Analysis of Inner Structure Changes of Concretes Exposed to High Temperatures using Micro X-ray Computed Tomography

Ac a Geodyn. Geoma e ., Vol. 12, No. 1 (177), 79–89, 2015 DOI: 10.13168/AGG.2015.0009 jou nal homepage: h p://www.i sm.cas.cz/ac a ORIGINAL PAPER ANALYSIS OF INNER STRUCTURE CHANGES OF CONCRETES EXPOSED TO HIGH TEMPERATURES USING MICRO X-RAY COMPUTED TOMOGRAPHY Libo SITEK 1)*, Lenka BODNÁROVÁ 2), Kamil SOUČEK 1), Lubomí STAŠ 1) and Lucie GURKOVÁ 1) 1) Ins i u e o Geonics o he CAS, . . i., S uden ská 1768, 708 00 Os a a, Czech Republic 2) B no Uni e si y o Technology, Facul y o Ci il Enginee ing, Ve eří 331/95, 602 00 B no, Czech Republic *Co esponding au ho ‘s e-mail: libo .si [email p o ec ed]as.cz ABSTRACT The X- ay Compu ed Tomog aphy (X- ay CT) ep esen s a p og essi e non-des uc i e me ho d o analysing he inne s uc u e o ma e ials. The me hod was used o moni o ing changes in inne s uc u e o conc e e samples o di e en composi ion be o e and a e hei exposu e o a ious he mal loads. Eigh ypes o conc e e samples we e p epa ed which di e ed by cemen and agg ega e ypes. We in en ionally used such composi ion o conc e e mix u es which inc eased hei esis ance o high empe a u es. The inne s uc u e o samples was analysed using he mic o X- ay CT be o e he he mal loading and a e exposu e o samples o empe a u es o 600 °C o 900 °C. Damages and cha ac e o changes in he inne s uc u e caused by high empe a u es we e in luenced by he cha ac e o cemen and agg ega e we used. The analysis o ob ained esul s is p esen ed in he pape . ARTICLE INFO A icle his o y: Recei ed 6 No embe 2014 Accep ed 16 Feb ua y 2015 A ailable online 2 Ma ch 2015 K eywo ds: Conc e e Cemen pas e Agg ega e High empe a u e in he inne s uc u e o conc e e a e he he mal exposu e is o cu a sample and p o ide consequen mac oscopic and mic oscopic analyses. As i is impossible o e alua e changes in he same sample in unloaded s a e and in s a e a e he he mal load wi h adi ional me hods, we used he mic o X- ay Compu ed Tomog aphy (mic o X- ay CT) o he quan i a i e and quali a i e analyses o he inne s uc u e o conc e e. The majo ad an age o his p ocedu e is i s non-des uc i i y du ing he 3D isualiza ion and analyses o he s udied ma e ials. USE OF THE X-RAY COMPUTED TOMOGRAPHY IN INDUSTRIAL APPLICATIONS Recen ly, p og essi e me hods o analysing a ious ypes o ma e ials ha e been inc easingly applied. One o hem is he X- ay Compu ed Tomog aphy which was o iginally used o he pu poses o medical diagnos ics. Based on success ul and apid de elopmen o his diagnos ic me hod encou aged by esea ch and de elopmen o he X- ay Compu ed Tomog aphy sys ems du ing he 1980s, i b egins o be used o indus ial applica ions in ma e ials enginee ing, including he esea ch in o b eha iou o bioma e ials, as epo ed by O ani (2004). In his ield o esea ch, i is especially ad an ageous o use he 2D/3D X- ay Compu ed Tomog aphy sys ems wi h high- esolu ion imaging ( he so-called mic o X- ay CT) based on he mic o ocal X- ay sou ce and app op ia e line and a ea (2D) de ec o s o X- ays which enable o esol e indi idual p ic u e elemen s o a CT slice, o olume elemen s ( oxels), small as uni s and ens o mic ome es in INTRODUCTION Conc e e has many ad an ageous p ope ies, e.g. in lammabili y o low empe a u e conduc i i y. Howe e , when a conc e e cons uc ion is exposed o high empe a u es, i loses i s load- b ea ing capaci y, especially in su ace laye s which a e loaded a mos . Se ious p oblem is he so-called explosi e spalling which is caused by he combina ion o inc easing p essu e in inne po es and inne comp essi e s esses o igina ing om ma e ial expansion a highe empe a u es. Spalling o a conc e e laye leads o weakening o he conc e e c oss-sec ion o a e- in o ced cons uc ion and, in some cases, e en o he exposu e o s eel ein o cemen o empe