5 Alkali-silica reactionUK experienceI.SIMSAbstractThe developing awareness and acceptance of alkali-silica reaction (ASR) in theUK is traced from the late 1940s to the present day. Seemingly extensive studiesattheBuildingResearchEstablishmentduringthe1950swereoptimisticallyinterpretedbytheconstructionindustryasreassurance,andthisledtomorethan a decade of near-complacency when ASR was widely regarded as a foreignproblem. In the early 1970s, however, major structures in Jersey and on the UKmainland, particularly in the south-west and Midlands of England, were foundto be damaged as the result of ASR. Some of the earlier examples of structuresrecognised as being damaged by ASR are described, and the particular types ofaggregatecombinationsinvolvedareexamined. Anattemptismadetoassessthe current extent of the occurrence of ASR in the UK and to analyse the scale ofdamagetypicallyfound;someunusualexamplesareincluded.Guidanceandspecifications developed in the UK to minimise the risk of ASR in future concreteconstructionareexplainedindetail,andtheprospectsofdevelopinganadequately predictive test are fully discussed. Finally, critical consideration isgiven to the most recent advice on diagnosis, the appraisal of affected structuresin the UK and the possibilities of management or repair. The outlook is concludedto be generally encouraging, but some newly discovered aspects, including theimportant role of externally derived alkalis, are briefly considered.5.1 IntroductionWhenalkali-silicareactivitywasdiscoveredbyStanton1a,binAmericain1940,Britainwasotherwiseengaged,fightingtheBattleofBritainandsufferingtheBlitz.Duringthewaryearsandtotheendofthe1940s,American researchers established most of the basic parameters of ASR and itsconsequences2. Thefirst ASRtestsonBritishaggregateswerecarriedoutintheUSAin1947(see5.3.2).Intheir13thAnnualReportin19483,theCopyright 1992 Blackie and Son LtdCementandConcreteAssociationintheUKannouncedaninitialresearchprogrammeonteststodeterminewhethertheexpansivereactionbetweenaggregates and cement so often reported from the USA is a possible cause ofdisintegrationinBritain.Itisnotcleartowhatextentthatprogrammewaseverpursuedatthattime,butcertainlyamajorprogrammeofresearchhadcommencedattheBuildingResearchStation(BRS)in1946,underthesupervisionofF.E.Jones,andcontinuedthroughmostofthe1950s.Thepreliminaryfindings,asNationalBuildingStudies(NBS)ResearchPapers(RP),werepublishedin195246,andthefinalconclusionsandrecommendations, also as NBS Research Papers, were issued in 19587,8.ThereassuranceprovidedbytheBRS/NBSResearchPapers,whichwasarguably more apparent than real (see 5.3.3), led to a decade of near complacency,when the UK concrete industry was generally content to regard ASR as a foreignproblem.InJanuary1971,however,amassconcretedaminJersey,ChannelIslands,wasfoundtoexhibitdisplacementsandcrackingwhichwereinduecourse attributed to ASR9, and this finding reactivated British awareness of ASR.Thereawakeningwasrapid,withtheCementandConcreteAssociationandQueenMaryCollege(UniversityofLondon)jointlyhostingtheThirdInternationalConferenceonASRinSeptember197610(thefirstandsecondconferenceshadbeenheldinCopenhagenandReykjavikrespectively).InDecember 1976 unexplained cracking was found to be affecting concrete basesatanelectricitysubstationinPlymouth,Devon,andbyearly1977thiswasknown to be the result of ASR11. Over more than 10 years since that discovery,severalhundredconcretestructureshavebeenidentifiedontheUKmainlandwhich are affected, to a greater or lesser extent, by ASR.Likeadiseaseforwhichthecureisuncertain,thealarmandconcerngenerated by ASR has been generally out of proportion to the low incidence ofoccurrenceandthetypicallylimitedscaleofstructuraldamageinvolved.However, the degree of interest in ASR has enabled good progress to be madein the UK in formulating measures to minimise the risk of ASR in new concretework12,13,inestablishingagreedproceduresforthediagnosisofASR14and,mostrecently,inprovidingguidanceforthestructuralassessmentofaffectedstructures15. Althoughthe50thanniversaryofthediscoveryof ASRhasbeenreached, it seems less likely that modern UK concrete will ever suffer ASR andevenalreadyaffectedstructureswillmostlyprovetoberepairable.5.2 Reassurance: studies at the BRS5.2.11952reviewandinitialfindings46TheBRSresearchcommencedwithadetailedreviewoftheexperiencestodateintheUSA,includingreferencestospecificstructures,anassessmentoftest methods and an appreciation of corrective measures to avoid ASR in newCopyright 1992 Blackie and Son Ltdworks4.InhisprefatorynotetoRP14,F.M.Lea,thenDirectorofBuildingResearch, warned: IntheUSAmanyseriousandextensivefailuresofconcretehavesince(1940)beenattributedtoalkali-aggregatereaction. Fortunatel y, noevidencehasyetbeenobtainedoflarge-scalefailuresinthiscountryarisingfromthiscause;nevertheless,someaggregatedepositswhichmaybereactiveinsomedegreedooccur, andunfamiliaraggregatesmayneedexaminationbeforeuse. ThesecondNBSResearchPaper,No.15,providedageneralpreliminaryconsiderationofBritishPortlandcementsandBritishaggregates5.Thetotalalkali content of British cements, as analysed at the BRS since 1928, was foundtorangefrom0.5to1.0%Na2Oeq.andtobetypicallyhigherthanthe0.6%maximumlevelsuggestedbyAmericanexperienceforpreventingASR.Furthermore, extraction tests indicated that all of the acid-soluble Na2O and K2OinBritishcementswouldbecomeavailableforreactioninconcrete.InBritishaggregates, flint in the south-east England gravels was recognised to be slightlyreactive,probablyasaresultofmicrostructureratherthanthepresenceofopalinesilica16,andtheacknowledgedfreedomfromtroubleinpracticewasthoughttobeassociatedwiththeuseofwholeflintaggregate.Thus,evenin1952, the potential importance of the flint pessimum was understood and led tothe cautionary statement in RP 15: There is a possibility that if flints are used inadmixtureswithinertaggregates,sufficientexpansiontocausetroublemightoccur.OpalineMalmstoneandsomeglassyacidtointermediatevolcanicrockswerealsoidentifiedaspotentiallyreactivematerials,butthesewerenotconsideredtobeimportantsourcesofaggregatesintheUK.The expansion bar test (a mortar bar test), which had been identified in thereviewasseemingtobethemostgenerallyusefulprocedurecurrentlyavailable,wasthenappliedtoalimitedrangeofBritishaggregatesusingacementofmediumalkalicontent(0.7%Na2Oeq.)andemployingstorageat20C.Theresultsupto4yearsweregiveninNBSResearchPaperNo.176,wheretheinterpretationwasbaseduponcomparisonwiththeresultsobtainedin the same test series for two opal-bearing aggregates imported from the USA(Californiansiliceousmagnesianlimestone10%in Thames Valleyflintsand-and Nebraska sand). The American control samples gave results of 0.38% and0.03% respectively at 6 months, with 0.72% and 0.18% at 4 years, whereas thehighestvalueobtainedforanyoftheBritishnaturalaggregateswasnotmorethan0.03%upto4years(Table5.1andFigure5.1).Thisseeminglyhopefuloutcomewasconfirmedbythestatementthattheconclusionsregardingtherocktypesconcernedarereassuringforaggregatesinthiscountry,althoughthisviewwasconditionedbytheneedtobemindfulofthepresenceofanyopalinesilica.Withthebenefitofhindsight,itisperhapssignificantthatlow-powermicroscopicalexaminationofbrokenmortarbarsurfacesaftertestingidentifiedsmallamountsofalkali-silicageldepositsassociatedwiththeCopyright 1992 Blackie and Son LtdThames Valleyflintsandsample,similartothosefoundoccurringwithinthecontrolspecimensmadewithAmericanaggregates.ThiswouldseemtoindicatethatASRhadoccurredwithThamesValleyflintsand,butthatexpansionhadnotresultedinthesetests.However,thisobservationwasnotemphasisedandtheimpressionleftwasthatBritishaggregateswerenotexpansivelyreactive,incontrasttosomeofthosefoundoverseas.5.2.21958conclusionsandrecommendations7,8TheseriesofexperimentsdescribedinRP17(see5.2.1)wererepeatedusinghigh-alkali cement, actually the 0.7% Na2O eq. cement with added NaOH toreach an equivalent 1.2%, and the findings reported in Research Paper No. 207.Also, further British aggregates were tested, including representatives of thosematerialsinitiallyidentifiedinRP14asbeingpotentiallyreactive:flint,Malmstoneandacidtointermediatevolcanicrocks.Somelimitedstudieswerecarriedoutintoaggregatemixturesanddilutions.Inthehigh-alkalicementrepeatexpansionbartests,theAmericancontrolaggregateexhibitedhigherexpansionsthanwererecordedfortheearliermediumalkalicementseries,especiallyat6months(Table5.1).After4years,noneoftheBritishnaturalaggregatesexhibitedanyTable5.1Selectedmortarbarresults(percentageexpansion)fromBRS/NBSResearchPapers17and20 6,7.Allresultsareforlow-qualitymortar-bars(higherwater-cementratiosandgreaterporosities)storedat20C.Copyright 1992 Blackie and Son Ltdsubstantialexpansion,thehighestvalueobtainedbeingonly0.035%.Anextensionoftestingto7yearsdidnotsignificantlyaltertheconclusion,withthehighestvaluemovinguponlyto0.042%.Tests were carried out on additional flint aggregates from four sources, twoin the Thames Valley, one in Essex and another in Cambridgeshire. None of theflintgravelsexceeded0.02%inthesetests,evenafter7years,butthesingleflint-bearingsand(composedofquartzandflint)exhibitedrathermorebutstillcomparativelylowexpansivity,reaching0.05%after4yearsand0.06%after7years.Onceagaintheexpansionbarspecimensmadewithflintaggregates exhibited some evidence of ASR, in the form of gel deposits visibletolow-powermicroscopicalexaminationaftertesting.Theinherentreactivityofflint,notwithstandingthelackofexpansioninthese tests, was demonstrated by some chemical tests reported in RP 20 (pattestandrapidchemicaltestdescribedinRP25,seebelow).Aseriesofdilutions using Clee Hill basalt (considered inert) with Thames Valley flintaggregatedidnotproduceanysignificantexpansionsinthetestsat20C,althoughearlierworkintheUSAhadsuggestedthatatleastsomeBritishflintcouldbeinducedtocauseexpansion,upto0.14%at12months,whenpresentasaround20%ofthetotalaggregate,atleastattheelevatedtemperature of 38C17.Figure5.1ComparisonofmortarbarexpansionsforreactiveAmericancontrolaggregatesandThamesValleyflintsand:ThamesValleysand;FNebraskasand;bsiliceousmagnesianlimestone10%replacementingradedriversand.Allresultsareformediumalkalicement(0.7%Na2Oeq.),low-qualitymortarbars(higherwater-cementratiosandgreaterporosities)storedat20C.RedrawnandadaptedfromRef.6.Copyright 1992 Blackie and Son LtdA range of acid to intermediate volcanic rock materials, including rhyolitesandandesites,wastestedusingthesameregime,variouslyaswholeaggregate or as 50:50 mixtures with Clee Hill basalt (considered inert). Noneoftheaggregatesorcombinationstestedproducedanyexpansionsgreaterthan 0.03%, even after 3 years. Chemical tests also showed a complete lack ofreactivity for these volcanic rocks.Malmstonewassubjectedtoextensivetestingasithadbeenfoundtocontain opaline silica, although it was not in practice a source of aggregate fornormalconcrete.Asexpected,thechemicaltestsindicatedconsiderablereactivityofMalmstone,buttheexpansionbartestsproducednoexcessivemovements,eveninvariousdilutionswithflintsandandCleeHillbasalt.ThisapparentanomalywasexplainedinRP20asbeingcausedbythehighlyporousnatureofMalmstone,ratherthanbyanyinherentinadequacyintheexpansionbartestprocedure.Jones and Tarleton7 considered that the findings reported in RP 20 confirmedthose given 6 years earlier in RP 