2086
RUSTUMROYAND E. F. OSBORN
Vol. 71
1 CONTRIBUTION FROM SCHOOL OF MINERAL INDUSTRIES, THEPENNSYLVANIA STATECOLLEGE]
The System Lithium Metasilicate-Spodumene-Silica' BY RUSTUM Roy2AND E. F. OSBORN
Introduction
Related to compositions in the system LizOA1203Si02 are the lithia micas of the lepidolite series. During recent attempts to synthesize members of this s e r i e ~the , ~ desirability of having phase equilibrium data on systems containing Li20, AlzOs, and Si02 became evident. Besides contributing toward a solution of the problems mentioned above, therefore, data for the system L~zO-A~~OS-S~OZ can serve as a basis for understanding the more complex mixed alkali-aluminasilica systems. I n the present work, phase equilibrium data were obtained for the ternary system, lithium metasilicate - spodumene - silica. The location of this system within the system Li~o-Al2OaSiO~ is shown in Fig. 1. The data of Kracek,' of H a t ~ hand , ~ of Bowen and Greig,6 have established the relations as shown along the joins LiD-SiOs, LizO.Al203, 3 0 2 , and A1203-Si02, respectively. Inasmuch as lithium is a monovalent cation as are sodium and potassium, but has an ionic radius similar to that of magnesium, i t might be expected that the system lithia-alumina-silica would bear a family resemblance both to the alkalialumina silica systems? and to the magnesia-alumina-silica sysSuch is the case. The series of compounds LiAlSiOd, LiAlSi200 and LiAlSia0~,~have analogs Weight per cent. in both the soda and DotFig. 1.-The system Li20-A1208Si02, after Kracek,' Hatch,' Bowen and Greigd and ash systems. In all o i t h e Roy and Osborn. alkali systems, the field of monotropic, and spodumene cannot be obtained mullite approaches closely but does not cross the from the high temperature form. Crystals of join on which the ternary compounds appear. The eucryptite composition have been made, but these B-spodumene solid solution series does not have a are apparently a high temperature form structur- counterpart in the other alkali systems but resem(3) Rusturn Roy, "Decomposition and Re-synthesis of the Micas." ally different from the mineral. Glass of the eupaper presented before the 50th Annual Meeting of the Am. Ceram. cryptite composition has a density and refractive SOC., Chicago, Ill., April, 1948. index greater than that of the corresponding crys(4) (a) F. C. Kracek, J. Phys. Chcm., 84, Pt. 11. 2645 (1930); tals-an interesting phenomenon still not satis- (b) F. C. Kracek, THISJOURNAL, 61, 2870 (1939). ( 5 ) R. A. Hatch, A m . M i n . , 28, 471 (1943). factorily explained. (6) N. L. Bowen and J. W. Greig, [4] 7, 242 (1924).
The ternary system Li20-AlzOs-SiOz is noteworthy for the many interesting, but still unsolved, compositional and structural problems which i t contains. The three minerals petalite, spodumene and eucryptite appear in the system. Of these petalite has not been synthesized, and its chemical formula is in doubt. Spodumene, likewise, has not been synthesized but a high temperature polymorphic form can be crystallized from a glass of the spodumene composition. The same form also appears on heating spodumene above about 700'. The transformation, however, is
(1) Abstracted from a thesis submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy at The Pennsylvania State College. (2) Government of Bihsr (India) scholar.
(7) J. F. Schairer and N. L. Bowen, A m . J , Sci., 240, 196 (1947). (8) G. A. Rankin and H. E. Merwin, A m . J . Sci., 141 46, 322 (1918). (9) LiAlSiaOi is a suggested formula for petalite.
June, 1949
THESYSTEM LITHIUMMETASILICATESPODUMENE-SILICA
2087
all samples with crystals were examined with:X-ray bles the cordierite solid solution series in the sys- most techniques. A General Electric XRL) unit with a phototem MgO-AI203-Si02. It is of interest in this con- graphic recording camera of diameter 143.2 mm. and Fe nection that one of our colleagues has found that &(A = 1.936) radiation, and a Xorelco Geiger counter extensive solid solution exists between P-spodu- type X-ray spectrometer with CuKa ( A = 1.53736) radiamene and cordierite.1° The wide application of tion were used, the latter much more than the former. mixtures in the MgO-Al203-Si02 system in the Properties of the Crystalline Phases and Correceramic industry would lead one to expect an insponding Glasses creasing use of mixtures in the L ~ ~ O - A I Z O ~ S ~ O ~ Lithium Metasilicate.-Lithium metasilicate system, probably in the same fields. (Li20?3iOZ) appears in this system as well-formed, uniaxial positive, negatively elongated, prismatic Method of Investigation crystals with parallel extinction. Indices measThe raw materials used in the preparation of the m k - ured were: ENa = 1.590, UNa = 1.610. tures were lithium carbonate, alumina and silicic acid. Lithium Disi1icate.