Crystallization of Silica Gel in Alkaline Solutions at 100 To

American Mineralogist, Volume 74, pages 1147-1 151, 1989

Crystallization of silica gel in alkaline solutions at 100 to 180 'C: Characterizationof SiOr-Y by comparisonwith magadiite Hlluro Muursnr Department of Chemistry, Fukuoka University of Education, Akama, Munakata, Fukuoka 8 I I -41, Japan

AssrRAcr Sior-f, an intermediate product of the transformation of silica gel to quartz, was synthesized from KOH or NaOH solution containing SiO, and NaCl with a molar ratio of SiOr:HrO:Na*:OH-equal to l:100:2-4:0.5at 100-180'C. In almostall reactionprocesses, SiOr-)z appearedas the first crystal phaseand then was transformed into SiOr-Xr, cristobalite, and f,nally quartz. The characterizationsbased on the chemical and physical data for SiOr-/ showedthat Sior-y is magadiite (NarSi,oOrr.xHrO,x: 9-ll), which was discovered at Lake Magadi (an alkali lake), Kenya, and has been considered to be an important precursor of inorganic bedded cherts.

INrnooucrroN et al. (1973).Lagaly et al. (1975) have sinceshown that Hydrothermal crystallization of amorphous silica was natural and synthesized magadiite is a sodium silicate investigatedin the 6inary systemSiOr-HrO at high tem- with the compositionNarSi,oOrn'11HrO. perature and high pressure,and the transformatlon via We also obtained the product identified as SiOr-/in some metastable phases (cristobalite, keatite) to quartz Naollt KgH, and NarCO, solutions containing NaCl at 100-180'C pressure was observed6eat, teS+; Carr and Fyfe, 1958).When at the autogenous of an autoclaveat alkali-metal or alkaline-earth salt was added in aqueous a given temperature.In this paper, the transformation of media, the transformation to quartz occurred at lower amorphous to crystal phasesunder hydrothermal condi- temperatures with various intermediate phasesthat al- tions was further examined on the basis of the run du- ration,temperature,andconcentrationofalkali-metaland ways contain some alkali ions and ro u.. not crystal- phasesin the one-componentSiO, system. alkaline-earth salts. The preparation condition of SiOr-Y Heydemann (1964) observedthat amorphous silica in was determined from the experimental observations,and KOH solution crystallizesvia SiOr-Xand cristobalite to its synthesiswas attempted. The nature of SiOr-I was qurirrzat 100-250 "C, and characterizedthe nature and determined and compared with that of the other silicates. propertiesof SiOr-X. Mitsyuk et al. (1973, 1976) de- Theresults showedthat SiOr-Yis indeedmagadiite. The details chemicalproperties and X-ray scribedmany intermediate phasesthat appearat 150-2g0 -ofthe the difraction .C when alkali-metal or alkaline-earth salt is added to an data of SiOr-Xr, from which it was determined to be a alkaline medium. Sior-y, which is one of the interme- quasi-kenyaite, will be reportedin a subsequentpaper. diate crystal phases,was formed in 0.5-2o/oKOH solutions containing NaCl in the temperaturerange 150-180 ExpnnrnanNTAl- DETAILS 'C at 0.5-4.0 kbar. This is a sodium silicate that has the Hydrothermal experiments on the transformation of compositionNarO'2lSiOr'l2HrO-NarO.29SiOr.l4HrO, silicagel werecarried out in the temperaturerange 100- dependingon the nature of the reaction medium. This 180 'C in l-2 wto/o(0.18M4.36M, M = mol/L) KOH compoundwas found to have a high capabilityfor cation solutionscontaining 0.2, 0.5, 1.0, and 2.0M NaCl. We exchange.Kitahara et al. (1982)and Muraishi and Kita- usedTeflon bottleswith a volume of 15, 50, and 100cm3 hara ( 1986) observedthe effectsof seedcrystals, some asreaction vessels. Usually a I 5-cm3bottle wasused, and metal oxides, and alkali-metal salts on the crystallization 0.3 g of starting silica (Wako's silica gel Q-22, through of silica gel in KOH solution at temperaturesbelow 300 200 mesh) was put in the bottle together with 7.5 mL of oC, under which conditions SiOr-X, SiO2-Xr,and SiOr-I the solution. The Teflon bottle was closed with a Teflon mainly formed as intermediate phases. plug, and it was placed in a stainless-steelautoclave that On the other hand, Eugster (1967) described new hy- was designedto accommodate the bottle. After the au- drous sodium silicates,the mineral magadiite [reportedly toclave was kept at a given temperature for a definite NaSi'O'r(OH)3'3HrO)l and kenyaite [NaSi,,O,o,(OH).. time, it was quenchedrapidly with water. The reaction 3HrO)] from Lake Magadi, Kenya. It is thought that ma- products were taken out of the autoclaveand washedwith gadiite precipitates from alkaline lacustrine brines and NaOH solution (pH 10), since SiOr-I has good cation converts to kenyaite and finally to quartz (chert). Its syn- exchangebelow a pH of 10. Some of the conditions for thesishas been reported by McCulloch (1952) and Lagaly preparing SiOr-y are describedin Table 2. 0003-{04xl89/0910-1 I 47$02.00 lI47 r 148 MURAISHI: CRYSTALLIZATION OF SILICA GEL IN ALKALINE SOLUTIONS

