American Mineralogist, Volume 67, pages 350-355, 1982

The thermodynamicproperties of fluor'

M. D. BentoNr Department of the GeophysicalSciences University of Chicago, Chicago, Illinois 60637

H. T. HnseLToN.Jn., B. S' HelarNcwav U. S. Geological SurveY MS 959, Reston, Virginia 22092

O. J. Kleppe, Department of the GeophysicalSciences University of Chicago, Chicago, Illinois 60637

eNo R. A. RouB U. S. Geological SurveY MS 959, Reston, Virginia 22092

Abstract

The standardthermodynamic properties of fluor-topaz,AlzSiOnFz, have been calculated from low- and high-temperatureheat-capacity measurements and from high-temperature, oxide-meltcalorimetry. Fluor-topaz (from TopazMountain, Thomas Range, Utah) con- taining 0.04 wt. percentwater was used in all the experiments.Adiabatic calorimetry performedfrom 10.6to 379.2K givesSi9E - Si of 105.4* 0.2 J/mol'K. Combinedheat capacitiesdetermined by adiabaticcalorimetry (200-380 K) and differentialscanning calorimetry(340-800 K) werefit to the followingpolynomial (equation valid 200-1,000K): c; (ymol.K)= 471.41-0.08165r + l.2695xlO6T-2- 5485.510'5eO.1%). The enthalpyof the reactionCaF2 + Al2O3+ SiO2: CaO + Al2SiO4F2was measured at 970K by oxide-meltcalorimetry and gave L'rtels= 9l'E8-f3'56 kJ' Fromour heat-capacity measurementsand ancillary data, we calculateLrt2es = 96.12!3.95kJ for the reactionand LI4.2e8:-3084.45t4.70 kJ/mol, and AGi,zqe : -2910.66-+4.74kJlmolfor fluor-topaz.

Introduction synthesized topaz containing more than 50% hy- Topaz occurs frequently as an accessorymineral droxyl-topaz component. (1972, 1978)reported synthesisresults in fluorine-rich granitic rocks and associatedhydro- Rosenberg solid solutions; his is the only work thermally altered rocks. As one of the principal involving topaz thermodynamic data might be derived. fluorine bearing minerals, it offers a key to under- from which was undertaken to measurethe thermo- standing the genesisof these rocks. Topaz is a solid This study properties of fluor-topaz as part of a proj- solution between fluor-topaz (Al2SiO4F and (hy- dynamic of topaz and other pothetical) hydroxyl-topaz (Al2SiOr(OH)z).Natural ect on the thermochemistry and their petrological application. topazes vary from nearly pure fluor-topaz to about fluoro-silicates (197 Al2SiO4F1.4(OH)0.0, althougtr Rosenberg 2) has Experimental methods and results Starting materials rPresent address: Geophysical Laboratory, Carnegie Institu- tion of Washington, 2801Upton Street, NW, Washington,D. C. Natural fluor-topaz from , 2flnE. Thomas Range, Juab County, Utah, was collected 0003-(M)vE2l0304-0350$02.00 350 BARTON ET AL,: FLUOR.TOPAZ 351