-Crystals of lithium disiliThe lithium carbonate was obtained from a freshly opened bottle of J. T. Baker Analyzed Chemicals with the cate (Li~0.2Si02)usually form in the mixtures only significant impurities being 0.200% Soland 1.200% studied as subrounded, subequant prisms. A other alkalies. It was dried a t 300' in a platmum beaker, distinct micaceous cleavage is noticeable especially and stored in a desiccator. The alumina used was anhydrous aluminum hydroxide, in large crystals. Elongation is negative and the C. P. (J. T. Baker) with an analysis indicating the presence extinction parallel. Indices of refraction were deof 0.28% heavy metals as the only significant impurity. termined as N g = 1.558 and N p = 1.546. The The material was boiled in ammonium chloride, washed thoroughly on a Buchner funnel, and dried. It was then, crvstals are biaxial positive, probably orthorhombi& ignited in platinum a t 1000" for twenty hours, and 1400 for another fifteen hours. Below 200" i t was cooled in a desiccator. Silica was prepared from anhydrous silicic acid C. P., an Eimer and Amend purity reagent, with the only significant immritv being 0.05570 C1. The imition loss was tested against the theoretical formula Toss and agreed within 0.025%. Water was driven off at 1000' for twenty hours and 1250" for twelve hours. The mixtures were made by weighing the silica directly into the crucible, and adding weighed amounts of lithium carbonate and alumina. Mixtures were thoroughly stirred with a platinum rod, and melted in an electric "Globar" furnace. After crushing the quenched glass in a steel mortar, the iron was removed with a magnet and the glass remelted. The entire procedure was repeated three or four times, the glasses being checked for homogeneity under the microscope. Part of each glass was crystallized for 24-36 hours at temperatures between 600800 Hatch6 used similar techniques and analyses made indicated that the melted glasses were within 0.2% of the batch composition. The refractive indices of the glasses were measured in Na light (D line). The regular change of the indices (and also of the liquidus temperatures) indicated an accuracy of the melted compositions within the desired limits. The well-known quenching technique, developed a t the Geophysical Laboratory of the Carnegie Institution of Washington and frequently described" was used throughout. The platinum-platinum 90%, rhodium 10% thermocouples were calibrated regularly at the gold (1062.6 ") and diopside (1391.5') points. For a more frequent check on the accuracy of the thermocouples, a calibrated standard thermocouple was used. In the long runs (longer than two hours) temperatures were automatically controlled. In general the limit of accuracy was approximately *2.5", over extended periods. For short periods of time up to an hour, a large heat capacity furnace wts used which could be controlled manually to within *l , Unless otherwise stated, the limit of error in runs less thaa two hours is approximately =t2" and for longer runs *3 Various samples were tested to find whether equilibrium had been attained in a certain length of time. Equilibrium was attained rapidly, usually in fifteen minutes. All runs were thirty minutes or longer. Identification of the phases was performed with the petrographic microscope and from X-ray diffraction patterns. The latter method was used extensively and al-
'.
.
(IO) F. A Hummel, personal communication. (11) E. S. Shepherd, G. A. Rankin and F. E. Wright, Am. J . Sci., 18, 293 (1909).
TABLE I X-RAYDATA Values obtained in this study Liz0.2Si01 &spodumene Rel. 'A' Re1 VaYues int. Vaiues int. 'At
int.
7.32 4.70 4.08 3.75 3.67 3.34 2.72 2.40 2.37 2.31 1.975 1.84 1.57 1.50 1.47 1.41 1.30 1,256
0.06 0.110 0.04 0.33 1.00 0.03" ,ii4 .03 04 .02 .24 .02 .04a .04 I
.I8 .03 .03 .06
These values Li20.SiOZ lines.
7.35 5.39 4.10 3.73 3.64 3.58 2.93 2.385 2.36 2.30 2.06 2.00 1.96 1.84 1.81 1.528 1.47 1.46 1.44 1.425 1.37 1.25 1.22 correspond
0.04 -12 .04 .80
.65 1.00 0.10
4.61 3.97 3.48 3.173 2.33 2.26 2.13 1.943 1.89 1.653 1.542 1.458 1.434 1.411 1.385 1.2415
0.10 0.40 1.0 0.12 .04 -01 ,015 .04 .10 .06
.3 .1 .05 .07 .Ol .05 .005 .15 .02 .04 .007 .06 (?) ,005 .06 (2) .02 .04 04 .04 .04 .03 .03 .02 with those of the strongest I
No changes in refractive index indicative of solid solution were noticed, although irregular shifts in the X-ray diffraction patterns were found. I n this connection it was found that the X-ray diffraction data were not the same as those recently published by Austin.12 I t appears that the data given by Austin contain the spacings for lithium metasilicate as well as those for the disilicate (see Table I). This might be expected inas(12)
A. E. Austin, J . Am. Ceram. SOL,80, 218 (1947).
2088
RUSTUMR O Y
AND
E. F. OSBORN
Vol. 71
TABLE I1 MOLARREFRACTION DATA Substance
Av. index
Density
Molar vol.
Total refracn., R
Rat01
&*-
Eucryptite Natural (low) Synthetic (high) Glass
1.545 1.5225 1.541
2.667 2.352 2.429
94.5 107.1 103.8
29.74 32.71 32.61
10.26 12.71 12.61
3.62 3.996 3.984
Reference
Dana, “System of Mineralogy” Winkler18 Hatch: Gmelin’s H. buch d. anorg. Chem., 8th ed.
