521

The C anadinn Mineralo gistVol. 33, pp. 521-535 (1995)

CRYSTALLIZAf,ION AND RE.EOUILIBMTION OF ZONED CHROMITEIN ULTRAMAFIC CUMULATES. VAMMALA Ni.BELT,

SOUTHWESTERN FINLAND

PETRIPELTONEN

Geologicd Survey of Finl.an[ FIN-02150 Espoo, Finl.and

ABSTRA T

Cotectic proportions of chromite and olivine coprecipitated in tholeiitic island-arc-type mamas that intruded theSvecofennian supracrustal rocks of southwestem Finland during the orogeny. Crystal growth ofchromite continued until grainsbecame trapped by olivine, or was trenninated because of the appearance of clinopyroxene. From grain to grain, Cr(Cr + Al)varies between 0.25-0.80, whereas Mg(Mg + Fe2+) varies only little for any grven Cr#. The large Cr# range has been relatedto the continuous depletion of Cr in the magma due to chromite crystallization, and to concomitant emichment of magma in Alas a response to assimilation ofpettic sediments and accumulation of olivine. Because fractional crystsllization took place infeeder conduits at moderate crustal pressures, the possibility remains thal some of the Cr# variation arises from polybaricgrystallization of chromite during magma ascent. Almost all chromite grains are zoned toward a more Al-rich rim. Suchchemical zonation cannot be explained by any postcumulus or subsolidus mechanism, but is interpreted as initial growth-induced zonation, which lnoduces a grain-scale "memory" of the incremental compositional changes that took place duringcrystal growth, During subsolidus cooli4g, the grains of chromite re-equilibrated extensively with enclosing silicates byditfrrsion-controlled Mg-Feh cation exchange. This is not only supported by abnormally low (50G600'C) olivine-spinelblocking temperatures, but also by difhrsion profiIes preserved in the adjacent olivine. Subsolidus re-equilibration betweenolivine aad spinel proceeded to an unusual exten! indicating s s1e\il soeling rate for the ultramafic cumulates ofthe VammalaNi-belL

Kewords: cbromite, cumulates, crystal zoning, fractional crystallization, assimilation, subsolidus re-equilibratioq difhrsion,hoterozoic, Vammala Finland,

Solvnlans

La chromite et I'olivine ont cristallis6 en proportions cotectiques b partir dun magma thol6iitique typique d'un milieu d'arcinsulaire, mis en place dans les roches supracrustales lors de I'orogenbse svdcofennienne, dans le sud-ouest de la Finlande.La croissa:rce des cristaux de chromite a progress6 jusqu'au point oi ils ont 6tf pi6gds par l'olivine, ou bien jusqu'h la formationdu clinopyroxbne. D'un grain I I'autre, la valeur de Cr(Cr + A1) (c'est-i-dire, Cr#) varie ente 0.25 et 0.80, tandis queM/(Mg + Fe2+) ne varie que peu potu une valeur de Cr# donn6e. On explique llntervalle important des valeurs de Cr# par unappauwissement trnogressif du magma en Cr i cause de fu 66lstallisatti6n de la chromite, et par un endchissement progressif dumagma en Al suite i l'assimilation de s6diments p6litiques et A I'accumulation de I'olivine. Comme la cristallisation fractionn6edu magma s'est d6roul6e dans des conduits nouniciers h une pression crustale mod6r6e, la possibilit6 existe qu'une partie de lavariation en Cr# r6sulte de la cristallisation polybarique de la cbromite au cours de la mont6e du magma. hesque tous les grainsmontrent une zonation vers une bordure alumineuse. Une telle zonation en composition ne pourait r€sulter de r6-fouilibragepost-cumulus ou subsolidus; elle indiquerait plut6t une zonation primaire, qui conserve d l'dchelle d'un grain une "m6moire" deschangements incrdmentiels du maena pendant la croissance des cristaux. Pendant le refroidissemen! et donc A un stadesubsolidus, les grains de chromite ont r6agi I grande &helle avec les silicates qui les engloben! par &hange diffirsionnel de Mget de Fd+. Cette interpr6tation d6coule non seulement d'une tempdrature dablocage-enfte olivine et spinelle anormalementfaible, de 500 n 600oC, mais 4lrggi de gofils de composition pr6serv6s dans lblivine contigue. I,g rd-dquilibrage entre olivine etspinelle s'est poursuivi i un stade avanc6, indication d'un refroidissement prolong6 des cumulats ultramafiques de la ceinturenickelifBre de Vammala

Clraduit par la R6daction)

Mots-cl6s: chomite, cumulats, zonation des cristaux, cristallisation fractionnde, assimilation, re-6quilibrage subsolidus,difhrsion, pot6rozolque, Vammala, Finlande.

522 THE CANADIAN MINERALOGIST

INraopucuoN

1Xs qemposition of chromite is strongly dependentupon the composition and oxygen fugacity of theparental magma (rvine 1965, L967, Htll & Roeder1974, Sigurdsson & Schilling 1976, Allan et al. L988,Roeder & Reynolds 1991). These parameters in turnare controlled by the conditions prevailing duringpartial melting, fractional crystalllzation, convectionand mixing in magma chambers, contamination, Iiquidimmiscibility, and gaseous transfer processes. Theeffect of the pressure of crystallization is not wellconstained, but it may exert some control on thecomposition of chromite by affecting the partitioncoeffrcients (Dick & Bullen 1984) and the order ofappearance of minerals (Ctre.en et al. L97L).

Straightforward deduction of igneous events on thebasis of chemical variability of cbromite is, however,impeded by the strong tendency for it to re-equilibrateunder subsolidus conditions (kvine 1967, Cameron1975, Medaris 1975, Wilson 1982, Hatton & vonGruenewaldt 1985" Kimball 1990. Johnson & Prinz1991, Yang & Seccombe 1993, Roeder 1994).Although in this respect unfortunate, this tendency isalso of value in providing an opportunity to evaluatethe postmagmatic thermal history of igneous bodies. Inaddition, a wealth of petrogenetic informationpertaining to the igneous, subsolidus, and even tectonichistory of the host rock is recorded by the chemicalzonation in chromite (e.g.,Ef, Goresy et al. L976, Allanet aI. 1988, Ozawa 1989, Michailidis 1990, Sack &Ghiono 1991).

This work presents a detailed decription of the modeof occurrence, chemical variability and zonation ofchromite from the ultramafic cumulates of theVammala y1-6elt, southwestern Finland. An attempt ismade to remove the overlapping effects of alteration,subsolidus reactions and fractional crystallization, andthus trace out the compositional evolution of thismineral. A considerable amount of data now existsconcerning the igneous petrology, subsolidus reactionsof silicates, aud ore-forming processes in cumulates ofthe Vammala Ni-belt (Peltonen 1995a, b), whichprovides a sound background with which to interpretthe commonly ambiguous chemical variability ofchromite.

RecroNAL Geolocv AND AN Oumnw oF CITMULATESOF TTIE VAMMALA Ni-BSLT

The major geotectonic units of the Fennoscandian(formerly known as Baltic) Shield are the Archeanbasement (3.1--2.5 Ga) and the early Proterozoic(1.9-1.8 Ga) Svecofennian terrane (Fig. 1, inset). TheSvecofennian schists in westem and southern Finlandare predominantly mica gneisses whose protoliths weregraywacke-slate turbidites (Simonen 1980). In addi-tion to the formation of abundant granitic rocks, the

cuhnination of the Svecofennian orogeny was accom-panied by less voluminous tholeiitic magmatism, nowrepresented by mafic-ulnamafic cumulates thoroughthe Svecofennian terrane. The Vammala Ni-belt insouthwestem Fidand is a zone of operating andabandoned Ni-Cu sulfide mines and promisingprospects that are associated with bodies of primitiveolivine-rich cumulate. These bodies are best describedas ultramafic plates, pods, pipes, and lenses that "float"in high- to medium-grade metapettes (Fig. 1). Onlyminor amounts of gabbroic cumulates occur in theregion. Thus these ultramafic cumulates cannot repre-sent tle basal units of larger fractionated intrusions.Nor do they represent "Alpine-type" peridotites or theproducts of ultramafic magnas (e.9., komatiites), buthave instead been interpreted @eltonen 1995b) as theremnants of larger synorogenic conduits for basalticarc-type magmas. Eventually, these conduits becamechoked by cumulus crystals and were boudinaged intosmall lenses and fragments by concomitant tectonicmovements, probably even before they were com-pletely solidified.

