Alkaline Activation of Metakaolin_ Effect of Calcium Hydroxide in the Products of Reaction

Alkaline Activation of Metakaolin:Effect of Calcium Hydroxide in the Products of ReactionMaria Luz Granizo,Santiago Alonso,Maria T.Blanco-Varela,and Angel PalomoEduardo Torroja Institute(CSIC),28033Madrid,Spain The alkali activation of metakaolin(MK)leads to the produc-tion of high-mechanical-performance network-structure mate-rials.Adding calcium hydroxide(Ca(OH)2)to the raw MKproduces a somewhat different reaction:a network structure and C-S-H gel form.In the present study,MK and(MK؉Ca(OH)2)mixes were activated with5M and12M NaOHsolutions and cured at45°C.A5M concentration,in theabsence of Ca(OH)2,did not produce MK activation within thetest time.An activator concentration of12M resulted in complete activation and the formation of a network structure.When Ca(OH)2was present in the raw mix,a small amount ofC-S-H gel formed.I.IntroductionW HEN metakaolin(MK)is activated by alkali hydroxide solutions,a polycondensation reaction occurs,leading to the formation of a cementitious material with high mechanical strength.1–4The starting aluminosilicate and the compositional parameters,as well as the curing time and temperature,directly influence the final properties of the material that forms.5The synthesis of this binder can be compared with zeolite synthesis, although,in contrast to zeolites,the product is X-ray amorphous.Based on infrared(IR)and magic-angle spinning nuclear magnetic resonance(MAS NMR)results,we concluded that the structure of the binder is a three-dimensional framework formedby SiO4and AlO4tetrahedra linked alternately by shared O atoms,with alkali cations from the activator solution present to balance the negative charge of the tetrahedrally coordinated Al3ϩion.6–11When Ca(OH)2is mixed with MK,the activation reaction isdifferent:the products formed consist of a mixture of the material with the network structure described and a C-S-H gel,in propor-tions that are dependent on the reaction parameters.This observa-tion is in good agreement with the thermodynamical and micro-structural data found in the literature.11–15Thus,the aims of the present work were(i)to study the effect of modifications in the alkaline-activation process of MK resulting from variations in the alkali concentration(both with and without the addition ofCa(OH)2)and(ii)to compare and chemically characterize theproducts obtained in both cases.II.Experimental Procedure(1)MaterialsThe present study used a Spanish kaolin(C1,Guadalajara, Spain).A complete chemical and mineralogical characterization of the kaolin was given in a previous work.5Table I shows a chemical analysis of the kaolin,which contained quartz as a secondary species.The MK used as a raw material for synthesis experiments was prepared by thermal treatment of the kaolin at750°C for 24h.16The chemical composition of the prepared MK is shown in Table II.The Ca(OH)2was a chemical grade appropriate for laboratory use.(2)ProcedureMK and mixtures of(MKϩCa(OH)2)with a1:1ratio were activated with5M and12M NaOH solutions.We used a conduc-tion calorimeter system(JAF from Wexham Developments,Berks, U.K.)to examine the reaction of MK with the NaOH solutions. Calorimetric data were obtained isothermally at45°C(at temper-aturesϽ45°C,reactions occur too slowly18)with a fixed solution: solids ratio of0.7(both5M and12M NaOH solution concentra-tions were prepared).Samples removed from the calorimeter were treated with a mixture of acetone and ethanol(for only a few minutes,to avoid any dehydration effect),dried under vacuum to a constant weight, and treated with diluted HCl(1:9)and a solution of5%Na2CO3. (The attack with HCl and Na2CO3dissolves the products of alkaline activation,whereas the initial MK remains insoluble. Previous tests were conducted to