You are on page 1of 11

AmericanMineralogist, Volume6j, pages970-980,1978

of lithium-richpegmatites
Petrogenesis

Devto B. Srnwenr
U. S. GeologicalSuruey
Reston,Virginia22092

Abstract

The experimentally determined liquidus of the system NaAlSiBO8(Ab)-SiOdQz)-


LiAlSiO.(Eu)-HzO at 2 kbar PHrO hasa eutecticnearAb34Qz50Eul6 weightpercentat640"
+ 10'C. The field boundaryol albite and quartzis depressed nearly l00oC from the Ab-Qz-
HrO sidelineby saturationin Eu component;the field boundarybetweenalbiteand lithium
mineralsvarieslittle in LirO content(19 + 3 percentEu) betweenthe eutecticand the Ab-Eu-
HrO sideline.Liquidustemperatures rise very steeplyas Qz increases in the quartzfield.The
field boundariesindicate that pegmatitemagma wiil crystallizefeldsparand quartz and
concentrate lithiumin the remainingmagma.Similarrelationshavebeenobserved for theOr-
Qz-Eu-HrO systemby Munoz (1971).Petalitewill form if the field boundaryis reachedat
high temperaturesand low pressures; spodumeneforms at lower temperaturesand higher
pressures.
The bulk compositionsof natural homogeneous Li-rich pegmatitesplot on a diagramfor
the system(Or,Ab)-Qz-Eu-HrOin the thermalminimumor slightlytoward(Or,Ab) from it.
The geologicabundanceand similaritiesof thesebulk compositionstogetherwith the experi-
mental data supporta magmaticorigin of homogeneous Li-rich pegmatites.Li-rich magma
probablyformsby partialmeltingof Li-bearingmetasediments at temperatures 75oCor more
belowthe minimum meltingtemperatureat the samePH,O of the simplified"granite" system
KAlSigOr-NaAlSirOE-SiOr-HrO. An origin by anatexisis mostconsistent with the irregular
distributionof Li-rich pegmatitescloseto the sillimaniteisogradbut farthestof all pegmatite
varietiesfrom granitesrelatedto the samethermalevent.The sparsityof Li-rich pegmatite
bodieswithin relatedgranitesis also in accordwith an anatecticorigin: further heatingin the
sourcearea would yield larger amounts of granitic rock and dilute the Li content to the
n o r m a l( 1 0 0 p p m .
Magmaticdifferentiation adequately explainsthe sequence of zonesin zonedpegmatites,
includingthe abruptappearance of lithiummineralsin quantityandthe observed limit on the
abundanceof lithium mineralsin common natural bulk compositions. The occurrence of
petaliteor spodumenecannot be relatedto bulk composition,as all Li-pegmatitefalls
compositionally into the alkali feldspar-quartz-petalite volume.The high temperature-low
pressureassemblage isochemicallyequivalentto zone5 of Cameronet al. (1949)is perthite-
plagioclase-petalite-quartz, and to zone6 is plagioclase-petalite-quartz. Thesezonesseemto
bethe lastthat canreasonablybe interpretgdto haveoriginatedfrom moltensilicatemagma.
Gasesmore siliceous than coexistinglithium aluminumsilicatemineralswereexperimen-
tally studiedat 575'C and 2 kbar PHrO. Precipitation of the solidsfrom suchgaseswould
yield compositions like thoseof silica-richpegmatitezones.Compositions that could have
formedfrom gashavequartz ) feldspar.Eucryptiteprobablyforms solelyin the presence of
suchgases.

Introduction percentat most of the massor numberof pegmatites


Geologicalmappinghas proved the existenceof present.The Li-rich pegmatitesoccur immediately
regionalcompositionalzoningof pegmatitesadjacent adjacent to or in sillimanite-grade metamorphic
to relatedgranitic bodies(Mulligan, 1962,p. 420; rocks, but not within related granites.If Li-rich
Norton, 1975,p. 135-140).Li-rich pegmatites
occur magmaformedthe Li-rich pegmatites, it could origi-
farthestfrom the granite,and constituteonly one nate by anatexisof Li-bearingmetasediments or by
0003-004x/ 78/ 09| 0-0970$02.00 970
STEWART: LITHIUM.RICH PEGMATITES 971