PsrRorocv oF TTD ClMuLArEs ANDRsanet Ni{u DsPosrrs

Four representative bodies of ultramafic cumulatesfrom the central part of the Vammala Ni-belt weresampled from drill cores that transect whole bodies(Frg. 1). The Vammala intrusion hosts a low-grademagmatic Ni-Cu deposit that has been exploiled since1978 (HaHi et al. 1979), and the Ekojoki intrusion isassociated with a subeconomic Ni{u-PGE deposit.The Posionlahti and Murto intrusions were chosen asrepresentatives of the cumulate bodies that are devoidof known concentrations of sulfides. The Yammala,Ekojoki, and Posionlahti cumulate bodies are enclosedby high-grade migmatitic metapelites, whereas thecountry rocks of the Murto intrusion were meta-morphosed at a somewhat lower gtade and did notundergo partial melting during peak regional meta-morphism.

Petrograplry

All the cumulate bodies sampled consist of ortho-and mesocumulates. Although approximately 757o ofthe igneous pyroxenes and olivine have been replacedby retrograde amphiboles and pseudomorphic mesh-textured serpentine, respectively, the cumulus texturesare well preserved. Olivine, the most cornmon cumulusmineral, usually occws as euhedral grains up to 3 mmin diameter that have well-preserved grain margins.Usually, olivine is the only cumulus minslal, forming5A-,70Vo of the cumulates. Close to the periphery of thecumulate bodies, the modal content of olivine maydrop to 40Vo, at which point clinopyroxene also appearsas a cumulus phase. Clinopyroxene (diopside-augite)

ZONED CTIROMITE IN TI{E VAMMAT.A, MCKEL BELT 523

CENTRAL PART OF VAMMALA NI.BELT

23"00' EI

2h

FENNOSCANDIAN SHIELD

AILAI{IICOCtsAN

m AnclcANE:] EAFLYPrcTErcZOCE PGrsvEcsEl.lNtAN=l PHANEmZOCCO/En\ vlaluluu.sEll

ffi uiernatiticmetapslif€sandminor(<5 7o)meravolcanicrmts t_] Graniroids

\* glatschists a UltramaficcumulatssFIc. 1. Simplified geology of the Fennoscandian Shield (inset) and the cenral part of the Vammala Ni-belt, southwestern

Finland.

is the dominant intercumulus mineral, formingoikocrysts up to several centimeters in diameter.Orthopyroxene, which occurs as large oikocrysts andsmall anhedral grains, is less common than clino-pyroxene (cpx:opx = 3:l). Intercumulus amphibole ispresent only in amounts of a few percent and occurs inthe latest intercumulus voids and in a resorption-induced rim around the clinopyroxene; it can generallybe distinguished from secondary amFhiboles only onthe basis of its composition. Most Fe-Ni{u sulfidespresent in the mineralized cumulates occur as inter-stitial disseminations or as net-textured ore. withpyrrhotite - pentlandite - chalcopyrite + cubanite tmackinawite the most common sulfide assemblaee.

Petrogenesis

These ultramafic bodies are cumulates derived froma tholeiitic basaltic magma. The order of crystallization

was olivine + chromite, followed by clinopyroxene andorthopyroxene, fhen amphibole. The entry of plagro-clase was delayed, probably as a consequence ofcrystallization at a high confining pressure (e.9.,Kushiro & Yoder 1966). Other evidence for crystal-lization at a moderate to high crustal pressure includesextensive redistribution of Ca between pJiroxenesduring cooling, high KDcpx/ol@g/Fe) values, and therelatively aluminous composition of the pyroxenes.The parental magma that formed these cumulates waspervasively contaminated by local turbidites prior tothe final emplacement. Most of the assimilationoccurred during ascent of the magma in the arc crust;in situ assimrlation played only a minor role. Mineraland sulfide compositions of associated magmatic Ni -Cu + PGE deposits require that sedimentary sffidesparticipated in ore genesis, and were the ultimale causefor sulfide saturation and segregation in VammalaNi-belt intrusions (Peltonen l995ub).

524 TIIE CANADIAN MINERAI,OGIST

Four major events, magmatic crystallization, a-Fh-bolite-facies prograde metamorphism, retrogrademetamorphism, and static serpentinization, arerecorded by the mineral assemblages of the cumulates.Owing to their synkinematic emplacemenl however,only the outermost margins of the cumulates weresusceptible to the prograde growth of minerals,whereas away from the margins, the cumulates largelyretained their igneous mineralogy and textures(Peltonen 1995b).

DrsrnmtmoN eNr PSTRocRAPHYOF THE CHROMITE

Euhedral to subhedral grains of chromite are evenlydisseminated throughout the cumulates. Its modalabundance does not exceed 2-37o in any sample.Originatly, virtually all grains were either included incrystals of cumulus olivine, except for some that wereenclosed by intercumulus minerals, or located at grainboundaries between olivine and intercumulus minerals(Frg. 2). In the following discussion, grains in thesediffering textural settings are referred to as included,intergrain and grain-boundary chromites, respectively(Cameron 1975).

Most grains of chromite are smaller than 50 pm. Theavenrge grain-size of grain-boundary and intergrainchromites exceeds that of included grains @ig. 2).Chromite did not reach saturation much before olivinein the magma, as indicated by the small grain-size(<10 pm) of the chromite in the core of the olivinecrystals. The nearly simultaneous saturation of

2A 40 60 80 100 L2n l4

Graindiameter [HmlFtc. 2. Grain-size distribution of chromite in the cumulates

studied. Note that tle relative amount of includedcbromite is lower in the categories of larger grain-diameter.

cbromite and olivine, followed by their precipitation incotectic proportions, was not conducive to the forma-tion of chromite-rich cumulus layers. Textural featuresimply that chromite growth continued until a) grainsbecame happed by growing crystals of olivine, or b)the appearance of clinopyroxene rapidly consumed thechromium still available.

Included and intergrain chromites have a well-preserved euhedral to subhedral outline, and areeither opaque or almost so, and dark brown in color@gs. 3A{). Other morphologies, e.9., skeletalo arecompletely absent. The reflectance of chromitegradually decreases and the transparency increasesoutward from the core. Compared to the core are4 therim is lighter brown in fransmitted light, and the outer-most edge may have a greenish hue. Significantly,those grains completely enclosed by Fe-Ni-Cusulfides also show similar optical zoning. Some grainsof grain-boundary chromite have distinct axhedralprotrusions of green spinel that are restricted to theintercumulus side of the crystals (Fig. 3D). Somechromite grains contain tiny inclusions of Fe-Nisulfide, indicating that immiscible sulfides becamesaturated in the magma as chromite nucleated.Fractures within chromite grains, related to deforma-tion or serpentinization-related changes in volume, alefilled with magnetite, hydrous phyllosilicates andremobilized sulfides.

Locally, zoned grains of chromite are surrounded byan oxidized outer mantle of ferrian chromile, ferrianchromite + magnetite, or magnetite. The reflectance offerrian cbromite is intermediate to that of the chromiteand magnetite. Ferrian chromite is most commonlyobserved around grains of intergrain chromite that areenclosed by secondary amphibole. Where ferrianchromite surrounds chromite grains embedded inlrzardits, it may also be associated with chlorite.