17 on a firmer and wider basis, so that it couldbeconcludedthatthenormalBritishaggregatessofartested,whenusedaswholeaggregates,arenotexpansivelyreactivewithhigh-alkalicementsatnormal temperatures. Crucially, however, they added some words of caution inrespect of flint aggregates, perhaps partly because of the observation of evidenceofreactionintheabsenceofharmfulexpansioninthetests: Thereremainsthepossibilitythereforethatsomeflintsmaybeencounteredwhich,underadverseconditionsofdilution,alkalicontent,watercontentandtemperature,maycausetrouble.Sofaraspresentevidencegoes,however,thisisconsideredtoberatherunlikely. ThefinalNBSResearchPaper,No.258,comprisedtwoparts,bothconcernedprimarily with test procedures and, as such, accessory to the substantive resultsandconclusionsofRP20,butneverthelessinfluentialoverfutureBritishtestingpractice.Inthefirstpart(actuallypart VoftheNBSseriesonalkali-aggregatereaction),theapparentdeteriorationinbendingstrengthofmortarbarswasfoundtobeusefulsupplementaryevidencebutwasnotconsideredto be a preferable alternative to expansion testing. Also, the observation of geldepositseitherwithinorontheoutersurfacesofmortarbarswasthoughttoprovideatleastvaluableevidenceofpotentialexpansivereactivity,althoughitwasconcededthatsuchobservationshadnotalwaysbeenaccompaniedbyexpansionintheBRStestprogramme.The second part of RP 25 (actually part VI of the NBS series) discussed the testmethods used at the BRS and gave recommended test procedures. The pat test (orgel pat test), developed from a method earlier described by Stanton et al.18, wasrecommended as a useful rapid test for reactive material and was perhaps usedsubsequentlymoreintheUKthanelsewhere.Arapidchemicaltestwasrecommended as a useful acceptance test, albeit a modified version of the ASTMC289 method19, and a maximum dissolved silicaalkalinity reduction ratio of 1.5Copyright 1992 Blackie and Son Ltdwassuggested,incontrasttothemaximumratioof1giveninASTMC3320.Finally, the expansion bar test was recommended as the standard reliable test fordeterminingthedegreeofexpansivereactivityofanaggregateinvariouscombinations with cements. A BRS method for the expansion bar test was describedwhich differed in several important details from the ASTM C227 method21; thecement alkali content was standardised at 1.2% Na2O eq., the aggregate-cementratio was 2 rather than 1.25 and the preferred storage temperature was 20C ratherthan 38C. Using this BRS test, the aggregate was to be classed as expansivelyreactive if measured expansion exceeded 0.05% at 6 months or 0.1% at 12 months,in contrast to the revised ASTM C33 similar limits at 3 and 6 months respectively.Somementionwasmadeofthepossibilityofapplyingtheexpansiontesttoconcrete specimens, but it would be more than 20 years before such a procedurewas developed at the BRS.5.2.3SignificanceandeffectAtthetimethatDrJoneswasfinalisinghisBRSresearchprogramme,therewerenoknownexamplesof ASRaffectingstructuresintheUK,andthiswillinevitablyhavebeenregardedasanimportantfactorbearingupontheemphasisoftheconclusionstobedrawnfromthelaboratoryresults.Thelaboratory findings were seemingly consistent with the experience in practice,thatisBritishnaturalaggregateswerenotexpansivelyreactive.Totheirconsiderablecredit,JonesandTarletonrecognisedthatflintaggregates,inparticular,hadatleastahypotheticalpotentialforexpansivereactionincertaincircumstances,notwithstandingthelackofexpansionproducedforflintaggregatesintheBRStests.However,thepropheticcautionscontainedwithintheconclusionstotheBRSworkinrespectofBritishflintaggregatesweregenerallysupersededinthereadersappreciationbytheoverallreassurancewhichappearedtoemergefromanapparentlythoroughsurvey.Petrographically,flint(adistinctivechertofUpperCretaceousage)isapotentiallyalkali-reactiveformofsilica,evenwithoutthepresenceofanyopaline material. As early as 1947, flint from Dover cliffs had been shown to beexpansivelyreactiveinAmericantestswhenformingaround20%ofanotherwisenon-reactivetotalaggregate17.IntheBRStestprogramme,flintaggregates had shown clear evidence of ASR within expansion bar specimens,aswellasinseparatechemicaltests,althoughnosignificantexpansionshadbeenrecorded.Forsomereason,theexpansiontestingapproach,modellingnormal concrete with small mortar specimens of high cement content, appearstohavebeenregardedasbeyondreproach,despiteitsfailuretocauseexpansionevenwithopal-bearingMalmstonematerial.SubsequenteventshaveshownjustificationforJonesand TarletonscautiousappraisalofBritishflintaggregates,butitwastobenearly20yearsbeforetheUKconcreteindustrywouldacceptthat ASRwasanythingotherthanaforeignproblem.Copyright 1992 Blackie and Son Ltd5.3 False security: complacency in the 1960s5.3.1BritishStandardsTheBritishStandardsInstitution(BSI)isanindependentcharteredbodydevotedtotheassuranceofuniformityandqualityofmaterialsandproductsintheUK.PublishedBritishStandardsarenotlegallybinding,exceptwhenthey are called up by a contract specification or similar document, but they areacceptedthroughouttheconstructionindustryasbeingdependableandauthoritativeguidance.TheconstitutionoftheBSIensuresthatallpublishedstandardsarecompiledbyrepresentativecommitteesandthattheyareregularlyreviewedandamendedorrevisedwhennecessitatedbynewknowledge, recent developments or changed circumstances. British Standards,includingCodesofPractice,arethereforeimportantandfairlyreliableindicatorsofthecontemporaryviewswithintheconstructionindustryatthetimes of their publication, although they do represent a consensus and may notadequatelycoverparticularlycontentiousmatters.Thematerialsusedinthemanufactureofconcretewerecoveredinthe1960s,asnow,byBSSpecifications,notablyBS12forPortlandcement22(ordinary and rapid-hardening), BS 882 for aggregates23 (natural) and BS 3148forthemixwater24.TestmethodsweregivenforcementinBS12andBS455022,25andforaggregatesinBS81226.GuidanceforthemakinganduseofconcreteforstructureswasgiveninCP114,CP115andCP1162729,latersuperseded by CP 11030 (now BS 811031), whilst test methods for concrete weregiveninBS188132. Withonesmallexception,noneoftheseBritishStandardpublications made any direct or allusive reference to ASR until the 1980s, andeventhenonlyBS8110containedanysubstantiveguidance.ThesingleexceptionwasBS314824,inwhichitwassuggested(inanappendix)thatthealkalicontentofsomemixwatersmightbeofsignificanceinthepresenceofalkali-reactive combinations of aggregates and cement, although it was furtherstatedthat,Sofar,nonaturallyoccurringexpansivelyreactiveaggregateshavebeenfoundintheUnitedKingdom.The1954editionofBS882,thespecificationfornaturalaggregatesforconcrete,wasextensivelyrevisedandaneweditionpublishedin1965.Acomparisonofthe1954and1965editionsshowsthatthestandardwasrearranged and, in places, completely rewritten. The final results of the BRSresearch programme had been published in 1958 (see 5.2.2) and the findingsmusthavebeenfamiliartoallmembersoftherevisingcommittee;indeed,althoughthefullmembershipofthatcommitteeisnolongeravailable,itisknownthattheBRSwasrepresentedbyTeychenn33.Therefore,itseemscertainthataconsciousdecisionwastakenbytheBSIcommitteetoomitany reference to ASR in the concrete aggregates specification. One survivingmemberofthatcommitteehasindicatedthatthematterwasdiscussedandthat,becauseBRSresearchhadonlyrecentlyconcludedthatASRwasCopyright 1992 Blackie and Son Ltdunlikely to be a problem with British aggregates and no actual examples hadbeenreported,itwasconsideredinappropriateorevenunhelpfultoraiseconcernover ASRinthenationalspecification33.Eveninthe1983revisionofBS882,mentionof ASRisrestrictedtoanon-mandatorynotewithinanappendix, and that is only a cross-reference to the Code of Practice, BS 8110.From the end of the Second World War to the discovery after 1971 that ASRhad damaged the Val de la Mare Dam in Jersey, and by 1977 that concrete ontheUKmainlandwasalsoaffected,therewasvirtuallynoreferencetothepossibilityofASRinBritishStandardsrelatingtoconcreteandconcretematerials. Therefore, any supplier or user dependent upon the provisions of therelevantBritishStandardsforassuringthequalityofconcretewouldnothavebeencausedtomakeanyallowancewhateverforthepossibility,howeverremote, of ASR occurring at some time after construction. It has been argued,albeit retrospectively, that the broad-brush quality requirements of BS 88223,includingtheprovisionthataggregatesshallnotcontaindeleteriousmaterialsinsuchaformorinsufficientquantitytoaffectadverselythestrengthatanyageorthedurabilityoftheconcrete,couldbeconsideredtohave embraced ASR, even though ASR was not one of the specific examples ofsuchdeleteriousmaterialslistedinthestandardtoclarifytheabove-quotedrequirement.Thevalidityofsuchanargumentdependsupontheextenttowhichalternativeauthoritativeguidance,otherthanBSIdocuments,wasavailableintheUKtoadvisesuppliersandusersofthepossibilityof ASR.5.3.2OtherauthoritativeguidanceandadviceIn the UK, apart from BSI, authoritative national guidance on construction materialsand practice is available from the Building Research Station (or Building ResearchEstablishment) and, where appropriate, from the Cement and Concrete Association(C&CA, now renamed the British Cement Association, BCA) and the Transport andRoad Research Laboratory (TRRL). Occasionally similarly authoritative and usefulguidance may be produced on specific topics by the relevant professional institutions,learnedsocietiesoreventradeassociations. Academicpublicationsandcertaintextbooks can also provide useful guidance, but, by their nature, are not necessarilyrepresentative of the consensus viewpoint.After the publication of Jones and Tarletons final NBS reports in 1958 (see5.2.2),theBRS(BRE)didnotproduceanyfurtherofficialguidanceuntil1982,whenDigest258wasissued34,some6yearsaftertheidentificationofASR on the UK mainland. However, BRE researchers had been involved in theinvestigationintotheValdelaMareDaminJersey(see5.5)andMidgleydescribedatechniquearisingfromthisinvestigationatthe1976ASRconferenceinLondon35.ThenextpublicationonASRbyBREstaffdidnotappearuntil197936.Itisthusapparentthatforthewholeofthe1960s,andalsoforagoodpartofthe1970s,theseemingreassuranceprovidedbytheCopyright 1992 Blackie and Son Ltdearlier 19461958 BRS research programme remained unchallenged. Indeed, aBREDigestpublishedin197137,onthesubjectofconcreteappearance,containedthefollowingstatement: Fortunately,therearenoknowndepositsofaggregateintheUnitedKingdomthathaveanysignificantdegreeofalkali-aggregatereactivity,butadesignerofconcretebuildingsforotherpartsoftheworldmightrunintoseriousdifficultiesifhedoesnotgetadviceonthesuitabilityoftheavailableaggregates,(presentauthorsemphasis) Clearly, ASR was still to be regarded as a foreign problem.AlthoughtheC&CAdeclaredanintentiontocarryoutresearchinto ASRin19483, there is no available evidence that any such work was carried out, and theearliestC&CApublicationonthesubjectappearedaslateas197738,withtheirfirst research work appearing a year later58. A C&CA advisory note on the subjectofaggregateimpuritieswaspublishedin197040andmadesomereferencetoalkali-reactiveminerals,butconcludedconfidently,Inthiscountry,thesemineralsarefortunatelynotpresentintheaggregatesusedformakingconcretebut,inanyevent,commerciallyavailablecementsseldomcontainenoughalkalis to be troublesome. In practice