very extensive differentiation of granitic magma. givenby Stewart(1960,p. 28) and Munoz(1971).A
Deposition from gasesemanating from the crystalliz- questionspecificallyaddressedin this paper will be
ing magma has also been hypothesized.In this paper, whetheror not lithium-richzonesarethe lastto have
geological evidencewill be combined with data from a compositionthat couldhaveformedfrom a magma
experimental petrology to elucidate the processesin- or are the first to form largelyfrom gas.
volved in the petrogenesisof Li-rich pegmatite. Clearly the magma that formed each assemblage
Pegmatitesmay be either homogeneousor inter- neednot havethe samebulk compositionasthe mass
nally zoned. The sequenceof I I mineral assemblages of crystalsprecipitatedto form that particular zone.
in zoned pegmatites reported by Cameron et al. The crystalscould haveaccumulated from a melt of
(1949, p. 6l) is still regarded as essentiallycorrect. muchdifferentcomposition, and not all the phasesat
The minerals of these assemblagesare generally a single place necessarilyprecipitatedat the same
named in decreasingorder of the abundance of the tlme.
minerals,and the sequenceof the assemblages is from
the contact with the country rock to the center or
The compositionof lithium-richpegmatiteandthe dis-
core of the pegmatite. Not all assemblagesoccur in tributionof lithium mineralsin it
all pegmatites or even in all pegmatite districts. Crys- Numerousestimatesof the bulk compositionsof
tallization proceeded from the country rock contact individual pegmatiteshave becomeavailablesince
inward becauseno outer assemblagecuts an inner the late 1940's.Estimatesfor simplegraniticpegma-
assemblage,but inner assemblagesdo cut outer as- tites are most numerous,and a few tensof estimates
semblages. The first four assemblagesare feldspar- are for the bulk compositions of lithium-richpegma-
rich and differ from granite only in the ratio of one tites, many of which are homogeneous. Only a few
feldspar to another and in the abundance of mus- estimates are available for the bulk compositions of
covite. Where muscovite is abundant, potassic feld- zonedpegmatites and for their individualzones.Esti-
spar (perthite) is ordinarily absent or sparse.These matesarereportedin variousterms,suchasvisualor
first assemblagesmay be richer in AlrOr than many measuredmodes,weighedmineralseparates, or par-
granites and may have reactedwith country rock or tial or completechemicalanalyses.Thesedifferent
lost alkalies to it, but the data are skimpy. Lithium kinds of estimatesare not easilycompared.In several
minerals do not become a major component until instancesthe tonnagesand compositions of individ-
assemblage5, but persistthrough assemblages 5 to 8. ual zoneshavebeenestimatedto calculatethe bulk
The content of quartz increasesin theseassemblages, compositions. Suchestimates are of particularinter-
and quartz first becomesdominant in assemblage7. est becausethey permit important portions of the
The core commonly is almost entirely quartz. processof differentiationto be discussed, although
The phasepetrology of granitic magmas is reason- the reliabilityof the estimatetendsto decrease asthe
ably well understood becauseof an extensivecollec- structuralcomplexityincreases.
tion of geological and experimental evidence (re- An estimateof the standarderrorof theestimateof
viewed in Luth, 1976). Granitic magma contains the LirO contentof homogeneous lithium-richpeg-
more normative feldsparcomponentsthan normative matite can be obtainedfrom the exemplarydescrip-
quartz over the whole range of its probable composi- tion of the very largeKings Mountain,NC, deposits
tion. Silica (quartz) in the presenceof steam always given by Kesler(1961)1.Assaysfor LirO from 226
melts at a temperature hundreds of degreeshigher core samples,mostly 20 feet long, representative of
than the melting temperatureof the mixtures of feld- many million tons of pegmatiteare givenin his Fig-
spar componentsand silica found in granitic magma, ure 6. A plot of theseanalyses showsa normaldistri-
which suggeststhat either the quartz core did not bution and only slight skewness toward low values.
form from a magma or the magma contained com- The meanof the array is 1.53weightpercentLirO,
ponents that were depositedelsewherein the pegma- and the standarddeviation,o, is 0.32percentLirO.
tite, possibly in the mica-rich zones bordering the Kesler(1961,Table 3) also gave weightedaverage
core. A thorough discussionof the complexities of LirO contentsfor analysesfrom 14 crosssections.
the crystallizalion of hydrous pegmatitic magma, The meanof the data for thesecrosssectionsis 1.54
with emphasis on textural development,is given by weightpercentLirO, and o : 0.15percentLirO, or
.Jahns and Burnham (1969). Subsolidusreactionsare
also common in pegmatites,and particular attention ' The distributedreprint of this articlecontainsadditionalmate-
to this phenomenon in lithium-rich pegmatite was riaf in Fies.4 and 6.
972 STEWART: LITHIUM-RICH PEGMATITES

q
50 percentas comparedto the theoretical8.04 percent.
340 Usingthis information,which is rarelyavailable,we
bgo can convertthe standarddeviationin LirO to that of
E.^
uLv
spodumene, 4.2percentspodumenein singlesamples,
Ern
=,u
z. or 2.1 percentspodumene in groupedsamples. Com-
U
12 16 20 24 parabledatafor othermineralsno doubtwouldshow
% SPODUMENE standarddeviationsas large.The standarddeviation
to be attachedto estimateslesswell founded can be
036S 12 15 18 21
% Li20 larger still, a factor to rememberin the comparisons
ANATYSES
GROUPED amongpegmatites below.The topic is alsodiscussed
SINGLE
ANALYSES
by Norton and Page(1956)and Norton (1970).
Kesler (1961,Fig. 2 and p. 1066)reportedthat
Fig. l. Distribution of spodumene (weight percent) in 168 quartzmonzonitepegmatitiesalsooccurredat Kings
pegmatite samples from Kings Mountain (Lovering, 1958) Mountain and that, where thesepegmatitesadjoin
compared to the mean and + two standard deviations of single
spodumene-rich pegmatite,their qvartzand feldspars
analyses and grouped analyses from data given by Kesler (1961).
The bimodal distribution indicate that most spodumene
are coextensiveand presumablyare contempo-
pegmatites contain l6 to 24 percent spodumene, and a second raneous.
group of pegmatites contains little or no spodumene. Modal data for 168samplesof homogeneous peg-
matite in the Kings Mountain district gatheredby
W. R. Griffittswereusedby T. G. Lovering(1958)as
half the standarddeviationfor singleanalyses.Pre- a geologicexampleof a bimodaldistribution,Figure
sumablythe bestapproximationof the standardde- l. By combiningthe aboveestimates of spodumene
viation would be that from the groupedanalysesfor abundanceand 95 percentconfidencelimits asshown
the crosssections,and thus the 95 percentconfidence by the figure,the existence of two typesof homoge-
limits (2o) for the LirO content of this pegmatite neouspegmatiteis shown.One type has a consid-
masscould be 1.23and 1.85percentLirO. Kesler erableamountof spodumene (-20 percent),and the
reportedthe weightedaveragegradeof all ore milled otherhaspracticallynone.In additionto information
from 1954to 1960was 1.56percentLi2O. concerningthe reliability of various estimatesof
Kesler(1961,Table l) accountedfor the distribu- spodumenecontent,someexplanationis apparently
tion of LirO among the principal minerals;97.6 neededfor the appearanceof so much spodumene,
weightpercentof the LirO waspresentin spodumene, just asis observed within zonedpegmatites. If spodu-
which had an averageLirO contentof 7.46 weight mene pegmatitefor economicor other reasonsre-

Table l. Modes of spodumene pegmatites rounded to nearest percent

Spodwene Quartz t Alkali Muscovite Total


Iocality md Reference (wt. c) (wr. r) Feldspds (wr. r) (wt. t)
(wt. 8)

Kings Moutain, NC 20 32 4L 6 99
Kesler (1961)

Kings Mountain, NC 18 33 44 5 ro0


Browning, Clemons, Mcvay (1961)

Val DrOr, Quebec 2L 4L .32 6 100


Browning, Clmons, Mcvay (1961)

Peg claims, Maine 22 27 44 7 100


SundeLius (1963)

Beardmore , Ontdio 20 30 45 5 100


Milne (1962)

Mateen Pegmtite, Hill City, SD IO 38 Jb I 98


Broming, Clmons, McVay (1961)
STEWART: LITHIUM.RICH PEGM ATITES 9't3

Table 2. Modes*, in percent,Etta pegmatitefrom Norton et al. (1964,Table 5)

