PREPARATION OF IRON RICH CEMENTS USING RED MUD

w

C

Pergamon

andCoocreteResearch, Vol. 27, No. 7, pp.1037-1046,1997 Copyright01997 ElsevierScienceLtd Printedio theUSA. All rightsreserved 0008-8846/97$17.00+ .00

PII SOO08-8846(97)O01 O1-4 PREPARATION OF IRON RICH CEMENTS USING RED MUD

Maneesh Singhl

TataResearchDevelopmentand DesignCentre, 1,MangaldasRoad,Pune411 001,India S.N. Upadhayay Departmentof ChemicalEngineeringand Technology, BanarasHinduUniversity,Varanasi221005, India P.M. Prasad

Centreof AdvancedStudy Departmentof MetallurgicalEngineering, BanarasHinduUniversity,Varanasi221005, India (Refereed) (ReceivedJuly 16, 1996;in final formMay 14, 1997) ABSTRACT

Possibility of producing calcium sulfoaluminoferrite(SAF) (Cq(A,F)3~)calciumaluminoferrite(C4AF)basedcementsusinglime+ red mud + bauxite + gypSUIDhas been investigated. The effectsof composition,firing time and firing temperatureon the propertiesof cementsproducedhas been studied. The characteristicsof the cementsproducedhave been found to be strongly dependent on the raw mix compositionand firing temperature but not so much on firingtime. Someof these cementspossessstrengthscomparableto and at times even more than ordinaryPortlandcement (OPC). Sincethe red mud used containssignificantamountof titania,effect of titaniaon pure sulfoaluminatephasehas alsobeen studied.01997 Elsevier Science Ltd Introduction

In the last few decades, calcium sulfoaluminate(C4A33)phase has attracted considerable attentionas a componentof cementsexhibitinghigh earlyday strengthsand havingultimate strengthas well as dimensionalstabilitysimilarto that of ordinaryPortlandcement (OPC) (1-17). Calciumsulfoaluminatecan includein its structurea numberof impurities,for example A13+can be replacedby Ti4+,Cr3+,Mn3+,Fe3+,Sf+,etc. Whenthe raw mix containsa significant proportionof iron oxide,apart fromthe formationof ferriteand aluminoferritephases, ‘Correspondingauthor.Presentaddress:Divisionof Process Metallurgy, Lulezi University of Technology,S-97187,Lule&Sweden. 1037

