Download pegmatites PDF

TagsGranite Magma Igneous Rock Rock (Geology) Minerals
File Size628.8 KB
Total Pages6
Document Text Contents
Page 1

ELEMENTS, VOL. 8, PP. 289–294 AUGUST 2012289

1811-5209/12/0008-0289$2.50 DOI: 10.2113/gselements.8.4.289

Granitic Pegmatites
as Refl ections of Their Sources

Tracing igneous rocks to their ultimate sources represents a
recurrent and contemporary theme in petrologic research.
To the extent that pegmatites are derived from granites, our
ability to recognize the provenance of pegmatites hinges
upon our capacity to relate granites to their parental rocks.

The sources of granitic magmas are debated as much today
as they were over 60 years ago, when N. L. Bowen and H.
H. Read argued over “the room problem” (Young 1998).
They contested, in part, the relative importance of mantle
versus crustal sources for granitic magmas. Using today’s
robust database of trace element and isotope chemistry,
petrologists still cite evidence for entirely crustal origins
for granites (e.g. Chappell and White 2001) or sources with
a large mantle component (e.g. Healy et al. 2004; Smithies
et al. 2011). The nature and extent of interaction between
granitic magmas and the various rocks they encounter en
route toward the Earth’s surface is another area of past and
present dissonance (compare Roberts and Clemens 1995
with Pignotta and Paterson 2007).

Granitic magmas refi ne their compositions by crystal
fractionation and by the separation of residual liquids
from their crystalline products. Hence, regardless of

their starting compositions or
assimilant, granitic magmas
evolve toward the bulk composi-
tion of the thermal minimum or
eutectic in the system NaAlSi3O8–
KAlSi3O8–SiO2, with slight devia-
tions to the peralkaline (molar
Al2O3 < [Na2O + K2O + CaO])
or peraluminous (molar Al2O3 >
[Na2O + K2O + CaO]) sides of that
compositional system. This liquid
line of descent applies to the major
and minor components of granitic
liquids but generally not to their
trace elements (or isotopes).

If the trace elements that are
imparted to the melt at its source
behave as perfectly incompat-
ible in all of the ensuing crys-
talline phases, then pegmatites

would carry an amplifi ed signature of that trace element
pattern. The petrologically important trace elements found
in granite–pegmatite systems display variable degrees of
incompatibility as functions of the pressure, temperature,
and mineral phases in the system. Depending on their
compatibility in the rock-forming minerals of granites,
the relative and absolute abundances of the initial suite
of trace elements may be modifi ed by the process of frac-
tional crystallization or via contamination by assimilation
of material from external reservoirs (e.g. Novák et al. 2012).
However, the idiosyncratic chemical signatures of granitic
pegmatites are manifested by those trace elements that
are highly incompatible in the rock-forming minerals that
crystallize from granitic magmas. Hence, the abundances
of these trace elements increase essentially without moder-
ation until they form their own distinctive minerals, such
as beryl, spodumene, tantalite, etc., in the most evolved
types of pegmatites.

The current system for classifying pegmatites (TABLE 1)
begins with a subdivision of pegmatite classes (Ginsburg
et al. 1979). The pegmatite classes are distinguished on
the basis of the metamorphic environment of their host
rocks (the abyssal class), mineralogy (the muscovite class),
elemental composition (the rare-element class), and texture
(the miarolitic class). Most of the pegmatite classes carry
an implied connotation of their environment of emplace-
ment, more or less equivalent to depth of formation. The
pressures (depths) at which pegmatites crystallize, however,
are poorly constrained by any chemical or textural features
of the pegmatites themselves. Most pegmatites are intrusive
bodies, and hence postdate their immediately adjacent host

egmatites accentuate the trace element signatures of their granitic
sources. Through that signature, the origin of pegmatites can commonly
be ascribed to granites whose own source characteristics are known

and distinctive. Interactions with host rocks that might modify the composi-
tion of pegmatites are limited by the rapid cooling and low heat content of
pegmatite-forming magmas. The trace element signatures of most pegma-
tites clearly align with those of S-type (sedimentary source, mostly postcol-
lisional tectonic environment) and A-type (anorogenic environment, lower
continental crust ± mantle source) granites. Pegmatites are not commonly
associated with I-type (igneous source) granites. The distinction between
granites that spawn pegmatites and those that do not appears to depend
on the presence or absence, respectively, of fl uxing components, such as B,
P, and F, in addition to H2O, at the source.