Unit Ai-bite Perthite and Spodwene Muscovrte Biotite Tournaline


Quartz
microcline

Microcline-biotite pegmatite 10

quartz{uscovite-albite
Pr9!tr4urLE 60 15

Perthite-qudtz-spodmene
PegileurLE 25 10

Quartz-cLeavelmdite-
spodmene pegmatite 35 35 I 3

Quutz-spodmene pegmatite 7O I 3 25 I

Quartz pegmtite 99 <1 <I <l <]

t By visual estimates of surface exposures, supplmented by measurements of spoduene content.

quired a grade of about 35 percentto be minable, entiationis not readilyinterpretable. The bulk com-
sucha grade would be most unlikely to be attained positions of such zoned pegmatites commonly con-
overany considerable volume of rock in this district. tain half or less
the LizOcontentof the homogeneous
Estimates of the modesof representative homoge- Li-rich pegmatites,eventhough somezoneshavethe
neous or nearly homogeneousLi-rich pegmatites samecompositionas homogeneous Li pegmatites.
from other areasare givenin Table l, alongwith the Modal data can be utilizedwith chemicaldata for
data for Kings Mountain.In general,thesecomposi- eachmineralpresentto yield an estimatedchemical
tions also containtracesof other minerals,most no- composition.Table 3 showsthat the calculatedbulk
tably 0.3-0.4 weight percentberyl. In view of the compositionsof homogeneous spodumenepegma-
accuracyof suchestimates,thereare enoughsimilar- tites and of somepetalitepegmatites are similar.In
ities to suggestthat somethingakin to a eutectic most compositionsNarO ) KrO. The exceptionin
compositionis involved. the table,Beardmore,may not be seriousin view of
The quartz-clevelandite-spodumene zoneof a fa- the imprecisionof suchestimates. Li-rich pegmatite
mousand structurallysimplezonedlithium pegma- containsabout 2 weight percentmore AlrO, than
tite, the Etta spodumenemine, SD, (Table2; from Nockolds'(1954)averagealkaligraniteand approxi-
Norton et al., 1964,p. 316)is the only spodumene- matelythe 2 weightpercentlesscombinedNazOand
bearingzonethat approximates the compositionof KzO. The narrow rangeof LizO contentsis unlikely
homogeneousLi-rich pegmatite.Inner zones are to be fortuitous. The appearanceof spodumene
quartz-rich,and outer zonesare rich in feldspar, rather than petalite seemsnot to be a function of
mica, or quartz,and the courseof magmaticdiffer- composition,but the two mineralsseemantithetic.

T a b l e 3 . M a j o r c h e m i c a l c o m p o n e n t so f s o m e L i - r i c h p e g m a t i t e s

with Petalite
Comlpnent
weight per cent Kings Mtn., NC Hirvikallio, Finl.

sio2 74.6 73.5 76.O 73 73.7

Nzog L6.4 L6.7 15.9 16 1 6. 9

Na2O 3.3 3.4 2.9 4.5 3.r


K2o 2.9

Li20 I.5 I-4 1.6 1.5 1.8

E 5 components 94.2 97.6 oo? 9'l .o 9 8 .4


974 STEI4/ART : LIT H I UM-RIC H PEGM ATITE S

0uarE
Feldspar+quarE+
LiAlsilicare:10070
2% Li20
Spodumenepegmatites
^ composition 1 8% ti20
of zone
. bulkcomposition

Petalitepegmatites
. bulkcomposition

Spodumene
LiAtsi206
5 800
E
= .+ 1 . . ' : .:' . : : . ' r ',' i i . : r j
t sp(d u m e n e s s . , at.,,.,|
I
l. . ,;:'ri..l pSpodumeness
Feldspar ------r[-.-.'.,.,.- i.l p.u.,ip,i,.
{ab+or)
WEIGH]
PEBCENT LiAtsi04

Fig. 2. A plot of the compositions of Li-rich pegmatites ss+


f Ouanz )tarire| *o,,Ntr'
recalculated to 100 weight percent combined alkali feldspars,
B spodumene
ss
q u a r t z , a n d e u c r y p t i t e ,i g n o r i n g m i c a a n d o t h e r m i n o r a m o u n t s o f
0uartzandpetalite )etalite+spodumenl
minerals present. The bulk compositions of homogeneous
spodumene pegmatites (Table I and other sources) are shown by
d o t s . T h e b u l k c o m p o s i t i o n s o f p e t a l i t e p e g m a t i t e s( T a b l e 3 a n d Spodumene
o t h e r s o u r c e s )a r e s h o w n b y s q u a r e s ,t h a t f o r B i k i t a b e i n g m a r k e d andquarE
Spodumene an0
w i t h a ' B ' . C r o s s e sa r e u s e d t o m a r k t h e c o m p o s i t i o n so f i n d i v i d u a l eucryptite
zones from zoned pegmatites, the suite from the Etta spodumene
4ooL
mine (Table 2) being connected by short dashes. The 95 percent 40.'
si02 Sp LiAtsi04
c o n f i d e n c el i m i t s f o r L i r O i n g r o u p e d a n a l y s e so f K i n g s M o u n t a i n
spodumene pegmatites are shown by labeled lines. All Li-rich Fig. 3. Tentative T-X phase diagram for the system Qz-Eu-
p e g m a t i t ec o m p o s i t i o n s ,w h e t h e r h o m o g e n e o u so r n o t , f a l l i n t h e HrO at 2 kbar PHrO Only phase relations below 840"C have been
triangle feldspar-quartz-petalite. i n v e s t i g a t e de x p e r i m e n t a l l y .A l l n a t u r a l p e g m a t i t e c o m p o s i t i o n s
project onto this sideline between Qz and Pe compositions. Note
the limited temperature interval for the stability of petalite.