a u es abo e he c i ical le el. De ec s can be so se ious ha he comple e cons uc ion ge s des oyed. Design o a conc e e mix u e wi h esis ance o he mal exposu e o a ce ain pe iod o ime has been al eady classi ied among he passi e i e p o ec ion me hods. Basic composi ion o cemen conc e e can be p oblema ic, as i is a composi e ma e ial consis ing o wo ma kedly di e en componen s, i.e. he cemen pas e and agg ega e. When exposed o high empe a u es, bo h componen s beha e di e en ly and hei physical, chemical and mechanical p ope ies change. Expe imen al esea ch in o he a o emen ioned ield o s udy consis s o p epa a ion o es samples acco ding o he p elimina y concep o he conc e e mix u e composi ion and hei es ing a high empe a u es o du ing di ec lame impingemen . The nex s ep is he p ope ies diagnos ics o samples which can be based on a wide ange o es s and measu emen s. The only way how o examine changes L. Si ek e al. 80 Kodu , 2008; Xing a al., 2011 and Scheyd e al., 2012). A GGREGATE When p epa ing he he mal- esis an conc e e mix u e, we used he basal agg ega e which shows a long- e m esis ance o high empe a u es due o i s cha ac e and o igin. In addi ion, we also used he ligh weigh a i icial agg ega e om he expanded clay (see Eu opean S anda ds EN 13055-1, EN 14063-1 and EN ISO 9001). Agg ega e om expanded clay has sui able p ope ies in e ms o olume s abili y when exposed o empe a u es up o 1050 °C. (Technical guide - Liapo ). This agg ega e is p oduced in o a y kiln a a empe a u e o 1100°C o 1200°C. Wa e abso p ion o expanded clay agg ega e anges om 2 o 7 % (by weigh ) a e 30 min and om 7 o 9 % a e 24 hou s o ull imme sion in wa e (Technical guide - Liapo ). Beha iou o conc e e wi h his agg ega e is signi ican ly a ec ed by he so-called mois u e con en . The mois u e con en in conc e e is a basic ac o which can cause c acks and explosi e spalling du ing he he mal loading (e.g. Kopinga and Pel, 1994). Due o he e apo a ion o wa e om expanded clay, po es in conc e e a e loaded by some addi ional p essu e. I he p essu e o apou is highe han he lexu e s eng h o cemen pas e, he cemen pas e b eaks and mic oc acks o explosi e spalling a e o med ( an de Heijden a al., 2007). In e ms o i e and hea esis ance o ligh weigh conc e e, he wo mos impo an ac o s a e he mixing p ocedu e and condi ions o which he cons uc ion is exposed. The wo possibili ies o how o dose expanded clay agg ega es in o conc e e mix u e a e he ollowing:  in soaked s a e, wi h posi i e e ec s on wo kabili y o conc e e and in e nal cu ing. Highe mois u e con en is b ough in o conc e e s uc u e and con inuously eleased du ing he cu ing o conc e e. The ime necessa y o wa e elease depends on en i onmen al condi ions. This echnological p ocess is especially sui able o eady-mix conc e e, high pe o mance ligh weigh conc e e and applica ions in oad cons uc ions. Ligh weigh conc e e p epa ed by his echnological p ocess has he mois u e con en up o 20 % a e 28 days o cu ing;  in d y s a e. This echnological p ocess is especially used in p e ab ica ion, as he moulding o conc e e is as (wi hin 10 minu es). The e is no ime o he agg ega e o soak up mixing wa e ; howe e , his p ocedu e educes he wo kabili y o ligh weigh conc e e. Ligh weigh conc e e p epa ed by his echnological p ocess has much lowe mois u e con en ( om 3 o 5 % a e 28 days o cu ing). Conc e e es specimens wi h expanded clay agg ega es we e p epa ed acco ding o he i s p ocess. The agg ega es we e s o ed in wa e o size. Al hough any speci ic di e ence be ween he X- ay CT and mic o X- ay CT has no been speci ied, i can be de ined in ela ion o he space esolu ion o oxel o he limi alue o app oxima ely 200 μm which he medical X- ay CT scanne s canno usually each. The size o a sample examined wi h he mic o X- ay CT me hod anges om app oxima ely 40 cm (la