Synthesis and characterization of belite calcium sulfoaluminate cements produced by oxyfuel combustion residues

Syn hesis and cha ac e iza ion o beli e calcium sul oalumina e cemen s p oduced by oxy uel combus ion esidues A. Telesca*1, M. Ma occoli2, N. Ib is3, T. R., Naik4, C. Lupiáñez5, L. I. Díez6, L. M. Romeo7 and F. Mon agna o8 1-3School o Enginee ing, Uni e si à degli S udi della Basilica a, ITALY. (E-mail: an onio. [email protected] , milena.ma [email protected] , nelu a.ib[email p o ec ed] ) 4Depa men o Ci il Enginee ing and Mechanics, Uni e si y o Wisconsin, USA (E-mail: a [email protected]) 5-7Mechanical Enginee ing Depa men , Uni e sidad de Za agoza, SPAIN. (E-mail: ca los.lupianez@uniza .es, luisig@uniza .es, luismi@uniza .es) 4 Depa men o Chemical Sciences, Uni e si à degli S udi di Napoli Fede ico II, ITALY. (E-mail: abio.mon agna [email protected] ) ABSTRACT In his wo k, he possibili y o eusing ashes issued by an oxy uel combus ion p ocess (OC) as a sou ce o ma e ial in he p oduc ion o beli e calcium sul oalumina e BCSA cemen s has been in es iga ed. OF p ocess is one o he mos p omising combus ion echnologies o CO2 educ ion om powe plan s. Combus ion es s we e ca ied ou in an oxy uel bubbling luidized bed pilo plan . Fou BCSA clinke -gene a ing aw mixes we e hea ed in a labo a o y elec ic o en in he empe a u es ange 1150°- 1350°C: one included only na u al ma e ials (limes one, clay, bauxi e and gypsum), he o he s con ained OC ashes as o al subs i u e o clay. X- ay di ac ion (XRD) analysis on he bu ning p oduc s showed high con e sion o eac an s owa d he main BCSA clinke componen s (C2S and C4A3$), especially a 1200° o 1250°C. Mo eo e , physical-mechanical es s associa ed wi h XRD and di e en ial he mal- he mog a ime ic analyses accomplished on all he cemen s (ob ained by adding na u al gypsum o he clinke s p oduced a he bes syn hesis empe a u es) gene ally displayed a simila hyd a ion beha iou . Keywo ds: Oxy uel combus ion esidues, clinke , beli e calcium sul oalumina e. INTRODUCTION The apid g ow h o Wo ld popula ion, he ene gy demand inc ease, he global wa ming (GW) ha e pushed go e nmen s and in e na ional au ho i ies o se on he pa h owa ds sus ainable de elopmen . The e is no doub ha GW ep esen s he mos signi ican challenge o achie ing sus ainable de elopmen ; his phenomenon is caused by he emission and accumula ion in he a mosphe e o g eenhouse gases * Co esponding au ho . T: +39 0971 205225. E-mail: an o[email p o ec ed]. Fi h In e na ional Con e ence on Sus ainable Cons uc ion Ma e ials and Technologies. h p://www.claisse.in o/P oceedings.h m (GHGs) which a e mainly eleased by he bu ning o ossil uels, land clea ing, ag icul u e- ela ed and o he human ac i i ies [Benhelal e al., 2013; De Rich e e al., 2016; Szulejko e al., 2017; Telesca e al., 2017]. Ca bon dioxide is blamed o be he main esponsible o GW; he e o e, sea ching o p omising app oaches o mi iga e CO2 emissions ep esen s he p io i y o s udies aimed a alle ia ing he h ea o clima e change. In his ega d, he ca bon cap u e and s o age (CCS) echnology has he po en