from the Pliocene topaz-rich alkali (Lind- Adiabatic calorimetry sey, 1979).This topaz is known to be very closeto Heat-capacity measurements the fluorine end member (Penfieldand Minor, lg94; were made from 10.6to 379.2K by adiabatic Ribbe and Rosenberg,l97I) and occurs as crystals calorimetryat the U.S. Geological Survey, as large as 3 cm in the lithophysae. Associ- Reston, Virginia. The cryostat and calorimeter ated minerals include , , plagioclase, have been described in detail by biotite, Robie and Hemingway (1972) and Robie er a/. fluorite, , , pseudobrookite, (re76). ,and (Ream, 1979).About 60 g The topaz sample (50.5654g, of transparent, inclusion-free crystals and corrected for buoy- ancy) was loaded into fragments were handpicked from a few hundred the calorimeter. After evacu- ation of air from the calorimeter, grams of rough material. Approximately one-third it was backfilled with dry_helium gas at pascals of the crystals were light brown; this color is 6x103 pressure (4.0x 10-5 mole of He) to promote apparently caused by electronic defect cenrers thermal equilibra- tion, and sealed.The reported (Dickinson and Moore, 1967),not chemical impuri- temperaturesrefer to the International Practical ties. As previously demonstrated(e.g., Nassauand Temperature Scale of 1968(rPTS-68). Prescott, 1975),the color is removedby heatingto Table2 gives the results, 500"C for several hours. We did not heat treat the corrected for curvature, for the four series (1-4) material used for calorimetry. We did not attempt to of experiments in their order of collection. synthesize fluor-topaz by the reaction 2AlF, + The formula weight for pure fluor-topaz of 184.043g/mol used 2AlzO3 + 3SiO2: 3Al2SiOaF2nor did we attempt to in the calculations is based on the 1975values for use this reaction for the solution calorimetry be- the atomic weights (Commission on Atomic cause of the tendency of AlF3 to hydrolyze which Weights, 1976).Althoueh the sample topaz contains weight percent could result in considerable uncertainty as to the 0.04 OH, no correction compositions of the phasesunder study. was made to the measured heat capacities because Several crystals were analyzed on an automated the heat capacity of hydroxyl- topaz is not known. From electron microprobe at the University of Chicago comparisonsof the heat capacities of hydroxyl- using a ZAF correction program written by L M. and fluor-apatite, and those of hydroxyl-phlogopite Steele of the University of Chicago. The beam and fluorphlogopite, we estimate that the difference current was 15 ma at an acceleratingvoltage of 15 in the heat capacity between fluor- kv. Andalusite served as the standardfor aluminum and hydroxyl-topaz would not ex- ceed approximately percent. and silicon. The water and fluroine were analvzed 7 The diference in SigrSo between hydroiyl-apatite at the U. S. Geological Survey with a perkin Eimer and fluor-apatite is 0.6 percent, and 2408 elemental analyzer (V2O5flux) and a specific between hydroxyl-phlogopite and fluor-phlogopite is -0.5 percent. ion electrode respectively. The water and fluorine On the basis analyses on this material are in good agreement penfield with the original analyses by and Minor Table l. Chemical and crystallographicdata (1894). Although we searchedfor other elements. none were detected. Crystallographic data were Chflistry" Cel I Paraneters collected using powder-diffraction methods (Ni-fiI- oxide wt g a = 0.4647s(3)nn tered CuKa radiation with a corundum internal !Sl-9!* Al 203 56.08 1.00 b - 0.87897(4)nn standarda : 0.47593(1)nm, c : 1.29917(5)nm).The Si02 32.74 0.99 c = 0.83920(4)n0 program of Burnham (1962)was used to refine the F 20.3 1.95 v = 0.34z8l(2)nn3 data. The chemical and crystallographic data are Hzo o.o4 o.orr given in Table 1 Other materials used in the high-temperature Total I 09. l6 oxide-melt calorimetry were (prepared CaO from Less F=0 roo.6i reagent CaCO3, sintered at 1,400.C for I week); optically pure natural CaF2 (southern Illinois, Uni- Al and Si analyses by microprobe, F by speclflc ion electrode, and versity of Chicago collection #1875); optically pure H20 by CHNelemental analJzer. natural quartz (locality unknown, University of Based on 5 oxygens. Chicago collection #2W9); and a-Al2O3 (prepared f As hydroxyl. from reagent AI(OH)3, fired at 1,300"Cfor 2 days). 152 BARTON ET AL.: FLUOR-TOPAZ of the above observations, we estimate that the Table 2. Experimental heat capacitiesfor fluor-topaz. Formula 184'M3 g/mol. correction to the measuredentropy of our sample, weight caused by the replacement of 10 percent (10 times Eee t what is actually present) of the fluorine by hydrox- TeuP' rcop. reDP. capeclty "";::i., ".li31., yl, would be less than 0.1 percent, which is some- R J/ (rol'K ) K J/ (ool'K ) K J/( ool'K) what less than our experimentaluncertainty. Figure 1 shows the measuredheat capacities. serles I Serl.eo 3 Serles 5

