Properties of concreteFifth Edition

A. M. NevilleCBE, DSc(Eng), DSc, FIStructE, FREng, FRSEHonorary Member of the American Concrete In-

stituteHonorary Member and Gold Medallist of the

Concrete SocietyHonorary Member of the Brazilian Concrete Insti-

tuteformerly

Head of Department of Civil Engineering,University of Leeds, England

Dean of Engineering, University of Calgary,Canada

Principal and Vice-Chancellor, University of Dun-dee, Scotland

President of the Concrete SocietyVice-President of the Royal Academy of Engin-

eering

Harlow, England London New York Bo-ston San Francisco Toronto Sydney

Singapore Hong KongTokyo Seoul Taipei New Delhi Cape Town Madrid Mexico City Amsterdam Munich

Paris Milan

Pearson Education LimitedEdinburgh GateHarlowEssex CM20 2JEEnglandand Associated Companies throughout the worldVisit us on the World Wide Web at:http://www.pearsoned.co.uk A M Neville 1963, 1973, 1975, 1977, 1981,1995, 2011The rights of A M Neville to be identified as theauthor of this work has keen asserted by him in ac-cordance with the Copyright, Designs, and PatentsAct 1988.All rights reserved; no part of this publication maybe reproduced, stored in any retrieval system, ortransmitted in any form or by any means, electron-ic, mechanical, photocopying, recording or other-wise without either the prior written permissionof the Publishers or a licence permitting restricted

copying in the United Kingdom issued by theCopyright Licensing Agency Ltd, 90 TottenhamCourt Road. London WIT 4LPFirst published 2011ISBN: 978-0-273-75580-7British Library Cataloguing in PublicationDataA catalogue entry for this title is available fromthe British Library.Library of Congress Cataloging-in-Publica-tion DataNeville, Adam M.

Properties of concrete / A.M. Neville. -- 5th ed.p. cm.

ISBN 978-0-273-75580-7 (pbk.)1. Concrete. I. Title.TA439.N48 2011620.136--dc23

2011019819Set by 35 in 10/12 Monotype Times

Printed in Malaysia (CTP-VVP)5 4 3 2 115 14 13 12 11

Contents

Preface to the Fifth EditionPrefaceAcknowledgements

1 Portland cementHistorical noteManufacture of Portland cementChemical composition of Portland ce-mentHydration of cement

Calcium silicate hydratesTricalcium aluminate hydrate and theaction of gypsum

SettingFalse set

Fineness of cement

Structure of hydrated cementVolume of products of hydration

Capillary poresGel pores

Mechanical strength of cement gelWater held in hydrated cement pasteHeat of hydration of cementInfluence of the compound composi-tion on properties of cement

Effects of alkalisEffects of glass in clinker

Tests on properties of cementConsistency of standard pasteSetting timeSoundnessStrength of cement

References

2 Cementitious materials of differenttypes

Categorization of cementitious materi-alsDifferent cementsOrdinary Portland cementRapid-hardening Portland cementSpecial very rapid-hardening PortlandcementsLow heat Portland cementSulfate-resisting cementWhite cement and pigmentsPortland blastfurnace cementSupersulfated cementPozzolanas

Fly ashPozzolanic cements

Silica fume

FillersOther cementsWhich cement to useHigh-alumina cement

ManufactureComposition and hydrationResistance to chemical attackPhysical properties of high-aluminacement

Conversion of high-alumina cementRefractory properties of high-aluminacementReferences

3 Properties of aggregateGeneral classification of aggregatesClassification of natural aggregatesSamplingParticle shape and texture

Bond of aggregateStrength of aggregateOther mechanical properties of aggreg-ateSpecific gravityBulk densityPorosity and absorption of aggregateMoisture content of aggregateBulking of fine aggregateDeleterious substances in aggregate

Organic impuritiesClay and other fine materialSalt contaminationUnsound particles

Soundness of aggregateAlkalisilica reaction

Tests for aggregate reactivityAlkalicarbonate reaction

Thermal properties of aggregateSieve analysis

Grading curvesFineness modulus

Grading requirementsPractical gradingsGrading of fine and coarse aggregates

Oversize and undersizeGap-graded aggregateMaximum aggregate sizeUse of plumsHandling of aggregateSpecial aggregatesRecycled concrete aggregateReferences

4 Fresh concreteQuality of mixing water

Density of fresh concreteDefinition of workabilityThe need for sufficient workabilityFactors affecting workabilityMeasurement of workability

Slump testCompacting factor testASTM flow testRemoulding testVebe testFlow table testBall penetration test and compactab-ility testNassers K-testerTwo-point test

Comparison of testsStiffening time of concrete

Effect of time and temperature onworkabilitySegregationBleedingThe mixing of concrete

Concrete mixersUniformity of mixingMixing timeHand mixing

Ready-mixed concreteRetemperingPumped concrete

Concrete pumpsUse of pumpingRequirements for pumped concretePumping lightweight aggregate con-crete

Shotcrete

Underwater concretePreplaced aggregate concreteVibration of concrete

Internal vibratorsExternal vibratorsVibrating tablesOther vibrators

RevibrationVacuum-dewatered concrete

Permeable formworkAnalysis of fresh concreteSelf-compacting (self-consolidating)concreteReferences

5 AdmixturesBenefits of admixturesTypes of admixturesAccelerating admixtures

Retarding admixturesWater-reducing admixturesSuperplasticizers

Nature of superplasticizersEffects of superplasticizersDosage of superplasticizersLoss of workabilitySuperplasticizercement compatibil-ityUse of superplasticizers

Special admixturesWaterproofing admixturesAnti-bacterial and similar admix-tures

Remarks about the use of admixturesReferences

6 Strength of concreteWater/cement ratio

Effective water in the mixGel/space ratioPorosity

Cement compactsInfluence of properties of coarse ag-gregate on strengthInfluence of aggregate/cement ratio onstrengthNature of strength of concrete

Strength in tensionCracking and failure in compressionFailure under multiaxial stress

MicrocrackingAggregatecement paste interfaceEffect of age on strength of concreteMaturity of concreteRelation between compressive andtensile strengths

Bond between concrete and reinforce-mentReferences

7 Further aspects of hardened concreteCuring of concrete

Methods of curingTests on curing compoundsLength of curing

Autogenous healingVariability of strength of cementChanges in the properties of cementFatigue strength of concreteImpact strengthElectrical properties of concreteAcoustic propertiesReferences

8 Temperature effects in concrete

Influence of early temperature onstrength of concreteSteam curing at atmospheric pressureHigh-pressure steam curing (autoclav-ing)Other thermal curing methodsThermal properties of concrete

Thermal conductivityThermal diffusivitySpecific heat

Coefficient of thermal expansionStrength of concrete at high temperat-ures and resistance to fire

Modulus of elasticity at high tem-peraturesBehaviour of concrete in fire

Strength of concrete at very low tem-peraturesMass concrete

Concreting in hot weatherConcreting in cold weather

Concreting operationsReferences

9 Elasticity, shrinkage, and creepStressstrain relation and modulus ofelasticity

Expressions for stressstrain curveExpressions for modulus of elasticityDynamic modulus of elasticityPoissons ratioEarly volume changesAutogenous shrinkageSwellingDrying shrinkage

Mechanism of shrinkageFactors influencing shrinkage

Influence of curing and storage con-ditions

Prediction of shrinkageDifferential shrinkageShrinkage-induced crackingMoisture movementCarbonation shrinkageShrinkage compensation by the use ofexpansive cements

Types of expansive cementsShrinkage-compensating concrete

Creep of concreteFactors influencing creep

Influence of stress and strengthInfluence of properties of cementInfluence of ambient relative humid-ityOther influences

Relation between creep and timeNature of creepEffects of creepReferences

10 Durability of concreteCauses of inadequate durabilityTransport of fluids in concrete

Influence of the pore systemFlow, diffusion, and sorptionCoefficient of permeability

DiffusionDiffusion coefficientDiffusion through air and water

AbsorptionSurface absorption testsSorptivity

Water permeability of concrete

Permeability testingWater penetration test

Air and vapour permeabilityCarbonation

Effects of carbonationRate of carbonationFactors influencing carbonationCarbonation of concrete containingblended cementsMeasurement of carbonationFurther aspects of carbonation

Acid attack on concreteSulfate attack on concrete

Thaumasite form of sulfate attackMechanisms of attackFactors mitigating the attackTests on sulfate resistanceDelayed ettringite formation

EfflorescenceEffects of sea water on concrete

Salt weatheringSelection of concrete for exposureto sea water

Disruption by alkalisilica reactionPreventive measures

Abrasion of concreteTests for abrasion resistanceFactors influencing abrasion resist-ance

Erosion resistanceCavitation resistanceTypes of crackingReferences

11 Effects of freezing and thawing and ofchlorides

Action of frost

Behaviour of coarse aggregateparticles

Air entrainmentAir-void system characteristics

Entrained-air requirementsFactors influencing air entrainmentStability of entrained airAir entrainment by microspheresMeasurement of air content

Tests of resistance of concrete to freez-ing and thawingFurther effects of air entrainmentEffects of de-icing agentsChloride attack

Mechanism of chloride-induced cor-rosion

Chlorides in the mixIngress of chlorides

Threshold content of chloride ionsBinding of chloride ions

Influence of blended cements on corro-sionFurther factors influencing corrosion

Thickness of cover to reinforcementTests for penetrability of concrete tochloridesStopping corrosionReferences

