American Mineralogist, Volume 64, pages 1133-1155, 1979

Temperature-compositionrelationships between naturally occurring augite, pigeonite,and orthopyroxeneat one bar pressure

MeIcoTu RoSS AND J. SrepHEhr HUBBNBR U.S. Geological Survey Reston, Virginia 22092

Abstract

Natural orthopyroxene crystals containing augite exsolution lamellae were heated at one bar and controlled oxygen fugacity to determine the temperature at which orthopyroxene + augite first react to form pigeonite + melt. Pigeonite was detected by X-ray precession pho- tography. The reaction to form pigeonite is initiated at the orthopyroxene-augite boundary, probably on the augite lamellae, which are structurally similar to pigeonite and where the or- thopyroxene and augite are in chemical equilibrium. The upper temperature lirnit for the minimum stability temperature of pigeonite at one bar, determined from heating experiments on natural pyroxenes,varies from 1284"C at Mgl(Mg + )Fe):0.92 (atomic), to 90loC and 0.04. Orthopyroxene which does not contain augite exsolution lamellae, and remains unsatu- rated with respect to augite component, transforms to pigeonite at the higher temperatures of the orthopyroxene-pigeonite stability field. Partial melting accompanies formation of pigeon- ite for orthopyroxene compositions more magnesian than Mg/(Fe+Mg) : 0.55 (atomic).

Introduction existenceof three pyroxenesat one bar and a particu- lar bulk composition.This work has previously The pyroxene minerals, becauseof their presence been reportedin abstractform (Rosset al., 1972; in a wide variety of igneousand metamorphicrocks, Rossand Huebner,l975a,b). are the subject of intense and continuous study by many petrologists, mineralogists, and crystallogra- The quadrilateral pyroxenes phers. An understanding of the stability fields of theseminerals in terms of temperature,pressure and Definitions chemicalcomposition is crucial to understandingthe For the purposesof this paper, we consider the crystallizationhistory of pyroxene-bearingrocks. Of "quadrilateral pyroxenes"to be augite,pigeonite, or- the seven common pyroxenes,jadeite, acmite, om- thopyroxene,and protopyroxene.The chemicalcom- phacite,aegirine-augite, augite, pigeonite, and ortho- position of these pyroxenesis given by the formula pyroxene,the last three are by far the most important M2T2O.. The M sites may contain Ca,Mg,Fe'*, as and are the particular subject of this paper. One or well as lesseramounts of Fe3*, Li, Na, Mn2*, Tia*, more of thesethree pyroxenesoccurs in many terres- Cr3*, and Al3* in natural pyroxenes.The T sitescon- trial and lunar igneous rocks and in metamorphic tain Sio*and Al'*. rocks of the pyroxene-hornfels, amphibole-gran- All augites have space-groupsymmetry C2/c. ulite, and granulite facies. Most augitesare in the compositionrange defined by In this paper we give the resultsof heating experi- the following occupancylimits of the M sites: ments at one bar on naturally-occurring ortho- (l) ca + Mg + Fe2"+ Na * Fe3*+ Al > 1.00 pyroxeneswhich contain augite exsolution lamellae. (2) l00Ca/(Ca + Mg + Fe'z*)> 20 From these experimentswe determined the temper- and aturesat which the transformation (3) 100(Ca+ Mg + Fe2*)/(Ca+ Mg + Fe2*+ Na + Fe3*+ Al) > 75 opx t augite-+ opx + augite * pigeonite+ liquid Pigeonitespossess space-group symmetry A/c, or is first observed. This temperature is equal to, or FLr/c at lower temperatures,and have M-site occu- greater than, the minirrrum temperaturefor the co- pancy fimits ot 00n.3-004X/79/ l l l 2- l l 33$02.00 l 133 I 134 ROSS AND HUEBNER: AUGITE, PIGEONITE, AND ORTHOPYROXENE

(l) Ca + Mg + Fe'z*= 1.40 gaard pyroxene trend," now firmly establishedin and the geological literature, commenceswith copreci- (2) 100Ca/(Ca + Mg + Fe'*) < 20 pitation of magnesium-rich augite and ortho- Orthopyroxeneshave spacegroup symmetry Pbca pyroxene followed by coprecipitation of augite and and M-site occupancylimits of: pigeonite; the pyroxenesincrease in iron content as (l) Ca + Mg + Fe2*> 1.40 fractional crystallizationproceeds (Brown, 1957,Fig. and 2). This crystallizationtrend has also been observed such 12) l}}Ca/(Ca + Mg * Fe'z*)< 7 in numerousother volcanic and intrusive rocks, Protopyroxenehas not been found to occur natu- as BushveldComplex (Atkins, 1969)and the Hakone rally but has beensynthesized by many workers,usu- tholeiitic andesites(Nakamura and Kushiro, 1970b). ally has the compositionMgrSi'O", and has been re- Some of the most extensivepyroxene fractionation ferred to as protoenstatite.tto (1976)was the first to trends are observedin the lunar basalts(summarized prepare good single crystals of protopyroxene; the by Papike et al., 1976).Some individual pyroxene composition he synthesized is Li'ScrMgr-r*SirOu. phenocrystsin the lunar basalts are chemically and The spacegroup of Ito's phase is Pbcn, based on a structurally zoned so as to reflect the entire pyroxene full three-dimensionalcrystal structure analysis (H. crystall2ation trend, the phenocrystcores being mag- T. Evans,Jr., personalcommunication, 1977;Smyth nesium-rich pigeonite, the rims iron-rich augite and lto, 1977),which verifiesthe symmetryproposed (Boyd and Smith, l97l). by Smith (1959) on protopyroxeneof the composi- The association augite-orthopyroxene is found tion MgrSirOu. most typically in granulite-faciesmetamorphic rocks, The compositionsof pyroxenesare commonly ex- but also in some pyroxene-hornfelsand transitional pressed in terms of mole percent CarSirOu(Wo), amphibolite-granulite rocks. The chemical composi- MgrSirOu (En), and FerSirOu(Fs). The amount of tions of the coexistingpyroxenes in granulite-facies these three componentsdepends on the method of rocks having a variety of bulk compositions have normalization of the chemical formula; different been carefully describedby a number of geologists, methods can give slightly different values.To avoid including Binns (1962)on granulitesfrom the Broken inconsistenciescaused by the use of different pyrox- Hill district of New South Wales, Subramaniart ene nonns to calculate "Wo," "En," and "Fs" we (1962) on charnockitesand associatedgranulites have chosen to express"quadrilateral" pyroxene from various localities, Davidson (1968) on gran- compositionsusing the relationships ulites of the Quairading district, Western Australia, (l) [Ca]: lOOCa/(Ca* Mg + zFe) Immega and Klein (1976) on iron formation, and (2) lMgl: l00Mg/(Ca * Mg + )Fe) Rietmeijer (19'19)on iron-rich igneousrocks in Nor- and way. Coexistingorthopyroxene and augite also occur (3) [Fe] : 100Fel(Ca + Mg + )Fe) in lunar metamorphic rocks such as the Apollo 15 where Fe : Fe2** Fe3*and Ca,Mg,Fet*,and Fe'* anorthosite 15415(Stewart et al., 1972),tllLe Apollo are the numbers of cations per formula unit. For 17 troctolite 76535(Gooley et al., 1974),and the no- "quadrilateral" pyroxenesthat do not contain large ritic breccia 77215(Huebner et al., 1975).Petrogra- amounts of AlrOr, TiOr, MnO, FerOr, and NarO the phic and chemographicevidence suggests that the or- in these [Ca], [Mg], and [Fe] values are close to the values thopyroxene and augite were equilibrated Wo, En, and Fs, respectively,calculated by the vari- rocks. ous normalization methods. Equilibrium assemblagescontaining ortho- pyroxeneand pigeoniteare apparentlyquite rare, for Occurrencesin igneousand metamorphicrocks thesetwo phasesin the presenceof melt coexistonly Cooling magmas crystallize quadrilateral pyrox- over a very narrow compositional range. Virgo and enesin a generally consistentsequence of composi- Ross (1973) reported on coexisting orthopyroxene tions such that the first pyroxenesto crystallize are and pigeonitephenocrysts in Mull andesiteglass but magnesium-richand the last are iron-rich. The stud- could not demonstratetrue equilibrium relationships. ies of Wager and Deer (1939),Brown (1957),and The three-phaseassemblage orthopyroxene-pigeon- Brown and Vincent (1963) establish quantitatively ite-augite is also quite rare: basaltic or andesitic the pyroxene crystallization trend in rocks of gab- magmas are in equilibrium with three pyroxenes broic composition-specifically those of the Skaer- only over very restricted temperature-composition gaard Intrusion, East Greenland. The "Skaer- ranges.In strongly fractionatingliquids, however,the nOSS,4.ryD HUEBNER: AUGITE, PIGEONITE. AND ORTHOPYROXENE I 135 melt may intersectthe three-pyroxenefield over an appreciable temperature-compositionrange where orthopyroxene,pigeonite, augite, and melt are in re- action relation (Huebner et al., 1972).Examples of three coexistingpyroxenes have been reportedin an- desitesfrom Weiselberg,Germany and Hakone vol- cano,Japan by Nakamura and Kushiro (1970a,b). Protopyroxenehas not been reported as occurring naturally; however, evidence exists that a magne- sium-rich pigeonite ("clinoenstatite") found in por- phyritic volcanic rocks from Cape Vogel, Papua 1 (Dallwitz et al.,1966)and in the ShawL-group chon- I 1000 E (Dodd drite et al., 1975)is an inversionproduct from E high-temperatureprotopyroxene. E ,o qoo Subsolidusreactions An exampleof the subsotduspyroxene phase rela- tions is depictedin Figure l, a planar sectionthrough the pyroxene ternary phase volume, and coincident with an orthopyroxene-augitetie line. Augite and pi- geonite which coprecipitate above the three-pyrox- ene stability region and then cool slowly will unmix pigeoniteand augite,respectively. Pigeonite on cool- ing into the two-phaseaugite-orthopyroxene stabitty volume may transform to the stableassemblage aug- ite plus orthopyroxene, thereby forming so-called 01020304050 "inverted pigeonite" (Huebner et al., 1975). "In- 100Cal(Ca+Mg+>Fe)+ verted pigeonite" generallyoccurs in the more slowly Fig. l. pyroxene cooled intrusive rocks such as those of the Temperature-composition section along the quadrilateral join ICa]e[Mg]uuIFe]3o-[Ca]5o[Mg].rIFe],, (insdt) Skaergaard,Bushveld, Duluth, and Stillwater intru- which is drawn to be coincident with an orthopyroxene-augite tie sions.Alternatively, pigeonitemay persistmetastably line at granulite-facies metamorphic temperatures. This join very and continue to exsolveaugite as in the more rapidly closely approximates a binary system at temperatures below the cooled lunar basalt 12021 (Ross et al., 1973). The three-pyroxene region (-1150'C). Elsewhere the section is pseudobinary. Fields shown are sections phase original augite on continuedcooling below the through ternary three- volumes Solid lines-stable subsolidus elements. Dashed line- pyroxenestability volume either may continue to ex- metastable subsolidus elements. Dot-dash lines-solidus and solve pigeonite metastablyor may nucleate and be- liquidus relationships. OPx:orthopyroxene, p:pigeonite, gin to unmix the stableorthopyroxene phase. Di:CaMgSizOo, Hd:CaFeSi2O6. Orthopyroxeneand augitecan crystallizestably to- gether below the three-phaseregion. Unless cooling augite and pigeonite from a lunar basalt (Rosset a/., is very rapid or the initial calcium content is very 1973).In that study, the augite-pigeonitesolvus and low, the orthopyroxenewill exsolveaugite. Even the temperature of first melting were delineated by low-calcium metamorphic orthopyroxenes(0.5 to 1.2 means of experimental techniques similar to those weight percent CaO) invariably contain some ex- usedin this work. solved augite (Jaffe et al., 1975).Augite that crystal- lized below the three-phaseregion may on cooling Orthopyroxeneheating experiments unmix stableorthopyroxene, metastable pigeonite, or both (Jaffeet al.,1975). Previouswork The availability of naturally-occurring unmixed Earlier investigationsdo not establishthe condi- pyroxenespresents an opportunity to study sub- tions at which naturally occurring orthopyroxene* solidus phase relationshipsby subjectingthe pyrox- augite react to form pigeonite. Neither of the two ene crystals to stepwiseheating experiments.Such previouscomprehensive studies took into accountthe experiments were carried out on single crystals of Ca-, Fe-, and Mg-bearing componentswhich must I 136 ROSS ll{D HUEBNER: AUGITE. PIGEONITE. AND ORTHOPYROXENE all be considered together when discussing natural these orthopyroxenes ranges from 0.917 to 0.046, en- pyroxenes. In the first study, Yang (1973) presented compassing nearly the entire range found in natural runs in the iron-free system CaO-MgO-AlrO.-SiO, orthopyroxenes. that bracket what is probably the orthopyroxene- Orthopyroxene crystals 0.1 to 0.5 mm long and of augite-pigeonite equilibrium. These results, which good optical quality were selected for the heating ex- will be discussed later, cannot be directly applied to periments. All crystals, except those from samples 29 the Fe-bearing natural pyroxenes. In the second in- and 30, contain augite exsolution lamellae oriented vestigation, Bowen and Schairer (1935) presented a on the (100) plane of the host (Table 2). The augite phase diagram (their Fig. 8) which depicts the phase lamellae can usually be detected optically, as well as relations of synthetic pyroxene on the join MgrSi'Ou in most single-crystal X-ray precession photographs -FerSirOu at one bar. Included in this diagram is an of the h\l net. Typical orthopyroxene crystals con- orthopyroxene-clinopyroxene transition loop which taining lamellar inclusions of augite are shown in was delineated from heating experiments on six natu- Figures 2A and 2B. The amount of augite in the rally-occurring orthopyroxenes. Their clinopyroxene single-crystal mode of the host orthopyroxene crys- is pigeonite as we now define it. The other parts of tals (Table 2) is estimated from the intensity of the their Figure 8 need not concern us here other than to 402 and 6-02reflections of augite which appear adja- mention that they report that the pyroxene solidus in cent to the 902 and902 reflections, respectively, of the pure system MgrSirOu-FerSirOuis 200 to 300oC orthopyroxene. The detection limit by the X-ray pre- above the OPx-CPx transition loop and that olivine cession method is about I mole percent augite ori- and tridymite may appear as reaction products upon ented within the host Phase. heating of very iron-rich orthopyroxenes. The formation of pigeonite in a heated crystal is Two other studies give a clue to the temperatures signaled by the appearance of the 002 and 002 pi- at which natural (Ca-, Mg-, and Fe-bearing) ortho- geonite reflections adjacent to the 202 and 202 (or pyroxene might be expected to react to form augite. 202 and 202) orthopyroxene reflections, respectively Smyth (1969) studied a heated single crystal of ortho- (Fig. 3A, B). The pigeonite, as will be shown later, pyroxene of composition [Fe] : 70 by X-ray film commonly grows epitaxially at the interface of the methods. At 1050+25"C, the crystal inverted slowly augite lamellae and orthopyroxene host, thereby pre- to pigeonite and did not react back to orthopyroxene cluding optical distinction of pigeonite lamellae at lower temperatures. We do not know whether he formed on heating from augite lamellae originally first heated the orthopyroxene at lower temperatures present. At higher temperatures appropriate to the and found no reaction. Brown (1968) found that at composition, all the orthopyroxene is converted to less than I bar, a natural pyroxene having an approx- pigeonite (Fig. 3C). As it is equally probable that pi- imate composition Cao,roMgorrrFe,oroSirOu, and con- geonite will orient with respect to the orthopyroxene sisting of orthopyroxene with coarsely exsolved aug- in either one of the two possible (100) twin configu- ite ("inverted" pigeonite), did not react to pigeonite rations, two sets of hll pigeonite reflections usually during runs of 7 to 17 days unless heated to 1200"C. appear in the precession photographs. Upon addition of andesitic melt flux, Brown was able The accuracy of the estimated single-crystal modes to lower the reaction temperature to l000oC, and to (Ross et al., 1973, p. 623) as obtained from visually- reverse the reaction at 990o; but changes in the py- read X-ray intensities is difficult to assess.We esti- roxene composition, which undoubtedly took place mate that as little as one mole percent pigeonite can by means of cation exchange with the melt, are un- be detected in the heated crystals. In a sequence of known. heating experiments on the same crystal, we can dis- tinguish between I and 2 percent pigeonite, 15 and Characterization of samples studied 20 percent pigeonite, or 40 and 50 percent pigeonite Thirty natural orthopyroxenes were used in our in orthopyroxene. We can also compare this method heating experiments. Their chemical compositions of estirnating crystal modes with values determined are given in Table l; all contain negligible amounts by measuring the areas of high-calcium (augite) and of NarO, CrrOr, and TiOr, and only moderate low-calcium (pigeonite ard/or orthopyroxene) py- amounts of CaO (0.2 to 2.5 weight percent), MnO (0 roxene in photomicrographs of coarsely unmixed to 3.6 weight percent), and AlrO, (0 to 2.0 weight per- augite and pigeonite crystals from the Wyatt Har- cent) with the exception of sample 30 which contains bour Complex, Labrador. Agreement between the X- 8 weight percent Al,Or. The Mg/(Mg + )Fe) ratio in ray and optical methods for four unheated crystals ROSS l.r/D HUEBNER: AUGITE. PIGEONITE. AND ORTHOPYROXENE tt37