The modal and chemicalinformationcan be use- pegmatitesfall within theselimits. Zonesof several
fully compared(Fig.2) by recalculation to 100weight zonedpegmatitescorrespondin compositionto ho-
percentof combinedfeldspars,quartz,and a lithium- mogeneous lithium-rich pegmatites, but many
bearing end member,arbitrarily calculatedas eu- strongly zoned pegmatitescontain very siliceous
cryptite,LiAlsiol. Mica has beenignored.Figure2 zones(commonlyfracturefillingsor smallzones)that
showsthat all bulk andzonecompositions fall within alsocontainlithium minerals.The sequence of inner
the trianglefeldspar-quartz-petalite, demonstrating zonesin the Etta pegmatiteshows both types of
againthat the occurrence of petaliteor spodumene is zones.
not controlledby compositionaldifferences. Almost The foregoingchemicalevidencecan be summa-
all bulk compositionscontain more feldsparthan rized briefly beforeseekingexplanationsfrom experi-
quattz, but compositionsof zonesscatterwidely, in- mentalpetrology:
cludingsomevery silica-richcompositions whichare l Many Li-rich pegmatites havealmostthe same
the only virtually monomineralic zonesfound.Con- composition.Theycontainmorefeldsparthanquartz
toursof LirO contentwill be parallelto the feldspar- and approximately1.5weightpercentLi2O.
quartzsidelinebecausethe Li contentoffeldspar and 2. The occurrenceof particularlithium mineralsis
quartzis negligible. The 95 percentconfidence limits not controlledby differencesof bulk composition,
established earlierfor groupedsamplesof the homo- becauseall compositions fall in a narrow range
geneousKings Mountain pegmatiteare shown,and within the trianglefeldspar-quartz-petalite.
almost all the bulk compositionsfor homogeneous 3. Somemechanism existsthat can producea bi-
STEWA RT: LIT H I UM.RICH PEGM A TITES 975

0z 0uarE
Jr(,
si02
11 3 0 Ratioof solidsin gas
575'c
Prrs=2Kb

/z' n
/- Petalite gas Spodumene
+ spodumene+

gas
+ eucryptite+
SDdurnene
Gas
nroffi si% m%
s35E- \-1100
Ab Eu PERCENT
WEIGHT LiASiq'
NaAlSi30s PERCENT
WEIGHT LiAtsi04
Fig. 5. Schematicrepresentationof the compositionsof the
Fig. 4. Liquidus of the systemAb-Qz-Eu-HrO at PH2O= )
gasesthat coexistwith lithium mineralsor quartzat 575oand PH,q
kbar.Temperature in oC. = 2 kbar. The gasesare more siliceousthan the coexistinglithium
minerals.For naturalcompositionsthat projectto betweenquartz
and petalitecomposition,the ratio of solidsthat could precipitate
modal distributionof compositionsof homogeneous from the gasis shown.This compositionis like that found in some
pegmatites amountof lithium complexly-zoned
suchthat a substantial pegmatites(Fig. 2).

mineral(s)appearswhereany is present,and yet the


total amountthat can be presentis limited.
4. The sequence of zonesfrom country rock in- componentalongwith LiAlSiOl (Eu) to describe the
ward appearsto be from plagioclase-quartz-mus- principal lithium mineralsexcept lithium micas.Con-
covite compositionsthrough compositionswith K- siderationof the excessAl2Osrequiredto form mus-
feldsparsto lithium-richcompositions and finally to covite is omitted to avoid anticipatedcomplexities
silica-richcompositions. The silica-richcompositions (Luth, 1976,p. 369-372).Shaw (1963)studiedthe
are distinctive,may contain lithium minerals, and four-phasecurve K-feldspar-quartz-liquid-gasand
must be accountedfor. determinedthat only smallamountsof AlrOg(equiv-
5. Someinternalzonesin zonedpegmatites corre- alent to about 3 percentmuscovite)could be dis-
spond to lithium-richhomogeneous pegmatite,but solvedin the silicateliquid. This amount of Al2Os
many zoned pegmatitescontain silica-rich inner depressed the liquidus20 to 30oC at 2 kbat (U. S'
zonesthat lackequivalents amonghomogeneous peg- GeologicalSurvey,1961,P.74).
matites. Either higherpressures or the presenceof Ab com-
ponent causes greater solution of AlzOa.The peg-
ExperimentalpetrologY matites tabulated in Table I contain 5 to 8 weight
Becausepegmatitic lithium minerals occur with percentmuscovite.J. J. Norton (personalcommuni-
qvartz,alkali feldspars,micas,and little else,part of cation, 1978)found a medianof 8 weight percent
the systemHrO-LirO-KzO-NazO-AlrOs-SiO,would muscovitein pegmatiticHarney Peak granite, SD.
describesuchassemblages. However,simplifyingas- Huang and Wyllie (1973,p.3) produceda homoge-
sumptionsare neededto makean experimental study neous melt from a sampleof Harney Peak granite
practical.Sodicfeldsparis almosttwiceas abundant containing 13.8 weight percent muscovite (and
as potassicfeldspar in Li-rich pegmatites, and the plagioclase with An 5.1),and found that its solidus
selectionof NaAlSisO(Ab) as the feldspar com- with excess HrO was identical within the limits of
ponent avoids the complexity caused by the appear- experimental error with that for the systemAb-Or-
anceof lithium mica (Munoz, l97l) and of leucite. Qz-HzO.
The CaAlzSirOdAn) contentof naturalalbiteis I to 3 Liquidus temperaturesfor the systemAb-Qz-Eu-
percentand has beenignored.SiOz(Qz) is taken as a HzO would also be lowered by the addition of
9'16 STE I4TART: LITH I UM. RIC H PEGM ATITE S

KAlSioOs(Or) component.Further depressions of not seriousand may well reflectminor differencesin