1038

M. Singhet al.

Vol. 27, No. 7

partial substitutionof aluminaby iron oxide in sulfoahuninatetakes place. This is possible becausethe ionicsize of Fe3+is only20% largerthan A13+(Fe3+= 0.675°Aand A13+= 0.535 ‘A). Such a replacementleadsto the emergenceof a solid solutionof CqAs~ (calciumsulfoaluminate,CSA)-C4FS~ (calciumsulfoferrite,CSF) which may be designatedas calcium sulfoalurninoferrite(C4(A,F)S ~, SAF)phase.The extentto whichFezOscan substituteA1203 (18-21). is not known with certaintyand differentresearchersclaim it to be between9-20’Yo Accordingto Krivoborodovet al. (20) in the systemCaO-AlzOB-FezOs-CaS04 apart horn the formationof Cq(A,F)3~(where maximumsubstitutionof A13+by Fe3+can be 9’%0), some amountof S042”also gets fixed to calciumaluminoferrites.This results in the formationof two other series of cements-highbasicitysulfoferrites(GF.C ~-2CzF.C~) and C4A,FI.~~n where x varies between 0.3-1.6and n between0.4-1.0. On hydration sulfoaluminateforms ettringite(CsA.~C~~2H) and monosulfatehydrate (CSA.C~.12H)dependingupon the proportionof C4ASS-CS-Ca(OH)z-HzO(11-13). Likewise, on hydrationboth the high and low basicity calcium sulfoferritegive two hydrated products: CsF.C~.12H (monosulfate) and CsF.3C~.32H (ettringite). According to Krivoborodovet al. (20) in case of Cq(A,F)s~the hydrationrate duringthe first three days falls with the increasein F/A ratio but later it becomesequal to the hydrationrate of sulfoaluminate,as a resulttherewouldbe no visibleeffectin the latterdays.On the otherhand, in the case of calciumaluminoferritesthe presenceof sulfateenhancesthe hydraulicactivities of the solidsolutionsas comparedto the basicaluminoferritephases(18-21). The sulfoaluminoferritebased cementshave a numberof advantagesover ordinaryPortland cement. These include-energyconservationdue to lower clinkeringtemperature and formationof sofierclinkers,abilityto use industrialwastes/lowgradetraditionalraw materials, good strength characteristics,good anti-seepageproperties,good corrosionresistance, goodbehaviorat low temperatureand rapid settingcharacteristics.Investigationscarriedout in the recentyears have shownthat admixingof C4(A,F)S ~ with OPC leadsto an improvement in strength,reductionin porosityand increasein corrosionresistancecharacteristicsof the cement(18-20). Red mud is the causticinsolublewasteresiduegeneratedat the aluminarefinerieswhich use the Bayer process(pressurecausticdigestionof bauxite).1.0-1.6tons of it is generated per ton of aluminaproduced,as a resultit is estimatedthat in Indiaand Worldover 2 and 70 million tons of it is generated respectively.It is a complexmaterial whose chemical and mineralogicalcompositionsvarywidelydependingon the bauxiteand the localplantprocess parameters.It containssignificantquantitiesof FezOj,AIzOS,SiCh,TiO,, NaZOand CaO and trace quantitiesof numerousotherelementssuchas V, Ga, Cr, Zr, U, Th, Sc, La, Y, Sr, Ba, Hf, K, Pb, Mn, Cd, Ni, Zn, B, Bi, Co and Li (as oxides).Moreover,red mudsconsistof 1421 mineralphases(22-23). Out of differentindustrialwastesavailable,red mud was chosenfor the presentwork due to its high alumina and iron oxide as well as low silica contents.Low silica is necessary since in presence of C~ silicareacts to form undesirable2C2S.C=. This paper deals with the preparationof specialcementsusing lime,red mud, bauxiteand gypsum.The effects of differentprocessparameterslikecompositionof raw mix, fwingtemperatureand firingtime on the phase formation,strength,density and color were investigated.Since the red mud used contains significantproportion of titania, the effect of titania on CAAs~ was also studied.

Vol. 27, No. 7

REDMUD,CEMENTPRODUCTION,SULFOFERRITE

r

1039

TABLE 1

ChemicalAnalysisof Raw Materials(wt.’XO)

e203

A1Z03 Si02

ao

Lime

Gypsum

Red Mud

Bauxite

0.65 0.65 1.00 67.13

0.13 0.07 0.89 37.41

33.1 18.2 8.8 2.7 19.6 5.8

16.5 48.0 3.0 0.5 8.5

i02 a20

Soq

0.27 53.35

Materials and Methods

. As receiveduncausticizedred mud and bauxitefromM/sHINDALCOIndustries Ltd., Renukootwere used. Red mud and bauxitelumpswere crushedand/orground, dried and sieved to -150 pm. Commerciallyavailablehydratedlime and gypsum,obtainedfrom localmarketwere of-200 ~m size.The chemicalcompositionof the fourraw materialsused is givenin Table 1.

To prepareeach cementsample, 100g of the raw materialsin predeterminedproportionsweretakenand ballmilledfor 45 min. for homogenization.The resultantmix was made into a thick pasteusing an appropriateamountof water and molded into a 5 cm cube. The cube was dried overnightin a hot air oven at 100°Cand then fired (clinkered)in an electrictimace. An averageheatingrate of about 200°C/hwas employedand the firingtemperaturemaintainedat a predeterminedlevel(1150-1350°C)up to 2 h duration.The clinkerso preparedwas cooledovernightin the fimace itself to room temperature,groundin agatemortarandpestleand sievedthrough150~ mesh. To determinethe effectof chemicalcomposition,two setsof sampleswere fired (Table2) at 1250”Cfor 2 h. The amountsof red mud and bauxitewere variedin stepsof 50A.To study the effect of firing scheduleon the qualityof cementproduced,two representativesamples of each set (Table 3) were fired at different temperatures (1150, 1200, 1250, 1300 & 1350°C)for fixed time (2 h) and for varyingduration(0.5, 1.0, 1.5 & 2.0 h) at a fixed temperature(1250’’C). TABLE2 Composition(wt.%)of VariousSamplesFiredat 1250”Cfor2h.