@OQO@F6FOO O _ d;" ooooooN+oN o o o @ VL.O@ OOOOO-FNOO O 69O + N @ H qO - 0 L .qOO P + O .P ! . r .O FOOOOFQOOO N F6O O N O O O O OP O .L^> - Oo o = c .Ocd "C .!O+O O O + O Q ^60O ZOONO L O O O P< . J F c ! Y 6 O-OOO@N6OO O O . o o o L.F o=Ep o o .L .Fv 9OOOF6oFOo O O N oooo-oo@oo o ooN + @ t F A Nr O ! ! O. .d E ! d Oop -d N NO PP Fp L O .OOOOOL6!pv.FE -OOOOFOOOO N FO@ O N 6 L F V N = X OP O C o OJO 6 CP O F _+o, o o F or o o.F osvdr xo F L o OO 6D O 9- O E !-= J QF X E O @ O - dq o! =FOaD @60000NooN o - 6a OO >E C-OFQ$. PL--E E @oooN@oooN Q o d F L o il o o ' = .F o L'F c 6H x € o oOOOOOO@OO O +o- + + < d d! I Or O >- FF O @ dld p .u@. cv I P! o o.Fp - ov o d -QOOO-OOOO N FNO O N S= O O U X LV h . OOOp.F d^g O F +6 6 oP u O c < 62 - O o c l= d! F O F .P o H o d<.F < 6 "E O @vL c 4 N

->CL po L oFe^6 EL > 6 o o .uN. PC L N -.F60CLOOXCdFO .O O E FOOOOOOFOO N -FA O N o 9F + POvdOpErO 6dc! .F c F 6+ - O c: H o ^q6 c @ 6 o L^> 'Fo q 6 o o = c .Fr Ep oHD . o o O> + O^. .60 OE - ^az. 2 @qnoooQNoo- o N o- o L . d ldJo o r^E.FE = + @ ooooQ@+€oN o 6 OLN o Ov ..F-U o6..OO. O - oOOOO@OOOO O -F6 o - dL OOF \ E >OE v^t=oO O. O O .L o t9 o d.--.@. Ldo.FL.p! -OOoOOOFOO N Foo O i\ @ 6 o + > = O^ 686,^v GE ..F L q O F [email protected] O+ N 6 o r6 "=coooo 6 o il pp Q @.-D - c .o H < t o F o F O c o - XO o p U^pO Fz -o >x6 F o -L p F @@OOOoF@O+ O . 6 Z o c o c 6 x -^ =o oN c 6N € . ! o = o ooooo@oooo o +oo 6 6 o t O x o O O o -66 oO 6 OO @ L O O . . 's . o .F o .F L E O CZ .o - !<6! o Cp p O FOOOOOO-OO N Ng O o cL poo o oo oo^@. ooN r€€L a . > d P = o O o- oY . LO O o O @ . d oo dFvN>..^oFco^o c > +O -E E 6 6-6-OE+.-OzH O - o @ooNg6@@oo o o N< . FEOoFO OvO L=v O O +aOO+N-OO- O @ U F 6 P c 6 O O c F E c . !=.- . E ooooo6oooo o rL ^@ZE ..O .- 6I) oN+ @ N CO L >6EO "OXP O o O O L o { = o >t Q @^Eq dA 6 6 O p -OOOOOO-OO N FNF O 6 C F F + o<= O O OF 6 = Op o F O - @O 6 E C O -099rOOO@^ 6-OO ! @ or oop E 6J q-oets6=^acdc o d o + op olos_vH o o = 6 +N6O@FFON- O 6 oo o L o o ^L @ t o do 6 v^x . o o oooo@6FNo6 0 0 - - E - L E O < O =u o - - Y Oo o : ! ooooo+o+oo o @+o N 6 6O OF \ vO ttCtvl^.vL.p" .z@ 6P 09 L € o O O O OO^ >O.F p 6 -OOOOOO-OO N NFa O o .F L O O ..OC "d [email protected] J C q - NN lo do o b-o! pdLpo@@ no o o o O r c O O Ovo cE o = O -FOo^ r O O ! I O F O Y O E OF>6 -+JO C ! @NO+N@N-F@ O @ €o c6o F__ v 6 Nv r! o _ 6-NOO6-6OO O + pE -c6 il oF6^ o.-= c.@ 26. d O ooooo+o*oo o @oo N o OL J -F I o-O E p O O.F <+.. o .= OO EEJ - A> 6O^[email protected] ^ -oF O F< -oooooo-oo:-RSo-h os L: E < 9F . I oO o.- O=+ O 6 x oo! o odo-6r^uo6Fzoc^o o @l U O u o O u d c r@=oO c 5 !Y40.FOF E .l X F ouvvvvdoov=-dvv Nl o@ @o@oNo+ @ - a @@ ooo-Noo o dt Nl ot +t ot al oo roo+o+oo o NN6 N < | N .6 -O OOOOFOO N FhN O = N NO r @ o@ +oNN-oo @ N + o-oo@Fo@No L o0 @Q OO@ONOO O O N o oo rooooooo o N9o N < @oooooo900

-O oOOOFOO N -@O O o FOOOOOO-OO ooo F6

@N OO@NOO@ N O F6OOOONOON @+ oo+oNoo N - OFOOo6ONO- N OOTOOoO6OOO@$@@<-O ooooooo@oo .N -O OOOOFoO N FoN O | 6 FOOOOOO-oO OFN oo -@Foo

@l 0ooo6go-++ o o N+o+oF@@@n N oNNOOOOOO6 O O O 6NOO6@@@OO oooooNo600 0 @+@ @ 6 q o oooooNoooo -OOOOOOFOO N +Fo O - @ FOOOOFoOOO - @H @ o NI oo +oN6000 N N @ NOO+OO@OOO @ o@ ooNoood - o N-OONo+oOo v -r-o @ | o oOOOONO-OO

-O OOOO-OO N -+$ O @ -ooooFoooo FNO F - @F o @l -@ oONO+O+ + @ @ NONON@$FON @@ ooNo+oo o o o-OOO+qNOo oorooNo@oooNoo@N6 to FO OOOOFOO N F++ O @ F @- F O o +6 NONNOOF N N @@oo@+6@oo Q oo@ooo+ N @ nFOO@@OOOO N oo tooNo@oo o -+o @ @ ooooo@oNoo FO OOOOFOO N N+o O - N -OOOOFOOOO - @- O a @ O@ OOONFOO O N @o-+oo6ooo @@ 00+00-6 O @ O-OON@N@ON F oOIOONO@OOOFOo@Fo ooooo+oooo N ZN -O OOOO-OO N O+N O - @ NOOOO-OoQO =@ -@-o