then, the question of aggregate-alkali (sic)reactionconcernsonlycementsandaggregatesusedabroad(presentauthorsemphasis).RoedercarriedoutsomelimitedresearchatC&CAintoflintreactivity as part of a wider study in 19747541, but the results were not formallypublished and he was careful to state that, More recently interest has centred onthepossibilityofreactionbetweentheaggregateandthealkalisinthecementbutnoexamplesofsuchreactionareknowninthiscountry. Again,therefore,untilthelate1970s,theclearimpressiongivenwasthat ASR,whenmentionedat all, was only a foreign problem.TheSandandGravelAssociationofGreatBritain(SAGA)organisedanimportantsymposiumin1968onthesubjectofsea-dredgedaggregatesforconcrete,whichwereenjoyingarapidlyincreasingrateofuse.Thevariousauthors,forexampleShacklock39,wereprimarilyconcernedwiththepossibleinfluences of sea salt and shell material, and ASR was only mentioned brieflyby two of the contributors. Haigh42 (from a firm of consulting engineers) statedthat,Alkali-reactiveaggregates,apparentlynotaproblemwithaggregatesinthis country, are to be found frequently overseas. (present authors emphasis).VandeFliert43,referringtosomeDutchconcernsovertheuseofflintaggregates,saidthatAbout10yearsago,simultaneousresearchinGreatBritain(presumablythatatBRS)andintheNetherlandsshowedthatthisexpansion reaction with crypto-crystalline quartz stone (flint) in concrete neednot be feared (present authors parentheses).Most subjects become covered by numerous textbooks over a period of time, butinevitably some take on a particular authority, either in deference to the eminence ofthe author and/or his affiliations, or because of popular demand and the consequentreprintingandperiodicrevisingofthework.Forillustrativepurposes,justthreesuch books will be examined here. Nevilles book, Properties of Concrete, was firstCopyright 1992 Blackie and Son Ltdpublished in 1963; a second edition was produced in 1973, and the third edition isdated 198144. In the 1963 and 1973 editions, a thorough account of ASR is concludedby the statement, The alkali-aggregate reaction of the type described has fortunatelynot been encountered in Great Britain, but is widespread in many other countries(present authors emphasis). The 1981 edition is reworded slightly and makes briefreference to examples identified in Great Britain in 1978, but, for nearly 20 years,this popular and influential book had appeared to support the notion that ASR wasa foreign problem.Leas important specialist textbook, The Chemistry of Cement and Concrete, wasreissued as the third edition in 197045, being a revision of the second edition by LeaandDeschpublishedin1956.SinceLeahadbeenDirectorofBuildingResearchthroughouttheperiodofJonessworkatBRSinthe1950s,hisrathercautioussummary of ASR in respect of the UK is now seen to be significant. Lea states: AsurveyofthecommonBritishaggregateshasfailedtorevealanycontainingalkali-reactiveconstituents,butcareisneededwhenusingaggregatesofgeologicaltypeswhichmightcontainsuchconstituentsandofwhichnopreviousserviceexperienceisavailable.Itmighthavebeenexpectedthatflint,whichissuchalargeconstituentoftheThamesValleygravels,wouldbereactive.Longexperiencewiththisaggregatehasshownnosuchtroubles,butitappearsthatitconsistsofamicroporousmassofsilicaanddoesnotcontainopal. AlthoughLeadoesnotimplythatASRisexclusivelyanoverseasproblem,andadvisescautionwithsomenewanduntriedsourcesofaggregate,heisneverthelessfairlyunequivocalinhisassessmentofUKflintgravelasanon-reactiveaggregate.Finally,theRoadResearchLaboratory(laterTRRL)publishedthewellknownConcreteRoads,DesignandConstructionin195546,anditremainedavailableforpurchaseuntil1975.Theadvicegiventhereininrespect of ASR is consistent with the findings emerging from the BRS, and theinterpretationembodiedbytheBSIandotherauthorities: Someaggregates,notablythosecontainingopalinesilicaorchert,maycombinewiththealkalisinthecementtoproducedisruption.ThistroublehasnotbeenobservedinGreatBritain,althoughitismetwithintheUSAandelsewhere.5.3.3 Signs of scepticismItisapparentfromtheprecedingsectionsthattheoverwhelmingimpressiongained by the UK construction industry during the 1960s and the early 1970s wasthat ASR was an interesting but, fortunately, a foreign problem. Geologically, thisperception was always tenuous, since the inherent reactivity of certain mineral androckvarietiesdoesnotchangeatpoliticalboundaries,andinfactproperresidualconcern was expressed to some extent both by Jones and Tarleton7 and by Lea45 inrespect of those aggregate materials which were not included in the BRS researchCopyright 1992 Blackie and Son Ltdprogramme.Bythelate1960sandearly1970s,therearesomesignsthatnoteveryoneintheindustrywascompletelyreassured,especiallythosewhohadexperiencedalkalireactivityinotherpartsoftheworld.Idorn was invited to present a lecture to the Concrete Society in London in May1968, on the subject of the durability of concrete47. Idorn had recently completedasubstantialdoctoralstudyof ASRandotherthreatstodurabilityofconcretestructures in Denmark48, and much of his lecture to the Concrete Society concernedASR. Whilst it is undoubtedly significant that Idorn did not presume to suggest inhis lecture that ASR was likely to become recognised as a problem in the UK, it isequally apparent that the materials he had studied were generally similar to manyof those found and used in the UK. It was this general similarity with Danish andGermanflintaggregates,andtheuncertainnatureofthedistributionandprovenance of marine aggregates in the North Sea, that caused Fookes to recommendtheregularexaminationand ASRtestingofthesea-dredgedaggregatestobestockpiled for use in concrete for the prestigious Thames Barrier project in London,preparations for which commenced in the early 1970s49.AppliedgeologicalresearchintoASRandotheraspectsofconcretecommencedatQueenMaryCollege(QMC),LondonUniversity,attheendofthe1960s,inspiredbyFookes(avisitinglecturerandlatervisitingprofessor)andalsopromotedbythegrowingawarenessofspecialconcreteproblemsbeingencounteredintheunprecedentedMiddleEastconstructionboom50.DoctoralresearchcommencedatQMCin1972toincludesomestudyofconcretefromtheValdelaMareDaminJerseyandalsothealkali-silicareactivityofBritishflintaggregates51.Rapidexperiencewasgainedbytheinvestigationofreactivematerialsfromoverseas,includingCyprus52,Bahrain53 and elsewhere54. The Geomaterials Group at Queen Mary College isnow rightly regarded as one of the foremost authorities on ASR in the UK andthey started out by being unwilling to accept that it was safe to dismiss ASR asbeingexclusivelyaforeignproblem.5.4 First warning: the Val de la Mare Dam in Jersey5.4.1BackgroundThe Val de la Mare (Figure 5.2) mass concrete gravity dam at St Ouens inJerseywasbuiltbetween1957and19619,51,55,56.Thestructure,whichis29mmaximumheightabovefoundationlevelandextendsfor168mstraightacrossthevalley,retains950millionlitresoutoftheislandstotalwaterstoragecapacityof1500millionlitres.Constructionwasinverticalbaysorblocks,each6.7mwideandmadeupofliftsectionseach1.2mhigh.TheconcretewasmadeusingUKPortlandcement,togetherwithlocalaggregatesandwaterfromthestreamthatwastobedammed,andthemajoritywasplacedin195960. Copyright 1992 Blackie and Son LtdFigure5.3PartofthedownstreamfaceoftheValdelaMareDam,Jersey,viewedfromtheparapetwalkway,showingmapcrackingandexudations.FromRef.52.Figure5.2ValdelaMareDam,Jersey,inApril1989(cf.Figure5.7).Copyright 1992 Blackie and Son LtdThe Jersey New Waterworks Company first reported in 1971 that four oftheconcretesectionshadbecomediscoloured,dampenedandcracked,andtheconcretecresthandrailwasinsomeplacesdisplacedbyupto13mm(Figures5.3and5.4).Thisledtoinvestigations,includingsonicspeedmeasurements57,petrographicalexaminationsandaggregatetests,andremedialactionincludingtheinstallationofanchorbars,whichwascompletedby1975. Monitoringoverthefollowing13yearshasdemonstratedaslowlycontinuingexpansionatarateof5075microstrains/year and the failure of an anchor bar, caused by seismic shockratherthan ASR,butconfidencehasgrownthatthedamcannowenjoyaworkinglifecomparablewithotherconcretegravitydams9.5.4.2ReactiveconstituentsThe coarse aggregate comprised mainly crushed diorite or granodiorite fromalargequarryatRonezonthenorthcoastofJersey,mixedwithasmallerproportionofbeachgravelcontainingroundedpebblesofJerseyshale,chertandvariousigneousrocksincludingdacite.ThefineaggregatewasafossilbeachsandfromapitatStOuensBay,comprisingquartz,quartziteandfeldspar,withsmallamountsofchert,shellandvariousigneousmaterials.Althoughthebeachaggregateswerefoundtocontainsomepotentiallyreactiveconstituents,principallychertanddacite(acidvolcanicrock),thereislittleevidencethatthesematerialsreactedwithinthedamconcrete.Thereactivesilicawaspresentassecondaryhydrothermalveiningandvughfillingswithinthecrusheddioriticrockaggregatesource.Theveining material, which is associated with shear zones and lamprophyre dykeFigure5.4Displacementbyupto13mmoftheconcreteparapetoftheValdelaMareDam,Jersey.FromRef.52.Copyright 1992 Blackie and Son Ltdzoneswithinthequarry,comprisedvariablemixturesofchalcedonyandopal.Theseveinandchalcedonisedlamprophyrematerialsshowreactionwithin a few days at 20C in the gel pat test (Figures 5.5 and 5.6)51, whereasnosuchrapidreactioncouldbeobtainedwiththechertanddaciteparticlesfromthebeachdeposits.Althoughthechalcedony/opalveiningaffectsmuchoftheRonezrockmass, it clearly forms much less than 1 % of the rock, and contemporary workbyBRSconsideredittobelessthan0.1%9.Thepresenceofsomebeachaggregateinthecoarsefractionwillhavedilutedthiscontentofchalcedony/opal still further in the concrete. The pessimum for opal is known to be verylow58,say15%,butexpansivereactionseemsunlikelywithopalcontentsaslowasthosecomputedfortheRonezveiningmaterial.Itmustthereforebepresumedthatvariationsduringquarryingwilloccasionallyleadtohigherconcentrationsofreactivesilicaincertainbatchesofconcrete.TheBRScarriedouttestsonaggregatesamplesusingthemortarbarmethod,whichhadbeenthebasisoftheirNBSresearchprogrammeinthe1950s.Thetestdetailsarenotavailable,butitisreportedthatnoneoftheirresultsexceededtheASTMC33guidancecriteria20at6months9andthatFigure5.5Gelpattestofchalcedony/opalveinmaterial,similartothereactiveconstituentpresentinconcretefromtheValdelaMareDam,Jersey,after14daysofthetestatroomtemperature.FromRef.52.Copyright 1992 Blackie and Son Ltdresults for the Ronez and beach aggregates were similar. In view of the relianceplaceduponmortarbartestresultsinprovidingreassuranceaboutUKaggregates,thismusthavebeenaworryingoutcome.Itseemslikely,frommorerecentexperience,thatexpansionswouldonlyhavebeenproducedinthelaboratorybythecontrolledadditionoftheopalveiningmaterialtotheaggregate,andperhapsonlythenusingcomparativelyhighalkalicontentsandconcreteratherthanmortarspecimens.Thesource(orsources)ofthecementintheUKisnotknownforcertain,anditispossiblethatcementsfromvariousworksweresuppliedovertheperiodofconstructionofthedam.Approximatevaluesfortwoconcretesamples have been quoted as 0.74% and 0.96% Na2O eq. by weight of cement9,andCoombesfoundthatcementsuppliesduringplacementofthemostseverelydamagedconcrete(June-August1960)mostprobablyaveraged0.95%Na2Oeq.55.Thus,themostaffectedportionsofconcretemighthavecontainedhigherthantypicalamountsofopalineveinmaterialintheaggregateandalsorelativelyhighconcentrationsofalkalisfromthecement.Additionalreactivealkaliscouldhavebeenderivedfromtherockaggregatesthemselves,orfromseawatersaltcontaminationofthebeachaggregates, and