Series A Series B

.

Lime

Gypsum

Red Mud

Bauxite

47.5 32.5

7.5 12.5

0-45.0 5.0-50.0

45.0-0 50.0-5.0

1040

M. Singhet al.

Voi. 27, No. 7

TABLE3 Compositions(wt.%)of RepresentativeSamplesFiredat Different Temperatures(for2h) and for VaryingDuration(at 1250°C).

A4 A7 B3 B8

Lime

Gypsum

Red Mud

Bauxite

47.5 47.5 32.5 32.5

7.5 7.5 12.5 12.5

15.0 30.0 15.0 40.0

30.0 15.0 40.0 15.0

for C~essive Stre@2. Smallsizedcylindricalpelletswere madeby compacting 5 g of each cementsamplemixed with 2.0 ml of water in a die of 1.6 cm I.D. using a 2 kg weight(put on top of the plungerfor 5 min.).Pelletsso preparedwere kept on and covered with wet piecesof clothfor a day and then immersedin water for 27 days.Each samplewas then ground to a height of 1.5 cm on emery paper. Compressivestrengthsof such pellets were tested using an AmslerUniversalTestingMachine.In orderto get some relative idea, pelletsmade of OrdinaryPortlandCement(OPC,IndianStandardGrade33)were alsotested in a similarfashion.Due to the non-standardmethodof the testingthe compressivestrength the valuesobtainedgaveonlya relativeideaand not the actualstrengthvalues. To determinethe phasesformedon firingof the raw mixes and on hydration,typicalsampleswere groundto -75 ~m andXRD patternsrecordedusingRigakuGeigerflexXRD machineemployingNi-filteredCu Ku radiationat 40 kV and 25 mA. The Fe203 /A1203,,:.

L

Fe203/A1203 -

+~

!lm’

G (A)

01 0

1

1

I

1

10

20

30

40

“ARED MUD

1 (B)

% RED MUD

FIG. 1. Effect of red mud proportionand F/A ratio on the 28-days strength of cement pellets of SeriesA and SeriesB compositions.

Vol.27, No. 7

REDMUD,CEMENTPRODUCTION,SULFOFERRITE

1041

complexcompositionof the samplesmadethe quantitativeevaluationof differentphasesby XRD quite difficult.The largenumberof phasespresentin any sampleresultedin overlapping peaks. Moreoverthe diffractionpeak shiftsalso occur due to dopingeffectscausedby the presenceof impurities.Therefore,onlya qualitativeidentificationof the phaseswas done and no attemptwas madeto quanti@the phasespresent. Mkmstnictural Studks. Test pieces of hydrated cement samples were fixed on bronze mounts and coated with a thin layer of gold-palladiumalloy using Edwardssputtercoater. The coated specimenswere then examinedemployinga JEOL-840ScanningElectron Microscope(ModelJSM-840A). Results and Discussion