KEYWORDS: pegmatite, granite, S-type, A-type, I-type, assimilation, contamination

Petr ̌Cerný,1 David London,2 and Milan Novák3

1 Department of Geological Sciences
University of Manitoba, Winnipeg, MB R3T 2N2, Canada
E-mail: [email protected]

2 ConocoPhillips School of Geology & Geophysics
University of Oklahoma, 100 East Boyd Street, Room 710 SEC
Norman, OK 73019, USA
E-mail: [email protected]

3 Department of Geological Sciences, Masaryk University
Kotlářská 2, 611 37 Brno, Czech Republic
E-mail: [email protected]

Černýite, Hugo mine,
Keystone pegmatite,
Pennington County,
South Dakota, USA.

Named in honor
of Petr ̌Cerný.


Page 2


rocks. The pressure and temperature at which pegmatite
crystallize, therefore, may have little or no direct relation-
ship to the conditions of formation and the mineral assem-
blages of their hosts. For these reasons, the application of
the pegmatite classes is fraught with contradiction and
ambiguity (Tkachev 2011).

̌Cerný (1991) and ̌Cerný and Ercit (2005) expanded the clas-
sifi cation of granitic pegmatites to include ten subclasses,
four of which are subdivided into thirteen types, and two
types are further broken down into seven subtypes. All
of these categories are based on the trace element signa-
tures of the pegmatites as refl ected in their mineralogy and
mineral chemistry. TABLE 1 shows the hierarchy of classifi -
cation, beginning, as did Černý and Ercit (2005), with the
pegmatite classes. Color bars illustrate how the pegmatite
subclasses and their constituents fi t into an overarching
classifi cation of two pegmatite families.

Large granitic batholiths are probably assembled from
multiple plutons, which may arise from different or hetero-
geneous sources, each contributing its own trace element

suite. To the extent that pegmatites acquire their trace
elements from granitic plutons, therefore, one might expect
that the trace element signatures of pegmatites would be
hopelessly variable. The fact is, the trace element signa-
tures of most rare-element pegmatites can be grouped into
just two distinctive families ( ̌Cerný 1991): one that is
enriched in lithium, cesium, and tantalum (LCT) and the
other characterized by enrichment in niobium, yttrium,
and fl uorine (NYF).

Most of the pegmatites with the LCT signature have
compositional affi nity with S-type granites (Chappell and
White 2001). The peraluminous nature of S-type granites is
expressed by assemblages that include some combination
of muscovite, garnet, cordierite, sillimanite or andalusite,
tourmaline, and gahnite (ZnAl2O4). These granites stem
from the anatexis of metamorphic schists and aluminous
gneisses of sedimentary origin. The original sediments
(pelites) consist mostly of clay-rich material produced
by extensive chemical weathering of continental rocks.
The trace element signatures of the granites, and of LCT
pegmatites derived from them, are imparted mainly by the
participation of micas and feldspars in the melt-forming

Most of the pegmatites that belong to the NYF family are
sourced from A-type granites, where “A” means “anoro-
genic” (e.g. Eby 1990). The origins of A-type granites are
varied and debatable. The source of such granites is gener-
ally thought to be gneissic granulites deep in the conti-
nental crust, with some contribution from the mantle in

Class Subclass Type Subtype Family

Abyssal HREE


Muscovite–– REE
rare element


Rare element REE allanite––monazite

Li beryl beryl–columbite

complex spodumene
lepidolite LCT


Miarolitic REE topaz–beryl

Li beryl–topaz








TABLE 1 The pegmatite classifi cation scheme of Černý and
Ercit (2005), modifi ed to show the correlation

between pegmatite classes and families. NYF = niobium–yttrium–
fl uorine family (green); LCT = lithium–cesium–tantalum
family (yellow); HREE = heavy rare earth elements; LREE = light rare
earth elements

Page 3


the form of basaltic melt or low-density carbonic fl uid.
Černý and Ercit (2005) now ascribe a small fraction of the
LCT and NYF pegmatites to I-type sources. I-type granites
are usually affi liated with subduction-related magmatism,
but they can be generated from the metamorphic products
of any mafi c to intermediate igneous rocks or volcaniclastic
sedimentary rocks.