check the effect of temperature and HCl concentration on MK.Insoluble residue was analyzed through IR spectroscopy until no alteration was observed.)This chemical attack at room temperature makes it possible to quantify the unreacted MK;the insoluble residue obtained is a measure of the degree of reaction attained.To determine silicon,aluminum,and sodium content,the products resulting from the alkali activation of MK,as well as the solids obtained from the acid treatment,were analyzed by induc-tively coupled plasma(ICP)spectroscopy(Model70ϩ,Jobin Yvon,France),after alkaline melting with lithium metaborate.The amount of H2O fixed by the samples was determined by differen-tial thermal analysis/thermogravimetry(DTA/TG)(Model STA 409,Netzsch,Selb,Germany).The amount of water-soluble sodium was analyzed by ion chromatography(Model ED40, Dionex,Sunnyvale,CA).The products obtained from activating the mixes of MKϩCa(OH)2also were treated with methanol and salicylic acid to dissolve the C-S-H gel while retaining the network structure material and the unreacted MK in the residue.The attack of the methanol-salicylic acid dissolves calcium silicates;17however,the network structure material and the unreacted MK remain unalteredM.Grutzeck—contributing editorManuscript No.189201.Received September22,1999;approved November22, 2000.Funding for this study provided by CICYT,through Project Nos.MAT93-0282 and MAT96-0598.Table I.Chemical Analysis of RawKaolin C1Oxide/component Content(%)SiO249.85Al2O336.34Fe2O30.57MgO0.15CaO0.00Loss on ignition,LOI12.65J.Am.Ceram.Soc.,85[1]225–31(2002) 225in this alcoholic medium(this was proved in the laboratorythrough preliminary tests).Measuring the SiO2in the solutionprovides the amount of SiO2that is fixed as C-S-H gel.At thesame time,measuring the compositional oxides in the residue provides quantitative data on the oxides ratio in the network-structure material.SiO2in the organic solvent,as well as the SiO2,Al2O3,and NaO in the residue,was measured by ICP.The sampleswere further analyzed by Fourier transform infrared(FTIR)spec-troscopy(ATIMATTSON Genesis,Madison,WI).Figures1and2 are schematic diagrams of the analytical procedure.III.Results(1)Metakaolin Activation in the Absence of Ca(OH)2 MK activation with the12M NaOH solution was followed by means of conduction calorimetry(see Fig.3).With[NaOH]ϭ5M, the paste did not harden(this seems to imply that activation was not achieved);consequently,no heat-evolution peaks appeared during the calorimetry test.The curve in Fig.3shows an initial sharp and intense peak on contact for the MK–NaOH solution anda second peak with a maximum atϳ0.5h.The small degree ofreaction at an activator concentration of5M was confirmed by theFTIR results:the IR spectrum of the samples showed bandscorresponding to the starting MK(at400,800,and1050cmϪ1)(see Fig.5).With an activator concentration of12M,the IRspectrum revealed the disappearance of the Al O O stretching bandat800cmϪ1,belonging to MK;the appearance of a new band at ϳ720cmϪ1,which is characteristic of a network structure;and a shift in the1050cmϪ1band of the MK toward lower frequencies(Ͻ1000cmϪ1)(see Fig.5).Table III shows the percentage of insoluble residue(see Fig.1,step(2))for samples activated with both5M and12M concentra-tions of NaOH.These values confirm that little MK activationoccurs when[NaOH]ϭ5M;however,when[NaOH]ϭ12M,MKdissolution is almost complete.For this reason,only productsobtained with12M concentrations were analyzed and character-ized in the present study.Table IV shows the ICP results for silicon,aluminum,andsodium(see Fig.1,steps(6B),(7B),and(8B));the insolubleresidue in the acidic medium(Fig.1,steps(3B)and(4B));water-soluble sodium(Fig.1,step(8C));and the amount of H2Ofixed in the sample(Fig.1,step(5)).Table V shows the relativecomposition of the alkali-network structure(see Fig.1,steps(9)–(12))formed when MK was activated with[NaOH]ϭ12M.