-6, 1971
systems dur-
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 concrete, 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 tricalcium 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 expansion 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 Report 90, Washington, 1966
40
Budnikov, P.P., Kravchenko, I.V., Expansive Cements, Proc. 5th Int. Symp. on Chemistry of Cement, Tokyo, Vol 4,319-330,1968
41
Building Research l%tablishmen~ Alkali aggregate reactions in concrete, Building Research Establishment Digest 258, Watford, England, 8, 1982
42
ButL Yu. M., Kolbasov, V.M., Timashev, V.V., High temperature curing of concrete under 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, August, 1952
Page 88
References
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 temperatures,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 fmmation 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 185188, 1975
58
ClarlG A.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, 229347,1985
59
Clastres, P., Murat M., Bachiorrini, A., Hydration of expansive cements - correlation behveen 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 microscope, 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 sulfoaluminate-type expansive cements: schools of though~ Cem. Concr. Res., V 14, n 4,
610-612,1984
.,
Page 89
References
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% 267276, 1985
68
Collepardi, M., Degradation and restoration of mssonry walls of historical buildings,Materialsand 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
pounds, Cemento, v 72, n 2, Apr-Jun, 53-60,1975
com-
71
Concrete Society, Alkali-silica reaction, minimizing the risk of damage to concrete: guidance 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 portland cements, Cemento, v 72, n 4, Ott-Dtx, 181-194, 1975
74
Coutinho, A. de S., As~ts
1,10-12,1979
75
CrsmmoncL N.J., Quantitative X-ray diffraction analysis of ettringite, thaumasite and gypsum 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 brickwork Cem. Concr. Res., v 15, n 6, Nov, 1039-1050, 1985
77
CrammonG N.J., Quantitative X-ray difllaction analysis of ettringite, thaumasite and gypsum 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
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
of sulfate attack on concrete, Cem. ConCr. Aggregate& V 1, n
bei 20”C, Zement-Kalk-Gips,
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-aggregate 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 individual 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 National 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 porposed test method for detection of potentially deleterious expansion of concrete, Resented 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, 132143, 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 nSnfodng 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 ettringite 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, 104109, 1986
105
Ghorab, H.Y., Abou El Fetouh, S.H., New approach to the hydration reaction of the tricalcium 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 calcium 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 sulfoalurninate 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. Highway 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., Mechanism of expansion in rapid test method for alkali-aggregate reactiom Final Progress Repo% 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 ettringite 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 Materials 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 engineering 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., Portland Cement Association Research and Development Laboratories, Vol. 2, No. 3, September, Also PCA Research DepL Bulletin No. 203, 1960
121
Grounds, T., Midgley, H.G., Newell, D.V., Carbonation of ettringite by atmospheric carbon 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 85,215218, 1985
123
Hadley, D.H., l%e nature of the paste-aggregate interface, PhD thesis, University of Purdue, 1972
124
Hampson, C.J., Bailey, J.E., On the structure of some precipitated calcium aluminosulphate 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 ettringite 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 Portland 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 ccmcretes subjected to heat treatment, American Concr. Inst., SP 100-105, Detroit, 20592071, 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 stability of ettringite, Proc. Intnl. Conf. on Cement Microscopy, 8th, Int. Cement Microscopy 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,
137
Hjorth, L., !%udieophold ved 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
November 7, 1990
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, Laboratory 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 Reactions, 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 influenced 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
SW,, 35,1484,1939
153
Jones, T.N., New interpretation of alkali-silica reaction and expansion mechanisms h concrete,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 Effect of Temperature and Humidity on Expansion and Significance of Ettrlngite Development 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 thesis, University of Maryland 1941
157
Kalousek, G.L., Analyzing SOs-bearing phases in hydrating cement, Mat. Res. & Std., 6,
292-304,1965
at 25 C, Trans. Faraday
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 interface, Materials and Structures, Vol 20, 19-26, 1987
160
Kelly, R., Solid-liquid reactions. Part 2: Solid-liquid reactions amongst the calcium aluminates 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-aggregate reaction by analytical electron microscope, Proc. 7th Intnl. Conf. on Concrete AlkAiAggregate 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 resolution 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, Improvement 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: hydration behaviowy synthesis, and characterization of trace metal oxyanion substituted ettringites, 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 substituted 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 Technique 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 ashPortland 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 ettringite 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, 915921,1986
179
Leifel& G., Muenchberg, W,, Stegmaier, W., Ettringke and thaumasite as causes of expansions 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 Research, 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, Zement-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 preventive 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 proposals 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. Assoc., 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
206
207
Page 98
Mehta, P.K, Scanning electron micrographic studies of ettringite formation Cem. ConCr.