+N--O6@@ON F O @@ooooo€o@ N+FOOOO-OF O @ @ oooooNoNoo o @+o @ o o ooooo6o+oo N EO -OOOOOO-OO N ONN O N @ F @- @ 6

66000NO@O+ @ @ FOOFOOOONO @oFOoAOOO- O - @-OO@+@OO+ OOOOO-ONOO O NOO O I N ooooo+o+oo rEd' -OOOOQOFOO N OO@ O q -OOOOFoOOO FOo

O€+O+NooOF O N +oON+OoN-O O @6FOOoO@O- O F 6OOOn@OOOO O N OOOOOFONOO O @+O 6 F O oooooNo@oo o dOOOOOOFOO N O-@ O = -OOOOFOOOO N -oo \ NI NI a@ o o L o \c o o\c Nl o H z 6l@ o z 6la o + ++ ! \\\ + \ Ec o NI L + !c o oN 6 ololol o +l E d o + ++ € ololol +t ! oo .F-L.FoOcodd c --- = . F =o o oN = JO EO 4

+ N@ I 6N oooo oooo oo ooo ooooooo oo

3333699 3639 R33: R*36 33R3 PE sSE SS336ER is ---tru-- -s-ci N-;ci NN-N NNNN FN ooN NooNNNN NF

JJ -gENEIRR&de,eeddded -UEsE o&Sp6 -xBgS -$BBR ob= - =qi ap&N$bC6RKS F@@F

6OO 6@O@o 6NNNOOOOOF@ R3=:3RR5 R=3=3 R53:* RRUSR RSUSR N6O NFFFo N@OO O60@6N N O-F-FFFONN FNNFFNN NFNF FFNN O--P o o F. slQ{\ | | | | r*ra{ | | l!\ | | Y o I tqJ | | | | | | | | | | | bo FoFFOOooooO +OOOFFOO OFOFQ +ONOO NNONN LOA LQooo ql

ooN66OOO OONOO oQNo6 ooooQ ooo6N oBg oooNn oo++933RR33 FF FNN

I o bo 6@ 6O O@ +N N+ o+ €6 oo N- +< @6 FO ON ON @6 sc NN NN O- $+ 9S ++ o xa OL d6

6Z F

o +o@@6 F@ ro

HH6N @ @ OF@ tsONoO OoOONO ONNF 6Q O +NON @oOONFts 6O +@6€ @ Od@FN NOHFN @FFF-N OF-O AFF @O6FOF O@OO FOF FN o N@@€ O OO@N@ dddFo F-FFFO FFNN NFO FNFFFN NFFFNNFFO

E d> deee OFOOO OO O@ONdH O@6ood ONO@FO- O66NN OSoF ON9@66- OAN6ONNobF <6NN @ dN@+N dNNNN NN@6NN NNOF +NN N+6NN< NNN+OFNNN o +H oHa -@

6+NS9 00 0NOHHN 40@+60 6+60++O oN6O 69NO+OF oO+@ooO+oO NON@€ N€ N$6N@O N+gNNH NFO+ONF NotsF NFO€N+@ NOF+6@NNNF NNNN N NNNNO NNNNO NNNNNO NNO NNNNNN NNNFFFNNO (J

N ql c.rG. | | | I q? | | qi q?6a | € rl c{ql I | | | | | | {Js+ | | aaa F

ONNNi ON ONNNNd oONiFO NFNNNFO oN-FF N--- F6O@66N ONNNNOOFFF

OOOON ON OO6OOO OHOOOo OOO-oOo OFOOO Q-Oo OOFOOQO OOF666o666 NNN

+l

ats oo o6 oo o@ o@ @@ oo @N @@ Nd +o +o oo

o!

q= ROSS llrD HUEBNER: AUGITE, PIGEONITE, AND ORTHOPYROXENE I 139

o+@N+ ++ooo o ov

.F dN c ! o oo 'F p oc s+6o+oo 6 0 0 xo OFO@ONN 6 C p OO o !- o@6@@oo 9-P:o o d c d@ L P O OA > o po > L < LP o.F LO c E @ *o -RS -Rb' -cd xF o c oo F L O FO E o I x o.F o>r o Eo L6 6 L s9 o > 6 o oo FO Q E qD oo oio oNN 6A E € CO N@ NO@ N@@ OL Q O L oo606 t o L o 60 d pL "d OL LO C opp r o€ o o $.F op 'FETJO F O ^J E a o6 . rLrtL FO. Oo oc d oL c u P.-P = +! Eo o P. +! o ! o €o q@ OF o OP O c L L SN - F.F v.F po f L s O I -a - O \C x O o F > OO OO O LO PL Od ^ 9O L OO .F> o op I o.F c d FO I Fq OO ov o-F ooo oo6 4 9ory O= OF + c = > I o! . g 3'd - o5 Eba + I I p . OO @ O O .F C 56 : O A OO- 5 F E Ov 5 € L O< OE @6 N@ @6 90 op pFo @FOt.FL oo oN N@ O6 L O p @ O= o F o@ 60 a E . o p c > .v O 6.F 6 p E c O O Et ! O = .F LN F @ oO d 6'- . =o P d O F L po tr o L I L O >ov EO FL o .-O9 POO OL O P9 o O.F L oo L -=o5o O.F -o L 5 tgHo Eo "f qz )rit An ;l I a.l @ o s oooo ooQ oooooo oooo F

+aoN@@ +o6\ +N oNNNN@ 600@N@ @NFN ++o O N o@66 Fo@ 6F@o-o o-oo @ooN o o -++OON +g+O 66$ N@OnO+ @€@O +O6O @ 6

OO@O+Ft Qoo6+ O-QO+ O+$+ OQNNsOo o@osN+ QNN<+ QNFFo o+o$ NNO$ N@@ NO@+@@ NOO$ +++N N

6N@O@OO NNOONF6 -ulQ9o@ 4Noo 6ONN6OO ooNoo 6N66O NNoON NOON NOOts@OO NO6@@ N4@@O NO@ @ NOA oooo 600 @ooooo oooq 0o60

\F a^ ^a FA v^vo PF PF pv pFpF pF lllttrl | | uv | | | | ov | | | | oo | | ov4v t | 6v L .a6 .F6 .FL .F€.F6 .F@ LL LL L !L LL L4 oooooo .6 LOOOOOO LOOOQ rOOOO pp FOOO LFOOOOO LOoOO o660Q doo a p oooo6

oo-6@oo oFo6o ooo!t@ oQ-o ooQN6@@ OoOF@ OOoNQ ooN o OFN OO F@O Fo6O6 6 OO@@O

-]ot NO oo bN @o +o NO !9r_- qr@ eo @@ @@ oo o6 @@ NN

OL

N N qz I 140 ROSS,4ND HUEBNER: AUGITE. PIGEONITE,AND ORTHOPYROXENE

indicating that the lamellae had thickened consid- erably. Phases other than pyroxene that may also form at the higher heating temperatures include olivine, cris- tobalite, protopyroxene, and melt. The crystalline phases can be easily detected in the precession photo- graphs. The melt phase, when present in amounts great€r than about I volume percent, can be detected by carefully examining thinned and polished crystals by means of conventional light optical microscopic techniques. The modes of the heated orthopyroxene crystals and the run conditions are given in Table 2.

Heat treatment Each orthopyroxen€ crystal was wrapped loosely in platinum, platinum-iron alloy, or iron foil, heated to the desired temperature in a gas-flow furnace at a controlled (and measured) value of the oxygen fuga- city appropriate to maintain the bulk composition of the sample, and quenched to about 5oC in chilled mercury, nominally in one second. The crystals were then photographed by the X-ray precession method. Platinum-wound quenching furnaces and CO, + H, gas trains were used as described previously (Ross et al., 1973) but, because iron diffuses rapidly from platinum foil when melt is present, addi- (B) sample to tional steps were taken to prevent loss of iron Fig. 2. Orthopyroxene containing exsolution lamellae of augite (Huebner, 1973; Huebner oriented on (100) of the host. (A) Optical photomicrograph of through the melt to the foil individual orthopyroxene crystals, crossed polarizers, oil mount, et al., 1976).Iron-platinum alloy foil containers were sample 8 (Tables l, 2). White bar is 0.1 mm. (B) Dark-field used in many of these runs to specify the iron activity transmission electron micrograph (200 kV) of ion-thinned of the system. Judiciously chosen alloy compositions Note the large number of periodic interface orthopyroxene. and furnace oxygen fugacity (./o,) values specified dislocations along the boundaries (marked by arrows) between nearly equal to that augite (dark center band) and orthopyroxene (upper and lower wiistite activity values that are gray areas) Sample 3; crystal was quenched frorn947"C. Scale is inherent in the silicate samples, preventing appre- o.lp. ciable gain or loss of iron. Uncertainties in furnace gas /o, determinations (see Huebner, 1975) and in the thermodynamics of the iron-platinum alloys (Heald, varying in augite content from 30 to 70 mole percent 1961), and the lack of thermochemical data for py- was within 5 mole percent. roxene solid solutions prohibited calculation of the In a few X-ray photographs of heated ortho- necessary wiistite activity values; such values were pyroxene crystals, diffuse streaks were observed in originally determined by trial and error for represen- the positions where the 002 and 002 pigeonite reflec- tative pyroxene bulk compositions. One notable fea- tions should be (Fig. 3D). We interpret these diffuse ture of this technique, not possible for runs in evacu- reflections as being due to the formation of pigeonite ated silica-glass tubes, is that f o,cun be treated as an lamellae I to 5 unit-cells thick in the a* direction.' independent variable, in general chosen to maintain When the crystals showing such streaking were re- the ratios F e" /F e'* , Cf* / Cft , or T7'* /T|ot close to heated to a slightly higher temperature the streaks their original values. were replaced by sharp pigeonite diffraction maxima, Orthopyroxenes were heated for at least 18 hours before quenching. By repeating runs on the same crystal for various lengths of time, we ascertained 'Discrete diffraction maxima rather than semicontinuous dif- fuse streaks begin to appear when the lamellar precipitates are that l8-hour heating periods were usually sufficient greater than 5 unit cells thick (Ross, 1968, Fig. 4). to establish whether phase changes would occur at ROSSAND HUEBNER: AUGITE. PIGEONITE, AND ORTHOPYROXENE r l4l

a* opx, pig I

a o '102

P'v ^ 'prgI ,./ a o oo2 a

a ------)- c* ' opx o opx a a E 002 -\r\ a u \a* pig II . pig II D 102 a

a pig II d pig II .a opx q b o pigl tr pig II

a opx

' opx

c* pig II

C

Fig. 3. (A) hOlX+ay precessionphotograph (unfiltered Mo X-radiation) of an orthopyroxenecrysral (sample 17,Tables l, 2) held at ll02"C for 43 hours. The crystal has reactedto form about 4 mole percentpigeonite twinned on (100).The two twin orientationsare designatedpigeonite I (P I) and pigeonite II (P ID. OPx:orthopyroxene. (B) Reproduction of (A) showing the indexing of the reflections.(C) ftOlX-ray precessionphotograph (untrltered Mo X-radiation) of an orthopyroxenecrystal (sample 17, Tables l, 2) held at 1226"C for 54 hours. The orthopyroxenehas reactedcompletely to form pigeonite twinned on (100) plus a small amount of liquid (glass).(D) ftOlprecessionphotograph (unfltered Cu X-radiation) ofan orthopyroxenecrystal (sample18, Tables l, 2) held at 1038.C for 168hours. Diffuse streakingabout the 002 reflectionposition ofpigeonite is shown by an arrow. lt42 ROSS,4ND HIIEBNER: A(IGITE. PIGEONITE, AND ORI'IIOPYROXENE the interfacebetween the orthopyroxenehost and the augitelamellae at a given temperature. oPt The microprobeanalysis procedure used is that de- sFe scribed by Huebner et al. (1976), "Method One." It must be emphasizedthat the microprobe analysesof unheated pyroxenes in Table I are, to the extent practical, orthopyroxeneexclusive of analysesofthe E the augite lamellae. The wet-chemicalanalyses gen- r50 erally give greatervalues for calcium contentbecause F^" the wet-chemicalmethod givesthe bulk compositions +'- of the orthopyroxene crystals including all inter- >70 grown augite. o o

Retention of composition IRON LOSS The representativeanalyzed runs listed in Table 3 and plotted in Figure 4 are evidencethat when melt was absentpyroxene single crystalsneither lost iron no to pure platinum foil, nor gained iron from iron foil, " convincing are the long o,o'''.o,oo r':r*".':**;:** during this study. Especially ";"r.j:", runs maintained at low oxygen fugacity values for Fig. . Mg/(Mg+)Fe) of heated pyroxene crystals in Table 3 the subsolidusinvestigation discussed later. The sets plotted against their respective bulk compositions listed in Table of runs in platinum and in iron foil each scatteruni- l. Circles indicate runs made in platinum foil; squares designate Points cluster closely about the theoretical line of formly about the line of constantcomposition in Fig- runs in iron foil. constant composition. Neither set of runs is biased toward either ure 4, indicating that bulk compositionwas retained. iron loss or iron gain, suggesting that the observed scatt€r 'is One might have expectedthe runs in platinum to related to the precision of the microprobe chemical analyses, not have shown systematicloss of iron, and the runs in to a change in bulk comPosition.