Spodumene Natural (low) 1.67 3.16 120.5 43.7 9.84 3.56 Dana, “System of Mineralogy” Synthetic (high) 1.5205 2.406 155.0 47.15 12.25 3.85 Hatch,6 Roy and Osborn Glass 1.518 2.37 156.0 47.27 12.37 3.88 Roy and Osborn Petalite (1 :1:8) Mineral 1.510 2.4 255.0 76.1 11.4 3.75 Dana, “System of Mineralogy” Glass 1.495 2.358 259.3 75.6 10.9 3.73 Hatch: Roy and Osborn Petalite (1:1:6) Mineral 1.510 2.4 205.0 61.3 11.5 3.76 Dana, “System of Mineralogy” Glass 1.495 2.36 60.8 208.5 11.0 3.73 Hatch,6 Extrapolated a The value used for Liz0 is 5.1, an average obtained from the total refractions of LizO, Liz0.AlzO3,LisO.SiO2, Liz0.2 Si02 and 2Li~O.Si02. * For the method of calculation see Safford and Silvermanl’J: 7.19 the value for quartz was used in the “low” forms, and 7.45 the value for silica glass was used in the high and vitreous forms, as the subtraction for the reRea- values were calculated according to the method of Fajans and Kreid1,”although a fraction contribution of silica. major change in Ro2- in the arrangement of one molecule of A1201is progressively diluted as one proceeds to higher molecular weights. It should be noted that refractive indices and densities especially, differ widely as given in the literature, often by as much as 5%.
much as Austin used a commercial source of silica, and furthermore he obtained the compound “merely by allowing the crucible to cool from 1200°.”13 An inspection of the equilibrium diagram for the system LizO-Si02 (Kracek4”) will suggest that such a procedure might give a mixture of metasilicate and disilicate. Spodumene, @-Spodumene and Eucryptite.In most cases @-spodumene (LizO.Al203.4Si02) appears as small anhedral grains. Rarely welldeveloped, tetragonal dipyramids form. The indices for @-spodumene were found to be ENa = 1.522, and U N ~= 1.516 (==0.002)which are essentially the same as those given by Hatch.6 The birefringence of the solid solutions, however, did not seem to decrease with increasing silica content to the same extent as shown by Hatch, for even in the highest silica member i t appeared to be considerably greater than 0.001. Spodumene, the natural, low temperature form of Li20,A120:,4Si02did not appear in this study, nor has its synthesis ever been confirmed. Spodumene inverts to the high temperature polymorph rapidly above 900°, but the change may be brought about a t 7OOo1* by extremely fine grinding followed by heating for periods of time of the order of two hundred hours. Some attempts were made to synthesize the mineral form. From analogy with the use of calcium vanadate as a “flux’‘ in the synthesis of wollastonite, lithium vanadate and tungstate were used as fluxes. Additions of 5, 10 and 15% of flux were made to @-spodumeneand the mixtures heated to 500, 550 and 600’ for lengths of time varying between twenty-four and two hundred hours. In no case was spodumene (13) Austin, 09. cif., p. 219. (14) First shown by Meissner, 2. anore. Chrm., 110, 187 (1920).
obtained. The synthesis and inversion were attempted by hydrothermal methods. Pressure vessels of the type used by Ingerson and Morey16 were used a t temperatures between 400 and 700°, the pressure being varied between 500 and 12,000 p.s.i. Both @-spodumeneand glass of the spodumene composition were used, and additions of 5% of K F , or of acid or alkali were tried. In none of the experiments did spodumene form. Data were obtained concerning the probable structures of the high and low temperature forms of both spodumene and eucryptite. The use of molecular refraction values as an index to structure in silicates is illustrated in two recent papers by Safford and Silverman,lGand by Fajans and Kreidl.17 With the use of the Lorenz-Lorentz equation and the best available values for the density and refractive indices of the polymorphs and of the glass of each composition, values of the total molar refraction, R, the refraction contribution of alumina, Ra1o,,and of each oxygen ion Rol- were calculated and are shown in Table 11. The calculations of Safford and Silverman give an average value of R A ~ ~ofo ,12.31 for the four network structure type minerals, albite, nephelite, carnegieite and anorthite. For corundum and kyanite where the Ala+ is in sixfold coordination, the alumina refraction contribution was found by these writers to be approximately 10.5. These data coupled with those in Table I1 suggest the follow(15) E. Ingerson and J. Morey, Am. Mineral., 14, 1121 (1937). (18) H. Saffard and A. S h e r m a n , J . Am. Ceram. Sac., SO, 203 (1947). (17) K.Fajans and N. J. Krcidl, ibid., 81, 105 (1948).
June, 1949
THESYSTEM LITHIUMMETASILICATE-SPODUMENE-SILICA
ing: (1) that both natural eucryptite and spodumene have different crystal forms from the synthetic crystals; (2) that on analogy with spodumene, synthetic eucryptite is a high temperature polymorph of the mineral; (3) that the coordination number of AI3+in the low temperature forms of both spodumene and eucryptite is 6, whereas the high temperature forms possess a network of A104and Si04 tetrahedra; and (4) that the structure of the high temperature form in each case is closely similar to that of its glass. The evidence of Winklerl* on the structure of synthetic eucryptite, and the accepted structure of spodumene are in accordance with these deductions. Peta1ite.-Hatch6 found that petalite has no existence above the solidus in the binary system p-spodumene-silica. It was not encountered in the present study. The chemical formula for petalite, usually written as LizO.Alt03.8Si02, has often been questioned. The most recent analysis of the problem by Mikkola and Wiiklgsuggests the composition 4Li20.5A120&0Si02. Ballo and Dittler20suggested the formula LizO.Al203.6Si02, and this composition is shown in Fig. 1 as "lithium orthoclase." The evidence from molar refraction data (Table 11) indicates that neither the 1: 1: 8 nor the 1:1:6 formula gives very satisfactory correspondence in the R A ~ ~values. o, Silica.-Silica appeared in all three polymorphic forms, quartz, tridymite and cristobalite. Quartz appeared as a metastable phase above 867'*l in practically every case where silica was a phase, if the starting material was all crystalline and the duration of the run was less than twentyfour hours a t a temperature below 1100'. It persisted for shorter periods of times in runs a t temperatures up to 1200'. I n no case was quartz found in any mixture which had been heated above 1200'. The explanation of the existence of quartz above 867' is found in the fact that quartz formed during the crystallization of the glass a t 600-800' and as has often been reportedz2failed to invert. I t was also found that many mixtures especially those in the vicinity of the 980' invariant point (Fig. 6) when introduced as a glass into a furnace a t about 1000°, and held a t this temperature for times ranging from fifteen minutes to twenty-four hours, usually contained quartz as the silica phase. It appeared therefore that quartz was crystallizing out of a melt in the stability range of tridymite. Further investigation showed, however, that the quartz crystallized from the glasses during the short time (usually fifteen to twenty seconds) during which the temperature of the sample rose from room temperature to (18) H. G. F. Winkler, Acta. Cryst., 1, 27 (1948). (19) T. Mikkola and H. B. Wiik, Bull. Comm. Geol. F i n . , 140, 281 (1847) (20) R. Ballo and E. Dittler, Z . nnorg. Chem., 76, 39 (1912). (21) Revised value for the quartz-tridymite inversion given by Kracek, ref. 4h. (22) F. C. Kracek, N. L. Bowen and G.W. Morey, J . Phys. Chcm., 41, 1188 (1937).