z 200'c a -+- c o s0 (h)

-+- I 80'c c -:-. o 50 (h) --:-

10 (d) -:- o z a 1400c o o '. 02168 0 s l0 (d) Reoctiontime / d Reoctiontime

Fig. L Effectof NaCl concentrationin l0loKOH solutionin Fig. 2. Efect of temperature for a l0/o KOH solution con- thecomposition of SiOr:OH-:Na*: 3:l:l.l-11.0at 160'Con taining NaCl of 2 moUL on the crystallization sequence defined the crystallizationsequence of SiOr-I (4, SiOr-X,(Xr), cristo- in Figure I as a function of reaction time. balite(C), and quartz(Q) as a functionof reactiontime.

Magadiite was synthesizedby heating mixtures of the lower temperatures(e.g., below 120 "C) in the l0loKOH- ratio of 9 mol of SiOr, 2 mol of NaOH, and 75 mol of NaCl or 0.5MNarCO, solutions.When NaOH is usedas HrO at 100 "C for 4 weeks,according to the method de- an alkali sourceinstead of KOH (i.e., using NaOH-NaCl scribedby Lagaly et al. (1975).As it is known that ma- solutions),a well-crystallizedSiOr-I is synthesized.When gadiite as well as SiOr-I exhibits a cation-exchangeca- KOH is used instead of NaCl (i.e., using KOH-KCI so- pability, synthetic magadiite was also treated with the lutions), SiOr-X, is directly formed and transformed into same method describedabove for SiOr-f. cristobalite, which then goesto quartz. This differenceis The reaction products were identified by the use of a understandablefrom the fact that SiOr-I is a sodium Rigaku xcc-20 X-ray difractometer. Chemical compo- silicate and SiOr-X, is formed as a combination of both sition of the products was determined by conventional sodium and/or potassium silicatesas describedbelow. gravimetric methods for silica and by atomic absorption Figure 2 shows the effect of temperature on the crys- for Na and K. The water content was measured bv a tallization of silica gel for a loloKOH solution containing thermogravimetric method. 2MNaCl. It has been reportedthat SiOr-yis formed in a temperaturerange of 150-180'C under high pressure Rnsur,rs AND DrscussloN (Mitsyuk et al., 1973).In our investigation,the period Preparation conditions of SiOr-Y and rnagadiite over which SiOr-Ywas stablewas found to increasewith It was observedthat SiOr- y, SiOr-Xr, and cristobalite, a decreasein the temperature, and it was synthesizedeven as intermediate crystal phasesappearing in the order list- at 100 'C. Comparatively well-crystallizedSiOr-I was ed, form in the 100-180 "C experimentsin lo/o(0.18M) obtained after a reaction time of I to 3 months at 100 KOH solutions containing various amounts ofNaCl. The 'C. Sior-y was stable for several months and slowly final product is quartz. The results of the reaction se- transformed into quartz. Thus, a lower temperature was quenceof 160'C experimentsare shown in Figure l. In advantageousfor the formation of SiO'-Icompared with this figure (and in Fig. 2 also), the schematicline breadth SiOr-Xr. In addition, the amount of solid phase in the for each phasewas determined on the basis of the inten- alkaline solutions was followed through the reaction sity of a main X-ray diffraction peak. Increasingthe NaCl term. The solid phasedecreased in the initial stageof the concentration induces the formation of SiOr-Y and cris- reaction and then increasedwith reaction time up to con- tobalite. Further addition of NaCl seemsto result in pre- stant amounts. These results suggestthat the transfor- venting SiOr-X, from appearing, in which situation mation processof silica gel to SiOr-I may be a dissolu- SiOr-Y is directly followed by cristobalite. The same ex- tion-precipitation process. perimental results were also observed for the reaction at The appearanceof SiOr-Ivaried as a function of NaCl MURAISHI: CRYSTALLIZATION OF SILICA GEL IN ALKALINE SOLUTIONS tr49