The pegmatite subclasses, types, and subtypes of Černý
and Ercit (2005) can be assigned with little ambiguity to
one of the families (highlighted in yellow or green in TABLE
1); as well, hybrids that arise by mixing between LCT and
NYF sources can also be recognized ( ̌Cerný and Ercit 2005;
Martin and de Vito 2005; Novák et al. 2012). Pegmatites
that carry the LCT signature greatly outnumber those of
the NYF family, and within the LCT family, the beryllium-
and lithium-rich subclasses and types are by far the most
common of the rare-element pegmatites.

The classifi cation scheme proposed by Černý and Ercit
(2005) hinges upon the rare-element signatures of pegma-
tites, as there is often little else in their composition that
serves to distinguish them. In most cases, the rare-element
signature is ascertained from the exotic mineralogy of these
pegmatites. It is important to bear in mind that the vast
majority of pegmatites do not possess exotic minerals (see
London and Morgan 2012 this issue). However, the concept
of the pegmatite family was meant to apply not to any
individual pegmatite, but to a large group of comagmatic
pegmatites, of which only a few evolve to develop the diag-
nostic mineralogy of the family, its subclasses, types, etc.
In the common pegmatites that lack more exotic miner-
alogy, the characteristics of the pegmatite family can be
ascertained from and followed through the trace element
contents of the common minerals, such as micas, oxides,
mafi c silicates, and others ( ̌Cerný et al. 1985).

In TABLE 1, Černý’s (1991) pegmatite families are identi-
fi ed by their diagnostic trace element signatures. If these
signatures were substituted for the classes, the classifi cation
of granitic pegmatites would be purely on their chemical
attributes, without genetic inferences for depth of emplace-
ment or an implied tectonic setting.

The LCT Family
Enrichment in the rare element lithium is the most preva-
lent characteristic of the LCT pegmatites. The predominant
lithium minerals include the silicates spodumene, petalite,
lepidolite, and elbaite, and the phosphate series amblyg-
onite–montebrasite [LiAlPO4(F,OH)] and lithiophilite–
triphylite [Li(Mn,Fe)PO4]. Cesium can be elevated in beryl
and micas, but Cs can achieve concentrations suffi cient
to precipitate pollucite, CsAlSi2O6. Although columbite
(a Nb-dominant oxide) appears early in the evolutionary
sequence of the LCT pegmatites, Ta-rich oxides predomi-
nate toward the end (see Linnen et al. 2012 this issue). The
important fl uxing components B, P, and F are elevated but
variably enriched (see London and Morgan 2012). Boron is
found in black tourmaline in the margins of pegmatites, but
also in gem-forming elbaite in the central zones (Simmons
et al. 2012 this issue). Many of the LCT pegmatites contain
a plethora of primary and secondary phosphates in addi-
tion to apatite, and phosphorus is a signifi cant component
of the feldspars (London et al. 1999).

S-type granites arise from crustal thickening that is usually
associated with subduction and continental collision. In
most occurrences, however, pegmatites derived from these
granitic sources lack the foliation or pervasive deformation
that is expected in a syntectonic environment. Because
of the high abundance of granitic pegmatites in orogenic
belts, Martin and De Vito (2005) link the LCT family of

pegmatites to the tectonic environment of subduction.
They may be more properly affi liated with a post-tectonic
phase in the development of continental-collision belts
(Tkachev 2011). The main characteristics of the LCT pegma-
tites, however, are derived from previously unmelted, mica-
rich metamorphic rocks, irrespective of the tectonic regime
in which their initial partial melting occurs.

White mica (muscovite–paragonite–phengite solid solu-
tions) and dark mica (biotite-group solid solutions) carry
most of the trace elements that defi ne the signature of
the LCT pegmatites (e.g. Dahl et al. 1993). The abundant
white mica in schist of marine sedimentary origin reacts
extensively at the onset of anatexis (London et al. 2012).
The initial extent of melting is small, because it is limited
by the low sodium content of the rock. Consequently, a
large fraction of the rare-element content of mica schist is
transferred to a small volume of partial melt. The melting
reactions of white and dark mica also produce K-feldspar
+ aluminosilicate + spinel as products, especially when
the concentration of H2O in the melt is well below that
of saturation (e.g. Acosta-Vigil et al. 2003). Rubidium is
slightly incompatible in K-feldspar, whereas Li and Cs are
almost perfectly incompatible. Hence, the formation of
K-feldspar at the source, and its continued crystallization
from the granitic melt, leads the liquid line of descent
toward a composition in which the Cs/Rb and Li/Rb ratios
become highly elevated. This fractionation trend results
in the pattern of rare-alkali enrichment found in the
LCT pegmatites.