ge samples) o a ew mic ome es (small samples). Fo geosciences and cons uc ion ma e ials esea ch, s anda d samples o he mic o X- ay CT a e o one millime e o 5 cen ime es in size (Cnudde and Boone, 2013). The applica ion o he X- ay CT me hod o quan i a i e and quali a i e analyses o beha iou o di e en kinds o geoma e ials and o he ela ed ma e ials, such as ocks, soil, cons uc ion ma e ials, ce amic ma e ials and geocomposi es (Ščučka and Souček, 2007), is g adually inc easing. This non- des uc i e me hod o in es iga ions o geoma e ials b eha iou is also sui able o he 4D isualiza ion and analysis o changes in objec s exposed o ex e nal ac o s ( o example, s ess, he mal exposu e, deg ada ion changes o ma e ials, s udy o liquid- low in ma e ials, e c. – see Des uess e al., 2006). E SSENTIALS OF COMPUTED TOMOGRAPHY Indus ial mic o X- ay Compu ed Tomog aphy u ilizes he di e ences o X- ays p ope ies du ing he in e ac ion wi h he mass o a s udied ma e ial. In case o he X- ay CT, i is especially he abili y o X- ays o pene a e a ious kinds o ma e ials wi h di e en le els o hei a enua ion in ela ion o p ope ies o analysed objec s, especially objec densi y. I is necessa y o ob ain se ies o 2D adiog aphs o each de e mined o a ion angle du ing he scanning p ocess o he analysed objec wi h o a ion o 360° o he consequen econs uc ion o he CT olume consis ing o indi idual oxels. The 2D adiog aph p ojec ions a e ealized using he X- ays a ea de ec o wi h he de e mined numbe o pixels. The adiog aph is a 2D map o pixels which shows he in ensi y o a enua ion o X- ay signals passing h ough he s udied objec . Di e en g ey le els (g ey le el scales a e de e mined by he ype o de ec o and i s bi scale - 8-bi , 16- b i , e c.) co espond o a ious in ensi ies o a enua ion. Based on su icien numbe o p ojec ions, he CT olume o he objec is econs uc ed using algo i hm calcula ions (e.g. he Radon ans o ma ion, he in e se Radon ans- o ma ion o Fil e ed Back-P ojec ion Me hod as epo ed by Kak and Slaney, 1987). EXPERIMENTAL PROCEDURE P epa a ion o a mixing concep ion o conc e es loaded by high empe a u es o i e is no an easily sol able p oblem. P epa a ion o ou samples was b ased on ou own expe iences and ecommenda ions o o he au ho s which ha e been occupied in ensi ely b y he simila p oblem (e.g. Khou y, 1992; Vyd a e al., 2001; Mindeguia e al., 2013; Sa a e al., 2005, ANALYSIS OF INNER STRUCTURE CHANGES OF CONCRETES EXPOSED TO HIGH … . 81 Table 1 Composi ion o mix u es wi h basal agg ega e. Mix u e componen  Quan i y pe m3 Cemen 350 kg Wa e 175 kg Agg ega e 0-4 mm basal 1070 kg Agg ega e 4-8 mm basal 1050 kg Supe plas i ie Mape luid N2006.3 kg Table 2 Composi ion o mix u es wi h basal and expanded clay agg ega e. Mix u e componen  Quan i y pe m3 Cemen 375 kg Wa e 125 kg Expanded clay agg ega e 0-4/5000.24 m3 Expanded clay agg ega e 4-8/6000.4 m3 Agg ega e 0-4 mm basal 500 kg Supe plas i ie Mape luid N2004.5 kg Table 3 Types o cemen used in mix u es. Po land cemen CEM I 42.5 R  Po land composi e cemen wi h limes oneCEM II/B-M (S-LL) 32.5 R  Po land slag cemen CEM II/B-S 32.5 R  Blas u nace cemen CEM III/B 32.5 N-SV  a numbe o cylind ical co es, wi h a diame e o 22 mm and heigh o abou 40 mm, we e d illed. Then, he con igu a ion o he used X- ay omog aph was op imised o ob ain he bes possible quali y o indi idual CT slices. Acco ding o he size o analysed es ed objec s, he size o a oxel ( he b asic composi ion elemen o he econs uc ed omog aphy olume) was abou 0.025 mm. The inne s uc u e o cylind ical samples was u he scanned wi h he mic o X- ay CT. Subsequen ly, samples we e exposed o he he mal s ess o 600 °C o 900 °C. Samples we e hea ed in acco dance wi h he s anda d empe a u e cu e ISO 834 (see ISO 834-1) and las ed a he de e mined empe a u e o he pe iod o 60 minu es. The s anda d empe a u e- ime