.. (a) Effectof Composltlcm The clinkersproducedof SeriesA compositionwere lightbrown in color, soft and friable. With increasingred mud content (hence F/A ratio) their color became darker and densityas well as strengthincreased(Fig 1). XRD studiesshowedthat there was an increasein the formationC4AFand C2Fphaseswith correspondingdecreasein the C4(A,F~S phasewithincreasein the F/A ratiO. Thoughthe raw mixes of SeriesA contained5.2-11.4’Yo TiOz,the XRD patternsof these cements did not indicatethe presenceof any titanatephases.To understandthe absenceof such phases raw mixes were made using analyticalgrade CaO, A1203and FezOJtaken in a molar proportionCaO:A1203:FeZOJ= 4:1:1to whichwas addedvariedamountof analytical grade TiOZ(O-1O?4O). They were then fired in the fashiondescribedearlier.The phase identificationstudies(XRD)indicatedno formationof titanatephases.Insteadthere seemedto be completeuptakeof TiOzin aluminateand ferritephases.The substitutionof AIZOJand FezOJ by TiOzwouldnot changethe basiccrystalstructuresof C,AF or CZFphasesbut only distort it resulting in a shift of peak positionsdependingupon the degreeof doping.This was observed in all the cases;e.g. the strongestpeak of CZFwas observedat d = 2.681, 2.685 and 2.688 for TiOz= O%,4°Aand 8°Arespectively.Simikirwork carried out using CaO, A1203, CaSO, (CaO:A120J:CaSO, = 3:3:1)and O-lOOA TiOzshowedthat till 3°Aadditionthere was absence of titanate peaks and beyond that CT peaks were present. In the case of C4Ax~ phase increasingshiftsin peak positionwith titaniadosagewas recorded(peak positionsof McI= 422 were at 3.748, 3.762 and 3.764 for TiOz= O%,4% and 8Y0respectively).This explainsthe absenceof CT phasein the seriesA, in whichall of the titaniagot doped in the ferriteand C,(A,F)J~ phases. The clinkers prepared out of Series B compositionswere hard and brownish-blackin color. The hardness, “glassy” nature and density of clinkers increased with the red mud content. Unlike the previous set, the strength of the cement samples (B1-B1O)decreased with the increasein red mud content(Fig. 1). Furthermore,the samplesB7-B1Oon ~uring exhibitedsignificantcrackingand deformationdue to excessiveexpansion.Cq(A,Fj3S and CT were found to be the major phases formed in these cements.The fall in strengthwith increasein red mud contentwas attributedto the increasein CT with correspondingdecrease in C4(A,F)33. In the case of cements of Series B there was no peak correspondingto ferrite or aluminoferrite phases, which is surprising. A part of Fe@J got used UP in the formation of C.I(A,F)J~. In this set of experimentsthe peak positionof hkl = 422 of B3 (F/A = 0,5), B5 (F/A = 0.7) and B8 (F/A=l.l) was observedat 3.764,3.762and 3.753respectively,implying

1042

M. Singhet al.

Vol. 27,No. 7

AL

A = C4A35

-L ?. 5 % LIME GYPSUM -7, S % RED MUD-i5 . 0’.4 BAUXITE -30.0%

B=

CLAF c a CT

B

D = C2 AS

A ‘..

B3,

A

LliiE -32 ‘ +5 ‘% GYPSUM = 12.5% RED MUD - 15.0% BAUXITE - 40.O”L

.,

.,

.,

. D

I

I

20

,,

,i

D

I

30

40

50

60

7)

2e FIG. 2. XRD patterns of A4 and B3 fwedat 1250”Cfor 2 h. a progressiveincreasein doping(18). Suchpeak shiftswere also observedin CT for which the highest intensitywas recorded at 2.705, 2.703 and 2.698 respectively.The bakmce of FezOJ was found to form “glassy”phase (incorporatingFe20J)which could be observed evenvisually. Further experimentscarried out using varyingquantitiesof lime, gypsum,red mud and bauxite showedthat the cementscontainingC2Fand C4AFformedtill gypsumcontentwas lessthan 10’%o in the raw mix.Beyondthat, cementwithoutthesephasesbut thosecontaining CT formed. Cements made using fly ash (as source of alumina) instead of bauxite too showed similartrend, on the other hand when red mud was replaced by iron ore fines (as source of iron oxide and aluminabut withouttitania)this was not observed.This suggests the criticalinfluenceof sulfatecontentin the presenceof titania. XRD analysisof hydratedsampleA7 attributedthe strengthdevelopmentto the formation of CJ(A,F).3C~ .32H(ettringite)and C3(A,F)H,5. The hydrationof B5 cementresultedin the formationof ettringiteand unhydratedCT, whereason hydration,B8 gave monosulfate hytite (C3(A,F).CE.13H)and unhydratedCT. The S. E. microscopicstudy on the hydratedcementsamplesA7 and B5 showed formation of ettringitecrystalsof differentstructures(Fig. 3). Whereasthe ettringiteformed on hydrationof A7 was in the formof longneedle-likecrystals,that formedon hydrationof B5 was in formof compactbundles.The formertypeof structureis obtainedwhen the ettringite crystals sprout in large cavities,thereby filling the water filled space by long and thinner needles.The latteris obtainedwhenthe ettringiteis formedin restrictedspace(3). This may explainthe largeexpansionobservedin the caseof the samplesof B series.