Within individual pegmatites, the Nb–Ta oxides fractionate
from Nb-rich at the margins to Ta-rich in the central units.
Linnen et al. (2012) explain that trend by the contrasting
solubilities of Nb versus Ta oxides in melt as a function of
temperature. However, the same general trend of increasing
Ta/Nb ratio is present from the start of granite fractionation
(e.g. Černý et al. 1985), when the distributions of these
elements are controlled by major and accessory minerals
in which Nb and Ta are nonessential trace constituents.
London (2008) reviewed the published data on partitioning
of Nb and Ta among rutile, ilmenite, titanite, amphi-
boles, and biotite. There was no consistent pattern in the
partitioning data; that is, the phases in question did not
consistently incorporate one element over another. Thus,
the factors that fractionate Ta from Nb in granites are not
yet fully known.

Chappell and White (2001) observed that an elevated phos-
phorus content is as diagnostic of the S-type granites as
is their peraluminous composition. Both chemical attri-
butes are positively correlated with the derivation of melt
from metapelite protoliths and with the H2O content of
those melts (London et al. 1999; Acosta-Vigil et al. 2003).
Phosphorus- and Cs-rich lithium pegmatites are truly diag-
nostic of S-type sources for the LCT family of pegmatites
(Martin and De Vito 2005).

The NYF Family
Pegmatites that fi t into the NYF family are notable because
they contain chemically complex oxides and silicates that
carry heavy rare earth elements (HREEs), Ti, U, Th, and
Nb > Ta. These include euxenite/aeschynite [(Y,Ca,Ce,U,Th)
(Nb,Ta,Ti)2O6], samarskite/fergusonite [(Y,Fe3+,Fe2+,U,Th,Ca)
(Nb,Ta)O4], gadolinite [(Y,Ca)2Fe3+Be2Si2O10], and allanite-
(Y) [CaYFe2+Al2Si3O12(OH)]. Abundant fl uorite or topaz
refl ects the enrichment in fl uorine. The NYF pegmatites
are depleted in phosphorus, and tourmaline is uncommon.
Their mafi c minerals include ferruginous biotite, aegirine,
and riebeckite, the latter two denoting peralkaline compo-
sitions for these pegmatites.

Page 4


Most of the NYF pegmatites bear a chemical affi nity to
A-type granites (Eby 1990; Černý and Ercit 2005; Martin
and De Vito 2005). As a general model, A-type granites are
sourced from combinations of pyroxene-bearing quartzo-
feldspathic rocks of the lower continental crust with varying
amounts of added mantle components (e.g. King et al. 1997;
Christiansen et al. 2007). The magmas are believed to be
poor in H2O, but F is imparted by the decomposition of
amphiboles and micas to pyroxene (Skjerlie and Johnston
1992). In some instances, these granites may be entirely
mantle derived (e.g. Haapala et al. 2007), as indicated by
their low initial 87Sr/86Sr ratios (e.g. van Breemen et al.
1975). Where their tectonic setting can be ascertained,
A-type granites and NYF pegmatites are usually associated
with hot spots or rift zones within continents.

Martin and de Vito (2005) state that NYF pegmatites carry
much the same trace element enrichment patterns as do
peralkaline igneous rocks that fractionate directly from
mantle sources. That is true for the high fl uorine and
niobium signatures of both rocks, but the NYF pegmatites
are enriched in the heavy rare earths, whereas alkaline
magmas derived from mantle sources mostly show a light
rare earth enrichment. In addition, the NYF pegmatites, like
their A-type granite sources, are highly depleted in phos-
phorus and are poor in calcium. The peralkaline magmas of
direct mantle lineage culminate in rocks that are not only
calcic (carbonatites) but usually also phosphorus rich (as
apatite). That does not mean that pegmatites do not arise
from alkaline mantle sources. They do, but most do not fi t
the category, sensu stricto, of the NYF family of pegmatites,
which are granitic in their overall composition.