in any water-retaining structure there is always the possibilityofalkalimigrationstoformlocalisedconcentrations.Figure5.6Gelpattestofchalcedonisedlamprophyredykerock,similartothereactiveconstituentpresentinconcretefromtheValdelaMareDam,Jersey,after14daysofthetestatroomtemperature.FromRef.52.Copyright 1992 Blackie and Son Ltd5.4.3Featuresanddamageattributedto ASRThe visible damage and correspondingly low sonic velocities were almost entirelyrestrictedtoconcreteliftscastduringthecriticalperiodJuneto August1960(Figure 5.7) and block 6 was worst affected. The downstream concrete surface ofthe dam was stained with heavy carbonate exudation from between the sections,but the most serious discoloration was mainly concentrated into those areas castin the critical period (Figure 5.3). The overall discoloration in the affected areaswasfromanoriginalgrey-bufftoarelativelybrightorange-brown. Affectedareas also displayed random map cracking, these cracks being frequently borderedbyorange-browndiscolorationofslightlysmoothersurfacetexturethanthesurrounding concrete. There were white, sometimes porcellaneous, exudationsfromthesecracks,especiallythoseorientatedhorizontallyornearlyso,andthesedepositsincludedsomealkali-silicagel.Nosurfacepop-outswereobserved, but there was evidence of significant expansion in the form of large-scalemisalignment(Figure5.4).Goodcorrelationwasfoundbetweensonicvelocities, date of casting and the visible condition of the concrete. Nearly all ofthe sonic velocity results for 47 concrete lifts cast during the critical period wereless than the overall average of 4.56 km/s for the 77 unaffected lifts measured57and, between 1972 and 1983, the maximum apparent reduction in sonic velocityfor the affected concrete was 14%.Apartfrommeasurementoftheobvioushandraildisplacement,theearlyreports55,56donotindicateanyattempttoquantifytheamountsofanyexpansivemovementwhichhadoccurredupto1971.TheBRSisreportedtoFigure5.7DownstreamelevationoftheValdelaMareDam,Jersey,showing(a)theliftscastJune-August1960and(b)theircrackingintensity.FromRef.9.(a)(b)Copyright 1992 Blackie and Son Ltdhave carried out some tests for latent expansion using core samples, obtainingresultsofonlyabout0.01%after1yearat20Cand100%relativehumidity(RH),andsimilarlyafter3monthsat38Cand100%RH.Althoughtheseresultsarelow14,15andmightindicatethatmostofthepotentialforreactionandexpansionhadbeenexhaustedbeforedrillingandtestingofthecores,thereisnocertaintythatthesampleswerekeptmoistpriortotestinginaccordancewithmodernpractice,andalsoitisnotclearwhetheranyofthetestswerepursuedforstorageperiodslongerthanthosereported.The petrographical examinations carried out by the BRS do not appear to havebeen published, although it is generally believed that microscopic evidence of ASRwas identified. Sims51 obtained four small samples of the concrete and carried outpetrographical examinations; he was able to identify some gel deposits within theconcrete, variously infilling voids or lining cracks and fracture surfaces, and sporadicpotentiallyreactiveaggregateparticles,includingchalcedony,chertandrhyolite,but he could not find any definite reaction sites in the four samples available. Simsalso reported less microcracking in the samples than might have been expected forexpansive ASR and suggested that perhaps the main damaging reactions had occurreddeep within the innermost parts of the mass concrete, perhaps there promoted byconditions of higher temperatures and increased pore water pressures. This suggestionappeared to be supported by the composition of the gel deposits, in which relativelyreducedalkalicontentsandthepresenceofcalciteandsulphatesindicatedconsiderable migration away from the reacting centres. It was further suggested thatthe conditions prevalent at the site of reaction might have favoured the productionof a comparatively fluid gel, able to migrate, rather than a viscous swelling gel. If so,the observed cracking and expansion in the structure may have resulted primarilyfrom the initial reaction, with continued gel-producing reaction not promoting anysignificant further expansion.5.4.4 Remedial measuresAlthoughitwasassumedthatreactionwouldcontinue,theinvestigationsin197174indicatedthattheconcretewouldnotbecomeincapableoftakingtherequiredcompressiveloads,butthatexpansivecrackingcouldleadtoinstabilitybecauseofincreasinginternalupliftpressures9.Remedialactionwasthereforerestrictedtomeasurestomaintainstabilityofthedamagainstincreasedinternaluplift,andalsoitwasdecidedthatonlytheworst-affectedpartswouldbedealtwithinitially,otherareasbeingsubjectedtoasimilarform of repair as and when further deterioration occurred.Three40-mmMacalloyhightensilesteelanchorbarswereinstalledthroughtheupstreamsideoftheworst-affectedblock6andintotheunderlyingrock;thesewerepost-tensionedtoproduceaminimumfactorofsafetyagainstoverturningof1.7forhydrostaticinternalupliftpressureof100% in the upstream face and zero on the downstream face (Figure 5.8). SinceCopyright 1992 Blackie and Son Ltddrillingequipmentwasavailableonsite,itwasdecidedalsotodrillpressure-relief drainage holes vertically at 3.4-m centres through the central blocks 1019tointerceptanyseepageflowandprovidesomecontroloverthedevelopmentofupliftpressures.A grouting trial was undertaken on block 6 to try to seal any ASR crackingontheupstreamsidetoinhibitwaterseepage.Althoughalow-viscosityoil-basedchemicalgroutwasused(Polythixon60/40DR),verylittlepenetrationwasachieved;itisnotclearwhetherthiswasbecauseofsomeinadequacyinthegroutingsystemorinsteadbecausetheconcreteinteriorcontainedfewercracksandmicrocracksthanhadbeenexpected.Aswellastheactualremedialmeasures,facilitieswereestablished,especiallyonblock6,forthefuturemonitoringofmovements(Table5.2).Thisinvolvedtheinstrumentalmonitoringofblock6,includingvibrating-wireanchorbarloadcells,electricallyoperatedpiezometersand24mechanical strain gauges across lift and vertical joints. However, as with mostconditionmonitoring,muchrelianceisstillplaceduponregularvisualinspections,supportedbycracksurveysandin-situsonicvelocitysurveys. Afurther44electricpiezometerswereinstalledonselectedliftjointsin1981,andlaterremovedandreinstalledin1983afterrecalibrationandtreatmenttoreduce creep effects, but a new design has proved less effective than the earlierinstruments in these conditions, and by 1988 only a third were giving reliablereadings9.Figure5.8DetailofthedownstreamelevationoftheValdelaMareDam,Jersey,showingtheanchorbarinstallationtoworst-affectedblock6.Thefailurepointreferstothe1977anchorbarfractureduetoseismicshock.FromRef.9.Copyright 1992 Blackie and Son Ltd5.4.5 Experience to date and the futureIt is now 14 years since monitoring of the Val de la Mare Dam commenced, andthefindingsgenerallyindicatethattherecanbelifeafterASRformassconcretegravitydams9.Concernhadarisenin1977,whenitwasdiscoveredthat one of the three anchor bars had fractured at a depth of 18.6 m below thetopofthedam.However,anintensiveinvestigationrevealedthatthefailurecouldonlyhavebeencausedbyahigh-velocitytensiledynamicload,probablyfromaseismicshock,andtheanchorwasnotreplaced.Thetworemaininganchorshavecontinuedtoperformsatisfactorily,andinfactactaslargeextensometerstogetherwiththevibrating-wireloadcells.Over7years(197582)theloadonanchor1increasedby2.5%and,evenaftercontrolleddestressingin1982,theloadhascontinuedslowlytoincreaseatarateofaround2.5kNperannum.Theflowsfromthepressure-reliefdrainsdrilledin1974have,withoneexception,remainedsmall,andencrustationwithcalciumcarbonatehasbeentheonlyproblem,necessitatingperiodiccleaning.Theexceptionwasadrainin block 12, which did develop unexpectedly high flows in 1980, as a result ofleakage through a particular lift joint. This was corrected by the installation ofanepoxy-sealedneoprenerubbergasketintothesuspectjointsontheupstream face of the dam.Therenowseemsnoreasontoquestiontheintegrityofthe ValdelaMareDamfortheforeseeablefuture,butinevitablyitremainsofspecialinterestasthefirstexampleof ASRrecognisedintheBritishIsles.ItseemslogicalthatsomeotherconcretestructuresonJerseymustalsobeaffectedby ASR,andalimitedsurveyhasbeencarriedout9,59.Concretesmadeusingthegranodioriteaggregatesource(withsporadicchalcedony/opalveins)exhibitedworseTable5.2ValdelaMareDam,Jersey:mainmonitoringactivitiesandfrequencies.AfterRef.9.Copyright 1992 Blackie and Son Ltdcrackingatallagesthanconcretemadewithonlybeachaggregates(Figure5.9); indeed only one example with beach aggregates was more than a case ofsuspected ASR. In the case of the granodiorite concretes, the water-retainingstructures, including Val de la Mare Dam, were more seriously affected than thedry concretes. Cole & Horswill9 claim that, for Jersey concretes, ASR damagedoes not worsen significantly after an age of 1520 years (the dam was 9 yearsoldwhenASRwasdiscoveredandisnow27yearsold),althoughfurtherdeteriorationcausedbyotherdestructiveprocesses,includingcorrosionofembeddedsteelinreinforcedstructures,remainsapossibility. Figure5.9SurveyofJerseyconcretes(1982,updated1987),showingthatASRdamageislargelyrestrictedtoconcretesmadeusinggranodioritecoarseaggregateandespeciallywater-retainingstructures.ThecrackingintensityscaleisdefinedbyRef.59.FromRef.9afterRef.59.Copyright 1992 Blackie and Son Ltd5.5 Fear of epidemic: the growth in mainland examples5.5.1 Plymouththe first examplesFor a while it was possible to regard the dam in Jersey as something of a specialcase:causedbyhithertounsuspectedopal-bearingsecondaryveinsintheotherwisewell-provenaggregatesourceandalsopossiblyrelatingtothequarrying of an unusually intensely veined body of rock and/or the use of a veryhigh-alkalicement.Between1976and1978,anysuchresidualcomplacencywas overcome by the discovery of ASR on the UK mainland, first in the south-westofEnglandandthenintheEnglishMidlandsandSouth Wales60.ThefirstUKmainlandexamplewasfoundatanelectricitysubstationatMilehouse in Plymouth, Devon, affecting unreinforced concrete bases set in theground with their upper surfaces exposed to the weather (Figure 5.10). A numberoforganisationswereassociatedwiththeinvestigationintoseverecrackingofthesebasesandtheidentificationofASR,includingQueenMaryCollege,theC&CA, Messrs Sandberg and Plymouth Polytechnic11,60. Before long, the CentralElectricityGeneratingBoardhadidentifiedanumberofsimilarinstallationsinthesouth-westandSouthWalesexhibitingbroadlysimilarrandomcrackingpatterns,concretecompositionsandexposureconditions.AtMilehouse,theconcretecomprisedcrushedlimestonecoarseaggregate,asea-dredgedfineaggregatecontainingasubordinateproportionofchertandalocallyproducedhigh-alkalicement(probablyaround1.01.2%Na2Oeq.).Reactionsiteswereobservedintheconcrete,involvingbrownandwhitechertparticles,withtheformationofreactionrimsandassociatedgeldeposits(Figure5.11). AlthoughFigure5.10ConcretebaseatMilehouseelectricitysubstation,Plymouth,showingmapcracking.FromRef.11.Copyright 1992 Blackie and Son Ltdsome other causes of cracking were apparent11, it has become generally