12 Testing of hardened concreteTests for strength in compression

Cube testCylinder testEquivalent cube test

Effect of end condition of specimenand capping

Non-bonded caps

Testing of compression specimensFailure of compression specimensEffect of height/diameter ratio onstrength of cylindersComparison of strengths of cubes andcylindersTests for strength in tension

Flexural strength testsSplitting tension test

Influence on strength of moisture con-dition during testInfluence of size of specimen onstrength

Size effects in tensile strength testsSize effects in compressive strengthtestsSpecimen size and aggregate size

Test coresUse of small cores

Factors influencing strength of coresRelation of core strength to strengthin situ

Cast-in-place cylinder testInfluence of rate of application of loadon strengthAccelerated-curing test

Direct use of accelerated-curingstrength

Non-destructive testsRebound hammer testPenetration resistance testPull-out testPost-installed testsUltrasonic pulse velocity testFurther possibilities in non-destructivetestingResonant frequency method

Tests on the composition of hardenedconcrete

Cement contentDetermination of the original water/cement ratioPhysical methods

Variability of test resultsDistribution of strengthStandard deviation

References13 Concretes with particular properties

Concretes with different cementitiousmaterials

General features of use of fly ash,ggbs, and silica fumeDurability aspectsVariability of materials

Concrete containing fly ash

Influence of fly ash on properties offresh concreteHydration of fly ashStrength development of fly ashconcreteDurability of fly ash concrete

Concretes containing ground granu-lated blastfurnace slag (ggbs)

Influence of ggbs on properties offresh concreteHydration and strength developmentof concrete containing ggbsDurability aspects of concrete con-taining ggbs

Concrete containing silica fumeInfluence of silica fume on proper-ties of fresh concreteHydration and strength developmentof the Portland cementsilica fumesystem

Durability of concrete containingsilica fume

High performance concreteProperties of aggregate in high per-formance concreteAspects of high performance concretein the fresh state

Compatibility of Portland cementand superplasticizer

Aspects of hardened high performanceconcrete

Testing of high performance con-crete

Durability of high performance con-creteThe future of high performance con-creteLightweight concrete

Classification of lightweight con-cretes

Lightweight aggregatesNatural aggregatesManufactured aggregatesRequirements for aggregates forstructural concreteEffects of water absorption by light-weight aggregate

Lightweight aggregate concreteAspects of the fresh state

Strength of lightweight aggregate con-crete

Lightweight aggregatematrix bondElastic properties of lightweight ag-gregate concreteDurability of lightweight aggregateconcreteThermal properties of lightweight ag-gregate concreteCellular concrete

Autoclaved aerated concreteNo-fines concreteNailing concreteRemark about specialized concretesReferences

14 Selection of concrete mix proportions(mix design)

Cost considerationsSpecificationsThe process of mix selectionMean strength and minimum strength

Variability of strengthQuality controlFactors governing the selection of mixproportions

DurabilityWorkabilityMaximum size of aggregate

Grading and type of aggregateCement content

Mix proportions and quantities perbatch

Calculation by absolute volumeCombining aggregates to obtain a typegradingAmerican method of selection of mixproportions

ExampleMix selection for no-slump concreteMix selection for flowing concrete

Mix selection for high performanceconcreteMix selection for lightweight aggregateconcrete

ExampleBritish method of mix selection (mixdesign)

ExampleOther methods of mix selectionConcluding remarksReferences

Appendix I: Relevant ASTM StandardsAppendix II: Relevant British andEuropean StandardsName indexSubject index

Preface to the Fifth Edition

The format, organization and style of this editionare the same as those of the previous editions.The justification is that those features have shownthemselves to be successful by the fact that saleshave continued to be strong right up to the year2011. The total sales in English, as well as in the12 languages into which this book has been trans-lated, exceed half-a-million copies over a period ofnearly half-a-century.

With the passage of time, standards evolve, be-come modified, withdrawn and replaced. This pro-duces a need for updating a technical book suchas Properties of Concrete and can be accommod-ated by minor changes in new impressions of anexisting edition, as was done in the 14 impressionsof the fourth edition, which I intended to be final.This is still the case with American standards,where ASTM has a strict policy of periodic re-views, confirmation or replacement.

On the other hand, the situation of the Britishstandards is far more complex. Specifically, thereexist now some new British standards, describedas also European standards, denoted by BS EN.There continue to be in force some traditionalBritish standards, denoted by BS. In some cases,the British standards are described as obsolete,obsolescent, and also as current, superseded.All this is highly confusing but is perhaps an in-evitable consequence of a piecemeal introductionof new standards, which do not simply replacethe old ones on a one-to-one basis. There is awell-known saying that a camel is a horse de-signed by a committee. The European standardsare designed by an international committee!

I have retained by way of tables and limits in-formation contained in a number of the old Brit-ish standards, even those that have been with-drawn, because they contribute to knowledge ofwhat is desirable in the understanding of a rel-evant property. I believe that such an approachis valuable in a scientific book, encyclopaedic in

character. This is especially so because a num-ber of the new BS EN standards lay down howto measure some property of concrete and thento declare the outcome but say nothing aboutthe interpretation of the result. Such an approachdoes not contribute to knowledge of what is de-sirable, let alone to understanding of the relevantproperty.

The new standards have been introduced inthis fifth edition of Properties of Concrete forthe purpose of informing the reader about the ap-proach to, or principles of, testing. However, giv-en continuing evolution of standards, for specificuse, the reader should refer to the text of the actu-al standards and follow them scrupulously. Afterall, this book is not intended to be a manual orhandbook, let alone a cookbook.

Furthermore, I have not deleted references toearlier publications. I did so for two reasons.Firstly, this is a new edition of a successful bookand not a new book. Secondly, the old referencescontain the development of our knowledge, much

of it basic. On the other hand, many of the recentpapers contain minutiae of specific behaviour un-der specific conditions, and contribute but little tothe pool of knowledge capable of generalization.

It may be a reflection of my age but I find thatour pool of knowledge of value to designers, con-tractors and suppliers is not greatly increased bya paper co-authored by six people without muchcoordination or generalization. Nor are papers de-scribing the behaviour of concrete incorporatingfly ash from a single source of value to the con-crete community at large, the main benefit beingcommercial or personal.

This edition contains some additional topics:delayed ettringite formation, recycled concreteaggregate, self-compacting concrete, thaumasitesulfate attack, and of course augmentation andmodification of various topics.

I have not included that topical subject, sus-tainability (which seems to be the flavour of thedecade). As I see it, if sustainability of concreteas a material (distinct from a structure made of

concrete) is to ensure durability, then of course,this is of great importance; for this reasonchapters 10 and 11 are devoted to the durabilityof concrete.

However, durability does not mean a servicelife as long as possible. What we should aim atis a desired service life, and this is governed bythe function of the structure: a garden shed is atthe lower end of the scale, and a large bridge ordam are at the upper end. Dwellings are a goodexample of social needs changing with time, e.g.lifts (elevators) or bathrooms. Likewise, officesmay be of the open plan type or they may consistof numerous separate rooms Where there is a so-cial usage change, the old style may be a dis-advantage in that a modification of the structuremay be more expensive than demolition followedby a new design. And paying in advance for amore expensive structure in the first place is un-economical and may also discourage construc-tion. But these issues are outside the scope of this

book. So, if I do not enthuse over sustainability, itis not because of my ignorance.

In writing the fifth edition, and especially inincluding references to new standards, I havebeen greatly assisted by Robert Thomas, Man-ager, Library and Information Services of the In-stitution of Structural Engineers and by RoseMarney, Library Manager and Debra Francis,Librarian, of the Institution of Civil Engineers.Their exceptionally efficient and friendly help isgratefully acknowledged.

I am grateful to Simon Lake for progressingthe contractual aspects of the fifth edition and toPatrick Bond, Robert Sykes and Helen Leech forlooking after the production aspects of the book.

And, as always, I have to thank profoundly mylifelong technical collaborator and severe critic(i.e. wife), Dr. Mary Neville.

May I wish the reader (or in the 19th centuryparlance, the gentle reader) good concrete anddurable concrete structures.

A. M. N.London 2011

Preface

Concrete and steel are the two most commonlyused structural materials. They sometimes comple-ment one another, and sometimes compete withone another so that structures of a similar type andfunction can be built in either of these materials.And yet, the engineer often knows less about theconcrete of which the structure is made than aboutthe steel.

Steel is manufactured under carefully con-trolled conditions; its properties are determined ina laboratory and described in a manufacturers cer-tificate. Thus, the designer need only specify thesteel as complying with a relevant standard, andthe site engineers supervision is limited to theworkmanship of the connections between the indi-vidual steel members.

On a concrete building site, the situation istotally different. It is true that the quality of cementis guaranteed by the manufacturer in a manner

similar to that of steel and, provided suitable ce-mentitious materials are chosen, it is hardly evera cause of faults in a concrete structure. But itis concrete, and not cement, that is the buildingmaterial. The structural members are more oftenthan not made in situ, and their quality is almostexclusively dependent on the workmanship ofconcrete making and placing.

The disparity in the methods of steel and con-crete making is, therefore, clear, and the import-ance of the control of the quality of concrete workon the site is apparent. Furthermore, as the tradeof a concretor has not yet the training and the tra-dition of some of the other building trades, an en-gineers supervision on the site is essential. Thesefacts must be borne in mind by the designer, ascareful and intricate design can be easily vitiatedif the properties of the actual concrete differ fromthose assumed in the design calculations. Struc-tural design is only as good as the materials used.