Arervrrcer PRocEDURES

From each sample, five to ten grains of chromite ofdifferent sizes, and enclosed by various host-minerals,were selected for electron-microprobe analyses. Grain-size measurements were approximated by the diameterof the "best-fit'o circle assuming that the plane ofthe polished thin section passes through the core of thegrain, even though in most cases this would beforhritous. The influence of ftis snmfling effect onresults is discussed below. Mineral analyses were doneusing Cameca SX-50 electron microprobe facilities atthe Geoanalytical Laboratory of the OutokumpuMining Services and at the Geological Survey ofFinland. Operating conditions were: 20 kV, 15 nA,beam diameter 5 itm, except for step scans (15 kV,30 nA, 1 pm). Standards included both syntheticcompounds and natural minerals. The peaks werecounted for 1G-15 seconds, which, for ZnO for exam-ple, correspond to a precision of about t 0.M ttrL%o.

n Includedchroniteffi Crrain-f<xrndarycbromixeil fntergrain chrmite

ZONED CHROMIIE IN TIIE VAMMAT.A. MCKEL BELT

Hc. 3. A. Euhedral to subhedral grains of cbromite (Chr) in mineralized olivine{hromite cumulat€. Most grains are enclosedby pseudomorphic liz,atdtte Q,z) after primary olivine (Ol), whereas some of the larger grains occupy olivine - intercumulussulfide @o + Pn) grain boundaries. B. lrtergrain chromite completely enclosed within pyrrhotite, Black line indicateslocation ofthe step-scan profile ofFig. 9. C. Included grain ofcbromite optically zoned toward arim ofhigher transparency.D. Iarge grain-boundary cbromite between olivine and calcic clinoamphibole (Cam). Note the green anhedral spinelprotrusions along their intercumulus sides.

525

Total iron was recalculated according to the method ofDroop (1987), using stoichiometric criteria. Repre-sentative results of the analyses are presented inTable 1.

RFSULTS

Chemical variarton of the chromite cores

Because the ferric iron content of the cores and rimsis uniformly low (Table 1), results can be displayedconveniently in plots of Cr(Cr + Al) (Cr#) versusMg(Mg + Fe2*) (Mg#). On the Cr#-Mg# diagram(Fig. ), the field of the Vammala Ni-belt suiteof chromite compositions is distinct from the fields ofchromite in Alpine-type and abyssal peridotites @ick& Bullen 1984), layered intrusions, southeasternAlaskan infiusive bodies (Irvine 1967), and komatiites(e.9., fuou & Kerrich 1992). Overall, the form and

size of the field resemble those of oceanic peridotilesand associaled lavas (e.9., Dick & Bullen 1984),although the Vammala suite is located at much lowerMg#. Magmatic chromite that has similal Q14-14tXsystematics to the Vammala Ni-belt suite has only beenreported from ulfiamafic cumulates of island-arc-rootcomplexes (e.g., DeBari & Coleman 1989, Jan &Windley 1990, Kepezhinskas et aI. 1993).

Several additional fundamental features are dis-played by detailed CriF-Mg# plots @g. 5). Such plotsindicate that, especially in the case ofincluded grains,a positive correlation exists between Cr# and grainsize. Some of this effect arises from the exposure ofrandom sections of zoned grains on the polishedsurface. However, at least two findings, namely thatzoning with respect to Cr# within one grain is muchless extensive than variation between samples, and thatthe smallest grains are indeed visibly euhedral in shape,suggest that this sampling effect is not responsible for

526 TTIE CANADIAN MINERALOGIST

TABLEl. REFR,E^SENTAITYBI)RESIILTSGELBSTRON-MICR.OPROBEANALYSESOFCIIROMIIE,VAMMALANi.BELT

Codez)Aain lmrion3)Dimecrtlml 65 50Poim locarima) w ftn @ in fe+fu

SiO2 (wt%)azQRzQFeOMsoCsOTtU2lv[rtovzQGzQl{o7iA

Total

c6n NusiAIIb3+R2/l4cCaItl&vGN78Mgft@YFels

w0s

IN

v61o6

v59ss150

wo

PlP

0.6 0.041X00 Bn6n 756

4.6 29.fiL6r 2870.05 0.08056 0.630.rl ogl0n 0J9

4635 44Sr0.m 0.000.s 0.4{)

9.92 L@.42

0J09 00113g2t 4&1.412 rJ596.875 6.780l.fp 1.1730.015 00?2,0.117 o.tD0.0t6 00850.170 0.174

t02A 9:tg20m 0.m0.60 0.0810.136 aJ47a:ra a.M0.Dl 0.1m

001 0.m 0,v)15s2 L6.y 2g736 7.10 2j..&

29.@ 8.65 ?0.C7\6t 26 0360.ff3 001 0.010J6 056 0.880.6 033 0.4061 0.63 0:t9

424) 41.6'' 36.01ofi) 0.01 0.001.6 rA Ag7

100.41 100.40 w53

0.003 0.001 0.0d/4.912 5.n5 rM21J08 r.W 59086.331 6.708 76711.060 1.$0 0336o.mE 0.004 00020.114 0.11:} 0.r%0.106 0.n4 0.1110.132 0.135 0.1S92t3 8.915 8.4600.m0 0.06 0.0010's7 0z.3 0.0820.136 0.89 0M29.62 0.62 03900096 0g) 03&

@e fun fe4b

0.00 001 0.00mn a.& !.0"996.89 7.15 fi24

25J0 25.49 D65461 5.6 LS}or2 a.az 0.m039 092 0i4aAr 038 0J8049 0.47 055

n.ca 34.n 3E.630.00 0.6 0.v2135 r.43 039

9946 9.m 99.&

0.000 0.@ 0.m6.469 72tt 3,6y1369 1394 34585199 5:74 6:W78t7 1950 08530.m7 0.m5 00000.0/8 0.62 0'0730.091 0.$5 0.1400.104 0.097 oJa.7W) 7.168 E.6180.m 000402,4' 0zl5 0.0810.85 02t? 0.1090550 049, o:tu0.67 0.m8 0w

olx) o.@,18.@ 18.415s3 6.@

n53 t.8393 4.r40.02 0.80.44 0.410.47 0.450.,18 0.41

4Lt6 47t30.tD 0.61.45 r37

0,01 0.04432 2n367.n 7g

2652 %24454 5.630.6 ag)031 0210.47 0384.49 045

3538 31960.01 0.@135 lrJ4

1002E 1002r

0.008 0.0116.U2 7.9@1550 r.&35J69 5J881.915 L1360.016 0.0050.060 0.0100.104 0.620.102 00g)7s79 6.840.001 0.0ffi0.69 0a.20,49 0m0519 0&90.098 0.093

feab mg

0.01 0.@tt47 0.0115.09 69.4a35 31.&,7' 0.L20.01 0'g,2030 0.000.63 0.(n052 0.6

3931 Ln0.05 0.04051 0.01

99.cl 1V253

0.002 00103JU 0.0@3.178 15.6p6s72 75430g2s 0,0520.003 0.m60.61 0.0010.150 0.0m0.116 o.(tr/8.790 ozn0010 0.@0.16 0.w20.119 0.00/0.69 0.9860,02 09E2

100.43

0.m0 0.m6s.654 s:tx1.189 \W6.131 6.0451560 1.648o0r, 0.0080.m9 0.620.106 0.1030.1@ 0.1018.rn 8.7100.m0 0.0160.n5 0tI0ofi) 02140.611 o.d)l0.o/6 0.fi8

the bulk of the Cr# variation, and that small grains donot simply represent edges of large zoned grains.

The field of Vammala Ni-belt chromite coincidesremarkably with the theoretical isothermal-isobaricisopleths of chromite in equilibrium with olivine of aspecific forsterite content. Because coexisting olivinecontains 77-83 mol.Vo Fo. the location of the chromitefield far to the right ofthe theoretical high-temperatureFo66 isopleth (Fig. 5) indicates that the chromite andthe coexisting olivine do not reflect high-temperatureequiJibrium Qrvine 1965). There is a weak tendency forsmaller gains at any given Cr# to have a lower Mg#.Compositional differences between included and inter-grain chromites are small but significant. The field ofthose intergrain crystals of chromite that are enclosedby silicates (pyroxenes and amphiboles) lies at higherMg# than the field for included chromite. Intriguingly,grains either completely or partly enclosed by inter-cumulus sulfides have a Mg# value as low as that ofgrains included in olivine (Fig. 5).