305.57 145.3 243.ar r2r-7 470.7 r87.3 Dffi rential scanningcalorimetry 3to,27 r48.0 249.36 124-O 480.7 r88.9 3 15 .93 l|s .7 254.E9 125.4 190.l r90 . 5 Heat capacity measurementswere madefrom 340 !2L.4a 151.4 260.41 r29.r 500.6 192.2 121.3r 153.5 255.9r r31.3 5r0.6 193.3 to 800 K with a differential scanningcalorimeter at 333.r 2 I 55.3 271 '40 133.4 520.6 194.5 338.92 r56.7 276.87 135.7 530.5 196.1 the U. S. Geological Survey. A cleavage flake of 344.72 158.3 282.34 137.8 540.6 196'1 350.49 160.2 2A7.79 139.7 550.6 198.0 topaz weighing 25 mg placed in an unsealed gold 356.26 162.0 293.23 142.O 560.5 r99 .2 pan and reduced 352.O2 163.0 298.65 144.0 570.5 200.8 was used. The data were collected 367.f7 164.4 304'(i7 145.8 580.5 200'9 using a computer program written by K. M. Krupka 373.50 156.3 309.47 147.6 590.5 20r.8 37g.21 167.7 314.87 L49.4 500.5 202.9 of Batelle Northwest Laboratories, which utilizes 320.25 I 50.9 serles 2 Serlca 7 scans made with a corundum disk over the same Serles I 54,20 6.547 770.2 213.3 temperature intervals as the unknown. The heat 50.48 E.687 lo .56 0.0189 780.2 213.1 given Ditmars and Douglas (1971)for 65.10 11.44 LL.77 o.026l 790.2 215.2 capacities by 70.54 14.92 t2.99 o.0356 800.2 215.6 corundum were used in the reduction. Additional 76.80 18.53 14.33 0.0538 82.86 22.23 r5.85 0.0768 s€rlea I details of the experimental method are given by 89.O2 26.11 17.s4 o.ll4r 95.18 30.r9 t9.42 0. t59r 470.7 87.1 Krupka et al. (1979).Because of dfficulties with lol.29 34.31 2L.52 o.2400 480.7 88.5 not be ob- 107.36 38.49 23.87 o.34rr 490.7 89.8 instrument stability, reliable data could 113.40 12,63 26.49 0.4993 500.7 91.1 tained at temperaturesgreater than 800K. The data 119.40 45.81 29.43 0,7533 510.6 91.7 t25.37 51.06 32.73 t.215 520.5 92.7 are given in Table 2, series 5-9, and are plotted in r3r.32 55.r8 36.41 | .a22 530.6 93.9 r37 ,24 59.25 40.56 2.723 540.6 94.7 Figure 1 143.I 4 63.22 45.22 3.982 550.6 96.0 t49.O2 67.12 50.44 5.527 560.5 97.1 56.27 7 .120 570.5 9?.7 Enthalpy of solution measurements SerleB 3 62.56 9 .835 580. 5 98.8 590.5 200.1 High temperature oxide-melt calorimetry was r54.82 7l .30 Serles 5 500.5 203.0 160.46 74.92 used to determine the enthalpy of formation of r65.70 78.35 57O.5 199.0 serle6 9 171.04 81.73 560.5 200.0 fluor-topaz according to the reaction: t76.7t 85.21 590.5 200.8 340.0 157.5 r82.35 88.84 500.5 201.5 350.0 159.6 (1) 187.98 92.O9 610.4 202.9 360.0 161.4 Al2O3 + SiO2 + CaF2: AlzSiOaF2-r CaO 193.51 95.40 620.4 203.7 370.0 163.9 199,23 98.51 630.4 204.3 380.0 166.3 Enthalpy of solution values were obtained for CaO, 204.82 r01.5 640.4 204.8 390.0 169.0 2to.42 r04.9 650.4 205.I 400.0 171.3 Al2SiO4F2,CaO + Al2SiO4F2(molar 1:1 mixture), 216.00 108.0 650.4 205.4 410.0 171.3 22t .58 I tO.5 570.3 206.3 420.0 175.0 and CaF2+ Al2O3+ SiOz(molar 1:1:l mixture) and 227.t4 1r3.2 580.3 207.6 430.0 176.6 232.70 1r6.3 690.3 207.O 440.0 179.1 238.25 119.1 700,3 209.7 450.0 181.3 460.0 183.2 470.0 185.0 480.0 187' 0 490.0 la9 .2 500.0 r91.2