From the above it must not be concluded thatmaking good concrete is difficult. Bad concrete

often a substance of unsuitable consistency,hardening into a honeycombed, non-homogen-eous mass is made simply by mixing cement,aggregate and water. Surprisingly, the ingredientsof a good concrete are exactly the same, and it isonly the know-how, backed up by understand-ing, that is responsible for the difference.

What, then, is good concrete? There are twooverall criteria: the concrete has to be satisfactoryin its hardened state, and also in its fresh statewhile being transported from the mixer andplaced in the formwork. The requirements in thefresh state are that the consistency of the mix besuch that it can be compacted by the means de-sired without excessive effort, and also that themix be cohesive enough for the methods of trans-porting and placing used so as not to produce se-gregation with a consequent lack of homogen-eity of the finished product. The primary require-ments of a good concrete in its hardened stateare a satisfactory compressive strength and an ad-equate durability.

All this has been valid since the first editionof this book appeared in 1963. In its three edi-tions and the 12 languages in which translationshave been published, the book seems to haveserved well those involved in concrete, whichcontinues to be the most important and wide-spread construction material. However, very sig-nificant changes in knowledge and in practicehave taken place in recent years, and this is whya fourth edition needed to be written. The extentof these changes has been such that a bolt-onapproach was not appropriate and, except for itsfundamental core, this is, therefore, a new book.Its coverage has been greatly widened, and itgives a broad as well as a detailed view of con-crete as a construction material. But there hasbeen no change for changes sake. The form,style, approach, and organization of the materialin the previous editions have been maintained sothat those readers who are familiar with the earli-er versions will have no difficulty in finding theirway in the new book.

The fourth edition contains much new materi-al on cementitious materials, some of which werenot used, or were little used, in the past. Know-ledge of these materials should now form part ofthe engineers stock-in-trade. Durability of con-crete under various conditions of exposure, in-cluding carbonation and alkalisilica reaction, istreated fully. In particular, the behaviour of con-crete under the extreme conditions existing incoastal areas of the hot parts of the world, where agreat deal of construction nowadays takes place,is discussed. Other new topics are: high perform-ance concrete, recently introduced admixtures,concrete under cryogenic conditions, and proper-ties of the aggregatematrix interface, to mentionbut the main ones.

It has to be admitted that the treatment of thevarious cementitious materials presented quite achallenge which has provoked the following di-gression. A very large number of papers on thesematerials and some other topics were publishedin the 1980s and continue in the 1990s. Many

worthwhile papers have elucidated the behaviourof the various materials and their influence onthe properties of concrete. But many more repor-ted narrowly construed investigations which de-scribed the influence of a single parameter, withsome other conditions kept unrealistically con-stant. Sometimes it is forgotten that, in a concretemix, it is usually not possible to change one in-gredient without altering some other property ofthe mix.

Generalized inferences from such piecemealresearch are at best difficult and at worst dan-gerous. We do not need more of these little re-search projects, each one chalking up a publica-tion in the authors curriculum vitae. Nor do weneed an endless succession of formulae, each de-rived from a small set of data. Some, seeminglyimpressive, analyses show an excellent correla-tion with the experimental data fed into the poolfrom which the expressions were derived in thefirst place: such correlation is not surprising. Butthen it should not be surprising either if the ex-

pressions fail dismally when used to predict thebehaviour in untried circumstances where thereexist factors ignored in the original analysis.

A further comment can be made about the in-fluences of various factors on the behaviour ofconcrete which have been determined by statist-ical analyses. While the use of statistics in theevaluation of test results and in establishing rela-tionships is valuable, and often essential, a stat-istical relation alone, without a physical explana-tion, is not a sound basis for claiming that a truerelation exists between two or more factors. Like-wise, extrapolation of a valid relationship mustnot be assumed to be automatically valid. This isobvious but sometimes forgotten by an enthusi-astic author who is under the impression that heor she has discovered a general rule.

Whereas we must consider available research,there is little value in collecting together a massof research findings or giving a general reviewof each topic of research. Rather, this book hasstriven to integrate the various topics so as to

show their interdependence in the making and us-ing of concrete. An understanding of the physic-al and chemical phenomena involved is an essen-tial basis for tackling the unfamiliar, in contrastto the ad hoc approach for picking up clues frompast experience, which will work only so far, andsometimes may result in a catastrophe. Concreteis a patient material but, even so, avoidable flawsin the selection and proportioning of the mix in-gredients should be avoided.

It has to be remembered that the various con-crete mixes now used are derivatives and devel-opments of the traditional concrete, so that know-ledge of the basic properties of concrete contin-ues to be essential. In consequence, a large partof the book is devoted to these fundamentals. Theoriginal work of the pioneers of the knowledgeof concrete which explains the underlying beha-viour of concrete on a scientific basis and theclassical references have been retained: they al-low us to have a proper perspective of our know-ledge.

The ultimate purpose of this book is to facil-itate better construction in concrete. To achievethis, it is necessary to understand, to master, andto control the behaviour of concrete not only inthe laboratory but also in actual structures. It isin this respect that an author with a structuralbackground is at an advantage. Furthermore, ex-perience in construction and in investigations oflack of durability and serviceability has been ex-ploited.

Because the book is used in so many coun-tries, it was thought best to use both the SI andthe Imperial units of measurement, now paradox-ically known as American. All the data, diagramsand tables are, therefore, conveniently presentedfor readers, progressive or traditionalist, in allcountries.

This book was written in its entirety duringthe period of one year and it should thereforepresent a closely-knit explanation of the beha-viour of concrete, rather than a series of some-what disconnected chapters. This cohesion may

be of benefit to readers who have often been ob-liged to consult collections of uncoordinated art-icles in a book with a nominal editor or editors.

In a single volume, it is not possible to coverthe whole field of concrete: specialized materials,such as fibre reinforced concrete, polymer con-crete, or sulfur concrete, albeit useful, are notdealt with. Inevitably, the author selects what heconsiders most important or most interesting, orsimply what he knows most about, even thoughthe scope of his knowledge increases with ageand experience. The emphasis in this book is onan integrated view of the properties of concreteand on underlying scientific reasons, for, as HenriPoincar said, an accumulation of facts is nomore a science than a heap of stones is a house.

A. M. N.

Acknowledgements

The copyright of the following illustrations andtables rests with the Crown and my thanks are dueto the Controller of HM Stationery Office for per-mission to reproduce: Figures 2.5, 3.2, 3.15, 3.16,4.1, 7.25, 8.11, 12.10, 12.39, 14.3, 14.10, 14.12,14.13, and 14.14, and Tables 2.9, 3.8, 3.9, 8.4,13.14, 14.9, and 14.10.

The following have made material from theirpublications available to me, for which I thankthem: National Bureau of Standards (Washington,D.C.); US Bureau of Reclamation; American So-ciety for Testing and Materials (ASTM); Cementand Concrete Association (London); Portland Ce-ment Association (Skokie, Illinois); NationalReady-Mixed Concrete Association (SilverSpring, Maryland); American Ceramic Society;American Concrete Institute; Society of ChemicalIndustry (London); Institution of Civil Engineers(London); Institution of Structural Engineers

(London); Swedish Cement and Concrete Re-search Institute; Department of Energy, Minesand Resources (Ottawa); Edward Arnold (Pub-lishers) Ltd. (London); Reinhold Publishing Cor-poration, Book Division (New York); Butter-worths Scientific Publications (London);Deutsches Institut fr Normung e.V. (Berlin);Pergamon Press (Oxford); Martinus Nijhoff (TheHague); Civil Engineering (London); Il Cemento(Rome); Deutscher Ausschuss fr Stahlbeton(Berlin); Cement and Concrete Research(University Park, Pennsylvania); Zement undBeton (Vienna); Materials and Structures,RILEM (Paris); Bulletin du Ciment (Wildegg,Switzerland); American Society of Civil Engin-eers (New York); Magazine of Concrete Research(London); The Concrete Society (Crowthorne);Darmstadt Concrete (Darmstadt); LaboratoireCentral des Ponts et Chausses (Paris); BritishCeramic Proceedings (Stoke on Trent); Concrete(London). Tables from BS 812, BS 882, and BS5328 are reproduced by kind permission of theBritish Standards Institution, 2 Park Street, Lon-

don W1A 2BS, from where copies of the com-plete standards may be purchased. The late Pro-fessor J. F. Kirkaldy kindly provided the data ofTable 3.7.

The full details of the sources can be found atthe end of each chapter; the reference numbersappear with the captions to the illustrations andthe headings to the tables.

I am grateful to my various clients in litigationand arbitration, and equally to their opposingparties, who enabled me to achieve a better un-derstanding of the behaviour of concrete in ser-vice, often by way of observing its misbeha-viour.

Very considerable help in finding referenceswas provided by the staff of the Library of the In-stitution of Civil Engineers, and especially by MrRobert Thomas who was indefatigable in track-ing down the various sources. Finally, I wish toput on record the enormous effort and achieve-ment of Mary Hallam Neville in cementing thesources and references into a cohesive manuscript

culminating in a concrete book. Without herprompting (a much better word than nagging) thisbook may not have anteceded the Authors de-cease.

Chapter 1. Portland cement

Cement, in the general sense of the word, canbe described as a material with adhesive and co-hesive properties which make it capable of bond-ing mineral fragments into a compact whole. Thisdefinition embraces a large variety of cementingmaterials.

