American Mineralogist, Volume 58, pages 343-345, 1973

New Data on Forsteriteand MonticelliteSolid Solutions

Hounc-Yr YeNcl Department of the GeophysicalSciences, Unioersity of Chicago, Chicago,Illinois 60637

Abstract

Portions of the solvi in the system forsterite-monticellitewere studied at 1440'C, 1463'C, 1481'C, and 1496'C by determiningwith an electronmicroprobe the compositionsof coexisting crystals of forsterite and monticellite solid solutions which were crystallized from the "dry" melts with compositions ofi the system forsterite-monticellite. The extent of th€ forsterite solid solution determined in this study is less than any of those reported by the previous investigators. The monticellite limb obtained with the present method appears to be more sensitive to the change in temperature than has been suggestedbefore.

fntroduction tallizes readily from glass or gel and persistsin the fail to dissolvemetastable The suggestionthat CaO content in olivines could run product. Fo." might from gel and stable Mou* be a po,tentialenvironmental indicator for the crys- Mo"* when crystallized from glass tallization of olivines and their host rocks (Simkin might fail to nucleate when devitrified (Biggar 1969). Biggar and O'Hara and Smith, 1970) inspiredthe presentstudy on the and O'Hara, (1969) the problemsof obtaining limits of forsterite (Fo*) and monticellitesolid solu- concludedthat overcomein their study even tions (Mo,"), concerningwhich there are serious equilibrium were not longer than those used by disagreementsamong the investigators.The limits with experimentaltimes the limit of detect- of solid solutionson the systemforsterite-monticellite other investigators.In addition, X-ray diffraction and the slight changes above 1400oC at atmosphericpressure reported by ability of with changing compositionsun- Ricker and Osborn (1954) differed greatly from in lattice spacings the of their analyses. thosereported by Biggar and O'Hara (1969) (Fig. doubtedly restricted accuracy it is clear that any attemptsto obtain better 1). The experimentaland analyticalmethods were Thus, of Fq"" or Mou* must include essentially the same in these two reports, except data on the limit the reactionsas well as a better that Ricker and Osborn used glassesas starting methodsto accelerate to determine the compositions material, whereas Biggar and O'Hara used gels. analytical technique of the solid solutions. In the present study, coexist- Using crystalline mixtures of the end members as ing Fo," and were crystallizedin the presence starting material and the same experimental and Mo* amount of liquid between l44O"C analyticalmethods, Warner (1971) reporteda limit of substantial 1496'C at atmospheric pressure and were of 27.5 -r 2.5 wt percent monticellite in forsterite and electron microprobe. solid solution and 22.5 t 2.5 wt percentforsterite in anaTyzedwith an monticellite solid solution near 1450oC.The former value was close to that of Ricker and Osborn Method of Study (1954), but the latter was close to that of Biggar The chemical reaction or crystallizationproceeds and O'Hara (1969). faster when a catalytic liquid is present than when Failure to attain equilibrium, resulting from the the system is dry. Thus, the kinetic difficulties in slow reaction rate in the solid state even at temper- crystallizing an equilibrium assemblageof Fq," and atures above 1400'C, appearsto be responsiblefor Mo* in a reasonablyshort period of time could be these conflicting results. Metastable periclase crys- overcome by crystallizing them in the presenceof liquid. While Fou, * MQ"" * liquid equilibrium can 'Present address: Code 644, Goddard Space Flight be obtained only at the temperature of invariant Center, NASA, Greenbelt, Maryland 20771. point from a starting material with a bulk comp+'