In most cases, core compositions ofgrain-boundarychromite plot within the same CrtlMg# field as the

included chromite. Some of them have anomalouslylow Cr# or compositionsl 2ening that is stronglypolarized toward their intercumulus sides @ig. 3D).The transition from dark brown cores to bright greenanhedral protrusions is accompanied by extremechemical changes Grg. 6).These protrusions exhibit agreat tendency toward stabilization of the spinel(MgA12O) end-member: levels of Al, Mg, Ni, 7n, andMg# are higher, whereas levels of Cr, Fe3+, Fe2+, Ti,Mn, V and Cr# are lower in protrusions compared tothe cores ofthe grain-boundary spinel.

If ideal spinel solid-solutions are assumed, the dis-tribution of Mg and Fe2* between coedsting olivineand chromite can be used as a geothermometer (Irvine1965). Owing to the extensive hydration of olivine,only apparent blocking temperatures of the pairolivine-spinel rrere obtained. Temperatures, calculatedaccording to the calibration ofFabribs (1979) from thecore compositions of included chromite and the meancomposition of olivine in the same thin sections, arelow (500-600'C) and show a weak dependence ongrain size @rg. 7).

ZONED CHROMITE IN TIIB VAMMAI.A, MCKEL BELT 5n

TABLE 1. (COI\NNIJED)

C r r i o f * r r i - O P O P @ O O Orx*;!-l 10 1m 50 r5o 4 50 7o

poilrl@d@ m we &r w ftn w-=A -ffi

s. --G-G

G=_A

$i0z (wt%)AlzQezQFOr4pc@n@lVlrOvzorauo:lso7fi

Total

CdodS2uygwsiAIIb3+F6

lr8CanIrbvCItt7^l.{ScdYFg+

0.@ 0.04 0r48.8 1537 17.12899 5:B 626

'32 2j.35 U.l6489 4SA 4ga0.06 0J8 0.0t055 032 038034 043 039ofi 051 0J0

853 4543 43350.05 0.05 0I)2131 1.01 1.08

9,J6

0,0f/1.WtJ435,EE41.S?90.0160.l(r/0.or50.0966,W40.011oza01420.4r70.111

9n 1@16

0,@ 0.0114EA7 5361t.t@ t,5.2594/ 5.E151,809 t9540.010 o.(El0.{b5 0.9150.099 0.(ts70.111 0.1fi9.@0 g.LtI0.012 0.0u0101 otlt

0.665 0.6D0m4 0.60

0.02 0.0r 0.6 0.05L'AL U.65 14 , 15397.08 ?.01 64E 5.W

29J2 8.4 ?',59 2As9L1r L72 1Zt 3320.13 0.13 0.04 0.@0.64 0.65 0.49 0.4038 032 039 0g056 058 0.49 0.44

43St 4225 4,41 44.060.05 0.06 0.00 0.@0.?t 0.86 090 0.&t

0.0025.Wtr.4L6:t3X1.10t0.o3E0.136o.cB0.12l'9.U0.0140.\n0.141o@0.@2

100.49 99J3 1@.08 9ps9 9930 9E 19 9950 100.12 E9.45

0.05 a.u 0.04 0.@b.89 54J0 n.10 ,/..45156 3.08 7.14 5.@

x.E9 1756 28.18 2486657 14.01 434 500r 0.01 001 0.m0.15 0ff2 0,44 031oJE 0.10 039 030020 0.06 0.40 0s

31.19 7:74 35BO 35.95ilyt 0.6 0.m 0.(B059 7.O2 0J6 1.05

0.6 0.04 0.01u.g ,,.gr ^.495.68 6.08 5X2

a:9 /lsl nso5.r3 4J0 4590.08 0.05 o.ts030 039 0.40a% 032 032033 034 036

35.04 33.85 ?8s20.6 001 0.04r$ 057 056

0.0!:, 0.013 000849fi 8,43 14.114rut 0.8@ 0509

1350 L457 45890025 0J@ 0.@0@t 0.@E 0.fl)4ornE 0.69 0019aw 0.&0 0.0109512 6.184 LW0007 0104 0.m9o.t6l 0.1(B 0.1650.172 0312 05870.658 0412 0.0870.078 0.054 0.(B2

0.0i0 0.006 00r6f47 7,4L3 75141.412 l-B 1.1(b6.195 5n8 fiylfim Lu7 19E10.q)4 0.0m 0.060.67 0.060 0.05E0.s8 0.065 0.0560.64 0.m 0.ft87fi2 13ll 1.lm0.000 0.007 0,0100.1@ 0.19 02190216 0,l/9 0255as2l 0.497 0.4EE0.090 0.069 0.(r,0

9993 9.4

0.009 0.m6.4t8 6.@L9.92 1.r'16J2t 6'J/.9t:tg l.E3E0.014 o.mEo.(r/E 0.oi9o.un g.wo.ml 0.('5tgt4 79680.0@ 0.060.110 0.1090.a4 0.430555 0ll,50.916 a.vn

characteristic of alteration to ferrian chromite (below),and presumably these rim compositions are affected bysrurounding ferrian chromite or magnetite.

Compositional profiles across chromite-olivineboundaries (Frg. 8) indicate that the olivine also iszoned. The forsterite content of the olivine increasestoward the contact, being up a L.2 mol.Vo htgheradjacent to chromite than away from it' Igneouspyroxenes are commonly uralitized, and attempts tomeasure zoning in pyroxene were unsuccessfirl.

Ferrian chrornite and magnetite tna.ntles

In several cases Out not all), the chromite grains aresurrounded by a mantle of ferrian chromite t mag-netite. Although some of the rim compositions are"tansitional", the compositional boundary betweencbromite and ferrian chromite is commonly sharp.Ferrian chromite has higher Fe3*, Fe2*, Ti, Mn and V,but lower Al, Mg and Zn concenfiations than theadjacent cbromile. Although the Cr content of fenianchromite can either fall below or exceed that of

o@74J0513266fi2ts60.0100.lm0.0E90.1069J480.m00.1E10.169

0.{n5

0.ffi4931tA46:t491095OIB90.1300l)r0.12193.200100r560.14'45520l)93

bntZilCan ^At" oailabb A@ thsb&mad6Suuof tho cological

Swry of Fntd, FIN{Dl50Es!oo, EDI@{ q 6 835" diskeco frm lhe ador.4p, V, p,fr4r.nno tUultua8oomlalsbodls of Pcimbhd.Vmala'Eojo&imlMuo'rcspetively'3)or: *hdeachmolpinels (€@tod€dbyotivbe) P adS: iilcgnbokmorpireloctoredbyrflicnts (m*lypyorc) c FeNi{u$ilndes'resp€rivel$ O} gqin

b"-d-j,rpi""l"66gsiioti"iu-d!t'6*(qdbibols); ss: gEinboudrychrrcspirelsbemagulEds ndrilim4)taan tcim aonnei nag! nagmrilst MgFMs(Mg+F +); cr$+/(&+Ar! Y&3f=Felr/01olr+Ahcr)

kming within included and intergrain chrotnites

Gradual and smooth composition6l 26ning is ubiqui-tous in chromite of the Vammala Ni-belt. The absenceof complex zoning agrees with pefrographic observa-tions that chromite crystallization preceded that ofplagioclase @l Goresy et aI. 1976). In most gnins,irrespective of whether they are included or intergrainspinel, A1, Mg, and Mg# increase, whereas Cr, Fez+,and Cr# decrease from core toward the rim (Figs. 5, 8,9). In addition, Ti and Mn contents show a weaklendency to decrease, and Fe3+, to increase, whereas no76ning of grains with respect to V and Ni wasobserved. The decrease in Cr toward grain edges isbalanced by a concomitant increase of both Al andFe3+. However, the increase in Fe3+ is insignificantcompared to that in Al, suggesting that the strongdecrease in Cr content was not caused by a suddenincrease of flO). The increasing Mg content towardthe rim is balanced by a sympathetic decrease in Fez+.In some grains, Al and Mg decrease, and either Fe2+ orFe3* or both increase. Such compositional trends are

s28

cbromite, the Cr# of ferrian chromite is invariablyhigher owing to the low Al content of ferrian chromite.Magnetite can be either pure magnetite, nickeloanmagnetite or chromian magnetite. Chromite grains thatare surrounded by a ferrian chromite t magnetitemantle have retained their original patterns ofCr-Al and Fe2+-Mg zoung (e.g., Y27 and V59 inTable l).