Cp are given in Table 3. Attempts to measure the enthalpy of solution of pure CaF2 failed due to . DSC sintering of the powder which prevented complete . AOIABATIC dissolution during the time of the experiments (50 minutes). Westrich and Navrotsky (1981)encoun- tered similar problems with the dissolution of fluo' rite. T KELVINS The calorimetric methods have been described Fig. l. Experimental heat capacities for fluor-topaz. recently by Charlu et al. (1978). Heat-treated BARTON ET AL.: FLUOR-TOPAZ 353

Table 3. Enthalpy of solution measurements

Materia I AH0zo Experimentaldata in chronologicalorder aH!79 (kJ/mol); Samplewt (mg) in parentheses KJ

Alzsi 04F2 76.57+ 1.67* 74.74(2s.88),74.2r(44.77),79.68(41.84), 80.16(47.96) 72.46(4t.28),78.64(3s.32),80.39(50.80), 71.54(28.32) 74.56(37.60),78.04(42.71),78.88(36.29), 79.s3(24.50)

Ca0 -55.06+ -62.10(13.78), 4.10 -57.59(12.71),-S+.gO(10.44), -45.68(15.1 1 ) -12.e7{16.22), -64.68(14.50), -63.32(13.10), -48.s4 ( 15.44 -60.33(13.06),-50.11(13.99), -+g.OZ(r+.AS) )

CaO+ Al2Si04F2 29.83+ 3.05 20.96(29.58),29.19(26.69'), 38.67(34.10), 31.67(s6.75) ?9.91(s9.10),26.60(47.48), 28.81(66.64), 3e.10(s0.63) 29.48(4s.00),23.11(s3.18), 32.37(73.24), 2e.25(44.86)

CaF2+A1203+Si02 l2l.7l + 1.80 122.31(3A.74),124.72(36. 39), ll8. s2(30.50) 127.65(38. 90), t22.20(40.78),120.26(46.13) 117.79(39.05),118.32(40.45) addition of CaOto 21.51+ 4.44 Al25t 04F2

CaF2+ A1203+ Si0Z = CaO+ At2Si04F2 aHiTo**=91.88 + 3.56kJ aHlge = 96.12+ 3.95kJ

' Twostandard errors of the mean;errors are propagatedassuming they are independent;possible systematic errors have been neqlected. ** Usingthe value for the mechanicalmix of CaO+ Al2Si04FZ;see text.