ial o almos comple ely elimina e CO2 [Boo -Hand o d e al., 2014; Leung e al., 2014; Telesca e al., 2014a; Telesca e al., 2017]. CCS indica es a g oup o echnologies de eloped o ob ain CO2- ich lue gases eady o be s o ed by injec ion in o geological s a a wi h speci ic ea u es. Among he di e en CCS p ocesses (e.g., chemical abso p ion, chemical and calcium looping), oxy uel combus ion (OC) is o g ea in e es due o i s concep ual simplici y [Diez e al., 2015]. In OC a blend o nea ly pu e oxygen and pa o exhaus gas (employed as O2 diluen o sa e y easons), is used o combus ion, hus lowe ing bo h N2 and NOx con en s in he exhaus gas [Buh e e al., 2005; Lupiáñez e al., 2013a; Lupiáñez e al., 2014]. In absence o he mos ele an diluen o CO2 and a e u he ea men s, i is possible o ob ain s eams >90% CO2- ich. The CO2- ich s eam is hus eady o inal p ocessing and geological s o age. Cemen p oduc ion is one o he mos la ge-scale aw ma e ials consuming as well as ene gy-in ensi e manu ac u ing p ocesses and ep esen s one o he majo indus ial sec o s gi ing ise o CO2 emission. In ac , he Wo ld’s annual cemen p oduc ion cu en ly accoun s o abou 4.7 billion ons [Ac i i y epo 2016, 2017] and i s con ibu ion o he global an h opogenic CO2 emission is es ima ed as high as 7% (abou 26% o he indus ial CO2 emission) [T egambi e al., 2018]. I is he e o e widely accep ed he need o inc ease he sus ainabili y o cemen mainly o ien ed a mi iga ing he CO2 impac . Po land cemen (PC) is he mos widely used binde all o e he Wo ld; i is ob ained by in e g inding PC clinke (PCC) wi h a ew pe cen o calcium sul a es (mos ly gypsum). Fo each kg o PCC p oduced, abou 0.87 kg o CO2 a e eleased [Ba celo e al., 2014]; ca bon dioxide comes om bo h limes one he mal decomposi ion (abou 60% o he o al CO2 emission) and uel combus ion. The e o e, one app oach o lowe CO2 emissions is ela ed o he limes one educ ion in he clinke -gene a ing aw meal [Telesca e al., 2016]. In his ega d, beli e ich-cemen s a e conside ed en i onmen ally iendly binde s inasmuch as hey can allow a CO2 educ ion as much as 10% [Pim aksa e al., 2009; Cube os e al., 2010]; mo eo e , he p esence o calcium sul oalumina e (C4A3$) in he cemen clinke can bo h compensa e he lowe eac i i y o beli e ( ela i e o ali e) and u he educe he limes one equi emen in he aw eed [Quillin, 2001]. Beli e calcium sul oalumina e (BCSA) cemen s gene ally display physical and mechanical p ope ies compa able o OPC [Selçuk e al., 2010]; in ac , hei echnical beha iou mos ly depends on he abili y o C4A3$ and C2S, o espec i ely gene a e, upon hyd a ion, C6A$3H32, e ingi e (a ea ly ages) and CSH (a medium and longe ages). Mo eo e , BCSA cemen s a e in e es ing hyd aulic binde s om he en i onmen al poin o iew inasmuch as hei manu ac u ing p ocess, compa ed o ha o OPC, displays a p onounced en i onmen ally iendly cha ac e mainly associa ed o (a) he ela i ely low syn hesis empe a u e, (b) he dec eased speci ic uel consump ion [Ma occoli e al., 2010a], (c) he easie g indabili y and (d) he g ea e usabili y o indus ial was es and by-p oduc s whose u iliza ion is