TIIE CANADIAN MINERALOGIST

Cr/(CY+

DtscusstoN

Any satisfactory interpretation of the chemicalvariation of cbromite in these ulfiamaflc cumulatesmust take the following important observations intoaccount:(1) The core compositions define an exfiemely longand narrow field on the Cr# versas Mg# diagram, the

0.9 Alpine-t''pep€ddotiEs

II

o:l m/.'

05Abyssalp€ridodtes

Me/(Me+F#)Flc. 4. Compositions of chromite (open circle), fenian cbromite (half-filled square) and

magnetite (black square) from the cumulates of the Vammala Ni-belt. Compositionalfields of cbromite in Alpine and abyssal peridotites @ick & Butten 1984), layered andAlaskan-type intusions (Irvine 1967), and in two island arc root complexes (Ionsinafield: DeBari & Coleman 1989, Kamchatka field: Kepezhinskas 1,993) arc indicatedfor comparison. The chromite from the Vammala Ni-belt is distinct from chromitereported from komatiites (not shown), which tends to plot within the field of layeredinfrusions Qkou& Kenich 1992).

AI

$

e

%""":

1 0.9 0.8 03 06 05 0.4 03 02 0.1

ZONED CHROIVtrTE IN T}IE VAMMALA MCKEL BELT 529

Vannalaliacluded chomib

G/(Ch+Al);aI

aaaa

taI

at

a!aaa

t o - a

aoa

aaaa

aaI

aa

aa

oo

E:-150 In

vammalalntagrain chomite

Crl(Cr+Al)

ff Encloedbypyoxenes\-Z mdarylftntes

@ Corufeefr eercfosel by nrlfirle.s

Q rarrff encfoseaby nrllirles

Ekojoki/included cbromite

aata

Crl(Cr+Al)

Fbldof inrgninclromite domd bymrcmmd \'daphibole

\a

aaa

aa

- arolo ..

aI

Ia

ta

aa

aa

a!

aa

P?

05 0.4 03 02 0.1 0J a.4 03 02 0.1 0J 0.4 03 02 0.1

Mg/(Mg+F&) Ms(Ms+F#) ulMe+Fd')

Frc. 5. Detailed Cr#-Mg# plots where core compositions of cbromite from Vammala and Ekojoki are represented by circleswhose diamete$ correspond to those of the grains, and where chemical zonation is illustrated by tie lines that connectthe core and rim compositions. In these representative cumulate bodies, cbromite coexists with olivine that contains77-33 moI.qo Fo. Theoretical Fo66 isopleths (after Dick & Bullen 1984) are drawn to semiquantitatively illustrate what thecomposition of chromite would be if it was in high-temperature equilibrium with adjacent olivine. Scale bar refers tothe grain diameter.

slope of this field being close to that of theoreticalisothermal-isobaric isopleths of chromite compositionsin equilibrium with olivine of a specific Fo content. (2)The chromite grains have an unusually Fe-rich compo-sitions (and a low olivine-spinel blocking temperature)regardless of whether they are included by cumulusolivine. intercumulus silicates or Fe-Ni{u sulfides.(3) Intergrain cbromite has (at any given Cr# and grainsize) a somewhat higher Mg# than included chromirc.(4) hactically all grains of chromite are zoned towarda more A1- and Mg-rich rim. (5) The forsterite contentof olivine increases toward chromite grains.

Role of metamorphism and. alteration

On the Cr#-Mg# plot, the compositional field of theVammala Ni-belt suite of chromite compositionsresembles the field of metamorphic chromite that formin prograde reactions @vans & Frost 1975, Dymeket aI. 1988). However, chromite in the cumulates awayfrom the margin of the body is not a member of anyprograde metamorphic mineral assemblage, but clearly

is an igneous phase, most of which coprecipitated witholivine or immiscible sulfrdes (Figs. 3A{). Cbromitegrains enclosed by pyroxene grains are susceptible toretrogade modifications (e.g., exsolution of spinelduring uralitization). However, because all chromitegrains, irrespective of whether they are enclosed byminerals like olivine, pyroxenes, sulfides, retrogndeamphiboles or serpentine, have the same variation inCr# and Cr-Al zoning (Figs. 5, 8, 9), regional meta-mor?hism cannot be the fundamental cause for theobserved compositional vmiation. Infrared (IR) specfraand X-ray-diffraction Q(RD) studies indicate that theserpentine associated with included chromite ismonomineralic lizardite. Pseudomorphic texturesconsidered together with the stability field of lizarditeindicate that the hydration of olivine was a staticprocess that occurred at low lemperahrres and waterpressnres. Llz,ardrte formation was not followed byreheating, since this would have led to either thereplacement of lizardite by non-pseudomorphicantigorite, or metamorphic forsterite + talc t chlorite(O'Hanley et al. 1989). The structure of cbromite most

530 TTIE CANADIAN MINERALOGIST

Grain-boundary chromite

o.7 0.6 0.5 0.4 03 0.2 0.1Mg,(Mg+Fe2)

FlG. 6. Cr#-Mg# plot for grain-boundary chromite in thecumulales of the Vammala Ni-belt. Scale bar refers tothe grain diameter.

likely is resistant to such static hydration ofthe enclos-ing olivine.

Some of the chromite grains are surrounded by amantle of ferrian chromite. Significantly, the boundarybetween ferrian chromite and the chromite rim isinvariably sharp, and oxidation of iron does not extenddeep into the chromite. Since the original description offerrian chromite material, i.e., *ferritchromit'o(Spangenberg 1943), several mechanisms of formationhave been proposed (e.g., Beeson & Jackson 1969,Bliss & Mclean 1975, Wylie et al. 1987, Shen et al.1988, Burkhard 1993). In the cumulates studied here"ferrian chromite occurs between chromite anda) lizardrte after olivine, b) secondary amphibole,c) chlorite, and d) magnetiie, and apparently has

formed by several distinct processes of alteration,detailed analysis of which falls beyond the scope ofthis contribution. It is worth emphasizing, however,that any reaction involving the formation of ferrianchromite is not a plausible explanation for variable Cr#in the Vammala Ni-belt chromite suite or the ubiqui-tous Cr-Al zoning, because these features are alsoinvariably present in those grains that are not asso-ciated with ferian cbromite.

I conclude that the variability in the VammalaNi-belt suite of chromite compositions was not gener-ated by metamorphism or alteration, but should beinterpreted on the basis of igneous nad sooling-relatedprocesses,

Origin of grain-boundary chromite

The cores of most grains of grain-boundarychromite have an Mg# as low as that of includedchromite, indicating that the extent of cbromite-olivinesubsolidus re-equilibration does not depend on whetherchromite is partly or completely surrounded by olivine.d similar observation was made by Yang & Seccombe(L993), who found that grain-boundary chromitebetween ohvine and plagioclase or between olivine andclinopyroxene re-equilibrated to a similar extent nsincluded grains.

Some of the grain-boundary grains of chromiteexhibit asymmetrical anhedral protrusions of greenMgAlrOo-rich spinel toward intercumulus mineralsGtg. 3D). On the basis of textural evidence, tlese arelikely to be a result of postcumulus growth of grain-boundary chromite that was exposed to trapped,extensively fractionated, intercumulus melt. Suchgrowth depletes the adjacent stagnant interstitial liquidin Cr and Fe, causing progressively more Mg- andAl-rich spinel to crystallize (Hamlyn & Keays 1979).Alternatively, the origin of the overgrowths could beattributed to the exsolution of aluminous spinel frompyroxene during cooling or metarorphism. However,because overgrowths are not common around grains of

chromite completely embedded in pyroxene,this is a less satisfactory explanation.