acceptedthatASRwastheprincipalmechanismresponsibleforthedamage.Itisnowconsideredthattheparticularconditionstowhichtheseconcretebaseswereexposed,beingburiedinthegroundexceptfortheupperhorizontalsurface,whichwassubjecttowettinganddrying,wasanimportantfactorinthepromotionofASR,butsomeearlierinvestigationsconsideredthatstrayelectricalcurrentsmighthavebeeninvolved61.InPalmers1978review60,concreteassociatedwithtworeservoirsintheMidlands and another in South Wales were mentioned as examples of ASR, aswellastworatheratypicalcasesrespectivelyinvolvingeastcoastwartimeantitankdefencesandpossiblereactivitywithgreywackeaggregatesinScotland (see 5.6.4). However, it was probably the identification of ASR in theCharles Cross Multistorey Car Park in Plymouth which was mainly responsiblefor causing practising structural engineers in the UK to become acutely awareofthepotentialthreatof ASR. Thecarparkwasbuiltin1970,usingahelicalrampdesign,andwasexhibitingslighttoseriouscrackingoftheconcreteatanageofjust7years.Somebeamsexhibitedseveresurficialcracking,thepatternsofwhichwereinfluencedbytheunderlyingreinforcementandstressfields,andsomecolumnheadsexhibitedsubstantialcrackingandfragmentation(Figure5.12).Theground-levelretainingwalls,someparapetwallsandpavementsinthetoplevelexhibitedclassicmap-crackingpatterns.Laboratoryexaminationoftheconcreteagainrevealedthepresenceofreactivechertparticlesinasea-dredgedfineaggregate,togetherwithanon-reactive coarse aggregate (variously crushed limestone or crushed granite) andFigure5.11DetailoftheconcretefromacrackedbaseatMilehouseelectricitysubstation,Plymouth,showingreactivechertparticles.FromRef.11.Copyright 1992 Blackie and Son Ltda high-alkali Portland cement. The car park has been used for a thorough studyofconcreteconditionclassificationandcrackmapping62.Loadtestswerecarriedoutonvariousconcreteunitsatthecarparkin1981,1982,1985and198663.Theresultsconfirmedthatthebeamswerestillabletocarrytheirdesignload,andtheiractualserviceloadplusasignificantmarginsafely.Consequently,themainin-situstructuralframewasstrengthenedin1983andthecarparkhasremainedinservice,beingmonitoredandthestructuralconditionbeingperiodicallyreviewed.Thereactivecombinationofachert(orflint)-bearingfineaggregate,anon-reactive(oftencrushedlimestone)coarseaggregateandahighalkaliPortlandcementwasbecomingestablishedasthecauseofdamaging ASRintheUKand,withonlyafewexceptions,thishascontinuedtobethecaseasmoreandmoreexampleshavebeenidentified.Chertorflinthadalwaysbeenrecognisedasapotentiallyreactivematerial(see5.3.1)anditwasthecomparativelysmallproportionofchertinthetotalaggregatecombinationwhichappearstohaveapproachedthepessimumforchert,especiallyinthepresenceofhighalkaliconcentrations.Thecombinationforconcreteoflocalcrushedlimestonesandcrushedgraniteswithsea-Figure5.12CharlesCrossCarPark,Plymouth,showingaseverelycrackedcolumnhead.Copyright 1992 Blackie and Son Ltddredgedsands,importedfromfurthereastintheEnglishChannel,wasanewandunfortunatedevelopmentinthesouth-westofEnglandinthelate1960sorearly1970s,tomeetincreasedconstructiondemandinanareatraditionallyshortofgood-qualityland-basedsand.Itwasthedilutionoftheoverallchertorflintcontentoftheaggregatebythepresenceofnon-flintcoarseaggregate,asopposedtotheflintgravelwhichhadmoreusuallybeencombinedwithsuchflintsandsinsouthandsouth-easternEngland,whichcoincidentallycreatedthereactivemixture.Thereisnoreasontosupposethattheuseofsea-dredgedaggregates,insteadofland-basedaggregatesofsimilarmineralogicalcomposition,hadanyadverseinfluence on the concretes, except of course for any cases where inadequatewashingmighthaveleftaresidualcontentofseasalt.5.5.2GeographicalcontrolsApart from the Jersey ASR caused by opaline veination of an aggregate source andsome other possible cases involving greywacke aggregate (see 5.6.4), nearly allUK examples of ASR have involved chert or flint as the reactive constituent, andthenonlywhenpresentinsubordinatepessimumproportionsandalsointhepresenceofhighconcretealkalicontents(Table5.3).Sincemostconcretesaremade using locally derived aggregates and cements, it is therefore not surprisingthatcertainregionsofthecountryhavetendedtoyieldmostofthedefiniteorrelatively serious cases of ASR. The mixture of imported flint-bearing marine sandwithlocallimestonesandgranitesinsouth-westEnglandhasalreadybeenmentioned.InSouth Wales,somewhatsimilarly,sanddredgedfromtheBristolChannelcontainssmallproportionsofchert(probablyCarboniferous)andiscommonly blended with local crushed limestone aggregate. In the cases of bothsouth-west England and South Wales, at least some of the locally made Portlandcements were high-alkali in character. In the English Midlands, especially in theEastMidlands,concretesarefrequentlymadeusingtheabundantfluvioglacialsands and gravels commonly termed Trent Valley aggregates. These materials aredominated by quartzites, but both chert and flint are typically present, in varyingratios and proportions, in both the coarse and fine fractions. Again, at least one ofthePortlandcementsmadeintheregionwasahigh-alkalitype,and ASRhassometimes occurred when the local aggregates and cement have been used together.These sands and gravels are somewhat polymictic, and other types of potentiallyalkali-reactiveconstituentshavealsobeenrecognised,includingsomemetamorphic quartzites and glassy rhyolite, although it is uncertain to what extent,if any, such materials have acually caused or contributed to concrete damage.InotherpartsoftheUK,todate,localcombinationsofaggregateandcementhavenotprovedtobesusceptibleto ASR,andonlysporadiccasesofstructures possibly affected by ASR have been reported. However, it would notbesensibletoregardanygeographicalregionasbeingcompletelyfreefromCopyright 1992 Blackie and Son LtdTable5.3SomeexamplesofreactiveaggregatecombinationsidentifiedinUKconcretesaffectedbyASR.AdaptedfromRef.15.*Additionally,alkali-silicatereactionmightbeinvolved.Copyright 1992 Blackie and Son LtdCopyright 1992 Blackie and Son Ltdthe possibility of ASR and, in particular, the nationwide distribution and use ofprecastconcreteunitsmustalwaysbeborneinmind.Aswellasthepresenceofareactiveaggregatecombinationandofasufficient amount of alkalis, a supply of moisture is necessary for damaging ASRtooccurinconcrete.ThedegreeandfrequencyofwettinghasanimportantenvironmentalcontroloverASR,anditisbelievedthatcyclicwettinganddryingmaybethemostconducivecondition14.Consequentlytheincidenceandseverity of ASR may sometimes depend upon geographical aspect; for example,Fookesetal.62foundasympatheticrelationshipbetweenexposureanddeteriorationattheCharlesCrossCarParkinPlymouth(Figure5.13).Insomecircumstances,regularmoisturemigrationmightcauselocalisedconcentrationsof alkalis, so enabling ASR to occur in affected parts of a structure64.5.5.3 The spread of discoveries and allegationsIt is not possible to state categorically the number of concrete structures thatareaffectedby ASRintheUK,muchlesstoprovideanydetailedschedule,because of the persistent climate of secrecy that has bedevilled the growth inawarenessofASRinthiscountry.Atfirstitwashopedthattheexamplesidentifiedinthemid-1970s,suchastheelectricitysubstationcases,wouldprove to be special cases and that ASR would continue to be an extremelyFigure5.13DiagramrelatingthedegreeofexposuretothedegreeofcrackingatCharlesCrossCarPark,Plymouth,showingthatexposurehelpsdeteriorationtooccur.ThescalesfordegreesofexposureandcrackingaredefinedinRef.62.FromRef.62.Copyright 1992 Blackie and Son LtdrareoccurrenceintheBritishIsles38.Unfortunately,thenumberofcasesofASRconfirmedorclaimedhasprogressivelyrisen;the1982BREDigest25834 stated that about 30 cases had been identified in addition to the dam inJerseyandthePlymouthelectricitysubstationbases,butby1988theDigest33065,whichsupersededDigest258,admittedtomorethan170cases.Amajorityoftheseaffectedexamplesappeartohavebeencivilengineeringstructures, rather than buildings, but this might be because such constructionis usually subject to a more rigorous programme of inspection and conditionassessment.Nevertheless,itdoesappearthattheincidenceofdamagediagnosedasbeingcausedbyASRintheUKisverysmallwhencomparedwiththetotalamount of concrete construction carried out13. One major study of a sample ofthenearly6000highwaybridgesforwhichtheDepartmentofTransportisresponsibleinEnglandhasapparentlyidentifiedsignsofASRinbetweenahalfandathirdofthestructuresexamined66.However,itisbelievedthatnomorethanabout10%weredamagedtosomeextentbyASR;moreoverhighwaystructuresareliabletobeparticularlyvulnerableto ASRbecauseoftheconditionstowhichtheyareexposed.Itisalsoimportanttodistinguishbetweenthosefewcasesof ASRinwhichstructuraldamagehasresultedandthosemorenumerouscaseswhereanysurficialdamageisonlycosmetic,orwhichatworstcouldprovideafacilityforlaterdeteriorationbyotheragencies, such as frost, if left untreated.Itisalsonowrecognised14thatmicroscopicevidenceofthepresenceofASR in concrete is not necessarily adequate proof that any damage has beencausedorevenaccentuatedbyASR,sothatmanystatisticsregardingthescaleoftheeffectof ASRintheUK,orindeedelsewhere,mustbetreatedwith some caution. In an admirably thorough review of his observations overadecade,forexample,French67reportedsome40%of300concretestructures to be exhibiting evidence of ASR, but these reactions ranged fromjustdetectabletoseriousdestructiveprocessesandheprovidednoclearindicationoftheproportionconsideredtoexhibitthelatterdegreeofseverity.Thereisnoeasyandunequivocalmeansofquantifyingsuchdegreesofdamagecausedby ASRand,inanycase,itislikelytobetime-dependent;aminorcaseofASRidentifiedtodaycouldpossiblydevelopinto a more serious case over a period of time. In one appraisal of a group of93UKstructures,usingacoreexpansiontestasameasureofseverity,only32%exhibitedexpansionsgreaterthanabout0.10%andjust6%exhibitedexpansionsgreaterthan0.15%(Figure5.14)15.IthasbeennoticedthattheconcretestructuresfoundtobeaffectedbyASR in the UK were built during a period from the 1930s to 1970s, but thenumberofaffectedstructuresappearstoshowapreponderanceinthe15years1960to1975,whichinterestinglypost-datestheBRSresearchworkof the 1950s (Figure 5.15). Some authors have suggested that this apparentupsurgein ASRintheUKmightbeattributed,inpartatleast,tochangesCopyright 1992 Blackie and Son LtdinthecompositionandnatureofPortlandcementbeingmanufacturedintheUK68.Changestothecementmanufacturingprocessandgreatercontrolsovertheemissionsintotheatmospherefromcementworksmighthavecausedthealkalicontentstorise,andmineralogicalmodificationstoachievegreaterratesofstrengthgainmighthaveledtotheformationofaFigure5.15DiagramshowingthefrequencyoftheidentifiedoccurrenceofASRinconstructionofUKstructuresbetween1920and1985.Thehistogramisnotweightedtoreflectthevariationsinintensityofconstruction.FromRef.15.Figure5.14Expansionofconcretecorestakenfrom93UKstructuresaffectedbyASR.FromRef.15.Copyright 1992 Blackie and Son Ltdmorepermeablehydratedproduct.Thereseemstobelittlescientificevidencetosupporttheseallegationsand,althoughalkalicontentrecordsbefore1960arenotexhaustive,theredonotappeartohavebeenanysubstantialchangesineithertherangeortheweightedaveragealkalicontentofUKcements69. AmoreprobablereasonfortheapparentupsurgeofASRbetween1960and1975isthatthisperiodalsocorrespondstoamajorgrowthperiodinUKconstruction,mostnotablytherapidexpansionofthemotorwaysystem.Also,likeanyothernewphenomenontobediscovered, itmightbeassumedthatsomeearlierexamplesweremisdiagnosedandeithersuccessfullyrepairedordemolished.WhateverthenumberofaffectedUKstructuresmaybe,veryfewcasestudies