2089
867'. It was, however, not proven that this quartz does not grow further a t temperatures above 867". The spacings of the lines in the Xray pattern of the quartz which persisted in these mixtures (patterns made a t room temperature) were quite definitely displaced, and corresponded more nearly to those of high quartz than low quartz. The pattern for high quartz used in comparison was obtained with a high-temperature unit2*for the Geiger counter spectrometer a t 600'. The remarkable persistence of quartz above 867' as well as the displaced X-ray patterns suggest that a low quartz solid solution was present or that high quartz is stabilized to room temperature by the presence of the lithium ion. It may be noted that Winkler18in a recent analysis of the eucryptite structure has demonstrated that the Lif ion is of just the right size to fit into the "holes" in the high quartz structure. Tridymite was the stable phase in the temperature range in which the work was done, and although it seldom crystallized out first, the other forms of silica tended to invert to tridymite on being held for a sufficiently long time. Cristobalite usually crystallized out in the silica field. Near the boundary with @-spodumene, and especially near the eutectic it appeared as fine octahedra. The refractive index though not accurately determinable was apparently that of high cristobalite; .the X-ray patterns showed the presence of both high and low cristobalite. This is another case of the quenching of high cristobalite to room temperature while imbedded in glass.24 On crushing very fine, the crystals of high cristobalite inverted to the low form as shown by the X-ray patterns. Glasses.-The refractive index of the glass of each of the mixtures, quenched usually from about 100-200' above the liquidus, was determined in sodium light (D line). From these data the isofracts in the system lithium metasilicate@-spodumene-silicawere drawn and are shown in Fig. 2. The index of refraction and density of a quenched glass of spodumene composition were found to be, respectively, 1.518 and 2.37 as compared with reported6 values for P-spodumene of 1.5205 and 2.406. With annealing of the glass, its index and density would increase and approach even more closely to those of P-spodumene. A glass of eucryptite composition has actually a higher index and a greater density than crystals of high eucryptite (Table 11). Binary Systems Lithium Metasilicate-Silica -This is part of the system lithia-silica worked out by K r a ~ e k ' ~ ; Fig. 3 presents the latest revision of the system by K r a ~ e k . ~ ~ (23) R. Roy, E. T. Middleswarth and F. A. Hummel, A m . Mincml., 80, 458 (1948). (24) First reported by J. W. Greig, THISJOURNAL, 64, 2846 (1932). All recorded cases are summarized by A. Grenall, ibid., 70, 423 (1848).
RUSTUMROY AND E.F.OSBORN
2090
Fig. 2.-Isofracts
Vol. 71
slope of the line indicating the limit of the solid solution series as the temperature is decreased below 1356'. From the data obtained on two mixtures, 26.54 and 39.27% silica, Hatchs estimated the limit of solid solution a t 1356" to be about midway between these two points, and he indicated the boundary of the solid solution field below 1356' as a vertical, dashed line. The data obtained on six mixtures in this study are summarized in Table 111. These did not serve to establish the limits of solid solution much more accurately than the data of Hatch because of the inability either to deterI5l7 mine the presence of small amounts of a second phase Weight per cent. or to determine accurately for the quenched liquids in the system lithium metasilicate-8- by Optical or X-ray means the comDosition of a Bspodumene-silica. spodum&e solid solution. On the basis of slight shifts of X-ray pattern lines, five differentsolid solutions were considered to be present in the mixtures, indicated in Table I11 as SS (l),SS (2), SS (3), SS (4), and SS (5). The limit of solid solution a t 1356' is taken as 34 * 2% silica, with no change indicated with decreasing temperatures. 1
1
1
I
l
l
1800 -
1450
t
1
I
A
/
TRIDYMITC
i 1300
30 50 70 90 Lii0.2SiOz SiOr Weight per cent. Fig. 3.-Equilibrium diagram of the system lithium metasilicate-silica, after Kracek.'
10 LizO.SiOz
10 30 50 70 90 LizO.A1~0~.4SiOz (spodumene) Si02 Weight per cent. Fig. 4.-Equilibrium diagram of the system spodumenesilica, after Hatch.6
Lithium
Metasilicate-6-Spodumene.-Thir-
teen mixtures were prepared and studied in this system. The equilibrium diagram constructed from the data obtained is shown as Fig. 5.