Variarton in the Cr# of includ.ed andintergrain chromites

Several postcumulus and subsolidus processes areable to change the Cr# of liquidus chromite. Evenincluded chromite is liable to such re-equilibration,because olivine may contain some cbromium @ums1975, Lehmann 1983, Sutton et al. 1993) anldaluminlm (Jurewicz & Watson 1988. Zou & Sleele1993), and thus does not act as a perfect barrier todiffrrsion for these elements. Some investigators haveargued that included chromite and intercumulus meltare able to exchange Cr and Al cations through olivineby volume diffrrsion (Scowen et al. l99l), or maintain

0.7

0.6

OJ

^f 0.4H

C)

0.1

ZONED CHROMIIE IN T}IE VAMMALA MCKEL BELT

650

550

500m 40 CI 80 lm t20 1N 160 180 2W 2n

Grain diameter (Pm)Frc. 7. Variation of olivine-spinel blocking temperature (Fabribs 1979) with the grain size

of the included chromite.

531

700

600

contact through fine fractures in olivine (Roeder &Campbell 1985). If volume diffirsion through the hostmineral were the fundamental cause of the variableCr# or ubiquitous Cr-Al zoning of chromite in theVammala Ni-belt cumulates, this would require thatdiffrrsion had proceeded exactly to the same extent in

olivine, pyroxene and sulfrdes, which is consideredimprobable. To be the ultimate cause of the Cr# varia-tion seen in this study, melt-filled fiactures, in turn,would have to have penetrated to all grains of includedchromite. Furthermore, reaction of included chromitealong such fractures with this melt should perhaps have

olivine I chromite50

I aniph-

FIc. 8. Step-scan analyses across chromite located between unaltered olivine (left) andAl-rich calcic amphibole (rgoeous). Length of the profile and individual steps are180 and 3.6 pm, respectively. Beam diameter was I pm.

T.c

ig"g rti''o%ooj oo

oo++ O O

o+o

o

o Vammala9 Ekojokio Poaionlahti+Mwto

t r t : - - -

J

A l uFd:

532 TIIE CANADIAN MINERALOGIST

pyrftorire I cbrcmitg I pynhotite

Ftc. 9. Step-scan analyses across chromite completelyenclosed by pyrrhotite. Length of the profile and indi-vidual steps are 190 and 3.2 pm, lsspggtitely, with beamdiameter of I pm. Location of profile is indicated inFig.3B.

caused smallo needle- or knob-shaped overgro'fr/ths,which are not observed.

The solubility of aluminum in pyroxenes is knownto decrease as temperature or pressure decreases (e.g.,Herzberg 1978, Gasparik & Newton 1984). Cationexchange between chromite and pyroxenes, inducedfor example by cooling of the inrusions or uplift ofthe crusg may thus decrease the CrlAl of intergraincbromite. This probably explains the presence of veryaluminous spinel in mantle lherzolites that re-equilibrated at a temperature well below the peridotitesolidus @oeder 1994). The opposite effect, however, isexpected for included cbromite because the content anddiffirsivity of Crin olivine generally exceed those of Al(Jwewicz & Watson 1988, Zou & Steele 1993).Furthermore, because also those chromite grainscompletely enclosed by sulfides (which contain someCr but no Al) have similar patlerns of Cr# variation andCr-Al zoning as are present in included and intergrainchromite, such pressure- or temperature-inducedsubsolidus processes cannot explain the variable Cr# ofchromites in the Vammala suite.

Therefore, I believe that the variable Cr# ofincludedand int€rgain chromites should not be interpreted interms of postcumulus or subsolidus processes, butrecords evolving maerna composition during crystalgrowth. Some investigators (Sigurdsson & Schilling1976, Dick & Bullen 1984) have argued that the Cr# of

cbromite on the liquidus in MORB systems variesinversely with pressure and that highly variable Cr# insuch chromite is an expression of its polybaric crystal-lization. Such an interpretation gains some supportfrom experimental work, which suggests that highconfining pressure stabilizes spinel (MgAlrO/ at theexpense of chromite (FeCAO+) (Greenet al. 1971), andfrom the Al-rich composition of chromite in xenolithsfrom the mantle (Haggerty 1979) and lower crust(DeBari et al. 1.987). Because Peltonen (1995b)suggested that cumulates in the Vam*ala Ni-beltcrystallized at mid-crustal pressures, and that theappeanmce of plagioclase was delayed, the possibilityremains that at least some of the variations in Cr# aredue to crystallization of chromite under a range ofpressure conditions.

Recently, Allan et al. (L988) and Roeder &Reynolds (1991) came to the conclusion that meltcomposition is the only significant variable thatcontrols the composition of cbromite at liquidustemperatures, and that a large range in Cr# is an expres-sion of magma mixing or assimilation. There is noevidence for magma mixing in the cumulates of theVammala Ni-belt. The parental magma of the cumu-lates was, howevero contaminated by Svecofenniangraywacke-slate turbidites (Peltonen 1995b). Frac-tional crystallization under variable pressure condi-tions, combined with concomitant assimilafiel sfaluminous sediments, were probably the main causesof the large variations in Cr#.

Origin of Cr-Al zoning

Cr-Al-zoned cbromite as is found in the Vammalasuite is uncortrmon, but has been reported fiom theultramafic Kempirsay inffusion, in the southern Urals(Pustovetov et al. 1993), and from a lherzolite nodulefound in the Ichinomegata crater, Japan (Ozawa 1983).Unfortunately, these authors do not discuss the originof the Cr-Al Tsnlng. Lack of details concerning theigneous and metamolphic history of the Kempinsayintrusion, and the nature of the mineral hosting thechromite, prevents further evaluation of that occur-rence. In the therzolite nodule. ttre chromite is found inaggregates with other Al-bearing phases, pyroxenesand green spinel-bearing symplectite (Ozawa 1983). Itis therefore probable that chromite zoning toward anAl-rich rim in the Ichinomegata nodule is due tometamorphic processes. A similar origin for the Cr-Alzoning of cbromite in the Vammala Ni-belt is unlikely,because patterns sf 2qning are equal for all grains,irrespective of whether they are enclosed by magmaticolivine and pyroxenes, their alteration products orsulfides. ffis 2ening must reflect some process that isindependent of the identity of the enclosing mineral.One such process is deformation (diffrrsion creep),which may induce stess-directed self-diffirsion of Aland Cr (Ozawa 1989). Deformation-induced zoning is,

'tIIt , -t ir i!lI

dt

/ '

. j c r, . I

i li lil

r4-'3

a\- - -

F3*- -1 ) ! . - - - - - - - -

ZONED CHROMITE IN TI{E VAMMAT.A MCKEL BELT 533

however. distinct in the directions of the axes ofprinclpal sftess. Therefore, Al should, dependiqg onaxial orientation, show both increases and decreasestoward the rim of randomly oriented grains ofchromite, which differs from the observed patternof Cr-Al zoning.

Finally, one is left with an interpretation that Cr-Al2ening is a grain-scale expression of the same pro-cesses that are responsible for the grain-to-grain varia-tion in Cr#. ffis 26ning thus records the incrementalcompositional changes of the host magma and providesa grain-scale "memory" of the melt evolution duringthe growth of chromite. Low diffrrsivities of Cr and Alin chromite, in combination with crystal growth in avigorously convecting magmq immediately followedby entapment by olivine, probably prevented furtherinteraction between chromite grains and the magmaand preserved the disequilibrium compositions.

Variation in thz Mg# of included andintergrain chromite

The fact that the chromite in the Yammala Ni-belthas a considerably more Fe-rich composition than whatwould coexist with the adjacent olivine at a magmatictemperature (Fig. 5) indicates that the present Mg# ofthe spinel cores is not purely of igneous origin. Suchlow Mg# and low olivine-spinel blocking temperaturessuggest that the chromite has extensively re-equili-brated with olivine below the solidus. Because the ratesof Mg-Fe2+ interdiffirsion of spinel and olivine arerelatively high (Buening & Buseck 1973, Freer &O'Reilly 1980, Wilson 1982, Lehmann 1983), suchre-equilibration is a common process in slowly cooledigneous bodies (rvine 1965, Wilson 1982, Hatton &von Gruenewaldt 1985, Johnson &Binz 1991, Yang& Seccombe 1993).