cu e ISO 834 (also known as he Cellulosic cu e o he s anda d nominal i e cu e) ep esen s a bu ning a e o ully de eloped i e in a compa men . I shows ha empe a u e inc eases wi h ime a a cons an a e (see Fig. 1). The cu e is de ined by he ollowing equa ion:   0345 8 1log    gT whe e  g is he empe a u e o gases in compa men i e (°C), T0 ep esen s he empe a u e in compa men i e be o e i e ou b eak (°C) and is he du a ion o he i e (min). A e hea ing a he desi ed empe a u e, he samples we e cooled a he labo a o y 48 hou s p io o manu ac u ing; expanded clay agg ega es we e used in he soaked s a e. The mois u e con en o he conc e e wi h expanded clay agg ega es was 12 %, while he mois u e con en o he conc e e wi h basal agg ega es only 4 %. CEMENTS In e ms o a ious le els o he mal esis ance, ou ypes o cemen s we e chosen as binde o cemen mix u e p epa a ion, i.e. he Po land cemen , he Po land composi e cemen wi h limes one, he Po land slag cemen and, inally, he Blas u nace cemen . The Po land cemen is he mos commonly used ype o cemen in conc e e mixing. I was he eason why o he cemen s we e compa ed wi h his ype o cemen . E en hough, due o low esis ance o he mal loads, his cemen is no gene ally used o p epa a ion o he mally s able conc e es. Fi s ly, 8 di e en conc e e mix u es we e p epa ed using a ious ypes o cemen s and agg ega es. Fou mix u es we e composed o he b asal agg ega e; hey di e ed by he used cemen . In case o o he ou mix u es, he mix u e o he basal agg ega e and he a i icial agg ega e om he expanded clay was used. The Table 1 and Table 2 show he composi ion o all es ed mix u es. The ypes o used cemen a e speci ied in Table 3. Secondly, es ing beams o dimensions o abou 400x100x100 mm3 we e p epa ed om each conc e e mix u e. F om each analysed ype o conc e e, L. Si ek e al. 82 Fig. 1 The s anda d empe a u e- ime cu e ISO 834. Fig. 2 De ini ion o he es ed space in a conc e e sample o de ec ion o he inc emen in po e space a e hea ing (le - sample be o e he mal exposu e, igh - sample a e he mal exposu e). ANALYSIS OF INNER STRUCTURE CHANGES OF CONCRETES EXPOSED TO HIGH … . 83 Fig. 3 Example o he analysis o changes in he olume o po e space in one o he CT slices o a conc e e es sample be o e and a e he he mal exposu e. Table 4 Speci ica ion o XT H 225 indus ial mic o X- ay CT sys em used du ing expe imen . Max. accele a ion ol age and powe o he X- ay sou ce ( e lec ion mode) 225kV / 225W Max. accele a ion ol age and powe o he X- ay sou ce ( ansmission mode) 180kV / 20W Size o he X- ay ube ocus ( e lec ion mode / ansmission mode) <3 µm / <1 µm Max. weigh , diame e and heigh o scanned objec s 50 kg / app ox. 0.5 m / 0.5 m Max. omog aphy hickness o analysed ma e ials 237 kg.m-2 Senso o he X- ay adia ion (16-bi dep h) – a ea de ec o 200 µm pe pixel, No. o pixels - 2000 x 2000 newly o med due o hea ing. The di e ence be ween hese wo olumes indica es, in a ce ain way, he a e o esis ance o he analysed conc e e mix u e agains he he mal loading. Based on he analysis o econs uc ed olumes, changes in po es space we e de e mined again using he mic o X- ay CT. Po es space was analysed in he ollowing s eps:  A es ed space o app oxima ely equal olume (abou 3300 mm3) o he analysis o changes i n he olume o po es space on an iden ical sample b e o e and a e hea ing was de ined (see Fig. 2). The es ed space was posi ioned in he middle p a o he es ed sample o elimina e he in luence o ex e nal bounda y condi ions. Resul s om he sample edge can be in luence d b y an excessi e loss o he agg ega e and cemen binde caused by hea ing. Thus, i can nega i ely a ec he alue o po e space inc emen .  