For constructional purposes, the meaning ofthe term cement is restricted to the bonding ma-terials used with stones, sand, bricks, buildingblocks, etc. The principal constituents of this typeof cement are compounds of lime, so that inbuilding and civil engineering we are concernedwith calcareous cement. The cements of interestin the making of concrete have the property ofsetting and hardening under water by virtue ofa chemical reaction with it and are, therefore,called hydraulic cements.

Hydraulic cements consist mainly of silicatesand aluminates of lime, and can be classified

broadly as natural cements, Portland cements,and high-alumina cements. The present chapterdeals with the manufacture of Portland cementand its structure and properties, both when un-hydrated and in a hardened state. The differenttypes of Portland and other cements are con-sidered in Chapter 2.Historical noteThe use of cementing materials is very old. Theancient Egyptians used calcined impure gypsum.The Greeks and the Romans used calcined lime-stone and later learned to add to lime and water,sand and crushed stone or brick and broken tiles.This was the first concrete in history. Lime mor-tar does not harden under water, and for con-struction under water the Romans ground togeth-er lime and a volcanic ash or finely ground burntclay tiles. The active silica and alumina in the ashand the tiles combined with the lime to producewhat became known as pozzolanic cement fromthe name of the village of Pozzuoli, near Vesuvi-

us, where the volcanic ash was first found. Thename pozzolanic cement is used to this day todescribe cements obtained simply by the grindingof natural materials at normal temperature. Someof the Roman structures in which masonry wasbonded by mortar, such as the Coliseum in Romeand the Pont du Gard near Nmes, and concretestructures such as the Pantheon in Rome, havesurvived to this day, with the cementitious mater-ial still hard and firm. In the ruins at Pompeii, themortar is often less weathered than the rather softstone.

The Middle Ages brought a general decline inthe quality and use of cement, and it was onlyin the eighteenth century that an advance in theknowledge of cements occurred. John Smeaton,commissioned in 1756 to rebuild the EddystoneLighthouse, off the Cornish coast, found that thebest mortar was produced when pozzolana wasmixed with limestone containing a considerableproportion of clayey matter. By recognizing therole of the clay, hitherto considered undesirable,

Smeaton was the first to understand the chemicalproperties of hydraulic lime, that is a material ob-tained by burning a mixture of lime and clay.

There followed a development of other hy-draulic cements, such as the Roman cement ob-tained by James Parker by calcining nodules ofargillaceous limestone, culminating in the patentfor Portland cement taken out by Joseph Asp-din, a Leeds bricklayer, stonemason, and builder,in 1824. This cement was prepared by heating amixture of finely-divided clay and hard limestonein a furnace until CO2 had been driven off; thistemperature was much lower than that necessaryfor clinkering. The prototype of modern cementwas made in 1845 by Isaac Johnson, who burnta mixture of clay and chalk until clinkering, sothat the reactions necessary for the formation ofstrongly cementitious compounds took place.

The name Portland cement, given originallydue to the resemblance of the colour and qualityof the hardened cement to Portland stone alimestone quarried in Dorset has remained

throughout the world to this day to describe acement obtained by intimately mixing togethercalcareous and argillaceous, or other silica-,alumina-, and iron oxide-bearing materials, burn-ing them at a clinkering temperature, and grind-ing the resulting clinker. The definition of Port-land cement in various standards is on these lines,recognizing that gypsum is added after burning;nowadays, other materials may also be added orblended (see p. 64).Manufacture of Portland cementFrom the definition of Portland cement givenabove, it can be seen that it is made primarlyfrom a calcareous material, such as limestone orchalk, and from alumina and silica found as clayor shale. Marl, a mixture of calcareous and argil-laceous materials, is also used. Raw materials forthe manufacture of Portland cement are found innearly all countries and cement plants operate allover the world.

The process of manufacture of cement con-sists essentially of grinding the raw materials,mixing them intimately in certain proportions andburning in a large rotary kiln at a temperature ofup to about 1450 C when the material sinters andpartially fuses into balls known as clinker. Theclinker is cooled and ground to a fine powder,with some gypsum added, and the resultingproduct is the commercial Portland cement sowidely used throughout the world.

Some details of the manufacture of cementwill now be given, and these can be best followedwith reference to the diagrammatic representationof the process shown in Fig. 1.1.

Fig. 1.1. Diagrammatic representation of: (a)the wet process and (b) the dry process of

manufacture of cementThe mixing and grinding of the raw materials

can be done either in water or in a dry condition;hence the names wet and dry processes. Theactual methods of manufacture depend also onthe hardness of the raw materials used and ontheir moisture content.

Let us consider first the wet process. Whenchalk is used, it is finely broken up and dispersed

in water in a washmill; this is a circular pit withrevolving radial arms carrying rakes which breakup the lumps of solid matter. The clay is alsobroken up and mixed with water, usually in a sim-ilar washmill. The two mixtures are now pumpedso as to mix in predetermined proportions andpass through a series of screens. The resulting ce-ment slurry flows into storage tanks.

When limestone is used, it has to be blasted,then crushed, usually in two progressively smal-ler crushers, and then fed into a ball mill withthe clay dispersed in water. There, the comminu-tion of the limestone (to the fineness of flour) iscompleted, and the resultant slurry is pumped in-to storage tanks. From here onwards, the processis the same regardless of the original nature of theraw materials.

The slurry is a liquid of creamy consistency,with a water content of between 35 and 50 percent, and only a small fraction of material about2 per cent larger than a 90 m (No. 170 ASTM)sieve size. There are usually a number of storage

tanks in which the slurry is kept, the sedimenta-tion of the suspended solids being prevented bymechanical stirrers or bubbling by compressedair. The lime content of the slurry is governed bythe proportioning of the original calcareous andargillaceous materials, as mentioned earlier. Finaladjustment in order to achieve the required chem-ical composition can be made by blending slur-ries from different storage tanks, sometimes us-ing an elaborate system of blending tanks. Oc-casionally, for example in the worlds northern-most plant in Norway, the raw material is a rockof such composition that it alone is crushed andno blending is required.

Finally, the slurry with the desired lime con-tent passes into the rotary kiln. This is a large,refractory-lined steel cylinder, up to 8 m (or 26ft) in diameter, sometimes as long as 230 m (or760 ft), slowly rotating about its axis, which isslightly inclined to the horizontal. The slurry isfed in at the upper end while pulverized coal isblown in by an air blast at the lower end of the

kiln, where the temperature reaches about 1450C. The coal, which must not have too high anash content, deserves a special mention becausetypically 220 kg (500 lb) of coal is used to makeone tonne of cement. This is worth bearing inmind when considering the price of cement. Oil(of the order of 125 litres (33 US gallons) pertonne of cement) or natural gas were also used,but since the 1980s most oil-fired plants havebeen converted to coal, which is by far the mostcommon fuel used in most countries. It is worthnoting that, because it is burnt in the kiln, coalwith a high sulfur content can be used withoutharmful emissions.

The slurry, in its movement down the kiln, en-counters a progressively higher temperature. Atfirst, the water is driven off and CO2 is liberated;further on, the dry material undergoes a series ofchemical reactions until finally, in the hottest partof the kiln, some 20 to 30 per cent of the materialbecomes liquid, and lime, silica and alumina re-combine. The mass then fuses into balls, 3 to 25

mm ( to 1 in.) in diameter, known as clinker. Theclinker drops into coolers, which are of varioustypes and often provide means for an exchange ofheat with the air subsequently used for the com-bustion of the pulverized coal. The kiln has tooperate continuously in order to ensure a steadyregime, and therefore uniformity of clinker, andalso to reduce the deterioration of the refractorylining. It should be noted that the flame temper-ature reaches 1650 C. The largest existing kilnin a wet-process plant produces 3600 tonnes ofclinker a day. Because the manufacture of cementby the wet process is energy intensive, new wet-process plants are no longer built.

In the dry and semi-dry processes, the raw ma-terials are crushed and fed in the correct propor-tions into a grinding mill, where they are driedand reduced in size to a fine powder. The drypowder, called raw meal, is then pumped to ablending silo, and final adjustment is now madein the proportions of the materials required forthe manufacture of cement. To obtain a uniform

and intimate mixture, the raw meal is blended,usually by means of compressed air inducing anupward movement of the powder and decreasingits apparent density. The air is pumped over onequadrant of the silo at a time, and this permits theapparently heavier material from the non-aeratedquadrants to move laterally into the aerated quad-rant. Thus the aerated material tends to behave al-most like a liquid and, by aerating all quadrantsin turn for a total period of about one hour, a uni-form mixture is obtained. In some cement plants,continuous blending is used.

In the semi-dry process, the blended meal isnow sieved and fed into a rotating dish calleda granulator, water weighing about 12 per centof the meal being added at the same time. Inthis manner, hard pellets about 15 mm ( in.) indiameter are formed. This is necessary, as coldpowder fed direct into a kiln would not permitthe air flow and exchange of heat necessary forthe chemical reactions of formation of cementclinker.

The pellets are baked hard in a pre-heatinggrate by means of hot gases from the kiln. Thepellets then enter the kiln, and subsequent opera-tions are the same as in the wet process of man-ufacture. Since, however, the moisture content ofthe pellets is only 12 per cent as compared withthe 40 per cent moisture content of the slurryused in the wet process, the semi-dry-process kilnis considerably smaller. The amount of heat re-quired is also very much lower because onlysome 12 per cent of moisture has to be driven off,but additional heat has already been used in re-moving the original moisture content of the rawmaterials (usually 6 to 10 per cent). The processis thus quite economical, but only when the rawmaterials are comparatively dry. In such a casethe total coal consumption can be as little as 100kg (220 lb) per tonne of cement.