gene ally complica ed [Ma occoli e al., 2009; Ma occoli e al., 2010b; Ma e al., 2014; Shen and Qian, 2015]. BCSA cemen s a e gene ally p oduced by bu ning a aw mix composed by limes one (L), bauxi e (B), clay (C) and gypsum (G) a empe a u es anging om 1250°-1350°C [Xue e al., 2016]. In his pape , esidues gene a ed du ing an OC p ocess, ca ied ou in a pilo -scale luidized bed (FB) eac o , we e es ed as subs i u es o clay in he BCSA clinke - gene a ing aw mix. Fou mix u es we e hea ed in a labo a o y elec ic o en in he empe a u es ange 1150°-1350°C: one included only na u al ma e ials ( aken as a e e ence e m), he o he s con ained OC esidues in a measu e comp ised be ween 24.5 and 56.0% by mass. The con e sion o eac an s owa ds C2S and C4A3$ was in es iga ed by means o X- ay di ac ion (XRD) analysis. XRD and di e en ial he mal− he mog a ime ic (DT−TG) analyses oge he wi h physical and mechanical es s we e employed as main cha ac e iza ion echniques o ca y ou he in es iga ion. EXPERIMENTAL Ma e ials and Oxy uel Pilo Plan The na u al ma e ials (L, C, B and G) used in his in es iga ion we e aken om qua ies loca ed in I aly. The OC ashes came om a 95 kW h pilo -scale bubbling oxy uel luidized bed (FB) eac o (2.5 m high wi h an inne diame e o 21 cm); i is schema ically epo ed in Figu e 1 and desc ibed elsewhe e [Lupiáñez e al., 2013b]. Figu e 1. Oxy uel luidized bed pilo plan . The FB eac o (R), cooled by means o a wa e jacke placed a i s bo om and ou wa e -cooled p obes uni o mly dis ibu ed, was cha ged wi h a bed in en o y o 5 kg silica sand, luidized a 0.8 ms–1 (wi h a 65% CO2–35% O2 mix u e) and hea ed up o 850°C o 925°C; i was equipped wi h wo di e en de ices o he emo al o ly ash (ba le chambe , BC, and cyclone, CY, in he o de ) om he lue gas s eam. A blend o ligni e and biomass co n s o e ed he appa a us in a 70:30 ene gy a io; mo eo e , o cap u e he SO2 gene a ed du ing he combus ion h ough he in si u desul u iza ion (calcium, Ca (in he so ben ): sul u , S (in he uel)), wo di e en limes ones (A and B) we e al e na i ely injec ed in o FBR. Th ee di e en es s we e pe o med: hei ope a ing condi ions, o he sake o cla i y, a e summa ized in Table 1. Du ing each es , h ee kind o combus ion esidues we e collec ed: a bo om ash (BA), ex ac ed om he bo om o FBR, and wo ly ashes (BCFA, CYFA) wi hd awn along he lue gas ea men line. Table 1. Ope a ing condi ions o he es s ca ied ou in he in he FBR. Tes no. Fuel Limes one Ca:S FBR Tempe a u e, °C 1 Ligni e+Co n s o e A 2 925 2 “ A 6 850 3 “ B 6 925 The esidues o each es we e hen mixed o p epa e h ee blends (B1, B2 and B3), p opo ioned consis en wi h he co esponding amoun gene a ed du ing he expe imen al ac i i y (40%BA-50%BCFA-10%CYFA). The chemical composi ion (in e ms o majo oxides) o he aw ma e ials, de e mined by X- ay luo escence echnique (wa eleng h dispe si e BRUKER Explo e S4 appa a us), is shown in Table 2. Table 2. Chemical analysis o he na u al and was e aw ma e ials, w % L B G C B1 B2 B3 CaO 55.16 - 30.10 10.40 49.92 27.13 16.84 SiO2 - 7.20 3.00 54.10 21.35 37.70 44.38 Al2O3 - 56.30 0.90 11.50 4.23 8.62 10.67 Fe2O3 - 6.30 0.30 4.50 3.86 10.54 11.63 SO3 - - 36.20 - 7.26 6.87 5.51 O he s - 2.30 3.70 6.40 1.89 3.28 2.97 L.o.i.