The origin of the NYF trace element signature is compara-
tively obscure. For example, it is not known if the predomi-
nance of Nb over Ta refl ects their relative abundances in the
source rocks, or whether some aspect of the mineralogy or
fl uid chemistry of their parental alkaline magmas fraction-
ates these two elements. The heavy REEs are associated with
rocks rich in fl uorine. In turn, the high fl uorine content
of the NYF pegmatites is believed to come from melting
reactions involving F-rich amphiboles and micas.

Some members of the NYF family of pegmatites (TABLE 1),
like their A-type granitic sources, possess rare-element
signatures that are indicative of more than one important
source for their elemental and isotopic components. The
radiogenic isotope systems of at least some A-type gran-
ites possess defi nitive evidence for mixed mantle–crustal
materials (e.g. van Breemen et al. 1975).

̌Cerný and Ercit (2005) suggested that some rare-element
pegmatites in both families might be sourced from I-type
granites. However, the I-type granites that are clearly
associated with subduction zones (mostly Phanerozoic in
age) tend to lack signifi cant pegmatitic aureoles at their
margins, which is a hallmark of the S-type and A-type
granites. As an example, the granitic porphyries of Tertiary
age that generated large hydrothermal cells and copper
mineralization in the western Cordillera of North America
are devoid of pegmatitic textures. These igneous bodies are
thought to have exsolved a saline aqueous fl uid early in
the history of their magmatic consolidation, and hence,
according to the Jahns-Burnham model (see London and
Morgan 2012), should have been prime candidates for
developing a pegmatitic facies (cf Nabelek et al. 2010).

Pegmatites that are locally present in the interiors of I-type
plutons of the Sierra Nevada batholith, USA, possess sharp
intrusive contacts with their host granites and textures
that are indicative of thermal quenching of the pegmatite-
forming melts against their hosts (FIG. 1; also see Webber
et al. 2001). In this association, the pegmatites are not
derived from their immediately adjacent igneous rocks.
Tourmaline-rich pegmatites reportedly are common in
the Cathedral Peak granodiorite and other Sierra Nevada
plutons (Lawford Anderson, pers. comm. 2012). The I-type,
tin-rich Mole Granite in Queensland, Australia, possesses
along part of its margin a meager pegmatitic facies
enriched in beryl, topaz, and lithian dark mica (possibly
zinnwaldite). In these cases, however, the probable source
of these distinctive and incompatible trace elements (Li, Be,
B, and F) is subducted sediment, which was incorporated
into the eventual I-type granites (e.g. Bebout et al. 2007).
Hence, these dominantly I-type granites appear to spawn
pegmatites to the extent that they have incorporated S-type
materials (marine sediments), which make these magmas
hybrids as well.

Pegmatite-forming magmas contain negligible heat to
promote melting of rafted inclusions of solid rock. However,
the fractionated compositions of pegmatite magmas and
their fi nal aqueous fl uids are highly reactive with other,
less-evolved, common host lithologies (e.g. Morgan and
London 1987; Novák et al. 2012). Local contamination of
pegmatites occurs principally along dike margins during
emplacement, and again at the transition into subsolidus
conditions. Alteration of host rocks by pegmatite-derived
fl uids occurs late in the history of consolidation.

Mafi c Components
The process of crystallization and separation of mafi c
minerals from granitic magmas leaves their derivative
pegmatites depleted in Fe and Mg. In LCT pegmatites, the
crystallization of tourmaline can reduce Fe and Mg in the
melt to trace levels (Wolf and London 1997). The A-type
sources of NYF pegmatites are also poor in Fe and espe-
cially in Mg. It is common, however, for evolved pegma-
tites of both types to contain spectacular concentrations
of biotite or tourmaline along their margins (FIG. 2A, B).