Res., Vol 6,169-182,1976
Mehta, P.K, Morphology of csJcium sulphoaluminate hydrates, Journ. American CeV 52, No. 9,521, 1969
llUlliC
Sot.,
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 portland 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 studies 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 rnicrographlc 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 anhydrous 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 compressed 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 formation, Mater. Const. (Bucharest), V 10, N 3,123-7,1980
223
Monosi,S., Collepardi, M., Effect of silica fume on the hydration products of C3ACS(bar)H2 system, Proc. 8th Intnl. Congress Chem. Cement, Rio de Janiero, V III, 120124, 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, 812818,1988
225
Monteiro, P.J.M., Gjorv, O.E., MehQ P.K,, Effect of condensed silica fume on the steelcement 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 expansive 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 interfac, 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 sulfate 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-prmsure, 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 Beton, 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-(ettringite) 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 structure, Microstructural Development During Hydration of Cements, Proc. Materials Research 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 mechanism, 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 mechanism, 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 mechanism: 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 Alkali-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 ettringite, 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 aggregate 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 laboratory study for plant-produced prestressed concrete, A Research Investigation by Wiss,
Janney, Elstner & Associates Inc., Ill., for Prestressed Concrete Instititute, Chicago, December, 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 Principle of Crystallization Pressure, private communication, 1991
257
Ping, X. and Beaudoin, J.J., Mechanism of Sulphate Expansion II. Validation of Thermodynamic ‘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-compensating cement concrete, Dumb. of Concrete, 79-87, pub]. by Amer. ConCr. Inst., Sp-47, Detroit, 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, NovDec, 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 Repofi 1981-1984, Translated by TEX Associates Inc., Philadelphia, 1990
269
Richards, J.D., The effect of various sulphate solutions on the strength and other properties 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-Aggregate 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 cement and aggregate, Roe. 1989 Trondheim Conf., Amer. Concr. Inst. publication%SP114, Detroit, p669-688, 1989
277
Satav& V., Vepre~ O., Thermal decomposition of ettringite under hydrothermal conditions, 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 cementitious 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, Presented at the Fifty-Second Annual MeW.ingof ASTM, June27-Julyl, ASTM Roe. V, 942953,1949
283
Schwiete, H.E., Ludwig, U,, AlbeclGJ., Combining of calcium chloride and calcium sulphate 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 cement 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 in conventional standard and new accelerated testing, 8th Intnl. Conf. on Alkali-Aggregate Reaction, 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 dehydration 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 hardened Portland cement pastes, SillikattechniL V39, N 4, 120-2, 1988
295
Stefani& V., Abbiati, G., Negro, A., Bachiorrini, A., Mathematical model for the ettringite 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 compositions, 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, Congress Chem. Cement Rio de Janiero, V III, 323-329,1986
Page 104
References
305
Tognon, G.P. and Cangiano, S., Strength and Interface Bond in Concrete, in Bond in Concrete, 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 hydration 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
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, Principal 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 different 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., 199209,1981
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 hydroxide 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 subjected 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 secmdary ettringite formation, XV Conference on Silicate Industry and Silicate Science, Budapes~ 1989.6.12-16,392-395, 1989
Tsuyuki,
N., (h,
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 (infiuence 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 chioride 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 formation 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. Congress Chem. Cem., V 2, Paris, 11/88-11/93,1980
327
Zimbelman, R., A method for strengthening the bond between cement stone and aggregates, 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, fibrous 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 manufacture, 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 (industrial 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 steamcuring 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 dedolornitisation des granulats clans le 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, Manufacture 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 expansion, 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 concrete 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 Cements, Cement Association of Japan, Tokyo, 1968, p. 252
22
Filatov, L.G., Kuz’mina, L.A., Kuznetsov, S.N., 117, GEO, Investigation of the conversion of ettringite and portkmlite in cement during reactions involving both negative and
positive temperature, in Eksperimental’nyye Issledovaniya Mineraloobrazovaniya v Sukhikh 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 formation 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 heattreated 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 cement and the morpholoy of the resulting ettringite, Nov. Effekt. Vidy Tsemen@ 1981,705
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 Industry 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. Mineral., 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 inorganic 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, Concrete/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 atmosphere on concrete, GesellschafC FortscMtte der Mineralogies,Beiheft (Deu) Aug 1984, V