THESYSTEM LITHIUMMETASILICATESPODUMENE-SILICA
June, 1949
Composition in wt. % SpoduSilica mene
Temp.,
54 54 54 45 45 45 37 37 37 33 33 33 30 30 30 35 35 35
1360 1350 975 1360 1350 975 1360 1350 975 1360 1350 975 1360 1350 975 1360 1350 975
46 46 46 55 55 55 63 63 63 67 67 67 70 70 70 65 65 65
OC.
Time, hours
9.5 5.8 5.8 5.8 5.8 2.8 2.8
Phases present
3 8 100 3 8 100 3 8 100 3 8 100 3 8 100 3 8 100
SS ( l ) , glass (?) SS (l), Tr
30 35 30 30 30 30 30 30 30 30 30 30 30 45 30 40 30 30 35 30 30 30 30 30 60 30 30 40 30 30 30
All glass Rare MS, glass All MS All glass Rare MS, glass MS, glass MS, SP V Rare &IS, glass MS, glass
90.5 90.5 94.2 94.2 94.2 94.2 97.2 97.2 97.2 97.2 100 100 100
9.5
TABLE I11 BINARYSYSTEM SPODUMENE-SILICA
SS (11, Tr SS (l), glass SS ( l ) , Tr SS (l), Tr SS (l), glass SS (1) SS (1), Tr (?) ss (2) ss (2) ss (2) ss (3) ss (3) ? ss (3) ? ss (4) ss (4)
2.8 2.8 0 0 0
1024 1018 1408 1400 1028 1013 1415 1408 1024 1018 1425 1421 1415
720 60 30 30 30 30 30 30 720 30 45 40 30
209 I
Sp, glass SP, MS All glass Rare Sp, glass Sp, glass (?) SP, MS All glass Rare Sp, glass Sp, glass (?) SP, MS (?I All glass Sp, little glass All Sp
A eutectic occurs a t 1026' a t the composition 54.6% spodumene, 45.4% lithium metasilicate. No appreciable solid solution is present. This is in contrast to the work of Ballo and Dittler2" who found a series of solid solutions extending from lithium metasilicate to 60% spodumene, and in addition a 1:l molar compound. The melting point of spodumene was determined as 1423 * 2 O , in agreement with Hatch,6 in contrast ss ( 5 ) ? to numerous other workers. A slight anomaly is found a t the spodumene end, the temperature of TABLEIT appearance of the second phase being about 5' BINARYSYSTEM LITHIUM METASILICATE-SPODUMEKE lower than 1026'. This is not a matter of underIn this and succeeding tables the following abbreviations are used: Cr = cristobalite, Tr = tridymite, QZ = cooling, for crystalline material containing both quartz, Sp = p-spodumene, DS = lithium disilicate, MS = lithium metasilicate and @-spodumenewas used lithium metasilicate, V = very. as starting material. No explanation was found Composition in wt. R for this phenomenon. The data from which the Temp., Time, MetaSoodumene Phases present silicate "C. minutes diagram was constructed are given in Table IV. 100 100 100 82 82 82 82 58 58 58 49 49 49 49 36 36 36 86 27 27 27 20.5 20.5 20.5 20.5 14 14 14 14 9.5 9.5
0 0 0 18 18 18 18 42 42 42 51 51 51 51 64 64 64 64 73 73 73 79.5 79.5 79.5 79.5 86 86 86 86 90.5 90.5
1202 1200 1199 1190 1186 1028 1025 1116 1028 1025 1065 1060 1030 1025 1116 1100 1030 1023 1200 1030 1023 1275 1265 1023 1018 1335 1325 1024 1015 1380 1372
MS, SP All glass Rare MS, glass MS, glass MS, SP All glass Rare Sp, glass Sp, glass SP, MS V Rare Sp Sp, glass SP, MS All glass Rare Sp, glass Sp, glass Sp, MS All glass Rare Sp, glass Sp, glass SP, MS All glass Rare Sp, glass
1400
1300
9
I
/
I
d,
g
w
1200
1100
1000 10 LtO.SiOZ
30
50 70 90 Li~O~ALOO~4SiO~ (spodumene) Weight per cent. Fig. 5.-Equilibrium diagram of the system lithium metasilicate-b-spodumene.