Vo].27, No. 7

REDMUD,CEMENTPRODUCTION,SULFOFERRITE

1043

(B)

(A)

FIG. 3. ScanningElectronMicrographsof hardenedpastes(A7 and B5) curedfor 28 days. The strengthsof cementsof SeriesA (A4 and A7) were observedto increasewith the firing temperaturetill 1300”C(Fig. 4). At 1350°Cincipient fbsionof clinkerswas noticed.The phaseidentificationstudies(XRD)indicatedthat such an effectwas dueto an increasedformationof C4AFat the expenseof CZF,whilethe proportion of C4(A,F)3~ remainedunaffected. The strength of cementsof Series B (B3 and B8) monotonicallydecreasedtill 1250”C and then increased.XRD studyof B8 showedthe reasonfor fall in strengthto be due to the increasedformationof CT whilethe amountof CA(A,F)3 ~ phaseremainedthe same. .. . ect of ~. All the cementsof both series(A4,A7, B3 and B8) had maximum strengthwhen fwedfor 1-1.5h (Fig.4) at 1250YlWhilethe amountof C4(A,F)3~ remained unaffectedin all the cases,in case of SeriesA the amountof C4AFand CJAwere maximum between 1-1.5h whereas in SeriesB the fall in strengthfor 2 h of firingdurationwas dueto the increasedformationof nonhydraulicCZAS(gehlenite)phase. 5oo-

1

500

1

0 “ A4

4

*E

.Q 400r

o o

I

A? B3

) ‘ /-

d A?

‘E <400

/’

Be

300 -

/

x

I

A4

o

-

0

I

o

B3 B8

I

300 -

/

< 200 -\

200 -

/

.~-

-%

--

< OPC

)

100

o

-1150

I 1200

A

.\\ liso

1300

FIRING TEMPERATURE (“C)

0 0.5

1.5

Flf&G

TIME (h)

FIG. 4. Effectof firingtemperatureandtime on 28-daysstrengthof cementpellets,

2.0

1044

M. Singhet al.

_

200

Vol. 27,

1

,

,

1

1

1

t 6

1 8

No. 7

‘g

‘A 100-

50

0 024

10

‘X TITANIA FIG. 5. Effectof TiO*on 28-daysstrengthof C.A,~. . . As the HINDALCOred mud containssignificant feet Of Tltama On Pure Pk. quantity of titania, the effect of titania doping in C4AJ~ (preparedfrom analyticalgrade reagents)was studied.The clinkersobtainedwere soft and friable.Whilethe sampleswithout titania were white in color, and those containingtitania exhibitedlight to deep cream color, the latter being the case with high-titaniacements. It was also observed that the strengthof cured sampleswas maximumfor the cementcontaining4°Atitania(Fig. 5). The XRD phase identificationstudy indicatedabsenceof peaks of CT till 3% TiOzand beyond that increasedformationof CT with increasein the titaniacontent.Titanianot only formed CT but also substituted the alumina to some extent in C4AJ~. On hydration while T1 (O%TiO,) and T3 (4% TiO~)hydratedto monosulfateand ettringite;the CT present in T3 remainedunhydrated.

I

600

500

1

400 —

I 300

TEMPERATURE

I

200

I

100

0

( “C)

FIG. 6.

DTG of hydrated cements of phase showing fall in the dehydration temperature of monosulfatehydrateon increasein titaniacontent.

Vol. 27,No. 7

REDMUD,CEMENTPRODUCTION,SULFOFERRITE

1045

FIG. 7. SEM of TI and T3 hydratedsamplesshowingreductionin the size of monosulfatecrystals on additionof titania. Fig. 6 shows the role of titania on the thermal behaviorof hydrated samples. (2) The samples showed rapid mass loss at two regions. The first in the range 50-110”Cand the secondin the range 205-255°C.The first one is characteristicof ettringiteand the secondis of monosulfatehydrate.Whilethe presenceof titaniadid not seemto significantlyalter the temperatureat which ettringitedehydrated,the temperatureat whichmonosulfatelost water of hydrationdecreasedslightlywith increasein titania content.This may be due to the decreasein the size of monosulfatecrystalson additionof titania(Fig.7). Conclusions