ct z

6 I Y,Xz E o

(J o z Y,Xz

o o X2 100 200 300 400 o Molorrotio HzO/OH- Fig. o 4. Chemical composition of the solutions required for u X2 c the formation of SiOr-Iat 100-150'C and magadiiteat 100'C. o The amount of silica in the solutions is SiO,/OH- : 2-6. (J x2 x2 x2 XXT X

100 120 1t0 160 180 Temperqture ('C ) and natural magadiitefrom Brindley (1969).As seenin Table l, there is little difference between the X-ray dif- Fig.3. Therange ofSiOr-Yappearance as a functionofNaCl fraction peaksof natural magadiite and those of the syn- concentrationand temperaturefor a l0loKOH solution.(The thesizedmaterial. The slight differencemay be causedby crystalsindicated with smallletters are minor phasesas deter- a variety of impurities included in natural magadiite. The minedby X-ray diffractionpatterns.) (X); SiOIX other abbre- patterns of the two kinds of synthesized silicates-ma- viationsas in Figure1. gadiite and SiOr-I-are similar to each other except for the differencesin intensity, 1, resulting from their differencesin crystallinity. concentration and reaction temperature for a 10/oKOH Thermal changesin SiOr- Y and magadiite were deter- solution,as shownin Figure3. SiOr-Yappearsseparately mined by meansof differential thermal analysis(ore) and in the dotted region and is more stable under conditions thermal gravimetric analysis (rce). The DrA-rcA curves of lower temperatureand higher NaCl concentrations.In for both silicatesdrying in air are shown in Figure 5, and the other region, Sior-f forms together with the other the resultsactually overlap each other. It was found from intermediatephases or doesnot form at all. the rc.l curves that the dehydration occurred mainly in Figure 4 shows the composition of the solutions in four steps. The dehydration at temperaturesbelow 100 which SiOr-Iand magadiitewere prepared.For the prep- 'C may be assignedto adsorbedHrO and some water of aration of SiOr-I, a relatively large amount of water was crystallization, which are easily released. When these used, and the concentration of Na* in the solution was sampleswere stored in a desiccator using silica gel for a made to vary independently againstthat of OH-. This is desiccant,the ignition loss below 100'C was no longer different from the procedure to synthesize magadiite, observed.At about 130 and 190'C, some 500/oto 90o/o which was done at 100-120 oC in suspensionstypically of the total ignition loss was released.The bulk of the containing I mol of NaOH,2 to 6 mol of SiOr, and l0 containedwater was lost by heatingto 550'C. The water to 60 mol of HrO (Lagaly et al., 1973). content in the preparations varied slightly with the con- The resultsof the conversion of crystalline phaseswith ditions ofsynthesisand storage.On the other hand, orn time (Figs. I afi,2) show the following reaction path: curveshave a small broad endothermic peak at about 680 'C amorphous silica - SiOr-I. * SiOr-& - cristobalite - and an exothermic peak at about 750 "C. The former qtafiz. Sometimesit was observedthat the crystal growth was assignedto the destruction of the lattice of SiOr-I of SiOr-Iand SiOr-Xr,however, did not proceedconsec- (or magadiite) and to the formation of quartz, and the utively, but simultaneously, and that cristobalite was latter was assignedto its reconstitution into tridimite. formed prior to the appearanceof SiOr-Xr. If theseresults Table 2 summarizes the molar ratio of Na and Si in haveany fundamentalsignificance, the reactionpath ought some of the SiOr-I and magadiite samplesprepared un- to be expressedby the processincluding the parallel re- der various conditions. From these results, the chemical actions in which Sior-y and SiOr-X, are independently formula of SiOr-Iwas determinedto be NarSi,oOrn.xHrO. rather than consecutivelyformed. It was deduced from ignition loss that the amount of water of crystallization,r, equaled 8.5-ll. It is worth X-ray diffraction and chemical compositionof products noting that Sior-y was usually free from K*. This char- The X-ray diffraction peaksof the SiOr-Yand synthetic acteristic of SiOr-y distinguishesit from SiOr-Xr, which magadiite prepared by us are given in Table I and com- can contain as structural atoms both Na and K in anv pared with those of Sior-y from Mitsyuk et al. (1973\ proportion. l 150 MURAISHI: CRYSTALLIZATION OF SILICA GEL IN ALKALINE SOLUTIONS