That such re-equilibration indeed occurred inthe Vammala Ni-belt cumulates is supported by theMg-Fe2+ diffusion profiles for the olivine side ofthe spinel--olivine grain boundaries @ig. 8). At themomentthe chromite grains were entrapped by olivine,they were in chemical equilibrium. Also at highsubsolidus temperatures, disseminated chrcmile andolivine remain in equilibrium owing to high Mg-Fe2+interdiffusion of both minerals, but at lower tempera-tmes difftrsion is unable to maintain Mg-Fd+ equi-librium firther, and olivine becomes zoned near thechromite. Lack of complementary Mg-Fs2+ diffrrsionprofiles @e decreasing toward olivine) in the chromitesuggest that Mg-Fe2+ interdiffusion in chromiteexceeds that in otvine. The extremely narrow range ofchromite Mg# (at any given Cr#) reflects the limilsdsempositional range of the olivine with which itre-equilibrated.

Those grains of intergrain chromite that are enclosedby clinopyroxene have systematically rather higherMg# (and blocking tempelatures) thau those included

in olivine (Fig. 5). Such a finding suggests thatMg-Fe2+ interdiffrrsion in clinopyroxene lg l6vrs1 rhanthat in olivine, but is similar to that in orthopyroxene(Henry & Medaris 1980, Wilson 1982, Hatton & vonGruenewaldt 1985). The low Mg# of intergrain spinelcompletely enclosed by sulfides is enigmatic (Fig. 5).Overall, composition and zoning are similqr 16 thes..1other textural types of chromiie, indicating that it doesnot have a distinct origin, and that it also re-equili-brated during the subsolidus stage. Divalent iron isobviously available from the pyrrhotite-dominatedsulfides bu! because there are no traces of Mg in thesulfide close to chromite (Fig. 9), the sink for Mg2+ andspecific reactions during the re-equilibration remainunclear.

CoNctustotts

The composition of chromite gains in the ultra-mafic cumulates of the Vammala Ni-belt, southwesternFinland, record the evolution of the magma composi-tion during their crystal growth, as well as the therrnalconditions during subsequent cooling ofthe cumulates.This was rendered possible by the distinct diffrrsionalbehavior of trivalent (Cr, Al) and divalent cations(Mg, Fe2*) within cbromite. The chromite becameshongly Cr-Al-zoned because it remained in disequi-librium with the melt that continuously changed itscomposition owing to fractional crystalliTation andassimilation. This growth-induced zonation was notmodified by subsequent cooling and alteration, imply-ing that the diffrrsion rates of Cr and Al withinchromite must be low. In contrast, the cbromite grainswere close to ideal Mg-Feh exchange equilibriumwith the melt and continued to re-equilibrate withenclosing silicates well below the solidus, owing torelatively high Mg-Feh interdiffirsion in chromite,olivine and pyroxenes, decreasing in this order. Theextent to which this re-equilibration proceeded isamong the most pervasive so far reported, and isprobably a distinct feature.of synorogenic mafic-ultra-maJic intrusions that cooled slowly because of the lowtemperature-gradient between intrusions and theirsurroundings.

AcKtTowrcDcEl,IHvrs

This work was funded by the Geological Survey ofFinland (GSD. Dr. K. Korsman (GSD is thanked forhis continuous support of my Ph.D. research, of whichthis contribution forms a part. Outokumpu FinnminesLtd. is thalked for providing access to their drill-corerepositories, and P. Lamberg (Outokumpu MiningServices) is thanked for running the microprobe andbeing enthusiastically involved at the earhest stagesof this work. This paper benefitted from the thoroughreviews of Drs. P. Roeder and P. Scowen, and theeditorial comments of Dr. R.F. Martin.

534 TIIE CANADIAN MINERALOGIST

REFERENcF.s

Au"AN, J.F, SAcK, R.O. &Berv,l', R. (1988): Cr-rich spinelsas petrogenetic indicators: MORB-type lavas from theLamont seamount chain, eastern Pacific. Am. Mineral.73,741-753.

BrrsoN, M.H. & Jecrsor, E.D. (1969): Chemical composi-tion of altered chromites from the Stillwater Complex,Montana. Arn. Mineral, 54. 1084-1 100.

Blrss, N.W & MecLsAN, W.H. (1975): The paragenesis ofzoned cbromite from central Manitoba. Geochim,Co smochim. Acta 39, 9'7 3 -990.

BUEItr{c, D.K. & BusBc& P.R. (1973): Fe-Mg latrice diffir-siou in olivine. J. Geoplrys. Res. 78, 6852-6862.

BITRIiUARD, D.J.M. (1993): Accessory chromium spinels:their coexistence and alteration in serpentinites. Geochim"Cosmochim. Acta 57, 1297 -1306.

Bunxs, R.G. (1975): Crystal field effects in chromium and itspartitioning in the mantle. Geochim" Cosmochim. Acta39,8s7-864.

CAm.oN, E.N. (1975): Postcumulus and subsolidus equili-bration of chromite and coexisting silicates in the easternBushweld Complex. Geochim. Cosmochirn Acta 39,1021-1033.

DnBenr, S.M. & ComuaN, R.G. (1989): Examination of thedeep levels of an island arc: evidence from the Tonsinaultramafic-mafic assemblage, Tonsina, Alaska. J.Ge ophy s. Re s. 94, 437 3 4391,.

Krv, S.M. & I(av, R.W. (1987): Ultramafic xeno-littrs from the Adagdak Volcano, Adak, Aleutian Islands,Alaska: deformed igneous cumulates from the Moho of anisland arc. J. Geol. 95,329-341.

DICK, H.J.B. & Buuan, T. (1984): Chromian spinel as apetogenetic indicator in abyssal and alpine-type peri-dotites and spatially associated lavar'. Contrib. Mineral.Perrol. 86, 54-76.

DRoop, G.T.R. (1987): A general equation for estimating Fe3*concentations in ferromagnesian silicates and oxidesfrom microprobe analyses, using stoichiometric criteria.M ine ral. Mag. 51, 43 1 435.

DwEK, R.F, Bnonrm.s, S.C. & Scmrrurs, C.M. (1988):Petrogenesis of ultramafic metamorphic rocks from the3800 Ma Isua supracrustal belt, West Greenland. J. Petrol.29. 1353-1397.

EL coREsy, A., PRnrz, M. & Reunonn, P. (1976): Zoning inspinels as an indicator of crystallization histories of marebasalts. Proc. 7th Lanar Sci, Conf.,1261,-12i79,

EvaNs, B.W. & Fnosr, B.R. (1975): Chrome-spinel inprogressive metamorphism - a preliminary analysis.Geochim. Cosmochim- Acta 39. 959-972.

FABRTBs, I. (1979): Spinel-olivine geothermometry in peri-dotites from ultramafic comolexes. Contrib. Mineral.Petrol. 69.329-336.

Fnren, R. & OR.strrv, W. (1980): The dilfrrsion coefficientof Fe3+ ions in spinels with relevance to the process ofmaghemitization. Mineral. Mag. 43, 889-899.

Gaseann<, T. & NEwroN, R.C. (1984): The reversed aluminacontents of orthopfoxene in equilibrium with spinel andforsrcrite in the system MgG-Al2O3-SiO2, Contrib.Mineral. P errol. 85. I 86-196.

Gnrex, D.H., RNGWooD, A.E., WARE, N.G., Hrss$,soN,w.O., MaoR, A. & Iftss, E. (1971): Experimental petrol-ogy and petrogenesis of Apollo 12 basalts. Proc. 2ndLunar Sci. Conf., 601 -6L5.

HAGGERTv, S.E. (1979): Spinels in high pressure regimes, /nThe Mantle Sample: Inclusions in Kimberlite and OtherVolcanics @.R. Boyd & H.O.A. Meyer eds.). Proc. 2ndInt. Kimberlite Conf. 2. American Geophysical Union,Washington D.C. (1 83- I 96).