have been published, even in part form, and most information remainsconfidentialoronrestrictedaccess.Someofthebest-publicisedexamplesarestructuresalongtheA38trunkroadinDevon,includingtheMarshMillsViaduct on the outskirts of Plymouth, an underbridge at Plympton and the VossFarmbridgefurthereast. TheMarshMills Viaduct,whichwasbuiltin196970, consists of two separate three-lane carriageways, each supported on a seriesofT-shapedpiers.ThedeckbeamswereprecastinEast AngliaandwerenotaffectedbyASR70,butthepiers,cross-beams,deckslabandpilecapswerefoundtobedamagedtovaryingextentsbyASR70.AsforothercasesinthePlymouth area, the reactive combination consisted of crushed limestone coarseaggregate,asea-dredgedsandcontainingasubordinateproportionofchertandthelocalhigh-alkaliPortlandcementfromPlymstockworks.Areasofconcretesubjectedtoseverewettingweremostseriouslyaffected,includingthe cross-beam ends beneath deck joints but most notably the buried pile caps,whichwereperpetuallywetinthemarshyground(Figure5.16).Loadinghasbeen reduced by permanently reducing the traffic flow from three to two lanesoneachcarriageway,andsteelstanchionshavebeeninstalledbeneaththecantileverheadsofthepiers(Figure5.17),partlytoreducebendingmomentsbutalsotoprovidemoredirectloadtransferenceontothepilestolimitshearstresses in the badly cracked pile caps. It is understood that these measures areintended to prolong the serviceability of the structure for up to 10 years, whenreplacementisplanned.TheVossFarmoverbridge,alsobuiltaround1970,wasfoundtoexhibitseverecrackingattributabletoASRinthecolumns,theabutmentsandwingwallsandthefoundations(Figure5.18).Theconcretemixwassimilartothatfound to be reactive elsewhere in the region, with a minor proportion of chertbeing the active constituent. The circular columns were heavily reinforced andthecrackpatternwaswhollyorientatedaccordingtothestressfield,sothatonlyverticalcrackswereapparentwithnotraceofthemapcrackingoftenconsideredtypicalofASR(Figure5.19).Correctiveactionincludedthereplacementofthecolumns,andthegroundanchoringandrefacingoftheabutments and wing walls.Copyright 1992 Blackie and Son LtdOfthecasesof ASRintheEnglishMidlands,probablytheSmorrallLaneoverbridgeontheM6neartotheCorleyserviceshasbeenmostwidelyreported63.Thetwo-spanbridge,whichwasbuiltin1969,exhibitedsomeseverecrackingofthecentralpier-topbeam,thedeckbeamsandthesoffitofthe deck slab; cracking patterns on the beams were reminiscent of flexural andFigure5.17ApieroftheMarshMillsViaduct,nearPlymouth,showingthesteelstanchionsbeinginstalledasameasuretosupportthecantileverheadsandmodifyloadtransference.Figure5.16MarshMillsViaduct,nearPlymouth,showinganexcavationtorevealaseverelycrackedconcretepilecap,largelybelowthelocalwatertable.Copyright 1992 Blackie and Son LtdFigure5.18VossFarmOverbridge,Devon,showingaseverelycrackedwingwall.Figure5.19VossFarmOverbridge,Devon,showingcrackinginaheavilyreinforcedcircularcolumn(steelstrapsinstalledastemporaryprecaution).Photographtakenafterremovalofthecolumn,showingthedevelopmentofsomehorizontalcrackinginadditiontothedominantverticalcracking.PhotographbycourtesyofG.V.WaltersofECCQuarriesLimited,Exeter,Devon.Copyright 1992 Blackie and Son Ltdshearcracking(Figure5.20).PetrographicexaminationsestablishedthepresenceofASR(Figure5.21)71,withthereactiveconstituentbeingsubordinatechertpresentinboththecoarseandfineaggregateswhichwereotherwisedominatedbyquartzite.Structuralengineersconsideredthatthetwosimplysupportedspanscouldbecomeanunpredictablehazardwithin5years,63and,consequently,theconcretedeckswerereplacedbynewsteelconstructions;theoldconcretebeamswereremovedandretainedforcontinuedmonitoringandfurthertesting.Anotherexampleof ASRintheMidlandshasbeendescribedindetailbyNixonandGillson72.In1982,crackingofconcretebasesatanelectricitysubstation were once again reported, this time at Drakelow Power Station, nearBurton-on-Trent.Themainreactiveconstituentwasthoughttobechert,present at around 3% of the otherwise sandstone and quartzite sand and gravelaggregate,althoughsomeparticlesofrhyolitemayalsohavereacted73.Therewereindicationsthattheconcretealkalicontents,whichwerenotoriginallyunduly excessive (see 5.6), might have been enhanced by alkalis derived fromtheaggregatesandalsobealkalimigrationintotheupperpartsoftheslabscausedbymoisturemovements.ThisreviewofknowncasesofASRintheUK,aswithearliersuchreviews60,74,isincomplete,unavoidablyeclecticanddestinedrapidlytobecomeoutdated.Identifiedcasesinsouth-westEnglandnowincludefoot,roadandrailbridges,carparks,office,shop,collegeandhospitalbuildings,andconcreteblocksatsubstations,TVtransmittersandmicrowavetowers.Jetty,bridge,riverwall,reservoirandelectricitysubstationstructureshaveallFigue5.20SmorrallLaneOverbridge,Warwickshire,showingregulartensile-stylecrackingofdeckbeamsacrosstheeastboundcarriageway.FromRef.71.Copyright 1992 Blackie and Son LtdbeenfoundtobeaffectedinSouthWales.IntheMidlands,identifiedcaseshaveincludedfootandroadbridges,motorwayjunctioncomplexes,reservoirstructures, colliery shafts and electricity substation bases. Sporadic or specialcases(see5.5.4)havebeenreportedfromotherareasand,regrettably,manyothercasesprobablyremaintobediscovered.However,itisimportanttorememberthattheseidentifiedcasesstillrepresentasmallproportionofthenationstotalstockofconcretestructuresandalsothattheseverityofanydamageisfrequentlynotcritical.5.5.4 SpecialcasesApartfromthe ValdelaMareDaminJersey,UKcasesof ASRhavealmostexclusivelyinvolvedchertasthereactiveconstituent,andthephysicalmanifestationsofreactionhaveusuallybeenrestrictedtosurficialcracking,somesurfacediscolorationsand,occasionally,smalldisplacementsoftheconcrete.Somespecialexceptionsaredescribedinthissection,withthequestion of artificial aggregates being considered separately in the succeedingsection (5.5.5).As early as 1978, Palmer60 described a dam in Scotland, built in 1936, whichexhibited evidence of an alkali-aggregate reaction involving a greywacke aggregatederivedfromanexcavationonsiteandapparentlynotusedforanyotherconstruction.Itappearedthatthereactionhadnotcausedanydisruptionanditwas considered that the reaction had proceeded to completion, so that no remedialaction was required. More recently, a number of structures in Wales and the northFigure5.21PhotomicrographofconcretefromSmorrallLaneOverbridgedeckslab,showingareactioncentreinvolvingchert.Pln.Pols,(approx.15magnification).FromRef.71.Copyright 1992 Blackie and Son Ltdof England, also made using greywacke or similar aggregates, have been identifiedas exhibiting alkali-aggregate reactivity. The Maentwrog Dam, near Dolgellau inNorth Wales, which was completed in 1927, exhibits extensive cracking and heavywhite exudations on the downstream face and has apparently experienced severalphases of repair during its lifetime. A so-called alkali-greywacke reaction hasbeen observed in the concrete under the microscope14,75, which is at least in part aconventional ASR involving fine-grained quartz in the rock matrix. At present it isuncertain to what extent, if at all, this example might represent a type of alkali-silicate reaction, similar to that originally described in Canada76. It is also unclearwhether the alkali-greywacke reaction could have been a significant cause of thesurface cracking, or the continual leakage through the dam which occurs mainlyat construction joints. Construction work has now started to replace the old dam atMaentwrog with a new dam structure a short distance downstream.In other parts of the world, surface pop-outs have sometimes been a commonsymptom of ASR, but this has rarely been the case in the UK, where neither pop-Figure5.22UnusualUKexampleofASRinwhichsurfacegelexudationsandpop-outsaretheonlydetrimentalfeatures.Chertparticlesreactinasurfacezonewhichisgreatlyenrichedinalkalis.FromRef.77.Copyright 1992 Blackie and Son LtdoutsnorcopiousgelexudationsaretypicalfeaturesofconcretesurfacesonstructuresdiagnosedasbeingaffectedbyASR14.SimsandSotiropoulos77,however,haverecentlydescribedsomespecialcasesinwhichsurfacepop-outs and associated gel exudations are the only defects (Figure 5.22), with nocracking, expansion or other damage being apparent. In these cases, of whichseveraldifferentlocationswereinvestigated,thereactiveparticlescomprisedchert which formed only a very minor percentage of an otherwise wholly ooliticlimestoneaggregate.LocalisedASRwasabletooccurbecauseoftheverysubstantialconcentrationofalkalisthatwasbroughtaboutbythecontinuousmigrationofmoisturethoughbasementretainingwalls(Table5.4),thebacksidesofwhichwereindirectcontactwithwetgroundandthefrontsidesofwhich were exposed to an internal drying environment. The alkali content wasfoundtobeenhancedbyuptofivetimesbythisprocess,anditwasthoughtpossiblethatsealingorpaintingofthefrontconcretesurfacemighthaveexacerbated the effect.5.5.5ArtificialaggregatesThepossibilitythatsomeartificialaggregatesinuseintheUKmightprovetobealkali-reactivewasfirstconsideredbyJonesandTarleton6,7,whoincludedsomeair-cooledblastfurnaceslagintheirBRSsurvey.JonesandTarletonalsoutilisedPyrexglassasaknownreactivematerial,andglasshaslongbeenrecognisedasbeingpotentiallyreactivewithPortlandcement;glassfibrereinforcementofconcretecanonlybesuccessfulusingalkali-resistant glass. Figg78 has described a case in which decorative glass aggregateinconcretecladdingpanelsforabuildinginnorth-eastLondoncaused ASR,resultingincracking,distortionandspallingoftheunits.Althoughthenon-structural panels had been made using a low-alkali white Portland cement, theTable5.4Theconcentrationofalkalis(Na2Oeq.,kg/m3)intheinternalsurfacezoneofconcreteretainingwallsastheresultofthecontinuousmigrationofmoisture;thesurfacezoneisgreatlyenhancedinalkalis,whilstthecentreofthewallisprobablydepletedinalkalis.AfterRef.77.