Table3. Subsolidus-t;ffff,lj5ating retentionof bulk iron to have gained iron as FeO, but they did not. The observedscatter is random and is insteadan in- dication of the analyical level of significance;analy- Sample No. InlEial OP; Time (h) Container alog f6r! Product oPx - us/ (Ms+tFe) us/ ftFfxEe) sis of the data yields a standarddeviation of the mea- 0.867 1246 24-5 PT -0.04 o.a12 suredcomposition equal to Mg/(Mg + )Fe) : 0.015. o.867 Lr68 \22,5 Fe -o.64 0.868 For bulk compositionsmore magnesianthan Mg/ 6 0.858 LI79 a12 -0.14 o.872 (Mg + )Fe) : 0.55,the reactionof orthopyroxene+ 8 0.855 -0.64 0.866 1168 I22.5 augite to form pigeonitesometimes was accompanied 8 0.855 1168.5 7t7 Fe -r.17 0.861 by partial melting. Becausesample mass was small 9 0.819 t17 9 812 Pt -0. 14 c 840 ..2/ to that of the metal container,the problem of o 77I 1310 2I Pt -0.08 o.761 relative potentially severe.The experi- 13 0.733 1182 50 Pt -0.11 o.111 iron loss(or gain) was

16 0 557 lo12 43 -0.68 o 512 ments reported in Table 4 and plotted in Figure 5

16 o.551 1188 1r5 ?t -0.13 0,540 demonstratethat the use of platinum-iron alloy foil 1l o.526 t130 PI -0- 03 o.547 and controlled /o, can effectively maintain the bulk 18 0.431 to42 A1 -0. o1 o.435 composition of pyroxenes that have undergone a 19 0.409 lo42 L3 -0.68 o.382 smaU amount of partial melting. Within analytical 2l o.325 1020 163.5 Pt -0.3 0.337 uncertainty [estimated to be 10.01 in Mgl(Mg + 2l 0.325 1057 89.5 -0.33 0.135 lFe) in solids,+0.015 in meltl, the tie line or tie tri- 21 o.325 r057 89.5 Pt -0. tl 0.340 angle connecting the projected compositionsof the -0,31 22 o.2r4 1000 89 Pt o.221 run products passesthrough or enclosesthe initial -o 25 0.148 935 121.5 PI -42 0,121 pyroxenebulk compositionused in each experiment,

ltLtoe F6, = [los tn2 (run)t -tIoC fo2 tron-dstltc buffer] indicating that the bulk compositionwas retained. was not practical to analyzeby elec- ?m.ry"i" of clear core of slng1e crystal. llarsin rost iron (xs/({s+tFe)=o.913) Obviously, it and rea.ted to forE protopyroxene. tron microprobe each crystal after each fusa{ingrun. ROSS AND HUEBNER: AUGITE. PIGEONITE. AND ORT:HOPYROXENE

Table 4. Repres€ntativ€experiments demonstrating retention of bulk compositionin presenceof partial melt

1/ 2/ Sample Toa Tirne(h ) log f9 aFeO Phase Mode Ica] IMs] IFeJ No. 2

I 300 44 -11.75 0.08 Pxn oq lo 86.1 12.0 Melt 5 3l.8 49.0 19.2

I 332 7l -8.?6 0.05 opx 85 0.6 87.9 ll.5 Pig t0 86.1 8.2 Melt 38.9 47.8 I 3.3

1226 -l1.59 0.26 Pig 90 1.0 52.t 46.8 0livine 0.3 44.2 55.5 Melt 7 17.4 73.4

I 178 52 - I 2.06 Pxn 95 1.3 52.6 46.1 Melt 14.0 I 6.8 69.l 1/ Pxn - orthopyroxeneor orthopyroxene-pigeonitemixture 2/ Volumepercent, estimated optically andby microprobetraverse

Besidesbeing tedious and consumingmuch time, the As a check for iron loss or gain, we determined preparation of a polished flat surface requires re- from the X-ray precessionphotographs the lattice pa- moval of pyroxene, which changesthe shapeof the rametersof the host crystal after eachheat treatment. crystal. In our opinion, changing the shape of the Comparison of the unit-cell dimensions of heated crystal would introduce an unacceptablyhigh level with unheatedorthopyroxene crystals (Table 5) sug- of uncertainty into our technique of comparing the geststhat there are no major changesin buft chem- X-ray intensity "modes" of a crystal subjectedto suc- ical compositionsduring successiveheating experi- cessiveheat treatments. ments. If the Mg/(Mg + )Fe) composition ratio of

100 tFe/(ca+tvtg+>Fe) Fig. 5. Composition data for selected orthopyroxene crystals that were partially melted in platinum-iron alloy foil at controlled oxygen fugacity, demonstrating retention of bulk composition (see Table 4). Squares: orthopyroxene or orthopyroxene-pigeonite mixture; triangles:pigeonite; diamonds: olivine; circles:melt compositions;solid symbols:initial orthopyroxenebulk compositions(from Table l). Compositions of individual phasesare projected onto the pyroxene quadrilateral plane on the basis of the total Ca, Mg, and Fe present. Calcium-poor pyroxene forms >90 perc€ot (by volume) of each crystal. Initial bulk compositions (same sample numbers as in Table l) plot close to the respective bulk compositions of the run products. If iron was lost or gained, the effect is so small that it is obscured by chemical analytical uncertainty. Table 5. Unit-c€ll parametersfor host orthopyroxene(Opx) beforeheating (25'C) and for orthopyroxeneand pigeonite(Pig) in crystals quenchedfrom temperaturesnear f1 (Table 2)

SampI e Sample Number Phase a(A) ! (A) !.(A) T"C Number Phase c(A) !.(Al e.(A) T'C Opx I 8.20 8.805 5.174 90 25 Opx I 8.28 8.866 5.218 90 25

Opx I 8.20 8.823 5.I90 90 I 284Rp Opx 18.27 8.865 5.208 90 l2l4 Pig 9.646 8.823 5.171 108.6 I 284Rp Pig 108.5 l2l4

Opx 18.24 8.807 5.r90 90 25 Opx 18.?7 8.863 5.209 90 25

Opx 18.24 8.804 5.179 90 1269 Opx 18,27 8.856 5.?04 90 ll75 Pig 5.179 108.2 1269 Pis 9.681 8.856 5.210 108.3 I 175

Opx I 8.25 8.827 5.191 90 25 Opx 18.32 8.865 5.215 90 I',199 Pig 9.716 8.865 5.213 108.3 ll99 opx I 8.25 8.843 5.195 90 1259 Pi g 9.667 5.202 I 08.5 I 259 l4 Opx 18.27 8.863 5.201 90 25

Opx 18.24 8.81 5.203 90 133? Opx 18.27 8.851 5.206 90 I 160 5.218 108.6 I 160 P'iS 9.64 8.81 5.2?7 108.9 1332 Pig

Opx 18.24 8.844 5.188 90 25 opx 18.30 8.880 5.21| 90 1172 Pi9 5.227 108.6 ll72 Opx 18.26 8.840 5.197 90 1243 Pi9 8.840 108.3 I 243 l5 Opx 18.28 8.899 5.?17 90 25

Opx 18.?6 8.835 5.199 90 1275 Opx 18.33 5.223 90 ll30 Pig 9.675 8.835 5.202 108.3 I 275 Pis 5.235 108.8 ll30

Opx 18.24 8.835 5.193 90 25 Opx I 8' 30 5.213 90 I 150 Opx 18.27 8.830 5.208 90 1243 Pig 5.219 r08.6 ll50 Opx 18.26 8.848 5.196 90 1214 PiS 9.691 8.848 5.199 I 08.3 I 274 Opx 18.30 8.915 5.221 90 25

Opx 18.26 8.827 5. lg',l 90 25 Opx t8.28 8.915 5.21? 90 I1',lll818 Pis 8.9]5 5.zil 108.7 Opx 18.24 5.193 90 1246 Pj9 108.I 1246 17 Opx I 8.31 8.91I 5.211 90 25

Opx 18.27 8.857 5.202 90 1274 Opx 18.31 8.926 5.224 90 I 102 Pig 9,673 8.8s7 5.200 I 08.3 I 274 Pig 8'926 5.227 108.6 ll02 '18.31 Opx 18.24 8.850 5.I 95 90 ?5 Opx 8.913 5.199 90 I 189 Pi9 9.661 8.91 3 5.213 108.6 ll89 Opx 18.26 8.830 5.197 90 1232 '18.35 Pig 8.830 I 08.4 I 232 Opx 8.945 5.?26 90 25

Opx I 8.25 8,830 5.r97 90 t275 Opx 18.31 8.928 5.208 90 1038 Pig 9.673 8.830 5.197 I 08.4 I 275 Pi9 5.30 108.6 1038

Opx 18.24 8.850 5.200 90 25 Opx 18.31 5.213 90 1076 Pig 9.63? 5.234 108.8 1076 Opx 18.24 5.200 90 1243 Pig 5.208 108.6 t243 Opx 18.33 8.967 q 2?a on 2r\ Opx 18.25 8.830 5.200 90 1214 Pi g 9.683 8.830 5.209 I08.3 1274 Opx 18.35 8.958 5.223 90 1022 Pi9 5.243 I08.1 1022

9 Opx 18.26 8.829 5.200 90 25 20 Opx I 8.37 8.977 5.235 90 25 '18.38 8.990 5.235 90 I 036 opx 18.24 5.200 't08.390 1?14 Opx Pis I 214 Pig 9.713 8.990 5.225 I 08.6 I036

opx 18.24 5.198 90 1246 Opx I 8.34 8.974 5.232 90 25 Pi9 9.660 5.214 I 08.3 1246 Opx 18.33 8.971 5.231 90 I 003 opx 18.27 8.833 5.202 90 1274 Pig 9.663 8.833 5.195 108.4 1274 Pig 9.680 8.989 5.234 108.6 1020

l0 opx 18.26 8.856 5.205 90 25 22 upx 18.40 9.030 5.244 90 25 opx I8.26 8.861 5.207 90 t2t4 opx 18.41 9.009 5.244 90 972 Pig 9.690 8.861 5.207 108,4 l214 opx 18.40 9.016 5.240 90 1000 ll opx 18.?7 8.853 5.204 90 25 Pi9 9.016 5.240 108.9 1000 '18.39 opx I 8.28 8.850 5.200 90 l2ll 23 opx 9.020 5.228 90 25 Pis 9.675 8.850 5.208 108.4 l2ll opx 18.39 9.003 5.223 90 960 opx I 8.26 8.831 5.199 90 I l 50Rv Pis 9.698 9.027 5.222 108.6 985 opx 18,29 8,865 5.202 90 1214