Pb2B2Osglass from a singlehomogeneous batch was however, was more than 200 mg of material dis- usedin all the experiments.The topaz,CaO, Al2O3, solved in any given 30.5 g batch of melt. -150 SiO2,and CaF2were sized to to *350 mesh Two diferent values were obtainedfor the heat of and kept in a desiccator. solution of the Al2SiO+Fz* CaO. The linear combi- A 2 month exposure to air in the desiccator nation of the heat of solution of Al2SiOaF2and the reduced the heat of solution of CaO by about 10 heat of solution of CaO (21.5!4.44 kJ/mol) differs percent, apparently because of the formation of significantly from the heat of solution for the me- small amounts of Ca(OH)z and/or CaCO3. Brief chanical mixture (29.83-F3.05 kJ/mol). Consecutive heating to 1,300'Cbefore each experiment eliminat- dissolution experiments of stoichiometric propor- ed this perturbation. The final value obtained for tions of Al2SiO4F2followed by pure CaO in the All"o1. at 970 K of CaO in this study (-55.06 same alliquot of melt confirmed the mechanical- kJ/mol) agrees very well with that reported by mixture value. Therefore, the value of the mechani- Newton et al. (1977)(-55.15 kJ/mol). None of the cal mixture is used in the evaluation of the enthalpy other materials showed a time dependencein their of fluor-topaz. We have considered the effect of heat effects. In addition, consecutivemeasurements systematic elrors, including those resulting from on the same material in a given batch of melt hydrolysis (of topaz, CaFz, CaO), reaction of the showed no significant variation with increasingcon- mechanical mixture before the experiments, and centration of solute. In no series of experiments, impurities (e.g., 0.04 percent H2O in the topaz). 354 BARTON ET AL.: FLUOR-TOPAZ

Table 4. Smoothed thermodynamic properties for AlzSiO+Fz Table 5. Standard thermodynamic properties offluor-topaz at I bar Pressure

IEMP. SEAT ENTROPY BNIHALPY GIEBS ENERGI CAPACITT FUNCTION FUIICTION aH?,zgg = -3084.45+ 4.70 kJ/mol ci tsi-sir tui-nirrr - rci-niI rr oGl,zgg = -2910.65+ 4.74 kJ/mol XELVIN J/(rol'r)

S2sA = 105.4+ 0.2 J/(mol'K ) 5 0 .002 o. o0l 0,002 0.00r lo 0.015 0.005 0.004 0 .001 - J/(mot'K)* l5 0.063 0.0r8 0.o14 o .004 c; = 471.41- 0.08165T+ l'2595x106T-2 548s.sT-0'5 20 0.r88 o.o5l o.040 0.011 25 0.403 0,113 0,088 0.025 0.219 0.17r 0 .04E 30 o.422 * Valid over the range 200-1000K. 35 1.575 0 ,400 0.3r6 0,084 40 2 .588 o.672 0.533 0.140 45 3.9r9 1.052 0 .833 0.2r9 50 5.391 1.54r 1.2r5 0.326 60 8.495 2.775 2.148 o.527 any dissolution experiment will reflect the cumula- 70 14.56 4.521 1.471 1.o50 80 20.52 6 .858 5.232 t.625 tive composition of that particular batch of solvent. 90 26.75 9.628 7.271 2.351 loo 33.44 r2.79 9.552 3.239 This problem has not been reported for molten-salt oxide systems and I l0 40,29 l6 .30 r2.03 4 .264 calorimetric studies of simple 120 17.25 20.10 14.68 5.424 has been argued against on the basis of experiment 130 5! .28 21.16 17.45 6.707 140 61.11 28.44 20.13 8.106 (Charlu et a1.,1978).In this study, the addition of 150 67.87 32,88 21.27 9 ,508 160 74 .66 37.14 ,6.27 11.2r the "unusual" component fluorine apparently has 170 81.07 12.20 29.31 12.89 solution, 180 87.33 47.01 32.35 14.65 changed the character of the calorimeter 190 93.30 51.E9 35.41 t6.48 possibly through the formation of one or more 200 98.92 56.42 38,45 r8 .38 fluoride complexes and changesof coordination of 2lo 104.5 6t.79 4t.46 20.33 220 109.8 66.78 r4.45 22.32 the cations. Similar excess heats have been found 230 114,8 7t.71 11.10 2e.36 (oral 240 119.9 ,6.76 50.32 26.44 by Stein Julsrud of the University of Chicago 250 124.3 8r.t5 53.19 28.56 Westrich and Navrotsky 250 r28.E 86.71 56.01 30.70 communication, 1980)and 270 132.9 91.65 5E.79 32.46 (1931)in experiments on other fluoride-oxysalt sys- 280 136.9 96.56 6r.51 35.05 290 140.7 r01.4 61.17 17.26 tems. 300 144.4 106.3 66.78 39.48