35-70"C are anticipateddependingon the Or con- composition(particularlyof tracesof FeOa) of the
tent, as estimatedby comparisonof the temperature startingmaterialsor in lengthof runs.Experimental
of the four-phaseassemblagealbite-quartz-liquid- runs,thermodynamic calculations,and observed nat-
g a s w i t h t h e a l k a l i f e l d s p a r - q u a r t z - l i q u i d - g a s ural pseudomorphs of spodumeneand quartz after
boundary curve in the systemOr-Ab-Qz-HrO at petalite indicate that at low temperatureand high
constantHrO pressure (TuttleandBowen,1958)and pressurethe assemblage spodumeneplus quartz is
Ab:Or x 2'.1. more stable than petalite.An experimentally-deter-
A constantHrO pressureof 2 kbar waschosenfor mined P-T diagramfor the Pe compositionis given
convenienceand becausealkali feldspar in Li-rich in eernj' and Ferguson(1972,p. 673).
pegmatitescrystallizesas two separateminerals,im- The four-phasepoint (Fig. a) on the sidelineAb-
plying substantialdepressionof liquidus temper- Qz-HrO was redetermined to be at 735 A 5'C and
aturesby moderateto high HrO pressure.All experi- Ab62Qz38 composition by using Schairer and
mentswere performedat saturationwith HrO, but Bowen's(1956,Table2) glasses N, M, LM, K, and
this conditionmay not occurthroughoutcrystalliza- 162 by courtesyof J. F. Schairer.Two additional
tion in nature. glasseswere prepared by Schairer's methods at
The Eu-Ab-HrO sidelinewasreportedby Stewart Ab60Qz40and Ab65Qz35.The positionof the four-
(1960),alongwith a descriptionof the reagentsused phasepoint is at the samecompositionfound by
in the presentwork, preparationof startingmaterials, Tuttleand Bowen(1958,p. 53) but is at aboutl5.C
and otherexperimental details.The albitesynthesized lower temperature,a trivial difference.
did not show the same (l3l-131) separations as Portionsof the liquidusof the systemAb-Qz-Eu-
would have beenexpectedfor pure NaAlSirO, under HrO at 2 kbar (Fig. a) weredeterminedby using ten
the sameexperimental conditions,as confirmedand glassesof ternary composition. Stability fields for
extendedby Edgar and Piotrowski(1967),so some albite, quartz,petalite,eucryptite,B-spodumene solid
difference in alkali contentor Al: Si ratio apparently solution,and B-eucryptitesolid solutionappearon
existsin the disordered syntheticfeldsparthat hasnot the liquidus,but spodumene was not synthesized in
beenfound in natural low albitefrom Li-rich pegma- the presence of silicateliquid.The liquidusis charac-
tites. terizedby a eutecticnearAb34Qz50Eul6at 640" t
The phaserelationson theQz-Eu-HrO sidelineare l0'C. This compositionfalls in the triangleAb-Qz-
complexand incompletely understood. The approxi- Pe and correspondsto Ab34Qz27.lPe38.9, or
mation given in Figure 3 represents current under- Ab34Q242.4Sp23.6. The field boundary betweenal-
standing,but metastability is notoriouswith someof bite and lithium mineralsfalls steadilyas Qz in-
thesephases,and determinationof the composition creases, from 725"Con the Ab-Eu sidelineto about
of the extensivehigh-temperature solid solutionsis 715'C where it crossesthe Ab-Sp join and 660.C
dependenton graphsto relateprecisemeasurements whereit crosses the Ab-Pejoin, but the LirO content
of unit-cellparameters to composition. Graphsdeter- varieslittle (Eu : 19 + 3 percentor LirO : 2.2 I
mined by B. J. Skinner,H. T. Evans,W. C. phinney 0.35weightpercent).The liquidusat the Ab-Qz side-
and me for 21 compositions wereusedto determine line is depressed nearly 100"C by saturationwith
compositions (Skinnerand Evans,1960,and personal lithium component,and the proportionof quartzto
communications). A comprehensive study of phase feldsparis increasedby 60 percent.Liquidustemper-
relationsfor LiAlSirO. given by Munoz (1969)is aturesrise very steeplyas Qz increases in the quartz
illustrativeof the difficultsynthesis of spodumene at field.
low pressuresexperiencedby everyonewho has at- The liquidus determinedfor the systemAb-Qz-
temptedto work with this system(Shternbergel a/., Eu-HrO has many featuresin common with that
1972; Edgar, 1968). The high-temperaturephases deduced by Munoz(1971,p.2078) for thesystem Or-
havenot yet beenfound in nature,however.and the Qz-Eu-HrO. The last liquids in both systemsat 2
discrepancies betweenthe reportedpositionsfor the kbar disappearwith the crystallization of alkalifeld-
a-B spodumene transitionare not critical.Although spar, petalite,and quartz. Munoz estimatedthe eu-
the reactionLiAlSioolo: LiAISLO. + 2SiO,(petalite tectic temperaturein the potassicsystemto be be-
: spodumene-f 2 quartz) hasbeenreportedat tem- tween 600 and 625oC,
only slightlylower than that
peraturesthat differ by tensof degreesat 2 kbar by found here for the sodicsystem.When both alkali
Stewart(1963)and Munoz (1971),this discrepancy is feldsparsare presentthe temperaturewill be further
STEWART: LITHIUM-RICH PEGMATITES 977