Pi9 9.963 8.865 5.208 t08.4 l2l4 opx I 8.23 8.825 5.190 90 I l88Rv ROSS IND HUEBNER: AUGITE, PIGEONIT:E, AND ORTHOPYROXENE tt45

Table 5. (continued)

Sample Sample Number Phase !.(A) !.(A) c(A) Number g(A) q(A) s(A) T'C '18.38 '18.39 opx 9.000 5.240 90 25 opx 9.029 5.242 90 25 Pig 9.695 9.000 5.240 108.6 25 Aug 9.784 9.000 5.253 I 05.6 25 opx 18.37 5.239 90 965 Pi9 9.709 5.241 108.4 965 opx 18.41 9.01 3 5.248 90 952 '18.44 Pig 9.707 9.013 5.237 I 08.6 952 21 opx 9.054 5.24t 90 25 Aug 9.791 9.0t3 5.250 I 05.7 952 28 opx 18.44 9,056 5.246 90 25 Pig 9.693 9.000 5.232 108.8 965 Aug 9.775 9.000 5.248 105.7 965 opx ]8.43 9.050 5.250 90 901 Pi9 5.?54 108.4 90t opx ]8.41 9.028 5.244 90 25 opx 18.251 8.845 5.198 90 25 opx I 8.41 5.244 90 960 Pis I 08.5 960 opx 18.257 8.868 5.192 90 1367 Pig 9.598 8.820 5.174 t08.4 I 382 opx 18.44 9.037 5.240 90 985 Pi9 9.712 9,037 5.248 108.6 985 opx 18.23 8.865 5.]89 90 25 opx 18.21 8.837 5.198 90 t3t0

Pig 9.636 8.822 5.191 108.5 1364

the orthopyroxeneschanged during successivemns, In only one experiment(sample l, Table 2) did a run this change would be indicated by corresponding repeatedat the same temperatureshow a changein changesin the unit-cell edges.The a and b dimen- the phasecomposition (mode) of the crystal. We at- sionsof heatedcrystals listed in Table 5 are, with few tempted to reversethe reaction by heating selected exceptions,within 0.024 of the corresponding di- crystals at lower temperaturesthan the previous mensionsof the unheatedcrystals. From the data of heating step, but in only two experiments(samples Kuno (1954) for natural orthopyroxenesand Turn- ll, 13) did the amount of pigeoniteor augite de- ock et al. (1973) for synthetic orthopyroxenes,a de- creasesignificantly from that observedin the same creaseof 0.024 in a or b would correspondto a l0 crystal at a higher temperature.We cannot be certain percent increasein Mgl(Mg + )Fe). However, in that eachof thesecrystals did not losesmall amounts our experiments,c and D corrun*onlyvaried inversely, of iron during heating, but available unit-cell mea- and in some determinationsorthopyroxene in pure surementssuggest that such loss is not great (see platinum foil even yielded dimensionssuggestive of Table 5). In 13 other attempts no change was ob- the gain of a small amount of iron. We interpret most servedin the crvstalmode. of the reported changesin a and b with temperature of heating to be merely random experimentalvaria- Pyroxenephase relations tions attributable to the preparation and measure- ment of the X-ray precessionphotographs. Unit-cell Reactionmechanisms dimensionsdetermined by this method have an un- The three principal subsoliduschemical reactions certainty of +0.2 percent,which accountsfor the ob- observedin the orthopyroxeneheating experiments servedvariations. In only five runs (Table 5) did we are: detectsignificant evidence for iron lossfrom the crys- (I) homogeneousorthopyroxenes tal. OPx --+ OPx * pigeonite Results (II) orthopyroxenescontaining augite lamellae The experimentalheating conditions and data per- OPx * augite + OPx + pigeonite * augite tinent to phasecharacterization of the orthopyroxene crystalsare summarizedin Tables 2 and,5. The runs and for eachsample are listed in order of increasingtem- (III) ortlopyroxenes with [Fe]>75 and containing perature, except for "repeat" and "attempted re- augite lamellae versal" runs which are listed immediately after the OPx * augite --+ 3 pyroxenes+ fayalite + SiOz "initial" run. The samecrystal was usedfor the "ini- tial" and the "repeat" or "attempted reversal" runs. Two of the three subsolidus pyroxene reaction Lt46 ROSS,4ND HUEBNER: AUGITE. PIGEONITE. AND ORTHOPYROXENE paths are shown on Figure 6. First, consider a ho- orthopyroxene-pigeonitestability volume, and then mogeneousorthopyroxene of composition A. On into the pigeonite stability region; and (4) in crystals heating (path I) no changewill take place until tem- slightly more calcic than D ("inverted" pigeonites) perature 4, wherepigeonite of compositionF (out of the reaction sequencewould be the same as (3) ex- the plane of Fig. 6) starts to crystallize. Heating to cept above the three-pyroxeneregion the reaction evenhigher temperaturescauses more orthopyroxene path would pass into the pigeonite-augite stability to react to pigeonite until the three-phaseregion or- volume and then into the pigeonite stability region, thopyroxene* pigeonite + melt is reached. unlessmelting intervened. Next, consider heating an orthopyroxene which In our experiments,only local equilibrium is likely containsa few augite lamellae.At the (100) interface to prevail. Thus, at the lamellar interface,where the the initial compositions of the host and exsolved orthopyroxene was saturated with augite, reaction phasesare assumedto be A and B, respectively.If lo- path II prevailed,whereas some distance from the in- cal equilibrium prevails at the interface, the ortho- terface, where the host orthopyroxene composition pyroxeneand augitewill readjustcompositions along remains unchanged,the reaction mechanism fol- the orthopyroxene and augite solvi (path II). The lowed path I. slower the rate of heating, the greater should be the A third type of reaction was observedin the iron- volume about the interfacein which equilibrium pre- rich orthopyroxenes(samples 22to28). At [Fe1: 75- vails. If the augite lamellae are completely dissolved 90, we first observedfayalitic olivine and cristobalite into the host orthopyroxenebefore temperatureZ, is in the runs in which we first noted pigeonite,or in the reached,the reaction path changesfrom type II to next heating step.Two very iron-rich orthopyroxenes type I, and pigeonite eventually forms from a ho- (samples27 and 28) with [Fe] > 90 ultimately reacted mogeneousorthopyroxene. If augite persists,pigeon- completely to fayalite and cristobalite. ite with a composition projected onto the plane of Figure 6 (see inset) will first appear at temperature Partial melting 7,, at the two-phaseinterface between orthopyroxene Partial melting of pyroxenewas observedin some and augite.(As will be discussedlater, we have under of the heating experimentsfor orthopyroxene with ideal conditions observedthe localization of pigeon- Mgl(Mg+)Fe) - 0.55(see also Huebner et al.,1972, ite along relict augite lamellae.)As heating continues 1976). Phase-equilibrium relations involving the above ?1,augite will react at the interface, and locally partial melting of natural pyroxene bulk the samplewill passinto either of the two-phasere- compositionsare complicatedby minor components gions OPx + pigeonite, or pigeonite * augite, de- which depressthe solidus by partitioning into a pending upon the local activity of the CaSiO3com- partial melt, enriching it in AlrO, and TiOr. The ponent. Augite in the central part of the lamellae,not CaSiO, componentis also strongly fractionated into adjacent to the orthopyroxene interface, may not the melt. The residue of, for instance, 30 percent equilibrate during the short heating-run periods that partial melting is a very calcium-poor, refractory we used. However, persistenceof some augite to pyroxene. Becausethe calcium content of the higher temperaturesneed not prevent the appearance pyroxene changeson partial melting, we have not of pigeonite at f. reported or discussedexperiments in which the If equilibrium were to prevail throughout those amount of melt, estimatedby optical methods, crystalsof orthopyroxenewhich are intergrown with exceededl0 percent by volume of the crystal. Only a augite lamellae, the following reaction sequences small amount of melt, probably lessthan 2 percent, would take place during stepwise heating: (l) in appearedin someof the crystalsat temperaturesnear those crystalspossessing a bulk compositionA (Fig. 4. We believethat this melting did not greatly affect 6), augite would immediately dissolveand then at t our determination of the temperature-composition orthopyroxenewould begin to react to form pigeon- relations for the three-pyroxenestability region. ite (path I); (2) in crystalshaving bulk compositions of reaction to pigeonite lying between A and C the reaction would follow Temperatures form path II until augite disappearedand then would fol- The run temperaturesat which pigeonite was first low a type I path; (3) in crystalshaving bulk compo- recognizedin the X-ray photographs(Table 2) are sitions lying. between C and D (for example, some given in Table 6 as 2"o".In two cases(samples 13 and "inverted" pigeonites)the reactionwould follow path 20), these temperatureswere lowered somewhat II, pass through the three-pyroxeneregion into the because a significant amount of pigeonite (10-35 ROSS AND HUEBNER: AUGITE. PIGEONITE. AND ORTHOPYROXENE percent) was estimated to be present at the lost or gained, the observed scatter would temperature (I"o") at which it was first observed. The undoubtedly have been greater, considering the use basis for the slight lowering of temperatures is that, of different containers, temperatures, /o, values, and other variables being equal, an increasing propor- run durations, each of which would affect differently tion of pigeonite is correlated with increasing the rate of iron loss or gain. temperature. These two adjusted values of ?io" are Augite was not observed in those runs in which listed in Table 6. A similar adjustment could have pigeonite was first detected in samples #29 and 30 been made for the other samples but was neglected (Table 2). These samples plot at 1382" and 1360"C, because the correction was only 2-5oC. Also respectively, temperatures that lie considerably included in Table 6 are the [Ca] and [Fe] values above the magnesian orthopyroxene * augite : (Table l) of the unheated orthopyroxenes as well as pigeonite reaction curve (Fig. 7). Because we expect the expected [Ca], and [Fe], values for augite- the minimum temperature of pigeonite stability to saturated orthopyroxene at T, (composition C, Fig. occur when pigeonite is in equilibrium with both 6). The [Ca], values are estimated from the orthopyroxene and augite, it is not surprising that composition data given in the literature for these two samples plot well above the reaction curve. orthopyroxenes believed to have crystallized in We have not proven that the reaction curve drawn equilibrium with augite at pressures less than 8 kbar as a solid line in Figure 7, separating fields labelled and temperatures at or close to temperatures of O+A and O+PIL(-FA), is an equilibrium curve maximum calcium saturation (f). The [Fe], values because we have not reversed the reaction. Indeed, are obtained from [Fe] values by extrapolating along the detectability limit of the X-ray method precludes the appropriate orthopyroxene-augite tie line (see observation of pigeonite unless the sample is heated Ross and Huebner, 1975b) from [Ca] to [Ca],, at least several degrees (we estimate 2-5) above the making the assumption that these tie lines do not reaction equilibrium. Several lines of reasoning, rotate appreciably as temperature changes. For the however, suggest that our curve might not lie far purposes of this paper-where we are more above the equilibrium value: (l) the scatter of points interested in the length of a tie line than in its about the reaction curve is smal! (2) melt, which orientation-this is a reasonable assumption.' would promote the achievement of equilibrium, was The values of 7"o" for 28 heated orthopyroxene observed to be present in some of the plotted runs, samples are plotted against [Fe], in Figure 7. The and suspectedin many runs using samples#l-15; (3) data for 15 magnesian pyroxenes, which coexisted within the limits of uncertainty, our results agree with augite when pigeonite was first noted, show with the reversed determination of the iron-free relatively small scatter. Such a small uncertainty in equilibrium by Yang (1973), discussed in the next compoqition suggests retention of crystal bulk sectron. composition during each experiment; had FeO been Most iron-rich pyroxenes, unlike the magnesian pyroxenes, lost all the augite that was initially present at lower temperatures than those at which 2Were the exsolved augite to change from Mgl(Fe+Mg) : 100 pigeonite was observed. In only three cases (samples to 0, the maximum change that is theoretically possible, the #17, 26, and 28) was augite present when pigeonite coexisting orthopyroxene would change a maximum of 0.06 in Fe/ was first observed, thus fulfilling one of the (Fe+Mg). Natural assemblages suggest that the actual degree of rotation is considerably less. We have plotted the orthopyroxene- requirements of the equilibrium orthopyroxene * augite tie lines for about 300 naturally-occurring pairs and find augite ? pigeonite, that all phases be present. In that the slope varies with pyroxene major-element bulk eight cases (samples #16, 18-23, 25) the observed composition but not with rock type (high+emperature basalts, reaction is orthopyroxene --+ pigeonite, which must kimberlites, gabbros; intermediate-temperature andesites, iron- lie at higher temperatures than the orthopyroxene + rich gabbros, meteoritic rocks; low-temperature granulites, + pigeonite charnockites, granulite-amphibolite transition facies rocks, augite equilibrium. Most iron-rich rhyolites, lunar metamorphic rocks). The observed statistical pyroxene runs plotted in Figure 7 represent scatter in tie-line orientation for each rock type obscures any temperatures above the 3-pyroxene equilibrium. systematic tie-line rotation with changing rock type (changing Thus the 3-pyroxene curve separating the fields of temperature). We do that flnd tie-line Iength increases orthopyroxene * augite and pigeonite is shown as a systematically with decreasing temperature. Thus, we believe that to the first approximation necessary for our purposes, tie-line schematic, dashed ltne below the runs that lost augite. slopes are nearly independent of the temperature of pyroxene Note that the three runs in which augite was retained crystallization (Ross and Huebner, 1975b). (Nos. 17, 26, 28) are, relative to their [Fe], value, I 148 ROSS AND HUEBNER: AUGITE. PIGEONITE, AND ORTHOPYROXENE