The followingconclusionsmay be drawnffomthe presentinvestigationson the preparation of iron-richcementsusingred mud as a componentof the raw mixes: 1) It is feasible to produce cementspossessingacceptable28 day strengthusing lime + gypsum+ red mud + bauxite,but the propertiesof cementsstronglydependedon the proportionof the ingredients. 2) The phasesformedon clinkeringdependon the percentageof gypsum.When it is 7.5% the major phasesthat form are C.4A,F)J~, C4AFand C2F,and when it is increasedto 12.5’%. the majorcrystallinephasesformedhavebeen foundto be Cd(A,F)J~and CT. 3) The strengths of the cement samples increasewith the firing temperatureand abnormally high strengths(comparedto OPC) are obtainedat 1300”C.At 1350”Cincipient fusionis noticed. 4) A f~ing durationof 1-1.5hat 1250°Cyieldscementshavingthe higheststrengths. 5) The presence of titania in C4AJ5 increasesthe strengthdevelopmentof cementpaste, lowersthe dehydrationtemperatureof monosulfatehydrateand effectsthe morphological structureof the hydrates. Acknowledgment The authors are gratefil to Dr. E.C. Subbarao (Executive Director) and Dr. Pradip of TRDDC,Pune for their help, constantencouragementand constructivecriticismduringthe

—...—.

.-

1046

M Singhet al.

vol. 27, No. 7

courseof this work. They are thankfulto the Head of fhe Departmentof MetallurgicalEngineering, Banaras Hindu University,Varanasifor providingthe laboratoryfacilitiesto conductthese investigations. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23.

P.K.MehtaandA. Klein,J. Am.Ceram.Soc.,48,435 (1965). V. Satavaand O. Veprek,J. Am. Ccram.Soc.,58,357 (1975). P.K.MehtajCementand ConcreteResearch,6, 169(1976). Y.B.Pliego-CuervoandF.P.Glasser,CementandConcreteResearch 8,455 (1978). C. Tashiro,J. ObaandK. AlmnNL Cementand ConcreteResearch,9,303 (1979). P.K.Meh@WorldCementTechnology,21.166 (1980). G. Sudoh,T. Ohta and H. Harad&Proc. 7th InternationalCongresson Chemistryof Cements,5, 152(1980). J. Deng, W. Ge, M. Su and X. Li, Proc. 7th InternationalCongresson Chemistryof Cements,5, 381(1980). K. Iked&Proc.7th InternationalCongresson Chemistryof Cements,3,31, (1980). K. Ogawaand D.M.Roy,CementandConcreteResearch,11,741 (1981). K. Ogawaand D.M.Roy,Cementand ConcreteResearcL 12, 101(1982). K. OgawaandD.M.Roy,Cementand ConcreteResearch,12,247(1982). W. Kurdowskiand F. Sorrentino,Structureand Performanceof Cements,Applied SciencePublishers,LondonandNew York,471 (1983). W. Yanmouand D. Jun’an,Proc. 8th InternationalCongresson Chemistryof Cements,2, 300, (1986). I. Kapralikand F. Hanic,CementrmdConcreteResearch,19,89 (1989). F. Hanic,I. KapratikandA. Gabrisov4CementandConcreteResearch,19,671 (1989). F. Hanik,I. KapralikandM. Drabik,Silikaty,33,203 (1989). A. Zdorovand V. Bemshtein,lst NCBInternationalSeminar,Proceeding>6,82 (1987). A.P. Osokin,Y.R. Krivoborodovand N.F. DukovL9th InternationalCongresson Chemistryof Cements,Proceedings,3,256 (1992). Yu R. Krivoborodovand S.V. Samchenko,9th InternationalCongresson Chemis~ of Cements, Proceedings,3,209 (1992). Y. Xi, 9th InternationalCongresson Chemistryof Cements,Proceedings,4,377 (1992). P.M. Prasad and J.M. Sharm~ Proceedingsof Electrometallurgy,Cent. ElectrochemicalRes. Inst.,Karaikudi,12,(1986). P.M. Prasad, Second InternationalConferenceon Aluminium(INCAL-91),Proceedings,AluminiumAssociationof IndiajBangalore,117(1991).