3r0 147.8 llr.q 69,34 41.7r Thermodynamic properties of fl uor'topaz 320 I 5t.O rr5.8 71.44 43.95 330 I 54.3 r20.5 74.29 46.20 state (298.15K, I bar) thermody- 3t0 157.1 t25,1 t6.69 48.45 The stdndard !50 160.0 L29.' 79,42 50.71 namic properties of fluontopaz can be derived frort 360 162,7 134.3 8r.31 52.97 370 165.2 138.8 83.54 55.22 the experimental data presented here. In order to -2 85.t3 57.48 380 168.0 t11 derive the entropy, the measureddatawere extrap- 400 112,4 152.0 89.9 62.L the temperature irrter- 450 182.3 t72.9 t9.1 73.2 olated smoothly to zero from 500 190.3 l9?,5 r08.3 84.2 vat 10-16 K. Over the rangefrom 0-10 K, valuesfor 5to 196.8 21r.0 115.2 94.8 600 202.0 228.3 Lzt.l r05.2 enthalpy and entropy were obtained by graphical 650 206.2 214.7 t29.3 115.4 700 209,5 250. r 134.9 t25.2 integration. From 10-380 K, the data were analyti t 50 2r2 .r 274.6 140.0 134.6 The 800 214.I 2A8.4 144.5 143.9 cally smoothed by using cubic spline functions. 850 215.5 301.4 148.7 r52.7 to SzqrSo causedby the extrapolation 900 216,6 t13.8 152.5 161.3 contributions 950 2r7.3 325.5 r55.8 L69,7 of the measurementsbelow 10 K is 0.018J/mol'K. 1000 217.6 335.5 159.0 L77.6 Smoothed values for Cn, S;s; (rt'-t{dlr, ana 273.15 r34.5 93.20 59.65 33.55 298,r5 144.0 105.4 66 .30 39.06 (Grlfdff are listed in Table 4. For ideal fluor-topaz, there should be no zero- point entropy because all the aluminum sites are These possible effors should be small and have octahedral, excess alumina (substituting for Si) is been neglected. Even if one or more of these effors not observed in natural topazes (Ribbe and Rosen- were important, it would not change the observa- berg, 1971), and all the (F,OH) sites are filled by tion of the excessheat of solution, althoughit would fluorine. change the absolute values of the heats of solution. A four-term polynomial was fit to the adiabatic The observed diference shows an excess heat of (200-380 K) and differential scanningheat capacity solution in PbzBzOsmelts for CaGAlzSiOaF2 mix- measurements, using PHAS2O (Haas and Fisher, tures. Consequently, the heat efrect observed for 1976).The standarddeviation ofthe fit to these data BARTON ET AL.: FLUOR-TO?AZ 355