Tm

rj T; E\ \

Ca2Si206 \

A CFD 100 Cal(Ca+Mg+IFe) Fig 6. Planar section through the pyroxene ternary phase volume, T-(Ca2SirOu-Mg2Si2O6-Fe2Si2O6), and coincident with the orthopyroxene-augite tie line AB (inset). The OPx-augite tie lines do not rotate appreciably with temperature, thus the T-X section below the three-pyroxene field closely represents the equilibrium compositions. Liquid compositions are taken from Huebner et a/. (1976). The field boundaries above the temperature ofthe three-pyroxene field are projected onto the ?-Xsection and thus do not depict the actual equilibrium compositions. OPx:orthopyroxenes; P:pigeonite. A, B, C, D, E, F:compositions of phases; Ii:temperature of first appearance of three-pyroxene assemblage (heating path II); 4:temperature at first appearance of pigeonite in homogeneous orthopyroxene of composition A (heating path I); T-:temperature of first melting of the orthopyroxene-pigeonite assemblage. Di and Hd have the same meaning as in Fig. l. lower in temperature than the other eight runs, with analysis suggests thdt calcic amphibole may be iron-rich samples. present in a few grains ofthe pyroxene separate, but amphibole could not be detected by X-ray precession Detailed discussionof two samples methods. Compositions of the run products for Two series of heating experiments (Samples 3 and representative samples, summarized in Table 4 and 29) will be discussed in detail, because these runs plotted in Figure 5, show that the bulk composition illustrate the interface reaction mechanism and the did not change (see also Huebner et al., 1976, Table relationship between temperature and the calcium 2a fot complete chemical analyses). Pigeonite was content of orthopyroxene participating in the observed in run products heated to 1259"C and reaction OPx -+ pigeonite. The interface reaction above. Melt was observed at 1305oand l332oC, and mechanism is illustrated nicely by sample 3, an may well have first formed at lower temperatures, yet orthopyroxene of approximate composition not have been detected optically. The run at 1332"C Caoo,rMg,rorFeorrrSirOuwhich contains lamellae of demonstrates that augite lamellae reacted with augite 7 to 20 micrometers thick. Microprobe adjacent orthopyroxene to form pigeonite + melt ROSS AND HUEBNER: AUGITE. PIGEONITE. AND ORTHOPYROXENE tt49

Table 6. Temperature-composition relationships for pyroxene and augite, are close to the mlnlmum orthopyroxenes temperatureof the volume in which three pyroxenes are stable at this composition. -t"c!/ sanple tculll E"l! Kai! tnetf Sample 29 consists of chemically homogeneous l':o. obs Pig gem-quality orthopyroxene(Ca. o,rMg,uruFeo.rrSirOu) 1 o.6 8.0 4.4 1.'1 12A4 no included augite or other phases as 2 o.7 8.3 4.4 B.O 1,269 containing 3 0.6 12.O 4.3 11.s 1259 determined by both light optical and transmission 4 ?l 12.9 13.0 1243 less magnesian 5 2.1 13.5 4.3 13.O L243 electron microscopy. The sample is 6 I.7 14.O 4-3 13.B 1246 than sample 3 and therefore would be expected,if 7 14.O 4.3 1232 pigeonite I 4.8 13.s 4.3 1232 augite were present, to form at an even 9 17.8 17.5 12I4 lower temperature,approximately 1220"C, conpared 10 r.7 19.9 4.2 t9.5 12II 1I 1.7 22.5 21.8 1211 with the 1259otemperature observedfor sample 3 12 1.8 24.9 4.r 24.3 I 199 (Fig. 7; Table 6). However, in the absenceof augite I3 2.6 26. O 4.r 25.6 r 165 I4 1.0 31,.4 4.O 30.2 r 160 lamellae (calcium saturation zones),sample 29 did 15 2.4 43.2 3.8 42.4 1150 not form pigeonite until 1375"C, 155" above the 16 t.t 43.9 3.8 42.4 1118 L1 r.4 46. A 3.8 IIO2 expectedtemperature. Obviously, the temperatureat I8 1.4 56.1 3.5 103I which pigeonitefirst appearsdepends not only on the T9 54.2 3.6 57.0 LO22 20 3.5 58.0 ro25 iron-magnesium ratio, but also on the local calcium 2I 66.5 €,5.2 1003 '7 content and on the presenceof favorable nucleation 22 6.6 3.3 76.O 9'72 '74.4 23 o.4 3.3 16.O 960 sites.At temperaturesas high as l382oC,melting was 25 I.6 a5.7 84.O 950 not observed in this pyroxene. Orthopyroxene 25 1.6 87.3 1) 85. 0 901 27 r.6 90. 6 3.1 89.5 none sample 30, which also did not contain any augite, 2A 1.8 93.7 901 behaved similarly in that pigeonite appeared at 29 0.6 I1 .6 4.2 16.5 13S2 30 o.2 19.4 4.2 18.3 I 364 1360'C rather than at l2l2"C. the minimum temperatureexpected for three-pyroxeneequilibrium 1/ = Irow temperature compositions (Table 1) at this bulk composition. The lack of favorable

:'Assumed)/ sites in these two precipitate-free [ca] value for calcim-saturated orthopyroxene nucleation To-i *"-ihd -I rfh^c^hava orthopyroxenes(29, 30) could also delay the 3/^-.. -'slight adjustnent. of IFel values gjven in Lhird coltm appearanceof melt, although we are certain that the for expected values at Tobs. low calcium concentration contributes to the high 4/ rTemperature at which pigeonite first observed in the melting point. x-ray photographs (Table 2). See text for coment on runs 13 and 20). Comparisonwith previous work We did not attempt to determine the minimum (Fig. 5). Optical, X-ray, and microprobe chemical stability temperature of iron-free pigeonite in this analyses, all performed on the same crystal, revealed investigation.Extrapolation of the data in Figure 7 to that pigeonite and glass (quenched melt) occur [Me] : 100suggests that -l300oC is the temperature exclusively in thin lamellae, lenses, or blebs oriented at which iron-free orthopyroxene * augite react to along or close to the (100) orthopyroxene-augite form pigeonite.This temperaturecompares well with interfaces (Fig. 8). the value of l278rl0"C (convertedto IPTS '68) Most orthopyroxenes with optically visible augite reported by Yang (1973) for the equilibrium exsolution lamellae contain interface dislocations assemblage orthorhombic pyroxene-augite- formed to relieve the strain between the host and pigeonite. Yang originally designated this precipitate built up during cooling. These dis- orthorhombic phase as "protoenstatite." However, locations, examples of which are shown in Figure 2B, his observationsby optical methods were sufficient are crystal defects and act as favorable sites or only to indicate an orthorhombic symmetry (oral surfaces for nucleation of new phases such as communication, 1973), which is not sufrcient pigeonite or melt (Nicholson, 1968). The predicted evidence to distinguish between orthopyroxene ease of nucleation of new phases on dislocations (Pbca) and protopyroxene (Pbcn). In view of the adjacent to the augite lamellae would suggest that absenceof cracks and polysynthetictwinning in his our observed temperatures of first appearance of quenched "protoenstatite" (Yang, 1973,p. 492) and pigeonite, in samples that contain both ortho- the relatively calcic nature of this phase (2.4 weight I 150 ROSS AND HT]EBNER: AUGITE, PIGEONITE, AND ORTHOPYROXENE

O+P*L(*A)

o

G

F z U O

o+P

V o+A V

lrel, or oRTHoPYBoxENE Fig. 7. Summary of experimentalresults using orthopyroxeneat one bar. Individual data points are listed in Table 6. Closed circles indicate orthopyroxene in which augite was observedto coexist with pigeonite. Open circles indicate orthopyroxenesin which augite was detected, but pigeonite was not observed.Triangles designate orthopyroxenes in which no augite was observedwith the appearanceof pigeonite; the triangle points in the direction in which the 3-pyroxene equilibrium must lie. The solid line, which is thought to be close to equilibrium, is not reversed.The dashedlines are not well located by our experirnents,and are thus intended to be schematic.

percentCaO; p. 493),we favor the interpretationthat polation to zero iron content and the fact that Yang the orthorhombic pyroxene in question is used pure, synthetic starting materials in the system orthopyroxene(Pbca).3 Comparison of the results is CaO-MgO-SiOr-AlrO3, whereas our starting then limited only by the aforementioned extra- materials were natural pyroxene crystals that contained small amounts of TiO2, MnO, NarO, elc. 3"Proto€nstatite"is rarely, ifever, quenchedwithout someor all Comparisonwith the work of Bowen and Schairer of the crystallites reacting to form pigeonite ("clinoenstatite"). In (1935) is difficult becauseof uncertaintiesin the na- our experience such crystallites always possessa very phases they produced in their experi- characteristic texture which includes polysynthetic twinning and ture of the large prominent cracks, the latter caused by the large volume ments. Enstatite from the Bishopville meteorite, difference between "protoenstatite" and "clinoenstatite." IFe] : 0.6, reacted at ll45"C to produce ROSS ll{D HUEBNER: AUGITE, PIGEONITE, AND ORTHO?YROXENE I l5l