Analysis o he olume o po e space was ca ie d ou in he es ed space o all es samples b y de e mina ion o a h eshold which enabled he di e en ia ion be ween he po e space and he empe a u e o 20 °C. Once indi idual es ed objec s we e cooled, compa a i e CT scans o he co esponding conc e e samples we e pe o med o he consequen e alua ion o e ec s o he he mal load on hei in eg i y. Using he VGS udio Max so wa e, sepa a e econs uc ed CT olumes o analysed samples we e isualised. Co esponding CT olumes, scanned be o e and a e he he mal exposu e, we e dimensionally o ien ed in a way ha he omog aphic c oss-sec ions wi h iden ical posi ion could be simul aneously compa ed. On so- p osi ioned and isualised c oss-sec ions, e ec s o high empe a u es we e e alua ed. Fo p ecise quan i a i e e alua ion o ailu e a e o empe ed conc e e mix u es, olume o po es space in iden ical sample be o e and a e hea ing a he desi ed empe a u e we e compa ed. The olume o p o e space o he sample analysed be o e hea ing is ep esen ed mainly by ai bubble po es ha a ise om he echnological p ocessing o conc e e mix u es du ing sample p epa a ion. A e hea exposu e, he olume o po es space is ep esen ed no only by ai b ubbles, bu also by c acks in he cemen ma ix L. Si ek e al. 84 Table 5 Se ings o pa ame e s o mic o X-Ray CT. X- ay sou ce se ings X- ay pene a ion [kV] 150 X- ay In ensi y [µA] 300 - 400 X- ay il e Aluminium, hickness 2.25mm Se ings o CT scanning No. o p ojec ions [--] 2 800 -3 000 No. o ames pe p ojec ion [--] 8 P ojec ion exposu e [ms] 354 Scanning ime [h] app ox. 2.5 Recons uc ion Voxel esolu ion [µm] 20 - 30 No. o oxels in CT olume app ox. 1.5 E9 econs uc ion ime [min] app ox. 10 oxel bi dep h / g ey le els 8-bi / 255 EXPERIMENTAL EQUIPMENT AND CONDITIONS Fo he s udy o s uc u al changes o conc e e es samples exposed o he mal loads, he N ikon Me ology XT H 225 indus ial mic o X- ay CT sys em was used. I is a ully au oma ed appa a us wi h a o a ing scanning sys em equipped wi h a mic o ocal X- ay sou ce which gene a es cone-shaped beams. The used sys em is speci ied in Table 4. S udied olumes we e econs uc ed using CT P o 3D and CT P o 2D so wa e (by Nikon Me ology NV). Visualiza ion was p o ided by a VGS udio Max so wa e (by Volume G aphics), e sion 2.2, wi h addi ional modules o p ope image analysis. Se ings o pa ame e s o he mic o X- ay CT sys em used o scanning he inne s uc u e o conc e e ma e ials a e p esen ed in de ail in Table 5. Tes samples we e hea ed o he desi ed empe a u e using an elec ically hea ed labo a o y es ing u nace Classic 1013L wi h inne space dimensions o abou 800 x 800 x 800 mm3. The u nace (Fig. 4) enables o each and sus ain he empe a u e up o 1100°C acco ding o a ious empe a u e cu es. Cu en empe a u e was moni o ed and eco ded using a da a logge Tes o 174-T4 wi h ex e nal empe a u e p obes connec ion. Tempe a u e cu e o es ing ollowed he s anda d ISO 834 cu e. ANALYSIS OF RESULTS Resul s o s uc u al changes in es samples a e he he mal load a e g aphically p esen ed as ollows. Example o wo iden ical omog aphic c oss-sec ions o one sample be o e and a e he exposu e o he he mal load o 900 °C is shown in Figu e 5. I demons a es how easily indi idual g ains o agg ega e, ai po es, c acks in cemen pas e, e c. can be iden i ied on a pa icula CT slice. Fo an easie compa ison o bo h inne s uc u es, some ypical c acks and hei a e age wid hs a e highligh ed in s uc u es a e he mal exposu e. Fig. 4 Tes ing u nace Classic 1013L and da a logge Tes o 174-T4 wi h ex e nal empe a u e p obe connec ions o eco ding eal empe a u e cu e. es o conc e e ma e ial (agg ega es and cemen binde ). Example o he inc ease in he olume o po e space in one o he es sample sec ions be o e and a e hea ing is shown in Figu e 3.  Finally, he inc ease in he olume o po e space, caused by hea ing a ious conc e e mix u es up o a empe a u e o 600 °C and 900 °C, was calcula ed and e alua ed. ANALYSIS OF INNER STRUCTURE CHANGES OF CONCRETES EXPOSED TO HIGH … . 