*Approximately050mmincaseoflocationAand025mminthecaseoflocationB.Copyright 1992 Blackie and Son Ltdglassaggregateitselfhadcontainedsufficientreservesofavailablealkalistoperpetuatetheexpansivereaction.Althoughmanyoftheaggregatesthataremanufacturedorderivedfromindustrialwasteproductsarebothsiliceousandpoorlycrystallineorevenglassy,thereappeartohavebeennoreportsofanyconcretestructuresintheUKbeingdamagedastheresultof ASRinvolvingartificialaggregates,otherthanthedecorativeglassalreadymentioned.TherehavealsobeenveryfewpublishedattemptstoassesstheASRpotentialofsuchartificialaggregatematerials.SimsandBladon79carriedoutalimitedprogrammetoinvestigatethe ASR potential of sintered pulverised fuel ash (PFA), which is widely usedasalightweightaggregate(Lytag).Nosignificantexpansionofmortarbarandsmallconcretespecimenswasrecorded,andyetmicroscopicalexaminationoftheconcreteatvariousagessuggestedthatsomegelproductionmighthaveoccurred,impregnatingthecementpastematrixbutseeminglynotgivingrisetoexpansiveforces.Itisalsopossiblethatthevoided nature of such lightweight materials might allow a considerable degreeofASRtobeaccommodatedwithoutanydisruptivestressesbeinggeneratedwithinthemortarorconcrete.FurtherresearchisneededbeforeartificialaggregatescanbesafelyregardedasbeingincapableofexpansiveASRinconcrete over a prolonged period of service in conducive exposure conditions.5.5.6 Urgent search for remediesAs the number of confirmed cases of ASR built up in the UK and it became clearthat the problem was no longer restricted to foreign or special case occurrences,it became urgent to identify the main causal factors and so to formulate measuresfor avoiding the reaction in future construction. At the initiative of the C&CA, anInter-Industry Meeting was held on 10 June 1981, at which it was agreed to inviteMichael Hawkins, as the engineer of the county most afflicted with cases of ASR,to form an independent working party of specialists to co-operate in the productionof some guidance notes for the practising engineer. The first report of the HawkinsWorking Party was published in 198312 (see 5.6.1) and later updated, as a ConcreteSociety technical report, in 198713 (see 5.6.2). BRE Digest 25834 was publishedshortly before the first Hawkins Report and heralded a number of the conceptsshortlytobeendorsedbytheHawkins WorkingParty,includingthenamingofaggregategroupclassifications(tradegroups)consideredveryunlikelytobesusceptible to attack by cement alkalis and the notion of using a threshold limiton concrete alkali content of 3 kg/m3 as an alternative to the use of low-alkalicement to avoid ASR in new construction.Cement alkalis could be established by analysis and were, in any case, usuallyobtainableuponrequestfromthemanufacturer.Bycontrast,BRSDigest258maintained that, There are as yet no British Standard tests for the susceptibility ofaggregates to attack by alkalis and ASTM tests have not been found to predict theCopyright 1992 Blackie and Son Ltdreactivity of UK aggregates accurately. This statement is significant in view of thereliance that had previously been placed on the mortar bar test by the BRE/BRS inthe1950sduringtheirpreviousinvestigationsintotheASRpotentialofBritishaggregates. Appropriately,thebatonwastakenupbytheBSIin1981,whentheaggregates committee, CAB/2, formed a new working group, WG10, to evaluatemethods for assessing the ASR potential of aggregates (see 5.6.4).5.6 Countermeasures: minimising the future risk of ASR5.6.1 Hawkins Working PartyITheHawkinsReportwaspublishedin198312andemphasisedthattheincidence of ASR in the UK was small, but the importance of minimising therisk of ASR in future work was embodied by the warning that, When ASR hasbeendiagnosedinastructure,therearenomethodswhichcanbereliablyrecommendedatpresentforeitherpreventingfurtherdamageorcarryingouteffectiveandlastingrepairs.Thereportwaspresentedaspracticalguidance,basedoncontemporaryinformationandintendedonlytoapplytomaterials,conditionsandpracticeencounteredintheUK.ItwasconsideredthatASRcausesdamageonlyifthethreefollowingfactorsareallpresent:sufficientmoisture,sufficientalkaliandacriticalamountofreactivesilicaintheaggregate.TherecommendationsgivenintheHawkins guidance notes were thus based on ensuring that at least one of thesefactorswasabsent.TheriskofASRforaparticularconcreteinaparticularenvironment could be assessed by reference to these factors (Figure 5.23) and,ifariskwasdeemedtherebytoexist,thesamefactorsprovidedarangeofpossibleoptionalprecautionstobeimplemented.Twolevelsofprecautionwereenvisaged:alevelofnormalconcreteconstructionwhichitwassupposedwouldapplyinthemajorityofcases,andalevelforparticularlyvulnerableformsofconstruction,suchasconcreteburiedinwater-loggedgroundwiththetopsurfaceexposed(e.g.electricitysubstationbases).A sufficiency of moisture for damaging ASR to occur was considered to beanyexposuretomoisture,includingrelativehumiditylevelsexceeding75%.Inpracticethismeantanyexternallyexposedconcretewasinanenvironmentdeemedconduciveto ASR.The amount of reactive alkali available was considered in three main waysaccordingto:(a)thealkalicontentofthecement,(b)thealkalicontentofablend of cement with a certain minimum content of a mineral addition such asground granulated blast furnace slag (GGBFS) or pulverised fuel ash (PFA), or(c)thealkalicontentoftheconcrete.Inrespectof(a)above,theHawkinsReportstated,TheuseofPortlandcementwithanalkalicontentof0.6%orless is accepted worldwide as the best means of minimising the risk of damagedue to ASR. Only sulphate-resisting Portland cement (SRPC) was available inCopyright 1992 Blackie and Son LtdtheUKwithaguaranteedalkalicontentofnotmorethan0.6%,butordinaryPortlandcement(OPC)couldbefoundtobelow-alkalibyexaminationofmanufacturersanalyticaldatamakingdueallowanceforvariations,orberenderedlow-alkalibythecontrolledreplacementofpartofthecementbyGGBFSorPFA,whichwereassumedtocontributenoreactivealkalitotheconcrete.Inrespectof(b)above,acombinationofGGBFSandanyUKPortlandcementwhichcontained50%ormoreGGBFSwasconsideredequivalenttoa low-alkali cement. Similarly, a combination of PFA and any UK Portlandcementwhichcontained25%ormorePFAwasconsideredequivalenttoalow-alkalicement,withtheaddedconditionthatthealkalicontentoftheconcreteprovidedbythePortlandcementinthecombinationwasnotmorethan3.0kg/m3.Finally,inrespectof(c)above,itwasconsideredthatdamagefrom ASRwasunlikelytooccuriftheconcretehadanalkalicontentof3.0kg/m3orless,whichwascalculatedfromthealkalicontentofthecementandthemaximumexpectedcementcontentoftheconcrete,makingdueallowancesforfuturevariationsincementalkalicontentsandconcretebatchweights.Thisapproachwasnotconsideredadequateforparticularlyvulnerableconstruction.Thevalueof3.0kg/m3,whichwascoincidentallyequivalentFigure5.23SimplifiedflowchartfortheassessmentofrisksandprecautionsagainstASR,accordingtotheguidancegivenbytheHawkinsWorkingParty.SeeConcreteSocietyTechnicalReportNo.30,198713,forfulldetails.RedrawnandadaptedfromRefs.12and13.Copyright 1992 Blackie and Son Ltdtousingahigh-alkalicementofsay1.0%Na2Oeq.inatypicalstructuralconcrete of say 300 kg/m3 cement content, seems largely to have been basedupon research by Hobbs at the C&CA80 (Figure 5.24), using Beltane opal andmortarbars,althoughitwasconsideredthattheresultsagreedwithearlierfindingsinGermanyandelsewhere81,82.The1983HawkinsReportwasnotabletobeparticularlyhelpfulintheguidanceonaggregates.Itwasstatedthat,ThereareatpresentnoBritishStandardtestsforthealkali-reactivityofaggregates,buttheAmerican(ASTM)chemicalandmortarbarmethods19,21commonlyusedinothercountrieswereconsiderednottobeapracticalbasisforthespecificationofBritishconcretematerials.ModifyingthegroupclassificationapproachputforwardinBREDigest25834,theHawkinsReportstatedthat,Atpresent,itseemsthataggregatesconsistingwhollyofrocktypesintheBasalt,Gabbro,Granite,Gritstone,Hornfels,Limestone,PorphyryandSchisttradegroups,would be classified in the UK as unlikely to be reactive. The word whollywas explained as meaning both coarse and fine aggregates, and that the sourcewas not contaminated by reactive silica (such as the secondary opal veining inthegranodioritesourceinJersey).ConsiderationofthecompositionofanaggregatecouldthereforeenableanaggregatetobeadjudgedunlikelytobeFigure5.24Relationshipbetweentheexpansionofmortarbarsat20Candtheacid-solublealkalicontentofthemortar,usingopalasthereactiveconstituent,showingnoexpansionatalkalilevelslowerthan3.5kg/m3.AfterRef.82.Copyright 1992 Blackie and Son Ltdreactiveorcontainsconstituentswhicharesometimesfoundtobereactive,but,inassessingtheamountofreactivesilicainthelattercase,theHawkinsReportwarnedthat,thecriticalproportionisnotcapableofeasypredictionandinanycasewillvarywiththeaggregatecombination.In addition to the three main controlling factors (moisture, alkali contents andreactive aggregate), the possibilities of alkali migration through moisture movementandthecontributionofalkalisfromsourcesotherthancement(fromotherconstituentsorfromexternalsources)wererecognised,althoughnoadditionalprecautions could be recommended, save by tanking the concrete or by the use ofdefinitely non-reactive aggregates. The working party was also anxious to ensurethat zealous efforts to achieve the avoidance of ASR did not lead to the neglect, oreven act to the detriment, of other considerations: ItisessentialthatadjustmentswhichmaybemadetothemixinordertoavoidASRshouldneverleadtotheuseofmaterials,cementcontentsorwater/cementratioswhichwillbeinadequatetoensuregeneraldurability. ThefirstHawkinsReportwaswellreceivedintheindustry,butitwasonlyguidanceand,intheabsenceofaBSspecification,itwasclearthatawiderange of approaches to minimising the risk of ASR would arise from differentspecifiers,includingextremecaseswherealloftheoptionalprecautionswould be required to apply instead of just one of the options chosen accordingtocircumstances.TheHawkinsWorkingPartythereforeremainedinbeing,nowaffiliatedtotheConcreteSociety,andcommencedworkonmodelspecificationclauses.5.6.2 Hawkins Working PartyIIThesecondsubstantiveeditionoftheHawkinsReportwaspublishedasConcreteSocietyTechnicalReportNo.30in1987,althoughanearlierconsultationdocumenthadbeenpublishedin1985.TheneweditionwasstyledGuidanceNotesandModelSpecificationClausesandtheopportunityhad been taken to revise and update the guidance notes, which were otherwisebased upon the same principles and followed the same general approach as theoriginalHawkinsReport.Inhisforewordtothenewedition,MichaelHawkinsreaffirmedthatdamagebyASRaffectsonlyasmallproportionofconcreteconstructionundertaken in the UK, but added, Nevertheless, when it does occur the cost ofremedialworkcaninsomecasesbeveryhigh.ThereportstillacceptedthatASR tended to occur most frequently in certain regions, but warned that, Theincreasingtransportationoverlongdistancesofbothmaterialsandconcreteproductsmeansthatthepotentialfordamagefrom ASRwillnotberestrictedto particular regions. For this reason, and others, the simple UK map depictingCopyright 1992 Blackie and Son LtdthelocationsofreportedcasesofASRwhichhadbeenincludedinthefirsteditionwasomittedfromtherevision.Themodelspecificationclauseswerebased upon the revised guidance notes and will not be separately described inthissummary,exceptwhereconvenienttoclarifytheadvice.Themainrevisionsintheguidancenotesconcernedthecontributionsofreactive alkalis from GGBFS and PFA, the derivation of additional alkalis fromsodiumchlorideandothersources,andsomechangesandimprovementstotheguidanceaboutaggregates.InthefirstHawkinsReport,theGGBFSandPFA components were, for the purposes of that report, considered to contributeno reactive alkalis to the concrete system. In the new report, the reactive alkalicontentofGGBFSandPFAisdefinedasbeingthewater-solublealkalicontent, when determined using the BS 81226 extraction method for aggregatesandtheBS455025analysismethodforcements.Inconsequence,whentheconcretealkalicontentisbeingcalculated,theGGBFSorPFAreactivealkalicontentsmustbetakenintoaccount(Figure5.25).Thisapproachhasbeendisputed,andsomeauthoritiesmaintainthatthereactivealkalicontentoftheseadditionsshouldbebaseduponanotionalproportionoftheacid(ratherthan water) soluble alkali determination: 50% in the case of GGBFS and 17%(or one-sixth) in the case of PFA65. The Hawkins Working Party has establishedatechnicalsubcommitteetocarryoutfurtherresearchintothiscontroversyand,untilthefindingsbecomeknownandareaccepted,opinionswillremaindividedastothebestapproach.In the first report, it had been recognised that sodium chloride in sea dredgedand some other aggregates might be a source of additional alkalis, but it was