In the dry process (see Fig. 1b), the raw meal,which has a moisture content of about 0.2 percent, is passed through a pre-heater, usually ofa suspension type; that means that the raw meal

particles are suspended in the rising gases. Here,the raw meal is heated to about 800 C before be-ing fed into the kiln. Because the raw meal con-tains no moisture to be driven off and because itis already pre-heated, the kiln can be shorter thanin the wet process. The pre-heating uses the hotgas leaving the kiln. Because that gas contains asignificant proportion of the rather volatile alkalis(see p. 9) and chlorides, a part of the gas mayneed to be bled off to ensure that the alkali con-tent of the cement is not too high.

The major part of the raw meal can be passedthrough a fluidized calciner (using a separate heatsource) introduced between the pre-heater andthe kiln. The temperature in the fluidized calcineris about 820 C. This temperature is stable so thatthe calcination is uniform and the efficiency ofthe heat exchange is high.

A part of the raw meal is fed direct into thekiln in the usual manner but, overall, the effectof the fluidized calciner is to increase the decar-bonation (dissociation of CaCO3) of the raw meal

prior to entry into the kiln and thus greatly to in-crease the kiln throughput. What is probably thelargest dry-process plant in the world produces 10000 tonnes of clinker a day using a kiln 6.2 m(20 ft) in diameter and 105 m (345 ft) long. In theU.S. more than 80% of cement production usesthe dry process.

It should be stressed that all processes requirean intimate mixture of the raw materials becausea part of the reactions in the kiln must take placeby diffusion in solid materials, and a uniform dis-tribution of materials is essential to ensure a uni-form product.

On exit from the kiln, regardless of the type ofprocess, the clinker is cooled, the heat being usedto preheat the combustion air. The cool clinker,which is characteristically black, glistening, andhard, is interground with gypsum in order to pre-vent flash setting of the cement. The grinding isdone in a ball mill consisting of several com-partments with progressively smaller steel balls,sometimes preceded by passing through a roll

press. In most plants, a closed-circuit grindingsystem is used: the cement discharged by the millis passed through a separator, fine particles beingremoved to the storage silo by an air current,while the coarser particles are passed through themill once again. Closed-circuit grinding avoidsthe production of a large amount of excessivelyfine material or of a small amount of too coarsematerial, faults often encountered with open-cir-cuit grinding. Small quantities of grinding aidssuch as ethylene glycol or propylene glycol areused. Information about grinding aids is given byMassazza and Testolin.1.90 The performance ofa ball mill can be improved by pre-grinding theclinker in a horizontal impact crusher.

Once the cement has been satisfactorilyground, when it will have as many as 1.1 1012particles per kg (5 1011 per lb), it is ready fortransport in bulk. Less commonly, the cement ispacked in bags or drums. However, some typesof cement, such as white, hydrophobic, expans-ive, regulated-set, oil-well, and high-alumina, are

always packed in bags or drums. A standard bagin the United Kingdom contains 50 kg (110 lb) ofcement; a US sack weighs 94 lb (42.6 kg); otherbag sizes are also used. Bags of 25 kg are becom-ing popular.

Except when the raw materials necessitate theuse of the wet process, the dry process is usednowadays in order to minimize the energy re-quired for burning. Typically, the burning processrepresents 40 to 60 per cent of the productioncost, while the extraction of raw materials forthe manufacture of cement represents only 10 percent of the total cost of cement.

Around 1990, the average energy consump-tion in the United States for the production of1 tonne of cement by the dry process was 1.6MWh. In modern plants, this figure is much re-duced, being below 0.8 MWh in Austria.1.96Electricity consumption, which accounts forsome 6 to 8 per cent of total energy used, is typ-ically of the order: 10 kWh for crushing the rawmaterials, 28 kWh in the raw meal preparation,

24 kWh in burning, and 41 kWh in grinding.1.18The capital cost of installation of a cement plantis very high: nearly US$200 per tonne of cementproduced per annum.

In addition to the main processes, there arealso other processes of manufacture of cement,of which one, using gypsum instead of lime, per-haps deserves mention. Gypsum, clay and cokewith sand and iron oxide are burnt in a rotary kiln,the end products being Portland cement and sul-fur dioxide which is further converted into sulfur-ic acid.

In areas where only a small cement productionis required or where investment capital is limited,a vertical kiln of the Gottlieb type can be used.This fires nodules of raw meal and fine coalpowder combined, and produces agglomeratedclinker which is then broken up. A single kiln, 10m (33 ft) high, produces up to 300 tonnes of ce-ment a day. China used several thousand of suchkilns, but now has a very large modern cement

industry, producing 1000 million tonnes per an-num.Chemical composition of PortlandcementWe have seen that the raw materials used in themanufacture of Portland cement consist mainlyof lime, silica, alumina and iron oxide. Thesecompounds interact with one another in the kilnto form a series of more complex products and,apart from a small residue of uncombined limewhich has not had sufficient time to react, a stateof chemical equilibrium is reached. However,equilibrium is not maintained during cooling, andthe rate of cooling will affect the degree of crys-tallization and the amount of amorphous materialpresent in the cooled clinker. The properties ofthis amorphous material, known as glass, differconsiderably from those of crystalline com-pounds of a nominally similar chemical compos-ition. Another complication arises from the inter-

action of the liquid part of the clinker with thecrystalline compounds already present.

Nevertheless, cement can be considered as be-ing in frozen equilibrium, i.e. the cooled productsare assumed to reproduce the equilibrium exist-ing at the clinkering temperature. This assump-tion is, in fact, made in the calculation of thecompound composition of commercial cements:the potential composition is calculated from themeasured quantities of oxides present in theclinker as if full crystallization of equilibriumproducts had taken place.

Four compounds are usually regarded as themajor constituents of cement: they are listed inTable 1.1, together with their abbreviated sym-bols. This shortened notation, used by cementchemists, describes each oxide by one letter, viz.:CaO = C; SiO2 = S; Al2O3 = A; and Fe2O3 = F.Likewise, H2O in hydrated cement is denoted byH, and SO3 by .

Table 1.1. Main Compounds of Portland Ce-ment

In reality, the silicates in cement are not purecompounds, but contain minor oxides in solidsolution. These oxides have significant effects onthe atomic arrangements, crystal form and hy-draulic properties of the silicates.

The calculation of the potential compositionof Portland cement is based on the work of R.H. Bogue and others, and is often referred toas Bogue composition. Bogues1.2 equations forthe percentages of main compounds in cementare given below. The terms in brackets representthe percentage of the given oxide in the total massof cement.

C3S = 4.07(CaO) 7.60(SiO2) 6.72(Al2O3) 1.43(Fe2O3) 2.85(SO3)

C2S = 2.87(SiO2) 0.75(3CaO.SiO2)C3A = 2.65(Al2O3) 1.69(Fe2O3)

C4AF = 3.04(Fe2O3).There are also other methods of calculating

the composition,1.1 but the subject is not con-sidered to be within the scope of this book. Weshould note, however, that the Bogue composi-tion underestimates the C3S content (and overes-timates C2S) because other oxides replace someof the CaO in C3S; as already stated, chemicallypure C3S and C2S do not occur in Portland ce-ment clinker.

A modification of the Bogue compositionwhich takes into account the presence of sub-stituent ions in the nominally pure main com-pounds has been developed by Taylor1.84 for therapidly cooled clinkers produced in modern ce-ment plants.

In addition to the main compounds listed inTable 1.1, there exist minor compounds, such asMgO, TiO2, Mn2O3, K2O and Na2O; they usuallyamount to not more than a few per cent of themass of cement. Two of the minor compoundsare of particular interest: the oxides of sodiumand potassium, Na2O and K2O, known as the al-kalis (although other alkalis also exist in cement).They have been found to react with some aggreg-ates, the products of the reaction causing disin-tegration of the concrete, and have also been ob-served to affect the rate of the gain of strength ofcement.1.3 It should, therefore, be pointed out thatthe term minor compounds refers primarily totheir quantity and not necessarily to their import-ance. The quantity of the alkalis and of Mn2O3can be rapidly determined using a spectrophoto-meter.

The compound composition of cement hasbeen established largely on the basis of studiesof phase equilibria of the ternary systems CASand CAF, and the quaternary system

CC2SC5A3C4AF, and others. The course ofmelting or crystallization was traced and thecompositions of liquid and solid phases at anytemperature were computed. In addition to themethods of chemical analysis, the actual compos-ition of clinker can be determined by a micro-scope examination of powder preparations andtheir identification by the measurement of the re-fractive index. Polished and etched sections canbe used both in reflected and transmitted light.Other methods include the use of X-ray powderdiffraction to identify the crystalline phases andalso to study the crystal structure of some ofthe phases, and of differential thermal analysis;quantitative analysis is also possible, but com-plicated calibrations are involved.1.68 Moderntechniques include phase analysis through a scan-ning electron microscope and image analysisthrough an optical microscope or a scanning elec-tron microscope.