The Ternary System Lithium Metasilicate@-Spodumene-Silica The phase equilibrium diagram for the system lithium metasilicate-8-spodurnene-silica is shown
2092
RUSTUMROY AND E. F. OSBORN
Vol. 71
Table V. The location of the maximum could not be obtained directly from liquidus data, except as a rough approximation because of the very low temperature gradient along the boundary AB (Fig. 7). Two other considerations were used t o fix the position of M. First, M must lie within the triangle D R T as all mixtures in this field crystallize only to two phases, lithium disilicate and a 0-spodumene solid solution. Secondly, M must lie a t the junction of the boundary curve with a three-phase boundary (vide infra) which when extended as a straight line passes through D. To understand relationI2 1423' ships in a ternary system of this type with solid solution it is essential that (spodumene) the positions of the threeWeight per cent. phase boundaries25 be Fig. B.-Equilibriurn diagram of the system lithium metasilicate-8-spodumene-silica. known.26 Accurate deterDots represent compositions of mixtures studied ; heavy lines are boundary curves; mination of most threephase boundaries in this light lines are isotherms. svstem is not Dossible bein Fig. 6, constructed from data obtained on the cause of the low tempeiature gradie2 along the forty-seven mixtures whose compositions are boundary curve AB (Fig. 7), the inability to indicated by dots, and data from the binary determine the composition of crystals of psystems. A summary of the essential data ob- spodumene by either optical or X-ray methods, tained for these mixtures is given in Tables V and and the difficulty of recognizing very small VI. Heavy lines in Fig. 6 represent boundary amounts of a third phase especially if it is glass curves; arrows shown indicate the direction of or some form of silica. It was possible, however, falling temperature. Lighter lines are isotherms. to establish with fair accuracy the positions of The ternary system contains two invariant the two most significant three-phase boundaries, points, both eutectics, and a maximum on the AT and BR. Other three-phase boundaries boundary curve joining these two points. At one were drawn in a t reasonable locations as shown in invariant point, lithium disilicate, tridymite and Fig. 7. The three-phase boundary AT (Fig. 7) a &spodumene solid solution (33% SiOz) are in joins liquid h with the @-spodumenesolid soluequilibrium a t 980' with a liquid of the composi- tion which is in equilibrium with silica a t 980" in tion 39.2% silica, 26y0 spodumene and 34.8% the binary system 0-spodumene-silica. Because lithium metasilicate (f 0.25% oxides). At the the extent of solid solution in this binary system second invariant point lithium metasilicate, lith- a t 980' could not be determined accurately (Table ium disilicate and a (3-spodumene solid solution I11 and Fig. 4), ternary data were necessary to (24yo SiOz) are in equilibrium with a liquid of the fix the position of T . These data along with composition 39.2% metasilicate, 28.2% silica and those used to establish the position of BR are listed 32.6y0 spodumene a t 975'. At the maximum in Table VI. The position of T was found to be (point M in Fig. 7 ) , lithium disilicate and a P- 33y0 silica. This is corroboratory evidence therespodumene solid solution containing approxi- fore that the dashed line in Fig. 4 delimiting the mately 29.5% silica are in equilibrium with a liquid @-spodumenesolid solution field is approximately (25) The concept of three-phase boundaries is defined and deof approximately the composition, 36y0silica, 3i'Y0 scribed by N. L. Bowen, Am. J . Sci., 98, 222 (1914). metasilicate and 27y0spodumene, a t 985'. (26) Discussion of methods of determination of three-phase Data for establishing the composition and boundaries are given by W. L. Bowen, ibid., 40, 151 (1915), and (b) temperature of the two eutectics are included in E. F.Osborn and J. F. Schairv, ibid., 38, 726 (1941).
THESYSTEM LIT HI^ METASILICATESPODUMENESILICA
June, 1949
vertical from 1356" to 980".
The silica content
of R, which is not determined by binary relations, was found to be approximately 24%. TABLE V ~
~~~
30 36 36 36 36
25 22 22 22 22
TERNARY SYSTEM LITHIUM METASILICATE-~-SPODU- 35 24 35 24 YENE-SILICA 35 24 Composition in wt. % Meta- SpoduTemp., Time, 24 30 silicate mene Silica OC. minutes Phases present 24 30 Silica Field 24 30 1350 120 60 20 All glass 20 24 30 1335 120 60 20 Rare Cr, glass 24 30 20 30 1050 60 20 20 Cr, glass 32 26 30 1030 60 20 20 Cr, Sp, glass 32 26 30 990 60 20 20 Cr, Sp, glass 32 26 45 985 60 20 20 Cr, Sp, DS 32 26 1360 100 60 30 Rare silica 10 34 26 1220 90 60 30 10 Cr, Sp, glass 34 26 30 990 30 60 10 Cr, Sp, glass (?) 34 26 985 30 Cr.. SD. 60 30 10 - . DS ( ? ) Glass, silica 30 1100 50 10 40 Tr, glass 1010 45 9 . 6 39.4 50 10 40 Tr, DS, glass 1007 45 50 10 9 , 6 39.4 40 Tr, DS, Sp (?) 60 980 50 10 9.6 39.4 40 Tr, DS, Sp 975 600 10 50 9.6 39.4 40 Rare silica, glass 30 1160 50 30 33 29 20 40 1110 Qz, glass 50 30 33 29 20 45 1100 50 30 Qz, SP,glass 33 29 20 30 980 50 30 SP,Qz, glass (?I 33 29 20 30 975 50 30 SP, Qz, DS 18.5 38.2 20 All glass 40 1110 48 18 1 8 . 5 38.2 34 Cr, glass 45 1100 48 18 18.6 38.2 34 Cr, glass 40 990 48 18 18.5 38.2 34 30 985 48 18 Qz, DS 5 . 0 55 34 30 Tr (?), Sp, DS 975 48 18 5 . 0 55 34 Silica, glass 30 1105 48 20 32 5.0 55 Tr, Qz, glass 30 990 48 20 32 Tr, DS, QZ 60 985 48 20 32 5.0 55 30 975 48 20 DS, SP, QZ 32 Cr, glass 30 1105 48 24 28 34.8 36.1 Cr, Qz, glass 40 995 48 24 28 34.8 36.1 30 985 48 24 28 Qz. SP 34.8 3 6 . 1 30 980 48 24 Qz, SP, DS 33 28 39 All glass 30 1105 47 19 33 34 39 Cr, glass 30 47 995 19 34 39 33 Cr, glass 45 47 985 19 34 38.3 35.7 120 DS, Cr, Sp 980 47 19 34 38.3 35.7 V rare silica, glass 50 1075 45 19 36 38.3 35.7 Tr, glass 50 990 45 19 36 38.3 35.7 Cr, DS 985 30 45 19 36 38.3 35.7 30 980 45 19 SP, DS, Qz 2.7 79 36 V rare silica, glass 40 1050 46 10 44 Tr, glass 1015 30 46 10 44 41 55 Tr, DS, glass 30 1012 46 10 44 41 55 Tr, DS, Sp (7) 30 983 46 10 44 41 55 Tr, DS, Sp 40 975 46 10 44 Rare Cr, glass 40 44 1050 22 34 49 10 Cr. elass 30 44 985 22 34 49 10 30 DS,Sp, Qz (?) 44 980 22 34 50 Cr, glass 49 10 1075 45 25 30 30 Qz, glass 40 20 985 45 25 30 980 60 Qz, Sp, glass (?) 40 20 45 30 25
45 42 42 42 42 41
41 41 46 46 46 46 46 42 42 42 42 40 40 40
2093
975 1015 990 985 980
60 30 30 30 180
Sp, DS, Qz (?) Rare Cr, glass c r , glass DS, Qz, glass DS, Sp, QZ (?)