JI+J 344 MINERALOGICAL NOTES

Tesrn 1. Quenching Experiments and the Compositions of Forsterite and Monticellite Solid Solutions Foss+Mgo+L Pbses* composltlons (wr Z) TenPerature Dulatlon CoDPosltlons of Cryslalltne sIo; uso cao ("c) (hours) wt Z Mo ln Foss !t Z ro 1n uoss U + Mo ss Mg0+L 43.8 23.3 32.9 L440 8 13.5-14 5 16.2-18.0 f 41.0 29.O 30.0 1496 a L7.2-78 7 2?.2-28.7 F 1481 8 16.9-17.3 23.5-25.5 E luo 1481 15 5-1l.5 23.8-24,8 = .l U \l quenchlng runs' \ * Substantial eounts of glass 'ere present ! aL1 iMer + Mo.. the melts. i0 20 30 40 50 60 70 80 90 Fo* or Mou" crystallizesfrom IVIONTICELLITE WEIGHTPERCENT FORSTERITEAs observedby others during investigationof the metastablecrystallization of periclase,the quenching of portion of the system Flc. 1. Phase relations the (Table 1) forsterite-monticellite determined by Ricker and Osborn runs at l463oc, 1481"C, and 1496"C (1954) are shown in light solid lines. Solidus and solvi contain a few crystals of periclase in addition to reported by Biggar and O'Hara (1959) are shown in Fo* and Mo*". Note that the crystallization of dashed lines, The results of the present study are shown in metastablepericlases does not necessarilyaffect the Fo"", forsterite solid heavy solid curves. Abbreviations: equilibrium compositionsof coexistingFo," and Mo* Mo"", monticellite solid solution; L, liquid; Mer, solution; com- merwinite. in this study. Contrarily, it might causetheir positionsto depart from that of the starting material in cases when the starting compositions used by sition on the join Mo-Fo, the phaserelations of the previous investigatorswere on the Fo'Mo join and systemCaO-MgO-SiOz (Ricker and Osborn, 1954) in the single-phasefield. indicate that equilibrium can be obtained at any The solvi here observed between l44O' and temperaturebetween l5O2"C and 1430"C from a 1496'C in the systemFo-Mo (Fig. 1) showa forste- starting material with a bulk composition removed rite limb very close to that given by Biggar and from the join Fo-Mo. O'Hara (1969). However,the monticellitelimb dif- Glasseswith bulk compositionslisted in Table I fers in position and in slope from those reported by were prepared, and powdered devitrified glasses the other investigators. were held at temperaturesbetween 1502oC and 1430'C for 8 or 48 hours.Coarse-grained, euhedral Conclusion crystals of Fo." and Mq." embeddedin glasseswere The solvi in the system forsterite-monticellite found in the quenchingruns and were analyzedwith better be determined with the method used an electron microprobe. The microprobe technique could the present study. Further application of this is capableof determiningthe compositionsof crystals in to determine the subsolidusphase relations in situ with high precision,and it is consideredbetter method the compositionsof complex solid solutions for than any current X-ray or optical method. The or the other systemsis thereforeencouraged. operating conditions of the microprobe and the proceduresfor the reduction of data were described Note Added in Proof. R. Warner reinvestigated elsewhere (Yang and Foster, 1972). Ten crystals Fo"" recently and found a limit of 19 wt percentMo of Mo", and five of Fo* were analyzed in each in Fo* near 1450'C at atmospheric pressure quenching run for Mg, Ca, and Si. The range of (Warner, personalcommunication). these analysesis reported in Table 1, together with other experimentaldata. Acknowledgments This investigation was supported by NSF Gtant GA-21123 Resultsand Discussion to J. V. Smith. The Fo.ucrystallized at 1481"C in two quenching runs of 8 and 48 hours respectively are virtually References identical in composition,as are the Mo*" crystallized Brcc.rn, G. M., eNo M. J. O'Hene (1969) Monticellite and in the same runs. This suggeststhat apparent equi- forsterite crystalline solutions. Amer. Ceram. Soc. J. 52, librium can be obtained within a few hours when 249-252. MINERALOGICAL NOTES 345

Rrcxrn, R. W., rNo E. F. Ossonx (1954) Additional phase tem CaO-MgO-SiOrHO. Ph.D. Thesis, Stanford Uni- equilibrium data for the system CaO-MgO-SiOe.Amer. versity. 165 pp. Ceram. Soc. 1. 37. 133-139. YrNc, H. Y., rNo W. R. Fosrrn (1972) Stability of iron- SruxrN, T., lllu J. V. Sr"um (1970) Minor-element distri- pressure. bution in olivine. I. Geol.7E,304-325. free pigeonite at atmospheric Amer. Mineral. WARNER,R. (1971) Experimental inuestigationsin the sys- 57, 1232-1241.

American Mineralogist, Volume 58, page 345, 1973

ApatiteGhemistry and PhosphorusFugacity tn a Differentiatedlgneous Intrusion: Gorrection W. P. Nnsn Department of Geological and Geophysical Sciences, Uniuersity ol Utah, SaIt Lake City, Utah 84112

Values of. LG,o/2.303 RT for two reactionsgiven 5.26 1-+ 2.4) and A log fn : 2730/T + 7.93 by Nash (1972) are incorrect. (! 2.3). It is clear, upon recalculation,that the For the reactionCao(PO+)sOH + 5/2 MgzSiO+* effects of oxidation state and silica activity on the KFesAlSLOtn(OH), + l/2 KAISiBOs+ 15/2 equilibria are not immediatelyevident, as previously SiO2 = 5 CaMgSizOo* 3/2FerSiOt + 3/2 thought. Although the data of Stormer and Car- KAlSi2O6+ l5/4 Oz + 3/2 H2O + 3P the correct michael (1,97l) indicate a decreaseof phosphorus value is fugacity with decreasingmagmatic silica activity, no such simple behavior is indicated by the new curve ac"o : !68-sos- s7.87(+7.2). for Shonkin Sag laccolith which crossestheoretical 2.303RT fr basalt and rhyolite curves. Contrary to prior con- clusions, reduction in oxygen fugacity apparently For the reactionCar(PO+)sOH + 3 MgAIzO+* increasesphosphorus fugacity values, as indicated 57/4 SiO2 + l5/2 Fe = 15/4 FezSiO+* l/2 by the equilibria for a basalticassemblage coexisting Mg2SiOr * 2CaMgSi:Oo* 3CaAlzSirOa* l/2 with metallic iron. H2O + 3P the correct value is References ac,o -szs : !l - 20.65(+6.e). phosphorus 2.303RT T Nesn, W. P. (1972) Apatite chemistry and fugacity in a differentiated igneous intrusion' Amer. Min- The calculated phosphorusfugacities are shifted eral. 57,877-886. eNo I. S. E. CenvrrcHeBr(1971) Fluorine- to higher values of fn. The equilibrium curves for Sronurn, J. C., hydroxyl exchange in apatite and biotite: a potential thesetwo reactions(Nash, 1972,Fig.1) areshifted, igneous geothermometer. Contrib. Mineral. Petrology, 31, respectively,by the values A log fn = 5478/T * r2t-13r.