Estimating the composition of cement is aidedby more rapid methods of determining the ele-

mental composition, such as X-ray fluorescence,X-ray spectrometry, atomic absorption, flamephotometry, and electron probe micro-analysis.X-ray diffractometry is useful in the determin-ation of free lime, i.e. CaO as distinct fromCa(OH)2, and this is convenient in controlling thekiln performance.1.67

C3S, which is normally present in the largestamounts, occurs as small, equidimensional col-ourless grains. On cooling below 1250 C, it de-composes slowly but, if cooling is not too slow,C3S remains unchanged and is relatively stable atordinary temperatures.

C2S is known to have three, or possibly evenfour, forms. -C2S, which exists at high temper-atures, inverts to the -form at about 1450 C. -C2S undergoes further inversion to -C2S at about670 C but, at the rates of cooling of commercialcements, -C2S is preserved in the clinker. -C2Sforms rounded grains, usually showing twinning.

C3A forms rectangular crystals, but C3A infrozen glass forms an amorphous interstitialphase.

C4AF is really a solid solution ranging fromC2F to C6A2F, but the description C4AF is a con-venient simplification.1.4

The actual proportions of the various com-pounds vary considerably from cement to ce-ment, and indeed different types of cement areobtained by suitable proportioning of the raw ma-terials. In the United States, an attempt was atone time made to control the properties of ce-ments required for different purposes by specify-ing the limits of the four major compounds, ascalculated from the oxide analysis. This proced-ure would cut out numerous physical tests nor-mally performed, but unfortunately the calculatedcompound composition is not sufficiently accur-ate, nor does it take into account all the relevantproperties of cement, and cannot therefore serveas a substitute for direct testing of the requiredproperties.

A general idea of the composition of cementcan be obtained from Table 1.2, which gives theoxide composition limits of Portland cements.Table 1.3 gives the oxide composition of a typicalcement of the 1960s and the calculated com-pound composition,1.5 obtained by means ofBogues equations on p. 9.

Table 1.2. Usual Composition Limits of Port-land Cement

Table 1.3. Oxide and Compound Composi-tions of a Typical Portland Cement of the

1960s1.5

Two terms used in Table 1.3 require explana-tion. The insoluble residue, determined by treat-ing with hydrochloric acid, is a measure of adul-teration of cement, largely arising from impur-ities in gypsum. British Standard BS 12 : 1996(withdrawn) limits the insoluble residue to 1.5per cent of the mass of cement. European Stand-ard BS EN 197-1 : 2000, which allows a 5 percent content of a filler (see p. 88), limits the in-soluble residue to 5 per cent of the mass of thecement exclusive of the filler.

The loss on ignition shows the extent of car-bonation and hydration of free lime and free mag-nesia due to the exposure of cement to the atmo-sphere. The maximum loss on ignition (at 1000C) permitted by BS EN 197-1 : 2000 is 5 percent and by ASTM C 150-09 is 3 per cent exceptfor Type I cement (2.5 per cent); 4 per cent is ac-ceptable for cements in the tropics. Because hy-drated free lime is innocuous (see p. 51), for agiven free lime content of cement, a greater losson ignition is really advantageous. With cementscontaining a calcareous filler, a higher limit onthe loss on ignition is necessary: 5 per cent of themass of the cement nucleus is allowed by BS EN197-1 : 2000.

It is interesting to observe the large influenceof a change in the oxide composition on the com-pound composition of cement. Some data ofCzernins1.5 are given in Table 1.4; column (1)shows the composition of a fairly typical rapid-hardening cement. If the lime content is de-creased by 3 per cent, with corresponding in-

creases in the other oxides (column (2)), a con-siderable change in the C3S : C2S ratio results.Column (3) shows a change of per cent in thealumina and iron contents compared with the ce-ment of column (1). The lime and silica contentsare unaltered and yet the ratio of the two silic-ates, as well as the contents of C3A and C4AF,is greatly affected. It is apparent that the signi-ficance of the control of the oxide compositionof cement cannot be over-emphasized. Within theusual range of ordinary and rapid-hardening Port-land cements the sum of the contents of the twosilicates varies only within narrow limits, so thatthe variation in composition depends largely onthe ratio of CaO to SiO2 in the raw materials.

Table 1.4. Influence of Change in Oxide Com-position on the Compound Composition1.5

In some countries in the European Union,there is a limit on soluble hexavalent chromiumusually of 2 ppm of mass dry cement. Contact

with excess chromium in fresh concrete may leadto dermatitis.

It may be convenient at this stage to summar-ize the pattern of formation and hydration of ce-ment; this is shown schematically in Fig. 1.2.

Fig. 1.2. Schematic representation of theformation and hydration of Portland cement

Hydration of cementThe reactions by virtue of which Portland cementbecomes a bonding agent take place in a wa-

tercement paste. In other words, in the presenceof water, the silicates and aluminates listed inTable 1.1 form products of hydration which intime produce a firm and hard mass the hydratedcement paste.

There are two ways in which compounds ofthe type present in cement can react with water.In the first, a direct addition of some moleculesof water takes place, this being a true reaction ofhydration. The second type of reaction with wateris hydrolysis. It is convenient and usual, however,to apply the term hydration to all reactions of ce-ment with water, i.e. to both true hydration andhydrolysis.

Le Chatelier was the first to observe, about130 years ago, that the products of hydration ofcement are chemically the same as the productsof hydration of the individual compounds undersimilar conditions. This was later confirmed bySteinour1.6 and by Bogue and Lerch,1.7 with theproviso that the products of reaction may influ-ence one another or may themselves interact with

the other compounds in the system. The two cal-cium silicates are the main cementitious com-pounds in cement, and the physical behaviour ofcement during hydration is similar to that of thesetwo compounds alone.1.8 The hydration of the in-dividual compounds will be described in moredetail in the succeeding sections.

The products of hydration of cement have avery low solubility in water as shown by the sta-bility of the hydrated cement paste in contact withwater. The hydrated cement bonds firmly to theunreacted cement, but the exact way in whichthis is achieved is not certain. It is possible thatthe newly produced hydrate forms an envelopewhich grows from within by the action of waterthat has penetrated the surrounding film of hy-drate. Alternatively, the dissolved silicates maypass through the envelope and precipitate as anouter layer. A third possibility is for the colloid-al solution to be precipitated throughout the massafter the condition of saturation has been reached,

the further hydration continuing within this struc-ture.

Whatever the mode of precipitation of theproducts of hydration, the rate of hydration de-creases continuously, so that even after a longtime there remains an appreciable amount of un-hydrated cement. For instance, after 28 days incontact with water, grains of cement have beenfound to have hydrated to a depth of only 4 m,1.9and 8 m after a year. Powers1.10 calculated thatcomplete hydration under normal conditions ispossible only for cement particles smaller than50 m, but full hydration has been obtained bygrinding cement in water continuously for fivedays.

Microscopic examination of hydrated cementshows no evidence of channelling of water intothe grains of cement to hydrate selectively themore reactive compounds (e.g. C3S) which maylie in the centre of the particle. It would seemthen, that hydration proceeds by a gradual reduc-tion in the size of the cement particle. In fact, un-

hydrated grains of coarse cement were found tocontain C3S as well as C2S at the age of severalmonths,1.11 and it is probable that small grains ofC2S hydrate before the hydration of large grainsof C3S has been completed. The various com-pounds in cement are generally intermixed in allgrains, and some investigations have suggestedthat the residue of a grain after a given period ofhydration has the same percentage compositionas the whole of the original grain.1.12 However,the composition of the residue does changethroughout the period of cement hydration,1.49and especially during the first 24 hours selectivehydration may take place.

The main hydrates can be broadly classifiedas calcium silicate hydrates and tricalcium alu-minate hydrate. C4AF is believed to hydrate intotricalcium aluminate hydrate and an amorphousphase, probably CaO.Fe2O3.aq. It is possible alsothat some Fe2O3 is present in solid solution in thetricalcium aluminate hydrate.

The progress of hydration of cement can bedetermined by different means, such as the meas-urement of: (a) the amount of Ca(OH)2 in thepaste; (b) the heat evolved by hydration; (c) thespecific gravity of the paste; (d) the amount ofchemically combined water; (e) the amount ofunhydrated cement present (using X-ray quant-itative analysis); and (f) also indirectly from thestrength of the hydrated paste. Thermogravimet-ric techniques and continuous X-ray diffractionscanning of wet pastes undergoing hydration1.50can be used in studying early reactions. The mi-crostructure of hydrated cement paste can also bestudied by back-scattered electron imaging in ascanning electron microscope.Calcium silicate hydratesThe rates of hydration of C3S and C2S in a purestate differ considerably, as shown in Fig. 1.3.When the various compounds are present all to-gether in cement, their rates of hydration are af-fected by compound interactions. In commercial

cements, the calcium silicates contain small im-purities of some of the oxides present in theclinker. The impure C3S is known as alite andthe impure C2S as belite. These impurities havea strong effect on the properties of the calciumsilicate hydrates (see p. 48).

Fig. 1.3. Typical development of hydration ofpure compounds1.47

When hydration takes place in a limitedamount of water, as is the case in cement paste,

in mortar or in concrete, C3S is believed to un-dergo hydrolysis producing a calcium silicate oflower basicity, ultimately C3S2H3, with the re-leased lime separating out as Ca(OH)2. There ex-ists, however, some uncertainty as to whetherC3S and C2S result ultimately in the same hy-drate. It would appear to be so from considera-tions of the heat of hydration1.6 and of the surfacearea of the products of hydration,1.13 but physicalobservations indicate that there may be more thanone possibly several distinct calcium silic-ate hydrates. The C : S ratio would be affected ifsome of the lime were absorbed or held in solidsolution, and there is strong evidence that the ulti-mate product of hydration of C2S has a lime/silicaratio of 1.65. This may be due to the fact that thehydration of C3S is controlled by the rate of dif-fusion of ions through the overlying hydrate filmswhile the hydration of C2S is controlled by itsslow intrinsic rate of reaction.1.14 Furthermore,temperature may affect the products of hydration

of the two silicates because the permeability ofthe gel is affected by temperature.