(2)Metakaolin Activation in the Presence of Ca(OH)2 Figure4shows the heat-evolution-rate curves,as a function of time,obtained when MK activation was conducted in the presence of Ca(OH)2.In this case,an activator concentration of5M led tocomplete activation within the test time(the samplegained Fig.1.Scheme of the analysis procedure of metakaolin activated with alkaline dissolution in absence of calcium hydroxide.Table II.Chemical Analysis of RawMetakaolin Mk1Oxide/component Content(%)SiO257.00Al2O341.55Fe2O30.58MgO0.17CaO0.00Loss on ignition,LOI0.00226Journal of the American Ceramic Society—Granizo et al.Vol.85,No.1rigidity),and the reaction was recordable by calorimetry.Both curves (5M and 12M NaOH)present only one peak,because the first signal is produced so quickly (at the time of the contact between solid and liquid)that there is not sufficient time for the system to achieve the thermal equilibrium.Consequently,the calorimeter is not capable of detecting most of this initial infor-mation,which is lost.The differences between these two calori-metric curves may indicate differences in the nature of the products formed,as a function of the activator-solution concentra-tion.Generally,the IR spectra of these samples (see Fig.6)con-firmed a shift in the 1050cm Ϫ1band to lower frequencies (atϳ980cm Ϫ1),already observed by Palomo and Glasser.6However,differences were observed between the materials obtained by the 5M activation and the 12M activation:the appearance of sharp bands between 660and 720cm Ϫ1—similar to those present in network structures such as zeolites —when the activator concen-tration was 12M ,and the presence of a shoulder at ϳ1080cm Ϫ1(indicating the presence of unreacted MK)when the activator concentration was 5M .In Table VI,the percentage of unreacted MK in the presence of Ca(OH)2is expressed as an insoluble residue (see Fig.1,step (2B)).The higher the activator concentration,the more the reaction progresses (the more MK is dissolved).The ICP results fortheFig.2.Scheme of the analysis procedure of metakaolin activated with alkaline dissolution in presence of calcium hydroxide.January 2002Alkaline Activation of Metakaolin:Effect of Ca(OH)2in the Products of Reaction227insoluble residue (see Fig.2,steps (3B),(3C),(4B),and (4C))are shown in Table VII.The residue amounts (in percent)obtained when the samples were treated with methanol –salicylic acid (Fig.2,step (6))are shown in Table VIII.Table IX shows the percentage of SiO 2extracted as C-S-H gel when the sample was treated with methanol-salicylic acid,the unreacted SiO 2,and the percentage of inorganic polymer formed for each activator con-centration (see Fig.2,steps (5B),(3C),and (3E)).The main reaction product was an alkali containing network structure when the activator concentration was 12M ,whereas the main product was a C-S-H gel when the activator concentration was 5M .Previous NMR studies 11,12confirm these results by interpreting the signals at Ϫ84and Ϫ82ppm,respectively,of the 29Si spectrum of 5M -activated sample such as Si [Q2]and Si [Q2(1Al)](the presence of NaOH increases the solubility of Al(OH)4Ϫand allows the substitutions).An intense peak at Ϫ86ppm,corre-sponding to a three-dimensional network [Q4(4Al)],appears in the 29Si spectrum of the 12M -activated sample.At the lower NaOH concentration,formation of the alkali material was negligible.Thus,quantitative analysis of the alkali-network structure was conducted only for those samples activated with the solution of 12M NaOH.An analysis of the residue obtained when the samples were treated with methanol-salicylic acid (Fig.2,steps (7B),(8B),(9B),and 6(B))is shown in Table X.The percentages of oxides constituting the alkali material (Fig.2,steps (10)–(13))are shown in Table XI.IV.Discussion(1)Metakaolin Activation in the Absence of Ca(OH)2The reaction steps describing the alkaline activation of MK were explained in previous investigations.18These steps may or may not overlap,and even are or are not recordable,as a function of the experimental parameters (temperature)and the nature of the raw MK (specific surface).The activation of the present MK with a 12M NaOH solution occurred in a way similar to that previously cited:(i)an initialdissolution peak that was not recordable because of the rapid MK dissolution,(ii)an induction period (almost nonexistent in this case because of the high rate of reaction),and (iii)a signal correspond-ing to the formation of the alkali-network structure.In the present case,the high [OH Ϫ]in the liquid phase accelerated the reaction,giving the aqueous phase a high ion concentration that provoked immediate product precipitation.When the activator concentration was 5M ,however,MK dissolution was very slow,the threshold concentration for precipitation of the reaction products was not attained within the test time,and the heat-evolution rate was not recordable.The material did not harden but rather remained in the plastic state.This slow dissolution rate was confirmed by the insoluble-residue result,which showed that only 20%of the starting MK had been activated after 24h.(2)Metakaolin Activation in the Presence of Ca(OH)2As in the absence of Ca(OH)2,the amount of insoluble residue was high;however,in this case,the paste hardened.Thus,the presence of Ca(OH)2favored the formation of some products that gave rigidity to the paste.ICP analysis of the filtrates obtained when the samples were treated with methanol –salicylic acid (see Table IX)demonstrated that C-S-H gel formed in the presence of Ca(OH)2with both high and low NaOH concentrations.Table IX reveals that C-S-H gel was the main reaction product when activation was conducted with the lowest NaOH concentration (5M ),whereas the amount of inorganic alkali material was very low.In contrast,when the NaOH concentration was 12M ,the main product that formed was the inorganic alkali material,although 20%of the initial SiO 2was fixed as C-S-H gel.This difference in the products formed when Ca(OH)2was present in the system,as a function of activator concentration,resulted from the high [OH Ϫ]in the aqueous phase under high [NaOH].The [OH Ϫ]impeded Ca(OH)2dissolution (during the first stages),and because an insufficient number of Ca 2ϩions existed to achieve saturation,no C-S-H gel formed.In that case,the prevailing reaction was the formation of the alkali-network structure.The formation of this hydrated aluminosilicate provoked a pH increase,favoring Ca(OH)2dissolution and C-S-H gel formation.Both processes overlapped,so their calorimetric signals were not separate,and only one signal appeared that was associ-ated with both processes.However,when [OH Ϫ]decreased (to 5M ),Ca(OH)2solubility was not as retarded.Consequently,the coexistence of sodium and calcium ions in solution resulted in an association between the reactions for C-S-H-gel formation and for alkali-network structure formation.In this particular case,condi-tions favored the formation of a C-S-H gel.The formation of an inorganic alkaline polymer in both the presence and the absence of Ca(OH)2was confirmed through IR spectroscopy (see Figs.5and 6).Adding Ca(OH)2to the raw MK resulted in a higher concen-tration of ions in the solution,which produced lower mobilityandFig.3.Rate of heat evolution of metakaolin activated with 12M NaOH vstime.Table III.Insoluble Residue Results of Metakaolin Activated with NaOHSampleInsoluble residue content (%)5M NaOH80.012M NaOH 11.2Table IV.Results of Analysis of the Sample as Well as theSolid Obtained in the Insoluble-Residue TreatmentSampleContent (%)Soluble Na 2O content (%)Na 2OSiO 2Al 2O 3H 2O12M (sample)184028.311.0 2.7612M (insoluble residue)64.225.5Table V.Chemical Analysis of Polymer Obtainedin the Absence of Calcium HydroxideSampleContent (%)SiO 2Al 2O 3Na 2OH 2O12M 34.1525.9717.9611.00228Journal of the American Ceramic Society —Granizo et al.Vol.85,No.1delayed the polycondensation reaction (calorimetric curves re-vealed a longer delay in the formation of the alkali material).The slowest product precipitation in this case indicated that the aqueous phase was saturated and the MK reaction slow,as shown by the insoluble-residue value.When the NaOH concentration was 5M ,MK dissolution was very slight in both cases;the [OH Ϫ]was not high enough to provoke high dissolution.When Ca(OH)2was present,the paste was rigid at 24h because of the small proportion of C-S-H gelformed,whereas it remained plastic when only MK was present because little alkali material formed.Tables XII and XIII show the chemical composition,in terms of molar ratio of oxides,for the phases that formed.The network structure formed was the same in both cases,although it formed in a minor proportion with respect to the initial reactant amount when Ca(OH)2was present,because an important excess of this hydrox-ide remained unreacted.The Al:Na ratio was ϳ1in the product,indicating that the network structure that formed had no OH Ϫgroups in its structure and that all of the Al 3ϩcharge was balanced by Na ϩcations fixed inside its structure.The Si:Al ratio also was ϳ1,indicating that the silicon environment corresponded to a tetrahedral silicate coordinated with four Al atoms.The general approximate formula for the alkali material formed both with and without Ca(OH)2is Si 2Al 2Na 2H 4O 10.The analytical methodology used in both cases consisted of a selective-solutions method that made it possible to separate the material formed from the unreacted MK to analyze the chemical composition of the alkali material.On the other hand,this material from the C-S-H gel that formed when Ca(OH)2was present in the raw mix made it possible to quantitatively determine the proportions of all of the products formed in the reaction,thus providing for their later characterization.V.Conclusions(1)An inorganic alkali-network structure material formed in the present study when the concentration of NaOH was 12M.Fig.4.Rate of heat evolution of metakaolin ϩCa(OH)2activated with 12M and 5M NaOH vs time.Table VI.Insoluble Residue Results in the Presence ofCalcium HydroxideSampleInsoluble residue content (%)5M 34.212M 7.8Table VII.Oxide Content in the Insoluble-Residue Solid inthe Presence of Calcium HydroxideSampleSiO 2content (%)Al 2O 3content (%)Step 3BStep 3CStep 4B Step 4C5M 20.280.611.362.112M 4.2617.0 2.4813.6Table VIII.Residue Content of Samples inMethanol –Salicilic AcidSampleResidue content (%)5M 39.612M 45.0Table IX.Breakdown of SiO 2Fixed in theProducts FormedSampleComposition (%)SiO 2(CSH gel)Unreacted SiO 2SiO 2(alkaline polymer)5M17.580.6 1.812M 20.417.062.6Table X.Oxide Content of the Methanol –SalicilicAcid ResidueSampleContent (%)SiO 2Al 2O 3Na 2OH 2O12M 19.613.9 6.910.6Table XI.Chemical Analysis of Polymer Formed in thePresence of Calcium HydroxideSampleContent (%)SiO 2Al 2O 3Na 2OH 2O12M 15.411.5 6.910.6January 2002Alkaline Activation of Metakaolin:Effect of Ca(OH)2in the Products of Reaction229In the presence of Ca(OH)2,C-S-H gel was produced readilywith a low NaOH concentration;however,the formation of an alkali material as the main product required a higher activator concentration.(2)The reaction rate of alkali material formation was very low in the presence of Ca(OH)2because the ions were less mobile when the [OH Ϫ]of the solution was higher.(3)The main product formed both with and without Ca(OH)2was the same network structure,with the general approximate formula:Si 2Al 2Na 2H 4O 10.