o U I 0 5 10 U 0.10M HCt ( mt ) 0 200 400 500 800 Fig.6. Typicaltitration curvesof SiO,-Y andmagadiite sus- Temperoture( .C ) p"nrionrwith 0.10MHCl. Suspension,0.5 g of silicateper 50 mllh; Fig. 5. Thermal behavior of synthesizedSiO,-Y (solid lines) mL of water;rate of additionof HCl, 1.0 temperature, and magadiite (dashed lines). (a) WeightJoss curves; (b) orn 25.0"C. diagram. most all of it at pH 4.4 + 0. l. This is closely analogous to that of crystalline sodium polysilicateNarO'8SiOr' Properties gHrO investigatedby Iler (1964).The result provesthat Sior-y and magadiite were suspended in water after these sodium silicates [abbreviatedas Na-(silicate)] have drying in air and grinding into fine powders through 200 high exchangeabilityin acidic solution and high selectiv- mesh to determine ion-exchange properties. The suspen- ity for H* and then suggeststhat the following reaction sions were titrated continuously with 0.lMHCl solutions goesto completion: using an automatic titrator with a rate of addition of 1.0 Na-(silicate)+ H+ - Na* * H-(silicate), mllh. The results are shown in Figure 6. The titration curves of both silicates showed that half of the alkali in where H-(silicate) is crystalline silicic acid from sodium the crystal is neutralized at approximately pH 7, and al- silicate.The pH value of 4.4, which is the pH of the

Trele 1. X-ray diffractionpatterns of SiO,-Y and magadiite sior-y Magadiite

d(A) d(A) d(A) d(A) d(A)

15.50 100 15.78 90 15.64 oc 15.56 100 15.64 b5 7.76 20 7.80 16 7.76 15 7.77 I 7.90 12 6.26 c 7.26 8 5.17 30 s.20 23 5.20 23 5.18 18 5.22 20 5.03 13 5.00 14 (4 44), 4.48 IJ 4.48 20 4.46 18 4.48 13 4.23 10 +.zJ a 4.O2 13 4.01 8 4.O2 10 3.877 10 3.920 10 3 637 30 3.641 21 3.648 23 3.626 13 3.663 23 3.550 J.CCU 33 3.544 19 3.569 23 3.450 90 3 453 100 3.445 100 3.434 IJ 3.460 100 3.304 80 3 308 oc 3.308 oc 3.298 40 3.314 oc 3.197 11 qn J. IOU J. ICJ 81 3.151 74 3.145 60 3.159 72 2.989 2.998 I 2.864 2.876 2.866 b 2.880 8 2.831 10 2 827 9 2.823 2.82',1 10 2.831 8 2 639 2.578 10 2.585 2.603 2 599 11 2.600 (2.36) z,\tJo b 2.353 n 2.353 12 2.350 I 2.289 10 2.287 5 2.273 5 2.012 10 2.067 12 1.937 10 1.833 30 1.830 1.826 18 1.822 30 1.833 Nofe;Columns are as follows:(1) 155'C, in KOH + NaClsolution (Mitsyuk et al., 1976).(2) 150 rc, 5 days,in 1% KOH + 1.5M NaClsolution (this 'C, work). (3) 100 'C, I weeks, in 1oh KOH + 1M NaCl solution (this work). (4) Natural magadiite(Trinity County, California)(Brindley, 1 969). (5) 100 8 weeks,in the solutionwith molar ratioof SiOr:NaOH:HrO: 9:2:75 (thiswork). MURAISHI: CRYSTALLIZATION OF SILICA GEL IN ALKALINE SOLUTIONS 1151