985 980 1110 1105 1075 985 980 1025 995 985 980 990 985 980
30 60 30 40 30 50 30 30 120 40 30 180 50 30 30
rare Cr' Qz, DS, Sp ( ? ) Qz, DS, Sp All glass Rare Cr, glass Cr, Sp, glass Qz, Sp, glass (?) Qz, Sp, DS RareCr, glass Qz, Sp (?I, glass Qz, Sp, glass (?) Qz, Sp, DS Rare Cr, glass Cr, Sp, glass Qz, DS, Sp
Spodumene Field 51 1250 60 51 1220 60 51 980 30 51 975 5000 38 lo00 30 38 985 30 38 980 50 38 975 720 43.3 1165 30 43.3. 1105 30 43.3 980 130 43.3 975 720 40 1360 100 40 1000 30 40 980 30
V rare Sp, glass sp, silica, glass Sp, Qz, glass (?) Sp,Qz, DS (?) Rare Sp. glass SP,glass
SP,DS Sp, DS, silica (?) V rare sp, glass sp, silica (?), glass sp, silica (?),glass Sp, DS, silica (?) Rare Sp, glass Sp, glass Sp, silica (?), DS ( ?)
40
975
720
29.1 1025 120 30 29.1 985 30 29.1 980 30 985 28 980 30 28 975 1000 28 35 1000 26 30 993 26 985 60 26 980 100 26 975 600 26 18.3 4 4 4 41 41 41 40 40
1050 1035 975
120 30 600
Disilicate Field 1020 30 30 980 30,600 975 987 30 980 30
Sp, silica (?), DS ( ?) Rare Sp, glass Sp, glass MS,DS,Sp Rare Sp, glass MS, Sp MS, DS, Sp V rare Sp, glass Sp, glass MS, Sp, glass MS, Sp, glass (?) MS, DS, Sp Only Sp detected below liquidus Allglass Sp, glass MS, Sp All glass DS, glass DS, Sp, Silica (?) All glass DS, glass
RUSTUMROYAND E. F. OSBORN
2094 TABLE V Composition in wt. % Meta- SpoduTemD.. silicate mene Silica 'Cy
40 38 38 38 36 36 36 47 47 47 41.4 41.4 41.4 41.4 44 44 44 44 38 38 44 44 44 45 45 45 39 39 39 41 41 41 40 40 40
20 22 22 22 25 25 25 16 16 16 20.6 20.6 20.6 20.6 18 18 18 18 26 26 21 21 21 22.5 22.5 22.5 29 29 29 29 29 29 31 31 31
40 40 40 40 39 39 39 37 37 37 38 38 38 38 38 38 38 38 36 36 35 35 35 32.5 32.5 32.5 32 32 32 30 30 30 29 29 29
53 53 53 53 53 44 44
11 11 11 11 11 27 27 27 35 35 35 35
36 36 36 36 36 29 29 29 23 23 23 23
(Continued) Time, minutes
975 30,600 990 30 987 30 980 30 985 30 980 30 975 600 1000 30 985 30 980 30 987 30 985 30 980 70 975 600 1000 30 985 30 980 30 975 2700 985 30 980 30,600 990 30 985 35 980 30 1000 30 980 30 975 30 990 30 980 30 975 30 990 30 975 30 100 970 985 30 980 30 975 600
TABLE VI DATAINDICATING POSITIONS OF THREE-PHASE BOUNDPhases present
DS, Sp, silica (?) All glass RareDS Sp, DS, silica (2) Rare DS, glass Sp, DS, glass (?) Sp, DS, silica (?) V rare DS, glass DS, glass DS, Sp Rare DS, glass DS, glass DS,Sp DS, Sp Rare DS, glass DS, glass DS,Sp (?) DS, Sp DS, little glass DS,Sp V rare DS, glass DS, glass NS, DS, Sp All glass DS, glass DS,MS, glass All glass DS, Sp, glass DS, MS, Sp All glass SD, glass DS, MS, Sp All glass DS, Sp, glass (?) DS, MS, Sp
Metasilicate Field
14 42 42 42 42
1023 1015 985 980 990 985 980 1012 1005 995 980
30 30 40 30 30 30 30 45 30 35 40 30
All glass R~~~ MS, glass MS, DS, glass MS, DS, glass MS, DS, Sp R~~~ MS, glass MS, glass MS, DS, Sp All glass Rare MS, glass MS, Sp, glass MS, DS, Sp
Courses of crystallization and phase relations for other silicate systems of a similar type have been considered by various authors.25*26m27 Consequently a series of isothermal planes and a detailed analysis of the crystallization and melting phenomena are not given. I n Fig. 7, however, an isothermal plane for the temperature 970' is shown. This diagram illustrates the phase (27) N. L. Bowen, J. E. Schnirer and E. Posnjak, A m . J . Sci., 96 (5th series), 193 (1933).