(4)The analytical methodology used in the present study provides a method for separating the products formed underdifferent conditions,to quantitatively determine the proportions of all of the products formed in the reaction,thus providing for their later characterization.References1J.Davidovits,“Minerals Polymers and Methods of Making Them,”U.S.Pat.No.4472199,1982.2J.Davidovits,“Geopolymers:Inorganic Polymeric New Materials,”J.Therm.Anal.,37,1633–56(1991).3M.L.Granizo,M.T.Blanco-Varela, F.Puertas,and A.Palomo,“Alkali,Activation of Metakaolin:Influence of Synthesis Parameters ”;in Proceedings of the Tenth International Congress on the Chemistry of Cement ,Vol.3(Gothenburg,Sweden,1997).Fig.5.FTIR spectra of metakaolin activation in the absence of Ca(OH)2((a)metakaolin,(b)metakaolin with 5M NaOH,and (c)metakaolin with 12MNaOH).Fig.6.FTIR spectra of metakaolin activation in the presence of Ca(OH)2((a)metakaolin ϩCa(OH)2,(b)metakaolin ϩCa(OH)2with 5M NaOH,and (c)metakaolin ϩCa(OH)2with 12M NaOH).Table XII.Elemental Molar Ratios,in theAbsence of Calcium HydroxideSampleMolar ratio Si/AlAl/Na Si/NaSi/H12M 1.110.870.980.46Table XIII.Elemental Molar Ratios,in thePresence of Calcium HydroxideSampleMolar ratio Si/AlAl/Na Si/NaSi/H12M 1.14 1.01 1.150.48230Journal of the American Ceramic Society —Granizo et al.Vol.85,No.14P.Krivenko,“Alkaline Cements:Terminology,Classification,Aspects of Dura-bility”;in Proceedings of the Tenth International Congress on the Chemistry of Cement,Vol.4(Gothenburg,Sweden,1997).5M.L.Granizo,M.T.Blanco-Varela,and A.Palomo,“Influence of the Starting Kaolin on Alkali-Activated Materials Based on Metakaolin:Study of the Reaction Parameters by Isothermal Conduction Calorimetry,”J.Mater.Sci.,35,1–7(2000). 6A.Palomo and F.P.Glasser,“Chemically-Bonded Cementitious Materials Based on Metakaolin,”Br.Ceram.Trans.J.,91,107–12(1992).7E.M.Flanigan,H.Khatami,and H.A.Szymanski,“Molecular Sieve Zeolites”; pp.201–229in Advances in Chemicals Series,Vol.101.American Chemical Society, Washington,DC,1971.8J.G.S.Van Jaarsveld,J.S.J.Van Deventer,and A.Schwartzman,“The Potential Use of Geopolymeric Materials to Immobilise Toxic Metals:Part II.Material and Leaching Characteristics,”Mineral.Eng.,12,75–91(1999).9E.Lipmaa,A.Samoson,and M.Magi,“High-Resolution27Al NMR of Alumi-nosilicates,”J.Am.Chem.Soc.,108,1730–35(1986).10J.Davidovits,“Geopolymers:Inorganic Polymeric New Materials,”J.Mater. Ed.,16,91–138(1994).11S.Alonso, A.Palomo,I.Sobrados,and J.Sanz,“Activacio´n Alcalina del Metacaolı´n en Presencia de Ca(OH)2:Efecto del Activador.Estudio por NMR,”pp. 79–80in VI Reunio´n Nacional de Materiales.Edited by M.Elizegui,I.Garuz,and F.Pen˜alba.San Sebastian,1999.12P.Faucon,T.Charpentier,A.Nonat,and J.C.Petit,“Triple-Quantum Two-Dimensional27Al Magic Angle Nuclear Magnetic Resonance Study of the Aluminium Incorporation in Calcium Silicate Hydrates,”J.Am.Chem.Soc.,120,12075–82(1998). 13P.S.De Silva and F.P.Glasser,“Hydration of Cements Based on Metakaolin: Thermochemistry,”Adv.Cem.Res.,3,167–77(1993).14M.Grutzeck,A.Benesi,and B.Fanning,“Silicon-29MAS NMR Study of Calcium Silicate Hydrates,”J.Am.Ceram.Soc.,72,665–68(1989).15X.Cong and R.J.Kirkpatrick,“29Si NMR Study of the Structure of the Calcium Silicate Hydrate,”Adv.Cem.Bas.Mater.,3[]144–56(1996).16M.Murat,“Hydration Reaction of Hardening of Calcinated Clays and Related Minerals.II.Influence on Mineralogical Properties of the Raw Kaolinite of the Reactivity of Metakaolin,”Cem.Concr.Res.,13,511–18(1983).17S.Takashima,Sement Giujitzu Nempo,Vol.XII;p.12,1958.18M.L.Granizo and M.T.Blanco-Varela,“Alkaline Activation of Metakaolin: Isothermal Conduction Calorimetry Study,”J.Therm.Anal.,52,957–65(1997).ⅪJanuary2002Alkaline Activation of Metakaolin:Effect of Ca(OH)2in the Products of Reaction231。