is 0.7 percent. This equation is considered to be Krupka, K. M., Robie, R. A. and Hemingway, B. S. (1979)High- accurate to within 1.4 percent to 1,000 K. Above temperature heat.capacitiesof corundum, periclase,anorthile, 1,050K (dCJdD ( 0 so this equationshould not be CaAl2Si2Or glass, muscovite, pyrophyllite, KAlSi3Os glass, used for extrapollition grossular, and NaAlSirOe glass. American Minemlogist, 64, to higher temperatures.The 86-10r. thermodynamic properties at 299.15 K (Table 5) Lindsey, D. A. (1979) Preliminary report on Tertiary volcanism were derived given the heat of reaction (1), the and uranium mineralization in the Thomas Range and northcrn ancillary data for the other phases (taken from Drum Mountains, Juab County, Utah. U. S. Geological Sur- Robie et al., 1979), and the entropy and heat vey Open-file Rept. 79-1076. Nassau, K. and Prescott, B. E. (197t Blue and capacities obtained here. The uncertainty in the brown topaz produced by gamma irradiation. American Mineralogist, 60, enthalpy and Gibbs free energy arises mostly from 705-1@. the uncertainty in the heat of solution of the CaO + Newton, R. C., Charlu, T. V. Anderson, p. A. M. and Kleppa, Al2SiO4F2mixture. O. I. (1977) Thermochemistry of high pressure garnets and clinopyroxenesin the systemCaGMgO-Al2O3-SiO2. Geochi- Acknowledgments et CosmochimicaActa, 41, 369-377. Penfield, S. L. and Minor, Jr., J. C. (lB%) On the chemical MDB has been supported by a National Science Foundation (NSF) composition and related physical properties of topaz. Ameri- graduate fellowship and by a McCormick Fellowship at can Journal of Science, 3rd Series, 47, 387:396. the University of Chicago. The high-temperaturesolution calo_ Ream, L. R. (1979)The Thomas Range, Wah Wah Mountains, rimeter was constructed with support from NSF grant DMRZS_ and vicinity, western Utah. Mineralogical Record, 11567to OJK. NSF grant EARTE-13675 lO, 261- to Julian R. coldsmith 278. (University of Chicago) provided additional support. We appre- Ribbe, P. H. and Rosenberg, p. E. (lql) Optical and X-ray ciate helpful reviews by J. J. Hemley and G. R. Robinson, Jr. of determinative methods for fluorine nbpaz. American Miner- the U. S. Geological Survey and by H. R. Westrich and A. alogist, 56, l8lLl82t. Navrotsky of Arizona State University who also provided us Robie, R. A. and Hemingway, B. S. (lg2) Catorimetersfor heat with a preprint of some of their recent work on fluorosilicate of solution and calorimetry. low-temperature heat capacity measurements. U. S. Geological Survey Professionalpaper 255. Robie, R. References A., Hemingway, B. S. and Fisher, J. R. (lg9) Thermodynamic properties of minerals and related substances Burnham, C. W. (1%2) Lattice constant refinement. Carnegie at 298.15 K and I bar (105 pascals) pressure and at higher yearbook, Institution of Washington 61, 13.}_135. temperatures.U. S. GeologicalSurvey Bulletin 1452(revised). Robie, R. A., Hemingway, B. S., and Wilson, W. H. (1976)The heat capacities of Calorimetry Conference copp€r and of muscovite KAl2(AlSi3)Oro(OH)2, pyrophyllite Al2Si4Oro (OH)2, and illite K3(At7Mg) (Si,4At )O4o(OH)sbetween 15 and 375 K and their standardentropies at29E.l5 K. U.S. Geologi- cal Survey Journal of Research,4, 631-W. Rosenberg, P. E. OnD Compositional variations in synthetic nance of metal ions and defect centers in topaz. Journal of topaz. American Mineralogist, 57, l6E-187. Physical Chemistry, 7 l, 231-240. Rosenberg, P. E. (1978) Fluorine-hydroxyl exchangein topaz. Ditmars, D. A. and Dorrslas, T. B. (1971)Measurements of the (abstr.) TransactionsAmerican GeophysicalUnion, EOS, 59, relative enthalpy of pure o-Al2O3 (NBS heat capacity and l2lE. enthalpy st4ndard reference material no. 720) ftom 2ii to ll23 Westrich, H. R. and Nawotsky A. (1981)Some thermodynamic K. U. S. National Bureau of StandardsJournal of Research. properties of fluorapatite, fluorpargasite, and fluorphlogopite. 75A,401420. American Journal of Science, 2E1, l09l-l 103. Haas, J. L. and Fisher, J. R. (1976)Simultaneous evaluation and correlation of thermodynamic data. American Journal of Sci- Manuscript received,August 20, 1961; ence,276, 525-545. acceptedfor publication, November 24, 1981.