depressed by 35'C or more,but accordingto Figure3 The liquidus diagram implies a more complex
of this work and Munoz' Figure5, petaliteis still the processthan magmaticcrystallization for the silica-
lithium mineral on the liquidusuntil temperatures rich zones,which would requirevery much higher
below -550oC are attained.Suchlow temperatures temperatures to be completelymoltenthananyof the
may not be reachedby the liquidus at 2 kbar evenif zoneswhich precedethem in formation.The siiica-
all of the additionalAlzos, BeO, and other com- rich zonesarefoundonly in complexly-zoned pegma-
ponents known in nature are present.The field tites,aresmallin volume to
relative the whole pegma-
boundaryin the presence of additionalcomponents tite, and occur as inner zones or fracture fillings. It
may shift position,but no estimatecan be madenow will be arguedbelow that these compositions were
of the amount. depositedfrom gas (steam)that was in equilibrium
Thefieldboundaryof the lithium-richphases limits with crystallineor partly crystallineassemblages.
to about 2.2 weightpercentthe amount of LizO that
can accumulatein a melt that beginsto crystallizein Incongruentsolubilityin the gasphase
the feldsparfield and in that part of the quartz field The compositionof the HrO-rich gaseousphase
wherefeldsparwill beginto crystallizebeforelithium that coexistswith lithium minerals,alkali feldspar,
minerals.The limiting amountof LirO is somewhat and quartz can be studiedonly with difficultyin
greaterthan the amount of LirO presentat the eu- conventionalcold-sealpressurevessels, becausethe
tectic(1.85percentLirO, 81.0percentSiOr),because 20-50 mg of quenched fluid is difficult to analyze
of the curvatureof the field boundary.The limiting accurately for so many constituents. The method
amount of LizO is equivalentto about 27 percenl usedin this study was to select an isothermal isobaric
spodumene of theoreticalcomposition,or 45 percent sectionat 575"C and 2 kbar, and attempt to deter-
petalite.Comparisonof the bulk compositionsof mine phaseboundaries in the systemEu-Qz-HrO by
homogeneous Li-rich pegmatiteand of the zonesof experimentson 6 compositionswith which succes-
zoned Li-pegmatitesto the liquidusresultsshould sivelylargeramountsof HzO wereequilibrated. Ex-
contributeto understanding the petrogenesisofthese aminationof the solid phasesafter the experiments
compositions. indicatedtheir compositions,and by differencethe
compositionof thegaswasqualitatively determined'
Petrogenetic significance of positionsof pegmatite Theseexperimentsindicatedthe phaserelations
compositions on the liquidus shown schematically in Figure 5. Equilibriumwas
Comparison of Figures 2 and 4 reveals that the not attained in all runs; B-spodumene waspresentin
bulk compositionsof Li-rich pegmatitesare closely most runs rather than spodumene,though spodu-
groupedin the regionof minimum meltingtemper- mene was synthesized in many runs. One or two
atureson the liquidus.This correspondence suggests eucryptitecrystalswere observedon the surfacesof
that the compositionsof homogeneous lithium peg- somerunsalsocontainingpetaliteor quartz,suggest-
matitescould havebeenmagmasat the temperatures ing metastablepersistence after rapid initial leaching
and pressures shown on the diagram.If crystalliza- of the surfaceof the glassby steam.Figure5 shows
tion startsfrom the feldspar-silica sideline,no lith- that as HrO is addedto glassof Pe compositionand
ium mineral will form until the field boundaryis allowedto equilibrate,the followingsequence of as-
reached.Thus the liquidusdiagramprovidesan ex- semblages shouldbe observed (all plus gases of differ-
planationfor the sequence of zonesobserved, for the ent compositions): petalite,petaliteplusspodumene,
abrupt appearance of lithium mineralsin quantity spodumene, spodumene plus eucryptite,eucryptite,
from compositions nearthe field boundary when the and finally complete solution into the gas. The
field boundaryis reached, and for the limit on the amount of petalite that can be dissolved in the gasat
abundanceof lithium minerals. The limit on the 575oC, 2kbar was about 0.64 weight percent, Greater
gradeof spodumene (or petalite) in Li-rich pegmatite amounts of solids would probably be dissolved from
derivedfrom its magmatic origin should be a useful compositions between petalite and quartz, which is
concept for guiding development of and prospecting the range of natural compositions. If the gas compo-
for lithium ores.The phasediagramfor the Qz-Eu sitionis held constantby bufferingin the presence of
sidelineand diagramsfor the P-7 stabilityof spodu- both quartz and petalite,the solid precipitatedfrom
mene and petaliteindicatethat petalitepegmatites it would havea very silica-richcomposition(Fig' 5).
shouldbe the hightemperature-low pressure equiva- A composition(Q259.5Eu40.5) slightly more si-
lentsof spodumene pegmatites. liceousthan petalite(Qz58.9Eu4l.l)with 3l percent
978 STEWART: LITHIUM.RICH PEGMATITES

HrO yieldedpetaliteand spodumene and no apparent quartz and of potassium feldspar.Orville (1960) and
quartz,proving the gas was very siliceous.The gas Norton (1970) conclusivelyshowed the asymmetric
compositionbufferedby spodumene and quartzun- distribution of potassiumin pegmatites,and relatedit
der conditionswherepetaliteis not stablewas not to gravitational forces such as would be involved
determinedby my experiments. Edgar(1968,p.223) when a gas and a silicate liquid are in contact with
reportedthat B-spodumene formedfrom spodumene each other. The geologic occurrencesof petalite in
was more siliceousthan the LiAlSirO. composition, near-surfacecontact aureoles of granites in Devon,
indicatingthat the gasshouldbe Li-rich, but where Elba, Japan, Wyoming, and elsewheresuggestthat
enoughB-spodumene was obtainedin my experi- petalitecomponentsare readily transportedin gases.
mentsfor reliableestimates of compositionfrom cell- The alkali composition of the gas phase might be
edge measurements, it was always as siliceousor postulatedto be responsiblefor the quartz-richcores.
slightlylesssiliceousthanthestartingglass.Thecom- However, according to Orville (1963), at 500-600"C
positionsof the solidsprecipitated from gasesassoci- and 2 kbar the ratio of Na+ to K+ in the gas that
ated with alkali feldspars,micas,spodumene,and coexistswith albite and microcline will be about 4:l
quartz are probablylike thoseplotted in the Qz-rich for a wide range of total alkali concentrations.Ac-
part of Figure2. cording to Scavnicarand Sabatier(1957),in the same
These experimentsalso suggestthat subsolidus temperatureinterval if the ratio of Li+ to (Na+ * K+ )
leachingfrom or exchangewith earlier-formedas- were as high as 0.5, the sodium in albite or the potas-
semblages containingspodumene or petaliteprovide sium in microcline will be exchangedfor Li+, and a
a mechanism by whicheucryptitecanform in nature. siliceousB-spodumene(possiblywith quartz) will re-
This topic wasreviewedby Stewart(1960,p.28-29). sult. Becauseno B-spodumenehas beenfound evenin
Burt et al. (1977)postulateda stabilityfield for the the petalite pegmatites where appropriate temper-
assemblage eucryptiteplusquartzat low P and Z. No aturesare probable,it is likely that the Li+ concentra-
bulk compositions of zonesreportedfall outsidethe t i o n d i d n o t e x c e e d0 . 5 ( N a + * K + ) . C o m b i n a t i o n o f
trianglepetalite-quartz-alkali feldspar,so magmatic these two results indicatesthat Na+ > K+ > Li+ in
crystallizationseemsto be incapableof producing the gas, and episodes of Li+-rich gases to cause
eucryptite-bearing assemblages. Subsolidusrecrys- quartz-rich cores are not probable, nor should they
tallizationof alkalifeldspars andspodumene or peta- causeextensiveLi-metasomatismoutside pegmatites.
lite with F-bearingsteam can yield lepidolite,as The regions of lithium amphibole (holmquistite) in
determinedexperimentally by Munoz (1971).Le- the contact rocks around some lithium pegmatites
pidolitemost commonlyoccursin abundance in the ( K e s l e r , l 9 6 l ; H e i n r i c h , 1 9 6 5 ) a l s o c o n t a i n p o t a s -
innermostzonesand involvesthe samegas phase sium-bearing lithium micas and albite, an expected
discussed above. consequencefrom this analysis.
Compositions that couldhaveformedfrom gascan
probablybe distinguished from thosethat couldhave Concluding discussion
formed from magmaby the ratio of quartzto feld- The hypothesisthat Li-rich pegmatitescould have
spar. Where quartz exceedsfeldspar,gas is much formed from a magma is sustainedfor both homoge-
involved;however,feldspar-rich compositions could neous pegmatitesand for the feldspar-richzonesthat
be from magmas.The association of gaswith quartz contain spodumeneor petalite in zoned pegmatites.
) feldsparseemsto accordwith the texturalcriteria The high temperature-low pressure isochemical
of Jahnsand Burnham(1969),as demonstrated by e q u i v a l e n to f a s s e m b l a g5e o f C a m e r o n e ta l . ( 1 9 4 9 )i s
Cernj' and Ferguson(1972,p.674)for petalite-bear- perthite-plagioclase-petalite-quartz,and of assem-
ing intermediate zonesof the Tancopegmatite.Very blage 6 is plagioclase-petalite-quartz.These zones
largepetalite-bearing pegmatites at Bikita, Rhodesia seemto be the last that can reasonablybe interpreted
(Cooper,1964),Varutrdsk,Sweden(Quensel,1956) to have originated from molten silicate magma. The
and the Tancopegmatite,Manitoba,werenotedby remaining zones, core, and fracture fillings require
Cern!'and Ferguson(1972,p.676)to havepotassium the participation of transport through a gas, or sub-
feldspar) atbite in the petalitezonesand to be s o l i d u sr e c r y s t a l l i z a t i o n .
enrichedin quartz.The bulk compositionof the Bi- Fractional crystallization is a superficiallyattrac-
kita pegmatite(Gallagher,1962)seemsnot to be tive process to invoke for the generation of Li-rich
unusuallyrich in Qz (Fig. 2). The involvementof a magma. However, granites associatedwith lithium
gas phaseprobablycausesboth the enrichmentof pegmatites typically contain less than 100 ppm Li
STEI4/ART : LIT H I UM. RICH PEGM AT ITES 9',t9