peratureswhere they are partly melted, without vis- ible transformation." [n our experience,pigeonite is commonly not readily detected without an X-ray precessionphotograph. The transformation temper- atures reported by Bowen and Schairerfor their two iron-rich orthopyroxenes,[Fe] : 47.3 and 80.4, are 1090' and 955"C; these values are slightly higher than those produced in our own work. Our method detectssmall amountsof pigeonitefirst formed at re- gions of augite saturation. Bowen and Schairerpre- sumably could not detectpigeonite until the transfor- mation had proceededto a greaterextent. Thus, their observed transformation temperatures are slightly higher than our predicted values. Ishii (1975) relates the minimum temperature of pigeonite stability to the Fel(Mg+Fe) ratio (X,.) of pigeonite by the single expression 7'C : 1270 - 480X... This expressionis basedon the experimental work of Yang (1973)and Brown (1968),and on two temperature measurementsof eruptive flows from Japanesevolcanoes. Pigeonite compositions were ob- tained from electron microprobe analysesof ejecta materials. Ishii's expressiongives temperatures -25"C lower than the temperaturesof our runs, in basic agreementwith our results.We do not believe, Fig. 8. Orthopyroxene (sample 3 run at l3.32"C), @ntaining however,that temperaturemeasurements of extruded bands (7-20 tr^tmwide) of pigeonite + blebs of glass (quenched lava clearly give a close approximation to the tem- partial melt; note the "bubbles" possibly formed by contraction of perature of phenocrystcrystallization. Furthennore, liquid on cooling or by evolution of volatiles). Compositions of probe phases are giv€n in Table 4 and plotted in Fig. 5. The observed the electron composition must be carefully pigeonite + glass are inferred to have formed by reaction of host chosento reflect the compositionof pigeonite which orthopyroxene with augite along augite lamellar interfaces (see coexistswith augite and orthopyroxene.The pigeon- text). Bar scale is 0.1 mm. ite compositionsgiven by Ishii (1975,Table l, sam- ples 3109,3lll, 3l14) for the Mihara-yamaoccur- a quenched product identified as twinned clinopy- rence are clearly too calcic to have been in roxene. This temperature is too low for the reaction equilibrium with the adjacentorthopyroxene. of orthopyroxene + augite to yield pigeonite. We suggest that the meteoritic orthopyroxene was never Effects of nonhydrostaticstress and hydrostatic at calcium saturation (contained no augite) and that pressure Bowen and Schairer observed twinned pigeonite that Recent experimentalstudies have cast uncertainty formed from protopyroxene'on quenching from upon the true stability relationshipsof orthopyroxene I145"C. Bowen and Schairer heated the two bronzite and pigeonite at low temperature.Riecker and samples,[Fe] : 13.5, 13.8,with NaF flux to assistthe Rooney (1967) and Coe (1970) show that under transformation of orthopyroxene to pigeonite. Their shearing stress, orthopyroxene transforms at low reported transformation temperature, I l20"C, is temperature to "clinoenstatite" similar to that ll7o below the temperature that we would predict observed by Sclar et al. (1964) and Boyd and from Figure 7 for the first appearance of pigeonite. England (1965). Raleigh et al. (1971) consider the We suspect that the observed reaction is the forma- "clinoenstatite" to be metastablewith respect to tion of monoclinic fluoramphibole, not a clinopyrox- orthopyroxene under hydrothermal conditions. ene. We are disturbed by the observation (p. 167) Grover (1972) "reversed" the ortho-clinopyroxene '.two that, in the absence of flux, the bronzites can be reaction in MgClr'HrO flux and found the opposite heated for hours at temperatures more than 300o result, that at temperaturesbelow 556oC "cli- above their true inversion temperature, even at tem- noenstatite" is stable relative to orthopyroxene.We l 152 ROSS AND HT]EBNER: AI]GITE, PIGEONITE, AND ORTHOPYROXENE assume that this low-temperature "clinoen- 1300 statite" is a very magnesium-rich,calcium-poor pigeonite with P2/c symmetry, but a thorough characterization of this experimentally-produced made.Coe and Kirby (1975)state phasehas not been o+P+L (iA) that shearing expands the "clinoenstatite" stability field. and in the "clinoenstatite" field increasesthe rate at which orthopyroxenetransforms to the stable "clinoenstatite." The above results are indeed 1100 o+P puzzling becauseorthopyroxene is the naturally- q occurring low-temperature,calcium-poor pyroxene, E formed in a variety of environments,many of which pZ ',8 were undoubtedly under resolved shear stress on (100) as they cooled through the purported "clinoenstatite" field. For example, enstatite from Norway, of compositionCaooooMgrsro Bamble, t;;t Feo*uSirOushows partial alteration to talc but no \7 -L]]t..t - low-temperature "clinoenstatite" phase appears in the single-crystalX-ray photographs. o Several efforts have been made to determine the o pigeonite stability field at hydrostatic pressuresof I o to 3 kbar H,O. Lindsley et al. (1973)established that 60 70 80 pigeonite with Fel(Mg+Fe) approximately equal to 4a n"lt 0.80 decomposedto orthopyroxene* augite at with previous 700oC and 3 kbar. Simmons et al. (1974), using a Fig.9. Summary of our results 6ig. ?) compared results. At one atmosphere: "Y" designates Yang with Fel(Mg+Fe) : 0.75, grew experimental bulk composition (1973), orthorhombic pyroxene to pigeonite (discussed in text). At pigeonite(+ olivine + quartz) from orthopyroxene+ l-3 kbar, "S" designates Simmons et al' (1974), and "P" indicates augite at 800"C (2 kbar), and decomposedpigeonite Podpora and Lindsley (1979). At 15 kbar "Gl" and "G2" into orthopyroxene+ augite at 800oC (2 kbar) and designate Grover e/ at. (1972) and "X" designates Smith (1972)' 750"C (3 kbar), thereby reversingthe reaction within a narrow but unknown range of Fel(Mg+Fe). Podpora and Lindsley (1979) used a mixture of these results should be optimum for determining the synthetic powdered orthopyroxeneand augite, with (oT/lP)ea of reaction, and the result is 8oC kbar-'. bulk Fel(Mg+Fe) :0.75, to crystallizepigeonite at 2 Possible dfficulties kbar and both 825" and 850oC.With bulk Fel : (Mg+Fe) : 0.60,they were able to form pigeoniteat Evaluation of the Clapeyron equation, (dT/dP) 900'C. Thus it appears possible to grow iron-rich (AV/LS) may help explain the relatively high pigeonites at temperaturesthat are lower than temperature that we found for formation of iron-rich suggestedby the runs in Figure 7. pigeonite, relative to other investigators. The molar At high pressure,under conditions believedto be volume of reaction calculated from unit-cell data of close to hydrostatic, the minimum temperature for the three pyroxene phases measured in the same the pigeonite field is known only at 15 kbar and for laboratory (Turnock et al., 1973) is only 0.05 cm' : Fel(Mg+Fe) > 0.6 (Fig. 9). For bulk compositions mole-' for the reaction orthopyroxene * augite -0.75 with Fel(Mg+Fe) : 0.84, the temperatureis in the pigeonite at Fel(Fe+Mg) (at standard P and range875o-900'C (Smith, 1972);fot Fel(Mg+Fe) : 7). This molar volume of reaction, combined with 0.60 and 0.75 the temperaturesare, respectively, the observed @f/dP) of 8oC kbar-', yields an 1005otlO"C and 900o+20"C (Grover et al., 1972). entropy of reaction (AS^) of only 0.15 cal deg-' The same starting materials were used for the Fe/ mole-'. That the AS^ value is small is undoubtedly (Fe+Mg) : 0.75 experimentsat 2 kbat (Simmonsel due to the similarity of the orthopyroxene, pigeonite, al.,1974) and 15 kbar (Grover et al.,1972),md the and augite structures. The AS^ would be dominated samecriteria were usedto interpret the results(D. H. by contributions from changing the cell size from Lindsley, personal communication, 1978). Thus, l8A in orthopyroxene to 9A in clinopyroxene and ROSS AND HUEBNER: AUGITE, PIGEONITE, AND ORTHOPYROXENE I 153

from rearranging atoms in the octahedral (M) sites. Epilogue Becauseblocks of atoms are rearrangedin changing We report theseresults as an attempt to outline the from Pbcab A/c (or P2,/c) symmetry,and because phase relations of calcium-poor pyroxenes.Rather only a small proportion of the cations redistribute than being the final word on the subject,we believe themselvesamong similar sites, we expect the the data for Mg-rich compositionsto be more secure entropy change associatedwith pigeonite growth than the data for more Fe-rich pyroxenes.It is clear from orthopyroxene + augite to be small.o In this that we advocate phase characterizationby single- situation the entropy associatedwith defectssuch as crystal X-ray diffraction methods, and encourage the dislocationswe showed in Figure 2b may be a greater use of electron microscopy-for ourselves significant component of the entropy of reaction. If and others. so, the nature and location ofdefects could influence Our experimentalresults show that it is possibleto the conditionsat which pigeoniteforms in laboratory retain the bulk composition of single crystals, even experiments,and in nature. Likewise the existence of when melt is present,during an experinent of at least a structurally disordered pyroxene phase or several days' duration. We have also demonstrated pyriboles, as proposedby Veblen et al. (1977),could the utility of the X-ray precessioncamera to detect influencethe results. small amounts of pigeonite in the run products- We conclude that the temperatureat which iron- amounts that cannot be proven by X-ray powder rich orthopyroxene > 0.45] firsr reacts [Fel(Fe+Mg) diffraction. We presentwhat we believeis closeto an with included augite is not known with certainty. We equilibrium between magnesianorthopyroxene, are confdent, however,that existing experimentation pigeonite and melt; however, we recognizethat the at low pressure(P < 2 kbar) provides limits for the orthopyroxene-pigeoniteboundary for our iron-rich reaction in the laboratory and in nature. The upper pyroxeneslies at lower temperaturesthan the runs in limit for the minimum temperature of pigeonite which we detected pigeonite. We hope that future formation must lie at lower temperaturesthan our investigators will better determine these reactions runs in which pigeonite,but not augite,was detected and evaluatethe many possiblevariables. in orthopyroxenehost. The upper limit cannot lie at greatertemperatures than those of our runs in which Acknowledgments both pigeonite and augite were detected in the We thank our colleagues at the U.S. Geological Survey for their orthopyroxene. The experiment of Simmons et a/. help in this study: Lovell B. Wiggins and Nelson Hickling assisted (1974) provides a lower temperaturelimit of 900"C in analyzing some of the pyroxenes; Mary Woodruff processed at Fel(Mg+Fe) = 9.75.It is possiblethat there is no much microprobe data through the computer and desk calculator, umque temperature-compositionrelationship for and assisted in assembly of the bibliography and typing of the the manuscript; Gordon L. Nord, Jr., provided reaction. There might the transmission be minor but unrecognized electron photomicrograph ofthe dislocations, Figure 28, and gave variations with changesin minor-elementchemistry, us much help with his knowledge of subsolidus reactions in and there is a good possibility that the temperatureat silicates.David R. Wones, G. L. Nord, D. H. Lindsley, and D. G. which the reaction is initiated will depend upon the Smith provided technical reviews of the manuscript. We are fine structure of the reactants. particularly grateful for the thorough and penetrating review by Lindsley. This research was supported in large part by NASA Uncertainty in the temperatureat which iron-rich contract T-2356A. orthopyroxene * augite react to form pigeonite prompts a note of caution. The 3-pyroxene assemblagesreported by Ross and Huebner (1975b) References at Fel(Mg+Fe) > 0.55 were based on the lowest temperatureat which pigeonitewas observed.Augite Atkins, F. B. (1969) Pyroxenesof the Bushveld intrusion, South did not always coexist with orthopyroxenesin these Africa. "/. Petrol., 10,222-249. (1964) runs at Fel(Mg+Fe) > Barker, F. Sapphirine-bearingrock, Val Codera,Italy. Am. 0.55; their reported Mineral.,49, 146-152. subsolidus geothermometershould be regarded as Binns, R. A. (1962)Metamorphic pyroxenes from the Broken Hill yielding maximum possibletemperatures. district, New South Wales.Mineral. Mag., 33,320-338. Bowen,N. L. and J. F. Schairer(1935) The systemMgO-FeO- aA similar relationship exists between the two-layer clay SiO1 Am. J. Sci.,29, l5l-217. mineral dickite and its one-layer polymorph kaotinite, which Boyd, F. R. and J. L. England (1965)The rhombic enstatite-cli- difer in S3e, rs by only 1.4+0.6 cal deg-t mole-t (King and noenstatiteinversion. CarnegieInst. Il'ash. YearBook,64, ll7- Weller, 196l). At least some of this enrropy change may be due to r23. rearrangement of the hydrogen bonds between the layers. - and D. Smith (1971) Compositional zoning in pyroxenes I 154 ROSS AND HI]EBNEK AT]GITE, PIGEONIT:E, AND ORTHOPYROXENE