onlySitecombinationsofPortlandcementwitheithergroundgranulatedblastfurnaceslag(GGBFS)orPFAWhencementtoBS12iscombinedonsitewitheitherGGBFSorPFAthereactivealkalicontentoftheconcreteiscalculatedfrom: where A=reactivealkalicontentofconcrete(kg/m3)C=targetmeanPortlandcementcontentofconcrete(kg/m3)a=reactivealkalicontent(%)ofcementE=targetmeancontentofeitherGGBFSorPFAintheconcrete(kg/m3)d=averagereactivealkalicontent(%)ofeithertheGGBFSorthePFAasprovidedbythemanufacturers.Figure5.25CalculationofthereactivealkalicontentofconcreteinaccordancewiththeConcreteSocietyTR30guidance,oneofseveralequationoptionsgiveninthereport.AfterRef.13.Copyright 1992 Blackie and Son Ltdsuggested that the CP 11030 limits on chloride for reinforced concrete should beapplied. By the time of the revised edition, it had become clear from experiencesin Denmark and from work at the BRE83 that the addition of sodium chloride to theconcreteincreasedthehydroxylionconcentrationandcouldpromotegreaterASR expansion. Consequently, provision was made in the calculation of concretereactivealkalicontentfortheadditionalsodiumassociatedwithanychloridepresentinthecoarseandfineaggregates.Norecommendationsweremadeinrespect of any reactive alkalis which could possibly be released from some rocktypes found in aggregates, although the opinion was expressed that, No significantamountsofreactivealkaliwillbederived(chloridesexcepted)fromnaturalaggregates in the UK.Theguidanceonaggregateswasamendedintwoimportantways:theclassificationofaggregatesunlikelytobereactivewasimprovedandextended, and a new concept was introduced regarding aggregates comprisinglargely flint or chert. The list of rocks and minerals considered unlikely to bereactivewasrevised(Table5.5)accordingtotheupdatedBS812:Part102(1984),inwhichthetradegroupshadbeenreplacedbyamorestraightforwardlistofcommonrocktypesfoundinaggregates,andanappendixwasincludedtoprovidesomeguidanceonthoserocktypeswhichhadnotbeenincludedonthattabulatedlist.Feldsparand(unstrained)quartzwereincludedonthelistofconstituentsunlikelytobereactive,enablingmanyfineaggregatestobeassessedmorereadily.Adilemmawaspresentedbychertandflint,becausealthoughtheseconstituents have been involved with most of the UK examples of ASR, they areextremelycommoninUKaggregates,andinthehighlybuilt-upsouth-eastofEnglandtheyformthemajorpartofmostaggregatecombinations.TherevisedTable5.5Rocksandmineralsunlikelytobereactive,accordingtotheguidancenotesgiveninConcreteSocietyTechnicalReportNo30,198713.ThelistofrocknamesisbaseduponthatgiveninBS812:Part102(1984)26.*Feldsparandquartzarenotrocktypesbutarediscretemineralgrainsoccurringprincipallyinfineaggregates.Nothighlystrainedquartzandnotquartzite.Chalkisincludedinthelistsinceitmayoccasionallybeaminorconstituentofconcreteaggregates.Copyright 1992 Blackie and Son LtdHawkinsReportadvisedthat, Experiencetodatefromcasesof ASRinvolvingsands and gravels indicates that up to a maximum of 5% by mass of flint, chertorchalcedonytakentogetherinthecombinedaggregatecanbetolerated.However,attheotherendoftherange,Itiscurrentlyconsideredthatacombinationoffineandcoarseaggregatewhichcontainsmorethan60%bymass of flint or chert is unlikely to cause damage due to ASR. In other words, intheabsenceofanyotherpotentiallyreactiveconstituent,aggregateseithercontainingnomorethan5%oralternativelycontainingmorethan60%chertandflintaretoberegardedasunlikelytobereactive.Thevalueof60%wasbasedpartlyuponobservation,wherebyconcretesmadewithpredominantlyflintaggregateswerenotfoundtobeaffectedbydamagingASR,andpartlyupon limited laboratory experiments at the BRE, the results of which have sincebeen confirmed by further work at the BRE and elsewhere (Figure 5.26).Nodetectableopalispermittedforaggregatesunlikelytobereactiveanditisimportanttorealisethatopalisconsideredbythereporttobeunacceptableunderanycircumstances: Sources(ofaggregate)knowntocontainopalinesilicashouldnotbeusedevenwhenthealkalicontentofeitherthecementitiousmaterialortheconcreteisbeingcontrolled. This is because there is some evidence that small amounts of opal in concretecan cause ASR damage even when the cement alkali content is less than 0.6%ortheconcretealkalicontentislessthan3.0kg/m3.Nospecificguidanceisgiveninrespectofalkalisderivedfromanysourcesother than the cementitious or aggregate materials, but the report requires that, Ifalkalisinexcessof0.2kg/m3ofconcretecomefromothersourcestheymustbetakenintoaccount.Ineffectthismeansthatmanyengineerswillrequireexhaustive alkali audits to be carried out for all the concrete mixes proposed for agivenproject.5.6.3 The Department of Transport and other specifiersBS811031supersededCP11030in1985astheCodeofPracticeforStructuralUseofConcrete.WhereasCP110hadnotmentioned ASR,BS8110(inclause6.2.5.4)includedrecommendationsthatfollowedthefirstHawkins Report12 and BRE Digest 25834. Other national authorities maintainconcretespecificationswhicharewidelyrecognisedandrespected,includinginteraliathePropertyServicesAgency(PSA),BritishRail,theCentralElectricityGeneratingBoard(CEGB)andtheBritishAirportsAuthority(BAA),butperhapsthemostimportantinthiscontextistheDepartment of Transport (DTp). Concrete and concrete materials are includedin the DTpSpecificationforHighway Works84, whichwaspublishedin1986tosupersedetheearlierSpecificationforRoadandBridgeWorks85. Copyright 1992 Blackie and Son LtdThepublishedSpecificationforRoadandBridge Worksdidnotcontainany reference to ASR. However, in July 1982 a new Clause 1618 was issuedbytheDTptoitsregionalofficestoamendthespecificationtoincludeMeasurestocontrolalkali-aggregatereactioninrespectofallstructuralconcrete86.TheDTpamendmentwasgenerallysimilartotheguidancewhichappearedintheBREDigest268andthefirstHawkinsReport.Thetotalalkalicontentderivingfromcementwasnotpermittedtoexceed3kg/m3ofconcreteunlessboththecoarseandfineaggregateswerenotFigure5.26Somerelationshipsbetweenflint/chertcontentofaggregateandconcreteprismexpansion,indicatingaflint/chertpessimumof1030%forflint-limestonecombinationsandnosignificantexpansionsformixturescontainingmorethan60%flint.RedrawnandadaptedfromRefs.103and104.Copyright 1992 Blackie and Son Ltdsusceptibletoalkali-aggregatereactionaccordingtocriteriastipulated.Low-alkali SRPC or blends of cement and GGBFS (in which GGBFS was atleast50%)werepermittedasmeansofreducingalkalicontentsthatwouldotherwise exceed 3 kg/m3. The concrete alkali content was to be calculatedfrom cement manufacturers data, adding 0.15% to allow for possible futurevariations,oralternativelyaddingtwicethestandarddeviationwherethatwouldamounttolessthan0.15%.Thecriteriagoverningthesusceptibilityofaggregateswereveryrestrictive,beingbasedupontheclassificationgiveninBREDigest258ratherthantheslightlymodifiedversiongiveninthefirstHawkinsReport.Only aggregates in which all of the significant rock components (i.e. > 5%by mass of the coarse or fine aggregate) were in the trade groups: granite orgabbroorbasalt(exceptandesites)orlimestonesorporphyry(exceptdacites,rhyolitesandfelsites)orhornfels,weretobeconsiderednotsusceptibletoAAR.OfthecommonestUKcoarseaggregates,onlythelimestonescouldsatisfytheserequirements,andalmostnoneofthenaturalfineaggregatescouldbejudgednotsusceptibleaccordingtosuchcriteria.AgainthepresenceofopalrenderedanyaggregatesusceptibletoAARwhatevertheremainingcompositionofthataggregate.AnextensivelyrevisedDTpspecificationwaspublishedin198684,andthisincludedclausesforthecontrolof ASR,whichwerebroadlysimilartothosecontainedwithinthe1985draftrevisionoftheHawkinsReport.TherequirementsareessentiallysimilartothoseprovidedinthemodelspecificationclausesoftherevisedHawkinsReportlaterissuedin198713(see 5.6.2), although the DTp is perhaps slightly more restrictive. AlthoughtheDTpspecificationallowsthat,Whentheproportionofchertorflintisgreaterthan60%(byweight)ofthetotalaggregateitshallbeconsideredtobenon-reactiveprovidingitcontainsnoopal,tridymiteorcristobalite,thelowerlimitforchert,flintorchalcedonytakentogetherisonly2%byweightinsteadofthe5%impliedbytheHawkinsReportintheguidancenotes. Also,whereastherevisedHawkinsReportconsidersthattheuseofthe certified average alkali content of cements eliminates the need for anydeliberatevariationfactorsinthecalculationofthereactivealkalicontentofconcrete,theDTpspecificationcontinuestorequirethattwicethestandarddeviationisaddedtotheaverageof25dailydeterminationsofcementalkalicontentbythemanufacturerandalsothat10kgisaddedtotheintendedcementcontentoftheconcretetoallowforpossiblebatchingerrors.Inpracticethedifferencecouldsometimesbeimportant,withtheHawkinsReportrecommendationsallowingconcretealkalicontentstorangeupto3.75kg/m3,whilsttheDTprequirementsseektomake3.0kg/m3anabsolutemaximumlimit.The1986DTpspecificationagreedwiththedraftrevisionoftheHawkinsReport in requiring only the water-soluble alkali contents of GGBFS and PFAtobetakenintoaccount.However,itisunderstoodthattheDTpspecificationCopyright 1992 Blackie and Son Ltdis shortly to be amended in line with BRE Digest 330, in which 50% and 17%proportionsoftheacid-solublealkalicontentsarepreferred65(see5.6.2).5.6.4 Quest for a testThetestmethodsadoptedfortheBRSresearchinthe1950s(see5.2)hadallbeen derived from American experience, and in the early 1950s several of thesehad evolved into ASTM standard methods19,21. In the UK, presumably because oftheapparentreassurancearisingfromtheBRSresearchprogramme,therewerenosimilarmovestocreateanyBritishStandardmethodsfortestingconcreteconstituentsforASR.OnceexamplesofASRhadbeenidentifiedontheUKmainlandinthemid-1970s,anewurgencywascreatedforfindingapredictivetest method which would enable the problem to be avoided in new construction.Atfirstitwasassumedthatthemortarbarexpansiontest(ASTMC22721)wasreliablebutthatthepracticalproblemconcernedthelongdurationofthetest:Palmer(C&CA)in197738saidthat,thereisnosatisfactoryshort-termtest.Only long-term tests on the aggregate can show with any certainty whether thereisanydangerinitsuse,andGuttandNixon(BRE)in197936saidthemortarbar method is the most generally accepted method of assessment available and ifthe limitations of time and sampling are borne in mind it has been shown to giveagoodguidetothereactivityoftheaggregate.Inathoroughreviewin1981,Simsemphasisedtheindicative,ratherthandefinitive,natureofthevarioustestsfor ASRandstressedtheimportanceofpetrographicalexamination87: Aknowledgeofthenatureofthematerialsbeingconsiderediscentraltotheassessmentprogrammeandsomeformofpetrographicexaminationshouldneverbeomitted.Insomecasesthepetrographicappraisalmayobviatetheneedforanyfurthertesting,andinmostcasesthemostpertinentindicativetestingsequencewillbeidentified. The inadequacy of the mortar bar test for UK reactive aggregate combinationshad not yet been realised, but Sims did warn that, In a few cases, mortar barsproduceanomalousresultsthataredifficulttoexplainandnotedthattheflint-bearingaggregatesofsouth-eastEnglandinvariablyproducemortarbarresults that give no cause for alarm, and yet the deleterious alkali-reactivity inthesouth-westofEnglandhasmostlyconcernedsimilarmaterialswhichareblendedwithotherrocktypesandaretherebypresentinverymuchsmallerproportions.Itwasthenconsideredthattestingactualcement-coarseaggregate-fineaggregatecombinations,ratherthantheindividualaggregateconstituentsseparately,mightavoidsuchamisleadingoutcome.However,by1986,Sims88couldrep