(Mulligan, 1973;J. J. Norton, personalcommunica- Cern!, P. and R. B. Ferguson(1972)The Tanco pegmatiteat


tions 1948-1978), and evenveryperfectfractionation Bernic Lake, Manitoba. IV. Petaliteand spodumenerelations.
Can. Mineral.,I l, 660-6'78.
that would yield a typicalLi-rich pegmatitecontain- Cooper,D. G. (1964)The geologyof theBikitapegmatite' In S.H.
ing -7100 ppm Li (:1.53 weight percentLirO) Haughton, Ed., The Geologyof SomeOre Depositsin Southern
would require at least 70 times as much parental Africa,p.441-461.Geol.Soc.SouthAfrica,2'
magmaas pegmatite.Extrememechanical problems Edgar,A. D (1968)The a-pliAlSirO' (spodumene) transition
arisein extractingsucha smallfractionof restliquid. from 5000to 45000lblin? Ps,o.In Int. Mineral. Ass.Pap. and
Proc., 5th GeneralMeet., Cambridge,Eng', 1966' p.222-231.
Alternatesuggestions havebeenmade(Norton, 1973, Mineral.Soc. London.
p. 369),and my presentopinionis that anatexisis the - and J. M. Piotrowski(1967)A20$1-;s1 for albitescrystal-
hypothesis mostconsistent with geologicalandexper- l i z e d i n t h e s y s t e m sN a A l S i a O " - 9 - L i A l S i , O u - H r Oa n d
imentalevidence.As Li-rich pegmatite and related NaAlStO'-LiAlSiO.-H,O. Mineral Mag., 36, 578-582'
granitesthat have normal (<100 ppm Li) contents Gallagher,M. J. (1962)Mineralogyof the Bikita pegmatites,
SouthernRhodesia,with specialreferenceto beryl depositsin
are associated in time and space,theymostprobably
Bikita Main Pegmatite.Geol. Suru. Gt. Brit', Atomic Energy
received lithium from the samesource.Partialmelt- Diu.,Rept.242.
ing of Li-enrichedmetasediments wouldyieldLi-rich Heinrich,E. W. (1965)Holmquistiteand pegmatitic lithiumexo-
magmaat temperatures 75oC(or more) lower than morphism.Indian Mineral.,6, 1-13.
Holdaway,M. J. (1971)Stabilityof andalusite andthe aluminum
thoseat whichgraniticmagmasof minimummelting
sificatephasediagram.Am. J. Sci.,271, 29-131.
compositionin the systemOr-Ab-Qz-HrO could
Huang,W. L. and P. J. Wyllie (1973)Meltingrelationsof mus-
form. Continuedheatingwould yield largeramounts covite-graniteto 35 kbar as a model for fusion of metamor-
of graniticmagmaand would dilutethe lithium con- phosedsubductedoceanicsediments. Contib. Mineral. Petrol,
tent. The irregulardistributionof Li-rich pegmatite 42, l-14.
Jahns,R H. and C W. Burnham(1969)Experimental studiesof
farthestof all pegmatitevarietiesfrom relatedgranite
pegmatitegenesis: I. A modelfor the derivationand crystalliza-
and the sparsityof Li-rich pegmatites within related
tion of graniticpegmatites. Econ.Geol.,64, 843-864.
granitesarein accordwith an anatectic origin.Li-rich Kesler,T L. (1961)Explorationof the Kings Mountainpegma-
pegmatites occurnearthe sillimaniteisogradin meta- tiIes.Mining Eng., 13, 1062-1068.
morphicrocks,indicatingtemperatures of )50loC at Lovering,T G. (1958)Handbookof Statistics for Geologists of the
any pressure andgreaterthan 625'C al2kbar (Hold- Mineral DepositsBranch.U.S. GeologicalSurvey.
Luth, W. C.(1976)Graniticrocks.In D. K. Baileyand R. Mac-
away, l97l), quite consistentwith liquidustemper- donald,Eds.,TheEuolutionofthe CrystallineRocks,p. 335-418.
aturesdeterminedexperimentally. Processes capable AcademicPress,New York.
of producingconcentrationratios of hundredsor Milne, V. G. (1962)Thepetrography and alteralionof somespodu-
thousandsto one relativeto granitesseemto have menepegmatilesnear Beardmore,Ontario.Ph.D. Dissertation,
operatedto createvaluableandgeochemically unique Universityof Toronto (Diss.Abst., 24, 1575-1576,1963).
Mulligan,R. (1962)Origin of lithiumand beryllium-bearing peg-
depositsof Rb, Cs,Nb, Ta, andSn in pegmatites, and
matites Can. Mining Metallurg. Bull., 65, 419-422.
magmaticdifferentation processes aloneseem not to - (1973)Lithium distributionin Canadiangranitoidrocks.
suffice. Can.J. Earth Sci, 10,316-323
Munoz,J. L. (1969)Stabilityrelationsof LiAlSi,O.at high pres-
Acknowledgments sures.Mineral.Soc.Am. Spec.Pap.,2,203-209.
I thank J. J. Norton, p Cern!, D. G. Cooper, and B J. Skinner - ( 1 9 7 1 ) H y d r o t h e r m a l s t a b i l i t y r e l a t i o n so f s y n t h e t i cl e p i d o -
for extensivediscussionsof the origin of Li-rich pegmatites and for lite. A m. M ineral., 56, 2069-2087'
j. N o c k o l d s , S . R . ( 1 9 5 4 ) A v e r a g e c h e m i c a l c o m p o s i t i o n so f s o m e
m o d a l d a t a a n d s p e c i m e n sf r o m m a n y p e g m a t i t e s , J . N o r t o n a n d
F. G. Lesure for critical and constructive reviews, and Dora von igneous rocks' Ceol' Soc' Am' Bull ,65, 1007-1032'
Limbach and Mary E. Woodruff for extensivelibrary researchand Norton, J. J (19?0) Composition of a pegmatite, Keystone, South
m e a s u r e m e n to f X - r a y p a t t e r n s . Dakota. Am. Mineral.,55, 981-1002.
- ( 1 9 7 3 ) L i t h i u m , c e s i u m ,a n d r u b i d i u m - t h e r a r e a l k a l i m e t -
als. In D. A. Brobst and W. P. Pratt, United States Mineral
References
Resources,p 365-378. U.S. Geol. Suru. Prof Pap S2O.
Browning, J. S., B. H. Clemmons and T. L. McYay (1961)Ben- - (1975) Pegmatite minerals. ln Mineral and water Resources
eficiating North Carolina Spodumene-Beryl Ores U S Bureau of ofsouth Dakota, p.132-149.South Dakota Geol. Surv. Bull. 16.
Mines, RI 5750. - a n d L . R . P a g e ( 1 9 5 6 ) M e t h o d s u s e d t o d e t e r m i n eg r a d e
Burt, D. M., D. London and M. R. Smith (1977)Eucryptitefrom
and reservesof pegmatites. Mining Eng , 205' 401-414.
A r i z o n a a n d t h e l i t h i u m a l u m i n o s i l i c a t ep h a s ed i a g r a m ( a b s t r . ) .
Geol Soc. Am. Abstructs with Programs, 9,917. - and others (1964) Geology and mineral deposits of some
C a m e r o n , E . N . , R . H . J a h n s ,A . H . M c N a i r a n d L . R . P a g e( 1 9 4 9 ) pegmatites in the Southern Black Hills, South Dakota. U.S'
Inlernal Structure of Granitic Pegmatites. Econ. Geol. Mono- Geol Suru. Prof. Pap., 297-E, 293-341.
graph 2. Orville, P. M. (1960) Petrology of several pegmatites in the Key-
980 STEII A RT: LIT H I UM. RICH PEGM ATI TES