HArr,r, T.A., VoR!trsro, K. & tl{MvnrsN, E. (1979):Vammala" a nickel deposit in layered ultramafite, south-west Finland. Econ. Geol. 74. 1166-1,182.

Hawyn, P.R. & KsAys, R.R. (1979): Origin of chromitecompositional variation in the Panton Sill, WesternAustralia. Contrib. Mineral. Petol. 69. 7 5-82.

HArroN, C.J. & vor,l GnruNrwau, G. (1985): Chromitefrom the Swartkop chrome mine - a:r estimate of theeffects of subsolidus reequilibration. Econ. Geol. 80,91t-924.

HsNny, D.J. & MEDARIS, L.G., JR. (1980): Application ofpyroxene and olivine-spinel geothermometers to spinelperidotites in southwestem Oregon. Am, J. Sci. ?'f,0-4,2 l l -231.

HEnznanc, C.T. (1978): Pyroxene geothermometry and geo-baxometT: experimental and thermodynamic evaluationof some subsolidus phase relations ilvelving pyroxenes inthe system CaO-MgO-Al2Or-$iOt Geochim. Cosmn-chim Acta 42.945-957.

I{u, R. & Romm, P. (1974): The crystallization of spinelfrom basaltic liquid as a function of oxygen fugacity. ./.Geol.4,709-'729.

IRvD{E, T.N. (1965): Chromian spinel as a petrogenetic indi-cator. 1. T\eory. Can. J. Eanh Sci.2,648-672.

- (1967): Chromian spinel as a petrogenetic indicator.2. Petrologic applications. Can. J. Eanh Sci.4,7l-L03.

JAN, M.Q. & WtNoLey, B.F. (1990): Chromian spinel - sili-cate chemistry in ultramafic rocks of the Jijal complex,nortlrwest Pakistan. J. Petol. 31. 667:715.

ZONED CHROMITE IN TI{E VAMMAI.A, MCKEL BELT 535

JoHxsoN, C.A. & PRN'4 M. (1991): Chomite and olivine intype tr chondrules in carbonaceous and ordinary chon-drites: implications fol thermal histories and groupdifferences. Geochim. Cosmochim. Acta 55. 893-904.

Junswcz, A.J.G. & WArsoN, E.B. (1988): Cations in olivine.2. Difftrsion in olivine xenocrysts, with applications topetrology and mineral physics. Contrib. Mineral. Petrol.w,186-20L.

Knranmsres, P.K., Tevron, R.N. & Tallaxa, H. (1993):Geochemistry of plutonic spinels from tle NorthKamchatka Arc: comparison with spinels from othertectonic settings . Mineral. Mag. {1,575-589.

Kn/aAt-t,, ICL. (1990): Effects of hydrothermal alteration onthe compositions of chromian spinels. Contrib. MiruraLPetrol. 105.337-346.

KusnBo, I. & YoDm., H.S., Jn. (1966): Anorthite-foatedteand anorthite-<nstatits reactions and their bearing on thebasalt-+clogite transformation. J. Petrol. 7, 337 -362,

LsttrtarN, J. (1983): Diffrrsion between olivine and spinel:application to geothermometry. Earth Planct. Sci. Izft, il,123-1,38.

Mrner.rs, L.G., Jn. (1975): Coexisting spinel and silicates inalpine peridotites of the granulite facies. Geochim.Cosmochim. Acta 39. 947 -958,

Mrcnuos, K.M. (1990): Zoned chromites witl high Mn-contents in the Fe-Ni-Cr-laterite ore deposits from theEdessa area in northem Gte*ce, Mineral, Deposita ?5,t90-t97.

OTIawLsr, D.S., CnRNosKy, J.V., Jn. & Wrcrs, F.J. (1989):The stability of lizardite and chrysotile. Can. Mineral. tl,483493.

Oznwe, K. (1983): Evaluation of olivine-spinel geothermo-metry arl an indicator of thermal history for peridotites.Contrib. MineraL. Petrol. 82, 52-65.

(1989): Stess-induced A1{r zoning of spinel indeformed peridotttex. Narure 338, 1,41,-1,44.

PELroNEl, P. (1995a): Magna - country rock interaction andtlle genesis of Ni-Cu deposits in the Vammala NickelBelt, Sw Finland. Mineral. Petrol. 52, 1,-24,

(1995b): Petrogenesis of ultramafic rocks in theVammala Nickel Belt: implications for crustal evolutionof the early hoterozoic Svecofenrdal arc tenare. Lithasu.2$-n4.

Pusrovrrov, A.A., MrrD{A, YE.A., UKHANoV, A.V.,Nmol'srava, N.YE. & Snqnr, V.G. (1993): hhomo-geneity in accessory cbrome spinel as a geothermometer.Geochem. Int. 30(5), 3 1, 41.

RoEDER, P.L. (1994): Cbromite: from the fiery rain ofchondrules to the Kilauea Iki lava lake. Can. Mirrcral.32"729JM.

& Cal,rBFrr, I.H. (1985): The effect of post-cumulus reactions on composition of chrome-spinels fromthe Jimberlana intrusion, J. Petrol. 26, 763-786.

- & REyNoLDs, I. (1991): Crystallization of chromiteand chromium solubility in basaltic melts. .L Petrol.32,909-934.

SACK, R.O. & GltroRso, M.S. (1991): Chromite as a pefioge-netic indicator. In Oxide Minerals @.H. Lindsley, ed.).Rev. Mineral. 25, 3'23-353.

ScownN, P.A.H., RoEDEn, P.L. & Hstz, R.T. (1991):Reequilibration of chromite within Kilauea Iki lava lake,Havtajj. Contrib. Mincral. Petrol. \M, 8-2O.

SrfiN, PuyAN, HWANG, SIryH-LL,Nc, CHU, HAo-Tsu & Jwc,Ruev-CHANc (1988): STEM study of 'Terrirchromit'' fromthe Heng-Chun chromitite. Ar4?. Minzral.73, 383-388.

SrcuRDssoN, H. & Scu-uqc, J.-G. (1976): Spinels in Mid-Atlantic Ridge basalts: chemistry and ogcwrence. EarthPlanet. Sci. Iztt. 29. 7 -20.

Snr,roNw, A. (1980): The hecambrian of Finland. Geol. Surv.Finland. Bull.3M.

SpeNonNssnc, K. (1943): Die Cbromerzlagerstiitte vonTampadel amTnbten. Z. Prald. Geol. 51, 13-35.

SurroN, S.R., IoNFs, K.W., GonooN, B., Rrwns, M.L., Ba:r,S. & Srvnnr, J.V. (1993): Reduced cbromium in olivinegrains from lunar basalt 15555: X-ray absorption nearedge structure CXANES). Geochim. Cosmochim. Acta 57,461468.

Wnsou, A.H. (1982): The geology of the Great 'Dyke",

Zimbabwe: the ultramafic rocks. J. Petrol.23,2N-292.

Wvr,n, A.G., Car.roele, P.A. & Buue, T.M. (1987):Compositional zoning in unusual Zn-rich chromite fromthe Sykesville district of Maryland and its bearing on theorigin of 'ferritchromif'. Am Mineral. 72, 413422.

YANc, KAI & SEccovmq P.K. (1993): Chemical variation ofchromite in the ultramafic cumulates of the GreatSerpentinite belt, Upper Bingara to Doonba New SouthWales. Australia , Can Mineral. 31,75-87 .

Znou, Mu-Fu & IGnrucq R. (1992): Morphology and com-position of chromite in komatiites from the BelingweGreenstone BelL Zmbabwe. Can. Mineral.30, 303-317.

7nou,Y. & Srm.e, I.M. (1993): Chemical zoning and diffir-sion of Ca, Al, Mn, and Cr in olivine of Springwater pall-asite. Proc. 24th Ianar Sci. Conf.,1573-L574.

Received. April 6, 1994, revised manuscript acceptedSeptember 20, 1994.