* 43.40 27.50 24.60 13.10 11.48 5.85 8.00 *Loss on igni ion a 950°C Mix u es Design and Cemen s P epa a ion Fou BCSA clinke -gene a ing aw mix u es we e designed acco ding o he modi ied Bogue equa ions [I in e al., 2011], assuming ha C2S, C4A3$, C4AF and C$ amoun ell in he ange 45%-60%, 20%-30%, 8%-20% and 4%-10%, espec i ely; mo eo e , a ee CaO concen a ion alue lowe han 1.5% was also conside ed. All he mix u es con ained L, B and G; a e e ence mix (MR) included also clay (C), whe eas he emaining h ee mix u es (M1, M2 and M3) al e na ely con ained B1, B2 o B3 as o al subs i u e o C. Table 3 and 4 espec i ely epo he modi ied Bogue po en ial mine alogical composi ion o he ou clinke s (CLIR, CLI1, CLI2 and CLI3) and he aw meals p opo ion. Table 3. Bogue po en ial mine alogical composi ion o he clinke s, w % C2S C4A3$ C4AF C$ C CLIR 56.0 25.9 8.3 5.1 0.1 CLI1 48.7 23.9 12.4 9.7 1.4 CLI2 52.6 20.7 17.9 4.9 0.1 CLI3 50.2 22.8 16.4 7.1 0.1 The aw mix u es we e hea ed in co e ed pla inum c ucibles in a labo a o y elec ic o en o wo hou s a 1200°, 1250°, 1300° and 1350°C and hen apidly cooled o oom empe a u e; he ob ained BCSA clinke s we e inely g ound in a F i sch Pul e ise e 6 (FP6) labo a o y plane a y mill o pass a 90 m sie e and a e wa ds submi ed o XRD analysis in o de o assess he bes empe a u e o maximizing bo h C2S and C4A3$ concen a ion. Table 4. Composi ion o he aw meals, w % L B G C B1 B2 B3 MR 53.0 13.0 12.0 22.0 - - - M1 50.0 13.0 5.0 - 56.0 - - M2 20.0 16.0 8.0 - - 32.0 - M3 52.0 13.5 10.0 - - - 24.5 BCSA cemen s (CEMR, CEM1, CEM2 and CEM3) we e hen p epa ed by g inding he bes ob ained clinke s wi h G in he FP6 labo a o y mill always o pass he 90 m sie e. The amoun o he added gypsum was de e mined conside ing bo h he C$ al eady p esen in he clinke (calcula ed wi h he modi ied Bogue equa ion) and he eac ion s oichiome y o he gene a ion o e ingi e (C6A$3H32, he main hyd a ion p oduc egula ing he echnical beha io a ea ly ages) and aluminium hyd oxide (AH3) [Manzano e al., 2012; Telesca e al., 2014c], namely: C4A3$+2C$+38H⇒C6A$H32 +2AH3 (1) XRD analysis was also employed o he e alua ion o BCSA cemen s hyd a ion p oduc s; i was pe o med wi h a B uke D2 Phase di ac ome e (CuKα adia ion and 0.02°2θs-1 scanning a e). Measu emen s on Mo a s BCSA mo a p isms we e p epa ed acco ding o he Eu opean S anda d EN 196-1 and cu ed a e demolding unde wa e a 20°±1°C; hey we e submi ed o comp essi e s eng h measu emen s a cu ing pe iods anging om 1 o 56 days. Measu emen s on Pas es BCSA cemen s we e pas e hyd a ed wi h a wa e /cemen a io equal o 0.5 by mass. The pas e samples we e cas in o 15-mm-high and 30-mm-diame e cylind ical molds, and inally placed in polye hylene bags inside a he mos a ic ba h (a 20°C and 95% ela i e humidi y) o cu ing imes anging om 4 hou s o 56 days. A he end o each aging pe iod he specimens we e i s c ushed and hen ea ed wi h ace one ( o s op hyd a ion) and die hyl e he ( o emo e wa e ); he pul e ized samples