stonedistrict,Black Hills, SouthDakota. Geol. Soc.Am. Bull.. s o l u t i o n so n t h e j o i n L i r O - A l r O r - S i O r . A m . J S c i . , 2 5 8 - A , 3 1 2 -


7t. 1467-1490. 324
- (1963) Alkali ion exchangebetweenvapor and feldspar Stewart, D. B. (1960) The system LiAlSiOo-NaAlSi,O,-H,O at
phases. Am. J. Sci.,261. 283-316. 2000 bars. Int Geol. Congr., XXI Session, Part 17, 15-30.
Quensel,P. (1956)The paragenesis of the Varutriisk pegmatite. - ( 1 9 6 3 )P e t r o g e n e s i as n d m i n e r a l a s s e m b l a g eosf l i t h i u m - r i c h
Arkiu Mineral. Geol.,2, no. 2,9-125. pegmatites (abstr.). Geol. Soc. Am. Spec. Pap., 76, 159.
Scavnicar, S. and G. Sabatier(1957)Actiondu chloruredelithium S u n d e l i u s ,H . W . ( 1 9 6 3 ) T h e P e g C l a i m s s p o d u m e n ep e g m a t i t e s ,
sur lesfeldspathsalcalins;donn6esnouvellessur le feldspath-Li, Maine. Econ. Geol., 58,84-106.
fe spodumene-Fe et l'a-eucryptite.Bull. Soc.fr.Minbral.Cristal- Tuttle, O. F. and N. L. Bowen (1958)Origin of granitein the light
1ogr.,80,308-317. of experimental studies in the system NaAlSirOr-KAlSirOr-
Schairer,J. F. and N. L. Bowen(1956)The systernNarO-AlrOr- SiOr-H,O. Geol Soc. Am. Mem.74.
SiOz.Am. J. Sci.,254, 129-195. U . S . G e o l o g i c a l S u r v e y ( 1 9 6 1 ) G e o l o g i c a l S u r v e y R e s e a r c h1 9 6 1 .
Shaw,H. R. (1963)The four-phasecurvesanidine-quartz-liquid- U. S. Geol. Suru. Prof. Paper 424-A.
gas between500 and 4000bars.Am. Mineral.,48,883-996. V e s a s a l o ,A . ( 1 9 5 9 ) O n t h e p e t a l i t eo c c u r r e n c e so f T a m m e l a , F i n -
Shternberg, A. A., T. I. Ivanovaand V. A. Kuznetsov(1972) land. Bull. Comm. Geol. Finlande, 184, 59-74.
Spodumene-amineraldepthindicator.Dokl. Akad.Nauk,202.
175-179.[transl.Dokl.Acad. Sci. USSR,Earth Sci.Section,202,
llt-l14 (1e73)1. Manuscript receioed, April 14, 1978; acceptetl
Skinner,B. J. and H. T. Evans,Jr. (1960)B-spodumene solid for publication, May 24, 1978.

You might also like