The C : S ratio has not been unequivocally de-termined because different test methods yield dif-ferent results.1.74 The variation can be as wide as1.5 by chemical extraction and 2.0 by thermogra-vimetric method.1.66 Electron-optical measure-ments also yield low values of the C : S ratio.1.72The ratio also varies with time and is influencedby the presence of other elements or compoundsin the cement. Nowadays, the calcium silicate hy-drates are broadly described as CSH, and the C: S ratio is believed to be probably near 2.1.19 Be-cause the crystals formed by hydration are imper-fect and extremely small, the mole ratio of waterto silica need not be a whole number. CSH usu-ally contains small amounts of Al, Fe, Mg, andother ions. At one time, CSH was referred to astobermorite gel because of a structural similarityto a mineral of this name, but this may not be cor-rect,1.60 and this description is now rarely used.

Making the approximate assumption thatC3S2H3 is the final product of hydration of bothC3S and C2S, the reactions of hydration can bewritten (as a guide, although not as exact stoi-chiometric equations) as follows.

For C3S:

2C3S + 6H C3S2H3 + 3Ca(OH)2.

The corresponding masses involved are:100 + 24 75 + 49.

For C2S:

2C2S + 4H C3S2H3 + Ca(OH)2.

The corresponding masses are:100 + 21 99 + 22.

Thus, on a mass basis, both silicates requireapproximately the same amount of water for their

hydration, but C3S produces more than twice asmuch Ca(OH)2 as is formed by the hydration ofC2S.

The physical properties of the calcium silicatehydrates are of interest in connection with the set-ting and hardening properties of cement. Thesehydrates appear amorphous but electron micro-scopy shows their crystalline character. It is in-teresting to note that one of the hydrates believedto exist, denoted by Taylor1.15 as CSH(I), has alayer structure similar to that of some clay min-erals, e.g. montmorillonite and halloysite. The in-dividual layers in the plane of the a and b axesare well crystallized while the distances betweenthem are less rigidly defined. Such a lattice wouldbe able to accommodate varying amounts of limewithout fundamental change a point relevantto the varying lime/silica ratios mentioned above.In fact, powder diagrams have shown that limein excess of one molecule per molecule of silicais held in a random manner.1.15 Steinour1.16 de-

scribed this as a merger of solid solution and ad-sorption.

Calcium silicates do not hydrate in the solidstate but the anhydrous silicates probably firstpass into solution and then react to form lesssoluble hydrated silicates which separate out ofthe supersaturated solution.1.17 This is the typeof mechanism of hydration first envisaged by LeChatelier in 1881.

Studies by Diamond1.60 indicate that the cal-cium silicate hydrates exist in a variety of forms:fibrous particles, flattened particles, a reticularnetwork, irregular grains, all rather difficult todefine. However, the predominant form is that offibrous particles, possibly solid, possibly hollow,sometimes flattened, sometimes branching at theends. Typically, they are 0.5 m to 2 m long andless than 0.2 m across. This is not a precise pic-ture, but the structure of calcium silicate hydratesis too disordered to be established by the existingtechniques, including a combination of the scan-

ning electron microscope and energy dispersiveX-ray spectrometer.

The hydration of C3S to a large extent charac-terizes the behaviour of cement and a descriptionof the latter may be appropriate. Hydration doesnot proceed at a steady rate or even at a stead-ily changing rate. The initial rapid release of cal-cium hydroxide into the solution leaves an outerlayer of calcium silicate hydrate, perhaps 10 nmthick.1.61 This layer impedes further hydration sothat, for some time thereafter, very little hydra-tion takes place.

As the hydration of cement is an exothermicreaction, the rate of evolution of heat is an indica-tion of the rate of hydration. This shows that thereare three peaks in the rate of hydration in the firstthree days or so, from the time when the dry ce-ment first comes into contact with water. Figure1.4 shows a plot of the rate of evolution of heatagainst time.1.81 We can see the first peak, whichis very high, and which corresponds to the initialhydration at the surface of the cement particles,

largely involving C3A. The duration of this highrate of hydration is very short, and there follows aso-called dormant period, sometimes called alsoan induction period, during which the rate is verylow. This period lasts one or two hours duringwhich the cement paste is workable.

Fig. 1.4. Rate of evolution of heat of Portlandcement with a water/cement ratio of 0.4.181The first peak of 3200 J/s kg is off the dia-

gramEventually, the surface layer is broken down,

possibly by an osmotic mechanism or by thegrowth of the crystals of calcium hydroxide. The

rate of hydration (and therefore of heat evolution)increases fairly slowly and the products of hydra-tion of individual grains come into contact withone another; setting then occurs. The rate of heatevolution reaches a second peak, typically at theage of about 10 hours, but sometimes as early as4 hours.

Following this peak, the rate of hydrationslows down over a long period, the diffusionthrough the pores in the products of hydration be-coming the controlling factor.1.62 With most, butnot all, cements, there is a renewed increase in therate of hydration up to a third, lower, peak at theage of between 18 and 30 hours. This peak is re-lated to a renewed reaction of C3A, following theexhaustion of gypsum.

The advent of the second peak is acceleratedby the presence of the alkalis, by a higher fine-ness of the cement particles, and by an increase intemperature.

Because of the similarity in the progress of hy-dration of neat calcium silicates and of commer-

cial Portland cements, they show similar strengthdevelopment.1.20 A considerable strength is pos-sessed long before the reactions of hydration arecomplete and it would thus seem that a smallamount of the hydrate binds together the unhyd-rated remainder; further hydration results in littleincrease in strength.

Ca(OH)2 liberated by the hydrolysis of thecalcium silicates forms thin hexagonal plates, of-ten tens of micrometres across, but later theymerge into a massive deposit.1.60

Tricalcium aluminate hydrate and the actionof gypsumThe amount of C3A present in most cements iscomparatively small but its behaviour and struc-tural relationship with the other phases in cementmake it of interest. The tricalcium aluminate hy-drate forms a prismatic dark interstitial material,possibly with other substances in solid solution,and is often in the form of flat plates individuallysurrounded by the calcium silicate hydrates.

The reaction of pure C3A with water is veryviolent and leads to immediate stiffening of thepaste, known as flash set. To prevent this fromhappening, gypsum (CaSO4.2H2O) is added tocement clinker. Gypsum and C3A react to forminsoluble calcium sulfoaluminate(3CaO.Al2O3.3CaSO4.32H2O), but eventuallytricalcium aluminate hydrate is formed, althoughthis is preceded by a metastable3CaO.Al2O3.CaSO4.12H2O, produced at the ex-pense of the original high-sulfate calciumsulfoaluminate.1.6 As more C3A comes into solu-tion, the composition changes, the sulfate contentdecreasing continuously. The rate of reaction ofthe aluminate is high and, if this readjustment incomposition is not rapid enough, direct hydrationof C3A is likely. In particular, the first peak inthe rate of heat development, normally observedwithin five minutes of adding water to cement,means that some calcium aluminate hydrate isformed directly during that period, the conditions

for the retardation by gypsum not yet having beenestablished.

Instead of gypsum, other forms of calciumsulfate can be used in the manufacture of cement:hemihydrate (CaSO4. H2O) or anhydrite(CaSO4).

There is some evidence that the hydration ofC3A is retarded by Ca(OH)2 liberated by the hy-drolysis of C3S.1.62 This occurs due to the factthat Ca(OH)2 reacts with C3A and water to formC4AH19, which forms a protective coating on thesurface of unhydrated grains of C3A. It is alsopossible that Ca(OH)2 decreases the concentra-tion of aluminate ions in the solution, thus slow-ing down the rate of hydration of C3A.1.62

The stable form of the calcium aluminate hy-drate ultimately existing in the hydrated cementpaste is probably the cubic crystal C3AH6, butit is possible that hexagonal C4AH12 crystallizesout first and later changes to the cubic form. Thusthe final form of the reaction can be written:

C3A + 6H C3AH6.

This again is an approximation and not a stoi-chiometric equation.

The molecular weights show that 100 parts ofC3A react with 40 parts of water by mass, whichis a much higher proportion of water than that re-quired by the silicates.

The presence of C3A in cement is undesirable:it contributes little or nothing to the strength ofcement except at early ages and, when hardenedcement paste is attacked by sulfates, expansiondue to the formation of calcium sulfoaluminatefrom C3A may result in a disruption of thehardened paste. However, C3A acts as a flux andthus reduces the temperature of burning ofclinker and facilitates the combination of limeand silica; for these reasons, C3A is useful inthe manufacture of cement. C4AF also acts as aflux. It may be noted that if some liquid were notformed during burning, the reactions in the kilnwould progress much more slowly and would

probably be incomplete. On the other hand, ahigher C3A content increases the energy requiredto grind the clinker.

A positive effect of C3A is its binding capacityof chlorides (see p. 571).

Gypsum reacts not only with C3A: with C4AFit forms calcium sulfoferrite as well as calciumsulfoaluminate, and its presence may acceleratethe hydration of the silicates.