we e subsequen ly s o ed in a desicca o o e silica gel–soda lime ( o ensu e p o ec ion agains H2O and CO2) and inally submi ed o simul aneous DT-TG (NETZSCH-Tasc 414/3 appa a us, hea ing a e 10°C min–1, ope a ing be ween oom empe a u e and 1000°C) and XRD analyses. Fo he expansion–sh inkage measu emen s, eigh pas e samples, shaped as small p isms (15X15X78 mm), we e i s ai cu ed a 20°C o 4 hou s and hen demolded. La e on, one se o samples was aged a 20°C unde ap wa e , he o he s o ed in a con olled humidi y (H) chambe a 70% R.H. and 20°C. The leng h changes we e de e mined as a e age alues o ou measu emen s wi h a calipe accu a e o ±1 m; he e e ence leng h o hem was ha e alua ed jus a e demolding [Telesca e al., 2014b; Valen i e al., 2012]. RESULTS AND DISCUSSION The in ensi y o he main XRD peaks o C2S and C4A3$ we e aken as a con e sion index o he BCSA clinke -gene a ing aw mixes owa d hese p oduc s. The “ex e nal s anda d me hod” [Culli y e al., 2001] was used and, as a e e ence line, he same e lec ion om pu e C2S and C4A3$ p epa a ions (gene a ed by high- empe a u e syn hesis o analy ical g ade calcium ca bona e and qua z o he o me and calcium ca bona e, alumina and gypsum o he la e ) was aken [Telesca e al., 2015]. The XRD in ensi y (cps) o he C2S and C4A3$ lines is plo ed as a unc ion o he hea ing empe a u e (Figu e 2 (a) and 2 (b), espec i ely); each cu e displays a maximum occu ing, o C2S, a abou 1200 °C (MR and M2) o 1250°C (M1 and M3). Conce ning calcium sul oalumina e, he bes syn hesis empe a u e was he same as ha o beli e o mix u es M1 and M2, while highe o MR (1250°C) and M3 (1300°C). Figu e 2. XRD in ensi y (coun s pe second, cps) o he C2S and C4A3$ main peak o he BCSA clinke s as a unc ion o he syn hesis empe a u e. F om an o e all examina ion o he XRD da a abou he syn he ic BCSA clinke s i was ound ha ha , a e e y in es iga ed empe a u e, he con e sion o eac an s was comple e and C2S and C4A3$ we e he main bu ning p oduc s. Fu he mo e, C4AF, C2AS and C$ equen ly occu ed as seconda y componen s. Tempe a u e, °C 1150 1200 1250 1300 1350 Peak in ensi y, cps 0 500 1000 1500 2000 2500 CLIR CLI1 CLI2 CLI3 Tempe a u e, °C 1150 1200 1250 1300 1350 Peak in ensi y, cps 500 1000 1500 2000 2500 3000 3500 4000 4500 CLIR CLI1 CLI2 CLI3 Figu e 3 illus a es he XRD pa e ns o mix u es MR, M1, M2 and M3 hea ed a he bes syn hesis empe a u es. The ou BCSA cemen s ob ained om he syn he ic clinke s gene a ed by RM, M1, M2 and M3 a hei bes syn hesis empe a u e we e espec i ely deno ed wi h he symbols CEMR, CEM1, CEM2 and CEM3. Figu e 3. XRD pa e ns o he BCSA clinke s ob ained om MR, M1, M2 and M3 a hei bes syn hesis empe a u e. Legend o symbols: Y=C4A3$, A=C$, B=C2S, G=C2AS, B =C4AF. Table 5 epo s he comp essi e s eng h de elopmen o BCSA-based mo a s as a unc ion o cu ing ime; a all he in es iga ed pe iods, he comp essi e s eng h alues o CEMR, CEM2 and CEM3 based-mo a s we e simila o each o he ; on he con a y, he comp essi e s eng h alues o CEM1 we e cons an ly lowe han hose o he o he cemen s. Table 5. Resul s o comp essi e s eng h measu emen s o