FIGURE 1 Granite and pegmatite dikes in tonalite along the Piute Pass Trail
between Piute Pass and Hutchinson Meadow, central Sierra Nevada,

California. The thicker central dike shown here possesses pegmatitic borders, in
which feldspar crystals are elongate and branch toward the dike center, followed
inwardly by aplitic texture, and then a return to coarse-grained pegmatitic texture
in the central zone. Pencil for scale. PHOTO: JAMES T. GUTMANN

Page 5


The common hosts for pegmatites, including greenstones,
amphibolites, mica schists, and gneisses, are the inferred
sources of the mafi c constituents (Van Lichtervelde et al.
2006; London 2008). Usually, the biotite- or tourmaline-
rich “fringe” ends abruptly inward, without any further
crystallization of either phase. London (2008) attributed
the sharp cessation of these mafi c silicates to crystallization
along the pegmatite contacts, which effectively seals off
chemical communication between the magma and the host
rocks. It is notable that although an infl ux of mafi c compo-
nents into pegmatite appears to be pronounced, there is
rarely any counterfl ow or diffusion of pegmatite-derived
components into the host rocks along their margins, except
locally and sporadically around the largest rare-element
pegmatites at the end stages of their consolidation (Morgan
and London 1987).

Alkaline Earths
Contamination of LCT pegmatites by metacarbonates
appears to have modifi ed the sources of some pegmatites
in central Madagascar and the Czech Republic (Novák et
al. 2012). These pegmatites possess strong enrichment in
Li and B (as spodumene, lepidolite, or elbaite) and locally
Cs (as londonite, CsBe4Al4[B11Be]O28), but they contain
primary assemblages that include diopside, danburite
(CaB2Si2O8), uvite (Ca–Mg tourmaline), and liddicoatite
(Ca–Li tourmaline).

Within a given large group of pegmatites (LCT or NYF),
a general trend in the fractionation of alkalis begins with
chemically primitive K-rich pegmatites closest to their
source, followed by Na-rich pegmatites at the distal end of
the most fractionated pegmatite types ( ̌Cerný 1991; London
2008). The possible infl uence of host-rock composition on
the alkali ratios of pegmatite-forming magmas, however,
has not been adequately considered. In the Middletown
district, Connecticut, Stugard (1958) found that the Na/K
ratio of feldspars in pegmatites correlates with the lithology
of the hosts: pegmatites hosted by metamorphosed grano-
diorite are dominated by sodic feldspar, whereas pegmatites
hosted by muscovite schists are primarily K-feldspar rich.
Metasomatic exchange of alkalis between host lithologies

and granitic melts can occur rapidly and over large
distances, such that the composition of a host rock could
dominate the resultant alkali ratio in small volumes of
intruded melt (London et al. 2012). The extent to which
this is true for pegmatites is not yet known. However, the
conditions in which pegmatites crystallize (see London and
Morgan 2012) are not conducive to an extended period
of open-system communication between pegmatites and
their hosts.

The compositions of pegmatites refl ect an association
mostly with two granite types: the S- and the A-types.
Pegmatites of the LCT family, especially those enriched
in Li, Cs, B, and P, greatly predominate over all others.
This indicates that the metamorphosed juvenile sedi-
ments from which S-type granites arise are particularly
prone to yielding pegmatite-forming melts. Considering
what makes S-type and A-type sources distinct from I-type
sources, the difference comes down to their abundance
of fl uxing components, that is, ligands other than silica
and alumina that profoundly infl uence the properties of
pegmatite-forming melts. S-type sources are enriched in B
and P, but also F, which is derived from the micas. A-types
are enriched in F, which is contributed by the eventual
melting of amphiboles and biotite. The archetypal I-type
granites found in subduction zones, as the sources of arc
volcanism and base-metal mineralization, are notably rich
in Cl and are hydrous, but they are largely devoid of the
fl uxing components noted here (see London and Morgan
2012). They generate enormous volumes of quartz veins but
lack pegmatites to any signifi cant extent. This distinction
points to an essential role for fl uxing components like B,
P, and F, along with H2O, in the formation of pegmatites,
as has been evident to most petrologists for over a century.

Many colleagues and students discussed with us the topics
covered here in the fi eld and laboratory, and Karen Ferreira
contributed to pulling the diverse contributions together,
including the reviewers’s comments.

FIGURE 2 (A) Tourmaline-rich “fringe” along the margins of a
thin pegmatite dike hosted by metaconglomerate,

from Capoeira 2, Borborema Pegmatitic Province, Brazil.

(B) Meter-scale crystals of biotite (dark; see arrows) radiate
down from the upper contact of the Ipê pegmatite, Governador
Valadares, Minas Gerais, Brazil. PHOTO: SKIP SIMMONS


Similer Documents