TABLE2. Reactionconditions and analysisof products

Molar ratioof solution Atomic ratio of oroduct Run' rfc) Time(d) OH Na* sio, H.o Na Si Y-1 140 0.25 1.36 50 1.91 14 0.03 Y-2 150 5 0.32 1.80 100 1.94 14 0.02 Y-3 140 0.64 1.36 100 1.98 14 0.03 Y-4 150 0.32 5.41 100 2.01 14 0.01 Y-5 140 0.341 1.70 5U 1.99 14 M-1 100 bU o.221 0.22 8 2.03 14 'Y, SiO,-Y;M, magadiite. t NaOHwas usedas an alkalinesource

solution at the equivalencepoint in Figure 6, corresponds Iler, R.K. (1964) Ion exchangeproperties ofa crystalline hydrated silica. to the concentration of hydronium ion formed by the Journalof Colloid Science,19, 648-657. (1954) reaction of the resulting silicic acid from Keat, P.P. A new crystalline silica. Science,120,328-330. ion exchange Krtahara,S, Matuie, T, and Muraishi, H. (1982)Some observation on with water: the crystallization of silica gel under hydrothermal conditions at 150 'C. H-(silicare) + (silicate) to 300 In S. Somiya, Ed, Proceedingof the First International + HrO H3O* + . Symposium on Hydrothermal Reactions,p. 480-495. Gakujutsu Bun- The ionization constant plq of the silanol groups on the ken Fukyu-kai, Tokyo. Lagaly,G., Beneke,K, and Weiss,A (1973)Organic complexes of syn- crystallinesilicic acid was estimatedto be about 6.9 on thetic magadiite. Proceedingsof the Intemational Clay Conference, the basisof the conceptof weak acid solutions.This is in Madrid 1972,p. 663-673. good agreementwith some estimatesbased on titration - ( 1975)Magadiite and H-magadiite: I. Sodium magadiiteand some curvesfor amorphoussilica, which are about 6.5 to 7.7 of its derivatives American Mineralogist, 60,642-649 (1952) (Schindlarand Kamber, 1968; McCulloch, L. A new highly silicious soda silica compound. Jour- Strazenkoet al., 1974). nal ofthe American Chemical Society, 74, 2453-2456. Both SiOr-Iand magadiite,synthesized by us, have a Mitsyuk, B.M., Gorogodskaya,L.I , and Rastrenko,A.I. (1973) Inter- refractiveindex of 1.481+ 0.003.This value is the same mediate forms of silica obtained in hydrothermal synthesisof quanz. as that reported for natural magadiite by Eugster(1967), Doklady Akademia Nauk SSSR,209,926-928. - who gavean averagevalue of 1.48. (l 976) The nature and propertiesofnew silica varieties.Geokhim- ia,1976,803-814. Therefore, from the results previously stated,it is con- Muraishi, H., and Kitahara, S. (1986) Efects of seedcrystals, some metal cludedthat SiOr-f is indeedmagadiite. oxides, and alkali metal salts on the crystallization of silica gel in the aqueoussolulions ofpotassium hydroxide at 160to 200'C. In S. So- RrrnnnNcns crtpo rniya,Ed , Hydrothermalreactions, vol. l, p. l-13. Uchida Rokakuho, Tokyo Brindley, G.W. (1969) Unit cell of magadiite in air, in vacuo and under Schindlar, P., and Kamber, H R (1968) Acidity of silanol groups. Hel- other conditions AmericanMineralogist, 54, 1583-1591 veticaChimica Acta, 51, 178l-1786. Cam,R.M., and Fyfe,W S. (1958)Some observations on rhe crystalliza- Strazhesko,D.N, Strelko,V.B., Belyakov,V.N., and Rubanic, S.C. (1974) tion of amorphous silica. American Mineralogist, 43, 903-9 I 6. Mechanism of cation exchangeon silica gels Joumal of Chromatog- Eugster,H P ( I 967) Hydrous sodium silicatesfrom Lake Magadi, Kenya: raphy,102,191-195. Precursorsof bedded chert. Science,I 57, I I 77-l I 80. Heydemann, A. (1964) Untersuchungeniiber die Bildungsbedingungen von Quarz im Temperaturbereichzwischen 100 "C und 250 "C. Bei- MlNuscnrpr RECEIvEDAucusr 9, 1988 fidge zur Mineralogie und Petrologie, 10,242-259 Maxuscnrpr ACcEPTEDMmcu 6. 1989