CEMR, CEM1, CEM2 and CEM3 a a ious aging pe iods. Days CEMR CEM1 CEM2 CEM3 1 12.70.4 8.70.5 10.90.3 12.20.3 2 19.20.3 14.50.2 16.10.5 19.70.2 7 26.40.2 20.40.6 24.50.5 27.90.5 14 32.50.1 24.60.3 30.50.8 33.50.2 28 37.10.5 26.90.1 34.50.7 36.80.4 56 36.90.7 28.40.5 35.10.8 37.10.3 The comp essi e s eng h esul s e e o h ee p isms (six de e mina ions) The esul s o he expansion–sh inkage es s a e illus a ed in Figu e 4. Figu e 4: Dimensional s abili y cu es o BCSA-based cemen s (ai and wa e cu ed). They indica e ha all he in es iga ed BCSA cemen s di e ed e y li le om each o he , bo h when subme ged unde wa e and cu ed in ai . In pa icula , unde wa e he maximum expansion alues, eached a e abou 14 days o cu ing, a e comp ised in he na ow ange o 0.17-0.23%. When cu ed in ai , he in es iga ed pas es showed a con inuous sh inkage ill 14 days when a minimum leng h change is eached (-0.06%, -0.05%, -0,08 and -0.07% o CEMR, CEM1, CEM2 and CEM3, espec i ely); since ha pe iod he alues emained cons an o all he in es iga ed sys ems. The change o mine alogical phases wi h ongoing hyd a ion was de e mined h ough XRD and DT-TG in es iga ions. The XRD analysis o all he BCSA cemen s e ealed, as expec ed, ha a e 4 hou s o hyd a ion e ingi e had al eady o med a he expense o pa o calcium sul oalumina e and calcium sul a es. A 28 days o cu ing e ingi e was s ill he main c ys alline phase o hyd a ed cemen s which also e ealed he p esence o some aces o C4A3$ and ine phases (e.g. gehleni e). A he same cu ing pe iod, s ä lingi e (C2ASH8), which s a ed o ming al eady a e 1 day o cu ing in CEMR and CEM1, was e iden oge he wi h ka oi e (C3AH6) in all he in es iga ed sys ems. These compounds had espec i ely o med om beli e (as silicon sou ce) and AH3 (as aluminium sou ce) and om he emaining beli e and C2ASH8 acco ding o he ollowing equa ions [Winne eld e al., 2016]: C2S+AH3+5H ⇒C2ASH8 (2) C2S+ C2ASH8 ⇒ CSH2+C3AH6 (3) Mo eo e , AH3 was no de ec ed a any cu ing pe iod due o i s amo phous na u e. No signi ican changes we e obse ed a 56 days o cu ing in he ou hyd a ed sys ems. As an example, XRD pa e ns o CEMR, CEM1 and CEM3 hyd a ed a 4 hou s, 1 and 56 days a e epo ed in Figu e 5. Tempe a u e (°C) 010 20 30 40 50 Leng h change (%) -0,10 -0,05 0,00 0,05 0,10 0,15 0,20 0,25 CEMR ai cu ed CEM1 ai cu ed CEM2 ai cu ed CEM3 ai cu ed CEMR wa e cu ed CEM1 wa e cu ed CEM2 wa e cu ed CEM3 wa e cu ed Figu e 5. XRD pa e ns o CEMR (le ), CEM1 (middle) and CEM3 ( igh ) hyd a ed o 4 hou s, 1 and 56 days. Legend o symbols: E=C6A$3H32, Y=C4A3$, S=C2ASH8, K=C3AH6, B=C2S. The DT−TG analyses almos con i med he mine alogical e alua ions made by XRD. Th ee endo he mal e ec s we e obse ed and a ibu ed, on he basis o li e a u e da a [Taylo , 1997], o he ollowing compounds: e ingi e (E) and, con a y o XRD da a, calcium silica e hyd a e (CSH) and aluminum hyd oxide (AH3); in pa icula , CSH, E and AH3 we e espec i ely de ec ed h ough he ollowing dehyd a ion endo he mal peaks: 103°±4 °C, 152°±5°C, 277°±3°C. S ä lingi e could no be iden i ied inasmuch as i s DT peak was o e lapped by he e ingi e one. (a) c (b)