from lunar rock 12021, Oceanus Procellarum. I. Petrcl., 12, 439464. ofpyroxenes and the origin ofmare basalts- Lunar Sci' IV'397- Brown, G. M. (1957) Pyroxenes from the early and middle stages 399. Lunar Science Institute, Houston, Texas. petrology of fractionation of the Skaergaard intrusion, East Greenland. Immega, I. P. and C. Klein, Jr. (1976) Mineralogy and Mineral. Mag., 3 l, 5l l-543. of some metamorphic Precambrian iron-formations in south- - (1968) Experimental studies on inversion relations in natu- western Montana. Am. Mineral. 6l' lll7-1144. ral pigeonitic pyroxenes. Carnegie Inst. Wash. Year Book,66, Ishii, T. (1975) The relations between temperature and composi' geo- 347-353. tion of pigeonite in some lavas and their application to - and E. A. Vincent (1963) Pyroxenes from th€ late stag€s of thermometry. Mineral. J.' 8, 48-57 - sub- fractionation of the Skaergaard intrusion, East Greenland. ,/. Ito, J. (1976) Experimental studies of Li+ + Sc+ coupled Petrol.,4,175-19'1. stitution in the Mg-silicates: olivine, clinopyroxene, ortho- phase with Butler, P. (1969) Mineral compositions and equilibria in the meta- pyroxene, protoenstatite and a new high temperature with Programs, 8, morphosed iron formation of the Gagnon region, Quebec, Can- c : 27A (abstr.). Geol. Soc. Am- Abstracts ada. J. Petrol.,10, 56-101. 93't-938. Cameron, E. N. (1963) Structure and rock sequencesof the critical Jaffe,H. W., P. Robinson,R. J. Tracyand M' Ross(1975) Orien- augite: zone of the Eastern Bushveld Complex. Mineral. Soc. Amer. tation of pigeonite exsolution lamellae in metamorphic phase Spec. Pap., I, 93-10'l . correlation with composition and calculated optimal Coe, R. S. (1970) The thermodynamic effect of shear stress on the boundaries. Am. Mineral., 60' 9-28' heat ca- ortho-clino inversion in enstatite and other coherent phase tran- King, E. G. and W. W. Weller (1961) Low-temperature sitions characterized by a finite simple shear. Contrib. Mineral. pacities and entropies at 298'15'K ofdiaspore, kaolinite, dickite Petrol., 2,247-264. and halloysite. LI.S. Bur. Mines Rept' Inv. 5810- - and S. H. Kirby (1975) The orthoenstatite to clinoenstatite Klein, C. (1966) Mineralogy and petrology of the metamorphosed 7' transformation by shearing and reversion by annealing: mecha- Wabush Iron Formation, southwestern Labrador. J. Petrol', nism and potential applications. Contrib. Mineral. Petrol., 52, 246-305. 29-55. Kuno, H. (1954)Study of orthopyroxenesfrom volcanicrocks' Dallwitz, W. B., D. H. Green and J. E. Thompson (1966) Cli- Am. Mineral., 39,3046. noenstatite in a volcanic rock from the Cape Vogel area, Papua. Larsen, E. S, Jr., and W. M. Draisin (1950) Composition of the 1n- J. Petrol., 7,375403. minerals in the rocks of the southern California batholith' Davidson, L. R. (1968) Variation in ferrous iron-magnesium dis- tern. Geol. Congress, lSth Session,Pt. ll' 66-79. (1973) Subsolidus tribution coefficients of metamorphic pyroxenes from Qua- Lindsley, D. H., J. E. Grover and C. E. Calabro near irading, western Australia. Contrib. Mineral. Petrol., 19, 239- phase relations of Ca-Mg-Fe pyroxenes wit! FelFe+Mg from 600- 259. 0.75 at pressures from l-5 kbar and temperatures Desborough, G. A., and H. J. Rose, Jr. (1968) X-ray and chemical 800'C (abstr.). Geol. Soc. Am. Abstracts with Programs, 5,71!- analysis of orthopyroxenes from the lower part of the Bushveld 714. of the Complex, South Africa. ln Geological Sumey Research 1968: Mason, B. and E. Jarosewich (1971) The comppsilion IJ.S. Geol. Surv. Prof. Paper,600-B,Bl-85. Johnstown meteorite. M eteoritics' 6' 241-245. of hy- Dodd, R. T., J. E. Grover and G. E. Brown (1975) Pyroxenes in Nakamura, Y. and L Kushito (1970a) Equilibrium relations micro- the Shaw (L-7) chondrite. Geochim. Cosmochim. Acta, 39, 1585- persthene, pigeonite, and augite in crystallizing magma: 1594. probe study of a pigeonite andesite from Weiselberg, Germany' Gooley, R., R. Brett and J. L. Warner (19'14) A lunar rock of deep Am. Mineral., 55, 1999-2015. - crustal origin: sample 76534. Geochim. Cosmochim. Acta, 38, - and (1970b) Compositional relations of coexisting 1329-1339. orthopyroxene, pigeonite, and augite in a tholeiitic andesite Grover,J. (1972)The stabilityof low-clinoenstatitein the system from Hakone Volcano. Contrib. M ineral. Petrol., 26' 265-27 5' Mg2Si2O6-CaMgSi2O6(abstr.). Trans.Am. Geophys.Union, 53' Nicholson, R. B. (1968) Nucleation at imperfections' ln Phase 539. Transformalions, p. 269-312. American Society of Metals Semi- D. H. Lindsley and A. C. Turnock (1972)Ca-Mg-Fe py- nar Publication, New York. roxenes:subsolidus phase relations in iron-rich portions of the Papike, J. J., F. N. Hodges, A. E' Bence, M. Cameron and J' M' pyroxene quadrilateral (abstr.). Geol. Soc.Am. Abstracts with Rhodes (1976) Mare basalts: crystal chemistry, mineralogy, and Programs, 4, 521-522. petrology. Rev. Geophys. Space Phys., 14' 475-540. Heald, E. F. (1967) Thermodynamicsof iron-platinum alloys. Podpora, C. and D. H. Lindsley (1979) Fe-rich pigeonites: mini- Trans.Met. Soc.AIME,239, l3t'1-134[. mum temperatures of stability in the Ca-Mg-Fe quadrilateral Huebner,J. S. (1973)Experimental control of wiistite activity and (abstr.). Trans. Am. Geophys. Union, 60,420421- mole fraction (abstr.). Geol. Soc.Am. Abstractswith Progrums, Raleigh, C. B., S. H. Kirby, N' L. Carter and H. G. Ave Lallemant 5, 676-6'17. (1971) Slip and the clinoenstite transformation as competing qll4022' - (1975) Oxygen fugacity values of furnace gas mixtures. rate processes in enstatite. J. Geophys. Res., 76, Am. Mineral., 60, 815-823. Ramberg, H. and G. W. Devore (1951) The distribution of Fe2* B. R. Lipin and L. B. Wiggins (1976)Partitioning of chro- and Mg2+ in coexisting olivines and pyroxenes. J. Geol., 59, mium betweensilicate crystalsand melt. Proc. Lunar Sci. Conf. 193-210. 7th, t195-1220. Riecker, R. E. and T. P. Rooney (1967) Deformation and poly- M. Ross and N. Hickling (1975)Significance of exsolved morphism of enstatite under shear stress. Geor' Soc' Am. Bull., pyroxen€sfrom lunar brec*ia77215.Proc. Lunar Sci.Conf. 6th' 78, t0/5-1O54. 529-546. Rietmeijer, F. J. M. (1979) Pyroxenes from lron-rich lgneous ROSS AND HIJEBNER: AI]GITE, PIGE1NITE, AND 1RTH1PYR1XENE l 155

Rocks in Rogaland, ph.D. S.W. Norway. Dissertation, Utrecht Smith, J. V. (1959) The crystal structure of protoenstatite, University, The Netherlands. MgSiO3.Acta Crystallogr.,I 2, 515-519. Ross, M. (1968) X-ray diffraction effects by non-ideal crystals of Smyth, J. R. (1969) Orthopyroxene-high-low clinopyroxenein_ biotite, muscovite, montmorillonite, mixed-layer clays, graphite, versions.Earth Planet.Sci. Lett.,6,4c(-407. and periclase. Z. Kristallogr., 126,80.97. - and J. Ito (1977)The synthesisand crystal structureof a - and J. S. Huebner (l975a) Estimation of the minimum magnesium-lithium-scandiurn protopyroxene. Am. Mineral., temperatur€ for existence of orthopyroxene, pigeonite, and aug_ 62. t252-t25'1. ite and its application to prediction of temperature of crystalli_ Stewart, D. B., M. Ross, B. A. Morgan, D. E. Appleman, J. S. zation of lunar pyroxenes. Lunar Sci. W,6g9_691. Lunar Sci_ Huebner and R. F. Commeau(1972). Mineratogy and petrology ence Institute, Houston, Texas. oflunar anorrhosite15415. Lunar Sci.III,726-729. -- - Lunar Sci_ and (1975b) A pyroxene geothermom€ter based on ence Institute, Houston, Texas. composition-temperature relationships of naturally occurring Subramaniam,A. P. (1962)Pyroxenes and garnetsfrom charno_ orthopyroxene, pigeonite, and augite (abstr.). Extended Ab_ ckites and associatedgranulites. In A. E. J. Engel, H. L. James stracts, International Conlerence on Geothermometry and Baro_ and B. F. Leonard, Eds.,Petrologic Studies: A Volumeto Honor metry. Pennsylvania State park, pennsyl_ University, University A. F. Buddington,p.2l-36. GeologicalSociety of America, New vania. York. --, - and E. Dowty (1973) Delineation of the one_atmo_ Turnock, A. C., D. H. Lindsley and J. E. Grover (19?3)Synthesis sphere augite-pigeonite miscibility gap for pyroxenes from lu- and unit-cell parametersof Ca-Mg-Fe pyroxenes.Am. Min- nar basalt 12021. Am. Mineral., J8, 619_635. eral., 58,50-59. -- and N. L. Hickling (1972) Delineation of rhe or_ Veblen, D. R., P. R. Buseckand C. W. Burnham (1977)Asbesti_ thopyroxene-pigeonite transition and its bearing on pyroxene form chain silicates:new minerals and stfuctural groups. Sci_ phase relations in lunar rocks. Lunar Sci. 1V,637_639. Lunar ence,198,359-366. Science Institute, Houston, Texas. Virgo, D. and M. Ross (19?3) Pyroxenesfrom Mull andesites. Sclar, B., C. L. C. Carrison and C. M. Schwarrz (1964) High_pres_ CarnegieInst. Ilash. Year Book, 72,535-540. sure stability field of clinoenstatite and the orthoenstatite_cli_ Wager, L R. and W. A. Deer (1939)Geological investigationin no€nstatite transition (abstr.). Trans. im. Geophys. Union, 45, East Greenland. Part III. The petrology of the Skaergaard t2t. In_ trusion, Kangerdlugssuag,East Greenland. Medd. om Grdnland, Simmons, E. C., D. H. Lindsley and J. J. papike (1974) phase rela- 105. t-352. tions and crystallization sequence in a contact-metamorphosed Yang, H. Y. (1973) Crystallization of iron-free pigeonite in the rock from the Gunflint petrol., Iron Formation, Minnesota. J. systemanorthite-diopside-+nstatite-silica at atmosphericpres_ 1t 539-65. stre. Am. J. Sci.,273,488497. Smith, D. (1972) Stability of iron-rlch pyroxene in th€ system CaSiO3-FeSiOr-MgSiO3. Am. M ineral., 5 7, 1413_1428. - (1974) Pyroxene-olivine-quartz assemblages in rocks asso_ ciated with the Nain Anorthosite Massif, petrol.. Labradot. J. Manuscript received, December 19, 1977; 1J. 58-78. acceptedfor publication, July 9, 1979.