85 Fig. 5 Changes in he inne conc e e s uc u e a e he exposu e o he mal load (le – s uc u e be o e he mal exposu e, igh – he same s uc u e a e he mal exposu e). Fig. 6 Conc e e wi h he Po land composi e cemen , basal agg ega e and ligh weigh agg ega e om expanded clay be o e and a e he exposu e o a he mal load o 900°C (wid hs o selec ed c acks a e indica ed in millime es). he cemen ma ix (see Fig. 9) we e p e ailing in s uc u es o cemen s exposed o he he mal load. I was gene ally ue ha ex ensi e c acks in he cemen ma ix occu ed especially in he ma ix consis ing o he Blas u nace cemen and he Po land slag cemen ( o example, Fig. 9). Sepa a ion c acks along agg ega e g ains we e mo e common in he cemen ma ix p epa ed om he Po land cemen and he Po land composi e cemen (see Fig. 6 and Fig. 11). The inne s uc u e o he conc e e wi h expanded clay was usually mo e damaged due o agg ega e p ope ies – mo e wa e was used in he p epa a ion I was p o ed ha he ype o used agg ega e, as well as he ype o cemen pas e, in luenced he inne s uc u e o s udied samples exposed o a he mal load. The mally loaded conc e e samples wi h expanded clay we e much mo e damaged (especially b y c acks along clay g ains) han he conc e e samples consis ing only o basal agg ega e (compa e Fig. 6 and Fig. 7). Na u e o occu ed damages we e signi ican ly in luenced by he ype o used cemen . I de e mined whe he sepa a ion c acks along he agg ega e g ains (especially along expanded clay – Fig. 6) o c acks in L. Si ek e al. 86 Fig. 7 Conc e e wi h he Po land composi e cemen and basal agg ega e be o e and a e he exposu e o a he mal load o 900°C (wid hs o selec ed c acks a e indica ed in millime es). Fig. 8 Conc e e wi h he Blas u nace cemen wi h basal agg ega e and ligh weigh agg ega e o m expanded clay be o e and a e exposu e o a he mal load o 600°C (wid hs o selec ed c acks a e indica ed in millime es). Fig. 9 Conc e e wi h he Blas u nace cemen wi h basal agg ega e and ligh weigh agg ega e o m expanded clay b e o e and a e exposu e o a he mal load o 900°C (wid hs o selec ed c acks a e indica ed in millime es). ANALYSIS OF INNER STRUCTURE CHANGES OF CONCRETES EXPOSED TO HIGH … . 87 Fig. 10 Conc e e wi h he Po land cemen wi h basal agg ega e and ligh weigh agg ega e o m expanded clay be o e and a e exposu e o a he mal load o 600°C (wid hs o selec ed c acks a e indica ed in millime es). Fig. 11 Conc e e wi h he Po land cemen wi h basal agg ega e and ligh weigh agg ega e o m expanded clay be o e and a e exposu e o a he mal load o 900°C (wid hs o selec ed c acks a e indica ed in millime es). The mal esis ance was isually e alua ed (in e ms o numbe , size and cha ac e o occu ed c acks) on es samples o all es ed conc e e mix u es hea ed o 600 °C and 900 °C. Mo eo e , he inc ease in po e space in he es ed olume o each sample b e o e and a e hea ing was e alua ed quan i a i ely. This can be in e p e ed as a o ma ion o new c acks and/o an inc ease in he olume o exis ing c acks. This dependence is illus a ed in Figu e 12. I is ob ious ha inc ease in he empe a u e loading o a sample also leads o a g ow h o he olume o c acks. Gene ally, samples wi h added expanded clay agg ega e a e mo e damaged by c acks (expansion o he po e space up o 8 %), which is appa en om he p ocess due o he p esence o po es in he clay agg ega e (see abo e). Consequen ly, he cemen ma ix s uc u e was damaged by he wa e e apo a ing om he agg ega e. As expec ed, he inne conc e e s uc u e was mo e damaged a he highe empe a u e (900 °C) han a he lowe one (600 °C). C acks occu ed in he cemen ma ix mo e equen ly and hey we e wide (compa e Fig. 8 and Fig. 9). Sepa a ion c acks along he expanded clay agg ega e only de eloped when he sample was hea ed o 900 °C. A a empe a u e o 600°C, any o samples did no exhibi ed c acks along he expanded clay (compa e, o ins ance, Fig. 10 and Fig. 11).