The amount of gypsum added to the cementclinker has to be very carefully watched; in par-ticular, an excess of gypsum leads to an expan-sion and consequent disruption of the set cementpaste. The optimum gypsum content is determin-ed by observation of the generation of the heat ofhydration. As already mentioned, the first peak inthe rate of heat evolution is followed by a secondpeak some 4 to 10 hours after the water had beenadded to cement, and with the correct amount ofgypsum there should be little C3A available forreaction after all the gypsum has combined, andno further peak in the heat liberation should oc-

cur. Thus, an optimum gypsum content leads to adesirable rate of early reaction and prevents localhigh concentration of products of hydration (seep. 362). In consequence, the size of pores in hy-drated cement paste is reduced and strength is in-creased.1.78

The amount of gypsum required increaseswith the C3A content and also with the alkalicontent of the cement. Increasing the fineness ofcement has the effect of increasing the quantityof C3A available at early stages, and this raisesthe gypsum requirement. A test for the optimumSO3 content in Portland cement was prescribedby ASTM C 543-84 (discontinued). The optimiz-ation is based on a 1-day strength, which usuallyalso produces the lowest shrinkage.

The amount of gypsum added to cementclinker is expressed as the mass of SO3 present;this is limited by European Standard BS EN197-1 : 2000 to a maximum of 3.5 per cent, butin some cases higher percentages are permitted.The chemically relevant SO3 is the soluble sulfate

contributed by gypsum and not that from high-sulfur fuel, which is bound in the clinker; this iswhy the current total SO3 limit is higher than inthe past. The maximum values of SO3 laid downin ASTM C 150-09 depend on the content ofC3A, and are higher in rapid-hardening cement.SettingThis is the term used to describe the stiffeningof the cement paste, although the definition ofthe stiffness of the paste which is considered setis somewhat arbitrary. Broadly speaking, settingrefers to a change from a fluid to a rigid stage.Although, during setting, the paste acquires somestrength, for practical purposes it is important todistinguish setting from hardening, which refersto the gain of strength of a set cement paste.

In practice, the terms initial set and final setare used to describe arbitrarily chosen stages ofsetting. The method of measurement of these set-ting times is described on p. 49.

It seems that setting is caused by a selectivehydration of cement compounds: the two first toreact are C3A and C3S. The flash-setting proper-ties of the former were mentioned in the preced-ing section but the addition of gypsum delays theformation of calcium aluminate hydrate, and it isthus C3S that sets first. Pure C3S mixed with wa-ter also exhibits an initial set but C2S stiffens in amore gradual manner.

In a properly retarded cement, the frameworkof the hydrated cement paste is established by thecalcium silicate hydrate, while, if C3A were al-lowed to set first, a rather porous calcium alumin-ate hydrate would form. The remaining cementcompounds would then hydrate within this por-ous framework and the strength characteristics ofthe cement paste would be adversely affected.

Apart from the rapidity of formation of crys-talline products, the development of films aroundcement grains and a mutual coagulation of com-ponents of the paste have also been suggested asfactors in the development of set.

At the time of the final set, there is a sharpdrop in the electrical conductivity of the cementpaste, and attempts have been made to measuresetting by electrical means.

The setting time of cement decreases with arise in temperature, but above about 30 C (85 F)a reverse effect may be observed.1.1 At low tem-peratures setting is retarded.False setFalse set is the name given to the abnormal pre-mature stiffening of cement within a few minutesof mixing with water. It differs from flash set inthat no appreciable heat is evolved, and remix-ing the cement paste without addition of waterrestores plasticity of the paste until it sets in thenormal manner and without a loss of strength.

Some of the causes of false set are to be foundin the dehydration of gypsum when intergroundwith too hot a clinker: hemihydrate (CaSO4.H2O) or anhydrite (CaSO4) are formed and whenthe cement is mixed with water these hydrate

to form needle-shaped crystals of gypsum. Thuswhat is called plaster set takes place with a res-ulting stiffening of the paste.

Another cause of false set may be associatedwith the alkalis in the cement. During storagethey may carbonate, and alkali carbonates reactwith Ca(OH)2, liberated by the hydrolysis of C3S,to form CaCO3. This precipitates and induces arigidity of the paste.

It has also been suggested that false set can bedue to the activation of C3S by aeration at mod-erately high humidities. Water is adsorbed on thegrains of cement, and these freshly activated sur-faces can combine very rapidly with more waterduring mixing: this rapid hydration would pro-duce false set.1.21

Laboratory tests at cement plants generally en-sure that cement is free from false set. If,however, false set is encountered, it can be dealtwith by remixing the concrete without adding anywater. Although this is not easy, workability will

be improved and the concrete can be placed in thenormal manner.Fineness of cementIt may be recalled that one of the last steps in themanufacture of cement is the grinding of clinkermixed with gypsum. Because hydration starts atthe surface of the cement particles, it is the totalsurface area of cement that represents the mater-ial available for hydration. Thus, the rate of hy-dration depends on the fineness of the cementparticles and, for a rapid development of strength,high fineness is necessary (see Fig. 1.5); the long-term strength is not affected. A higher early rateof hydration means, of course, also a higher rateof early heat evolution.

Fig. 1.5. Relation between strength of concreteat different ages and fineness of cement1.43

On the other hand, the cost of grinding to ahigher fineness is considerable, and also the finerthe cement the more rapidly it deteriorates on ex-posure to the atmosphere. Finer cement leads toa stronger reaction with alkali-reactive aggreg-ate,1.44 and makes the cement paste, though notnecessarily concrete, exhibit a higher shrinkage

and a greater proneness to cracking. However,fine cement bleeds less than a coarser one.

An increase in fineness increases the amountof gypsum required for proper retardation be-cause, in a finer cement, more C3A is availablefor early hydration. The water content of a pasteof standard consistency is greater the finer thecement, but conversely an increase in finenessof cement slightly improves the workability of aconcrete mix. This anomaly may be due partly tothe fact that the tests for consistency of cementpaste and workability measure different proper-ties of fresh paste; also, accidental air affects theworkability of cement paste, and cements of dif-ferent fineness may contain different amounts ofair.

We can see then that fineness is a vital prop-erty of cement and has to be carefully controlled.The fraction of cement retained on a 45 m (No.325 ASTM) test sieve can be determined usingASTM C 430-08. (For size of openings of dif-ferent sieves see Table 3.14.) This would ensure

that the cement does not contain an excess oflarge grains which, because of their comparat-ively small surface area per unit mass, would playonly a small role in the process of hydration anddevelopment of strength.

However, the sieve test gives no informationon the size of grains smaller than 45 m (No. 325ASTM) sieve, and it is the finer particles that playthe greatest part in the early hydration.

For this reason, modern standards prescribe atest for fineness by determination of the specificsurface of cement expressed as the total surfacearea in square metres per kilogram. A direct ap-proach is to measure the particle size distributionby sedimentation or elutriation: these methodsare based on the dependence of the rate of freefall of particles on their diameter. Stokes lawgives the terminal velocity of fall under gravity ofa spherical particle in a fluid medium; the cementparticles are, in fact, not spherical. This mediummust of course be chemically inert with respect tocement. It is also important to achieve a satisfact-

ory dispersion of cement particles as partial floc-culation would produce a decrease in the appar-ent specific surface.

A development of these methods is the Wagn-er turbidimeter used in the United States (ASTMC 115-10). In this test, the concentration ofparticles in suspension at a given level in ker-osene is determined using a beam of light, thepercentage of light transmitted being measuredby a photocell. The turbidimeter gives generallyconsistent results, but an error is introduced byassuming a uniform size distribution of particlessmaller than 7.5 m. It is precisely these finestparticles that contribute most to the specific sur-face of cement and the error is especially sig-nificant with the finer cements used nowadays.However, an improvement on the standard meth-od is possible if the concentration of particles5 m in size is determined and a modificationof calculations is made.1.51 A typical curve ofparticle size distribution is shown in Fig. 1.6,which gives also the corresponding contribution

of these particles to the total surface area of thesample. As mentioned on p. 7, the particle sizedistribution depends on the method of grindingand varies, therefore, from plant to plant.

Fig. 1.6. Example of particle size distributionand cumulative surface area contributed by

particles up to any given size for 1 g of cement

It must be admitted, however, that it is notquite clear what is a good grading of cement:should all the particles be of the same size orshould their distribution be such that they are ableto pack densely? It is now believed that, for a giv-en specific surface of cement, early strength de-velopment is better if at least 50 per cent of theparticles lie between 3 and 30 m, with corres-pondingly fewer very fine and fewer very coarseparticles. An even higher proportion of particlesin the range of 3 to 30 m, up to 95 per cent,is believed to lead to an improved early strengthand also to a good ultimate strength of concretemade with such a cement. To achieve such a con-trolled particle size distribution it is necessaryto use high-efficiency classifiers in closed-circuitgrinding of clinker. These classifiers reduce theamount of energy used in grinding.1.80

The reason for the beneficial effect of middle-size particles may be found in the test results ofAtcin et al.1.91 who found that grinding of ce-ment results in a certain amount of compound se-

gregation. Specifically, particles smaller than 4m are very rich in SO3 and rich in the alkalis;particles coarser than 30 m contain a large pro-portion of C2S, while the particles between 4 and30 m are rich in C3S.

It should be noted, however, that there is nosimple relation between strength and cementparticle size distribution: for examp