Vol. 71
ARIES Composition in wt. % Spodusilicate
mene
Silica
20.8 20.8 20.8 20.8 20.8 26.9 26.9 26.9 26.9 25 25 25 25 26.9 26.9 26.9 26.9 26 26 26 26 26
56.7 56.7 56.7 56.7 56.7 55.6 55.6 55.6 55.6 52 52 52 52 37.1 37.1 37.1 37.1 36 36 36 36 36
22.5 22.5 22.5 22.5 22.5 17.6 17.6 17.6 17.6 23 23 23 23 36 36 36 36 38 38 38 38 38
Temp., Time, OC. minutes
1220 1210 990 986 975 1160 1150 998 995 1105 990 986 975 1100 985 980 975 1105 1075 990 985 975
60 30 75 60 600 60 50 30 30 50 75 60 720 30 60 100 960 30 50 60 60 960
Phases present
All glass Rare Sp, glass Sp, MS (?), glass Sp, MS, glass SP, MS, DS 0) All glass Rare Sp, glass Sp, glass Sp, MS, glass V rare Sp, glass Sp, MS (?), glass Sp, MS, glass
SP,M8,DS, (?I V rare Sp Sp, glass SP, DS
SP,DS All glass Sp, glass Sp, silica, glass Sp, silica, glass (7) Sp, DS,silica
relationships existing 5 O below the lowest temperature at which liquid can exist under equilibrium conditions. The ternary diagram is divided into four triangular areas by heavy lines. At equilibrium all mixtures within the triangle LKR will consist of lithium metasilicate and a Pspodumene solid solution whose composition will lie between K and R, the exact composition of the final solid solution depending on the original comDosition of the mixture. All mixtures in the t r i k g l e LRD will be composed of the three crystalline phases, lithium metasilicate, lithium disilicate and P-spodumene solid solution R. Within the area RDT, mixtures will have crystallized to only two phases, lithium disilicate and a P-spodumene solid solution of a composition between R and T. Finally mixtures lying within the area DTS contain the three phases, lithium disilicate, P-spodumene solid solution T, and tridymite. During crystallization if equilibrium were not maintained, silica might appear in other forms, and i t has already been mentioned that both auartz and cristobalite were mesent in the &bility range of tridymite, invert'ing to the stable phase only after a very long time. Furthermore, failure to maintain equilibrium throughout the Crystallization process may result in incompleteness of the continuous reaction between Pspodumene solid solutions and liquid. Thus, for example, although under equilibrium conditions mixtures in the area LRK crystallize t o only two phases, if equilibrium is not maintained the liquid may not disappear until the eutectic B is reached
June, 1949
2095 THESYSTEM LITHIUM METASILICAT~CSPODUMENE-SILICA
a t 975'. Hence some lithium disilicate will be present in the final mixture and furthermore the 0spodumene crystals may appear zoned, the outer zones being higher in silica content. It is possible for certain mixtures in this system to exhibit a reversal TURE6 PUASCS of zoning, even though the mixtures are cooled cont i n u o ~ s l y . ~ 'This ~ is true for all mixtures in the area BRCM (Fig. 7), M being the position of the maximum on the boundary curve AB, and M C being the three-phase boundary intersecting the boundary curve a t M. During the cooling of these mixtures from the liquidus temperature until they meet the boundary curve, the composition of the solid solution in equilibrium with ' 40 30 liquid changes toward silica. On further coolWeight per cent. ing, 8s the Of Fig. 7.-An isothermal plane indicating phase equilibrium relations a t 970". Heavy the liquid toward solid lines bound four composition triangles within which either two or three phases the eutectic the are present as indicated; light solid lines are boundary curves; light dashed lines are boundary curve, the three phase boundaries; A and B are eutectics; M is a maximum on the boundary position of the solid s o h - curve AB. tions in equilibrium with Evidence is presented for believing that the these liquids moves away from silica. If reaction of crystals with liquid is not complete, zoned 0- coordination number of A13+ in the high temperaspodumene crystals may appear having an inner ture forms of spodumene and eucryptite is four and outer zone both poorer in silica than an inter- whereas in the low temperature, natural modificamediate zone. tions the coordination number is six. Neither petalite nor the natural form of spodumene was Summary Phase equilibrium relations in the binary system synthesized, and the composition of the former is lithium metasilicate-/3-spodumene and in the still in doubt. ternary system lithium metasilicate-6-spoduQuartz commonly appears in the ternary mene-silica are presented. I n the binary system system a t temperatures above its stability range. no solid solution or binary compound is present, X-RaY data obtained on this quartz a t room a eutectic occurring a t 1026' and a t the composi- temperature suggest that ions other than s i 4 + tion 45.4% lithium metasilicate. In the ternary and 02- are Present in the quartz in sufficient system two eutectics and a maximum appear, quantities either to shift the X-ray lines of IOW as shown in Figs. 6 and 7. The positions of three- quartz Or to stabilize the high form. High phase boundaries and the phase relations below cristobalite appears a t room temperature in Some the temperature a t which liquid is present in the mixtures. ternary system are shown in Fig. 7. RECEIVED JANUARY 3, 1949