62, No, 1, p 127-128
42
Kollmann, H,, 199, CA, Mineralogical studies on efflorescence and the appearance of expansion 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 solution 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 aluminate 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 monosulfoaluminate 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 clans l’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 mecanisme 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 development in the area of railroad ties of prestressed concrete), Betonwerk + FertigteilTechnik 52 (1986) H. 6, S. 381/383
56
Ludwig, N.C., Pence, S.A., 237,23, Properties of Portland cement pastes cured at elevated temperatures and pressures, Journ. Amer. Concr. Inst., 27, 6, 1956
57
Ludwig, U., 238, 10, Concerning the setting and hardening of cements, Tonindustriezeitung, 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 determination 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 linear 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, Platres, Chaux, V 756, 1985,319-27
69
Negro, A., Stafferi, L., 313, CA, Ettringites and cement expansion, Cemento, V 64, n 1112, 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 hardening 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 Zhelezobetons., 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 (concerning 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 formation 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. Aleksandrovskii, 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 hydrating 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 between additive and cement matrix, Deutscher Ausschuss fur Stahlbeton, 1977, No. 283,537
87
Reinsdorf, S., 362,23, Unter suchungen uber &n Einfluss von Portlandzementen unterschiedlicher kilnkerabstimmung und verschiedener mineralischer poro ser Zuschlagstaffe
.... Unpublished Doctor’s dissertation, The University for Architecture and Building, Weimar, 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 products 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 ettringite, 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 temperature 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, Principal 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 stability of 3Ca0.A1203.3CaS04.30-36H20, Amual Mtg. of German Mineralogical Sot.,
Gesellschaf$ Fortschritte der Mineralogies,Beiheft (Deu) Aug 1984, Vol 62, No 1,258259
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 thermoturbulent 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
Aggregate
limestone . . . . . . . . . . . . . . . . . . . . . . . . . . 6,39
andtransition zone, . . . . . . . . . . , . ...6.7,39
Airentrainment . . . . . . . . . . . . . . ...52
Alkali-aggregate reaction . . . . . . . . . 4,6,71
Alkalis
effect onettringite formation . . . . . . . . . . . . 25
Attack mechanism . . . . . . . . . . . . . . ...4
Blast furnace slag . . . . . . . . . . . . . . ..39
Calcium silicate hydrate
andsulphates . . . . . . . . . . . . . . . . . . . ...23.33
Calcium aluminates
nXardationof
. . . . . . . . . . . . . . . . . . . . . . ...18
Calcium hydroxide . . . . . . . . . . . . . . .,4
orientation in transition zone . . . . . . . . . . 3942
Carboaluminates . . . . . . . . . . . . 21,28,39
carbonates
effect onettringite formatb . . . . . . . . . . . . 25
Carbonation . . . . . . . . . . . . . . . . . ...4
Cement
llneneas . . . . . . . . ..o . . . . . . . . . ..o .$.24.45
hydradonchemistry . . . . . . . . . . . . . . . . . . . . 16
Cement chemistry
andettringite formation . . . . . . . . . . . . . . 60-68
chlorides
diffusion of . . . . . . . . . . . . . . . . . . . . . . . . ,4,6
penetration of . .,,..,.0..
. . . . . . . . . . . ,,, . 4
Chloroaluminate . . . . . . . . . . . . . . .. 6,28
Corrosion . . . . . . . . . . . . . . . . . .. 4,21
Creep . . . . . . . . . . . . . . . . . . . . ..10
Crystal growth
andpressure exerted . . . . . . . . . . . . . . . . ...29
Duggantest,
. . . . . . . . . . . . . . . .69-82
Ettringite
wdmwtitimofdudnti,
. . . . . . . . . . . 18
andthe transition zone . . . . . . . . . . . . . . . 39-42
chemistry of . . . . . . . . . . . . . . . . . . . . . . . . . . 17
Crystal structure of . . ..o . . . .. o.... . . . ...17
decomposition of. . . . . . . . . . . . . . . . . . . . 6,22
deposition of . . . . . . . . . . . . . . . . .4.6.7.9.39
expansion mechanism . . . . . . . . . 29-32,44-68
formation atTectedby alkalis ..,.....,
,.. .25
formation a.fkcted by temp. . . . . . . ..8.22.24
formation of . . . . . . . . . . . . . . . . . . . .16.20.36
morphology of. , . ., . . . . . . . . . . . . . . . . ...28
precipitation of . . . . . . . . . . . . . . . . . . . ...4.36
stability of . . . . . . . . . . . . . . . . ..2O. 24,25,35
Expansion . . . . . . . . . . . . . .4.6.24.44.74
andtype of cement, . . . . . . . . . . . . . . ...57.73
due to ettringite formation. . . . . . 30,47-51,71
typeof . . . . . . . . . . . . . . . . . . . . . . . . . . . ...56
Flyash . . . . . . . . . . . . . . . . . .. 39,45
Frost durability, . . . . . . . . . . . . . .. 8,36
Gypsum, , . . . . . . . . . . . ..l9.24.33.46
effect of percent on expansion. ,....46,58,
62
Heat curing
guidelines for. ,., . . . . . . . . . . . . . ..lO.64-67
Humidity
of storage, effect on expansion . . . . . . . . ...51
Hydration
intransition zone, . . . . . . . . . . . . . . . . .. 41-42
Ion substitution
andeffectoftemperature. .,......
. . . . ...24
inettringite crystal structure . . . . . . . . . . ...18
Jouravskite .,,
. . . . ...,.,...
. ..18
Microcracks . . . . . . . . . . . . . . . . .. 6,33
at aggregate/paste interface . . . . . . . . . . . ..7, 8
infilling of . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
Monosulphoalumi nate
formation of . . . . . . . . . . . . . . . . . . . . . . . ...16
stability of . . . . . . . . . . . . . . . . . . . . . . . . 20,35
Permeability .,.......,...,,.
. ..12
pH
andstability of aluminates, ...,...
. . . . ...21
effect ofalkalis on. , .,.,.....
. . . . . . . ...25
Pore solution
ionsin . . . . . . . . . . . . . ..o. o. . . . . . . . . . ...33
Subject Index
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
Page 115
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),
1 In,
= 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
m
II
An organization of cement manufacturers to improve
and extend the uses of portland cement and concrete
through market developmen~ engineerin~ research,
education, and public affairs work.
Printed in U.S.A.
RD108,O1T