Titanium, aluminum and interlayer cation substitutions … · American Mineralogist, Volume 6E,...

American Mineralogist, Volume 6E, pages 880-399, 1983

Titanium, aluminum and interlayer cation substitutions in biotitefrom high-grade gneisses' West Greenland

Rosenr F. DYtvtBx

Department of Geological SciencesHarvard (Jniversity, Cambridge, Massachusetts 02138

Abstract

A detailed electron microprobe study has been carried out on biotite in pelitic to maficmetamorphic rocks from two localities in West Greenland. Comparison of analyses fromsimilar rock types indicates higher Ti, K/(K+Na), and possibly (K+Na), similar AlIv, andlower AlvI in granulite us. amphibolite grade samples. Normalization of the analyses on thebasis of7 octahedral and tetrahedral cations results in calculated total positive charge thatexceeds the theoretical maximum of 22.0 in all cases, assuming an anion framework of l0oxygen and 2(OH+F+Cl). In granulite grade biotite this charge excess correlates well with2Ti, indicating that vacancy- or oxy-substitution [Ti4* + n : 2R2* or Tia+ 1 2gz- : 3z++ 2(OH)-l predominate over ones in which octahedral Ti is balanced by tetrahedral Al.The vacancy substitution is preferred as the best explanation ofthese data. Many analysesyield a charge excess above that accounted for by Ti, and contain Alvl not balanced byAlIv. In amphibolite grade biotite, charge excess correlates with 312 AlvI, suggestingadditional vacancies associated with dioctohedral substitution [2,{13* + n : 3R2*].Vacancy-formation does not appear to violate any crystal-chemical relationships in biotite,and may be necessary to accommodate substitution of the smaller Al and Ti for Fe and Mgin octahedral positions.

These results lead to an iterative normalization procedure for microprobe analyses ofbiotite [Total Cations - (K+Na+Ca+Ba) + Ti + %Aryl :7.0, where Al]"t : (Al+Cr)vI -

Alrv + (K+Na+2Ca+2Ba)1, which allocates a vacancy for each Ti and 2 AlYi. Thisiterative normalization eliminates charge excess, and permits Fe3* to be estimated(exclusive of oxyannite).

A model chemical system for micas is discussed which illustrates how many complexsubstitutional mechanisms can be reduced to linear combinations of four simple types

lTschermak: MgSi : AlAl; Dioctohedral: 3Mg : 2Al!; Ti-spinel: MgTi :2Al; Talc: lSi: KAll. On the basis of these relationships, a method is described for recasting biotitemicroprobe analyses into selected end-members.

Introduction

The composition of biotite can deviate considerablyfrom the theoretical range of the phlogopite-annite series:KA(Mg,Fe2*)vIlSirAl;rvo,o(oH)2. The mosr commoncation substitutions involve replacement of Mg2* andFe2* by Al3+, Fe3+ and Tia+ on octahedral sites, and Sia*by additional Al3* on tetrahedral sites. In some situa-tions, replacement of K by Na on the interlayer or "A"site may also be important.

Biotite crystal chemistry is complicated further by theapparent occurrence of vacancies. Foster (1960a,b) dem-onstrated rather convincingly that octahedral occupancyin biotite is generally less than 3 cations/formula. Similar-ly, many recalculated biotite analyses yield A-site occu-pancies less than l.

The presence ofthese "other" cations causes a distor-tion of the biotite structure due to differences in ionic0003-(Mv83/0910-0880$02.00 880

radius and charge (e.g., Hazen and Wones, 1972), andmay exert a strong influence on both the stability range ofbiotite (Rutherford, 1973; Forbes and Flower, 1974;Hewitt and Wones, 1975; Robert,l976a,b\, and on Mg-Fe partitioning between biotite and other silicates such asmuscovite (Guidotti et al., 1977\ and garnet (Dahl, 1969;Dallmeyer, 1974a). In addition, Goldman and Albee(1977) discussed some efects that Ti- and Al-substitutioncould have on biotite-garnet geothermometry.

The effects of possible vacancies on biotite stability areunclear, but in all probability are linked to the nature ofthe cation(s) with which they are associated (c/., Forbesand Flower, 1974). More important, however, are theproblems posed for calculations of mineral equilibriainvolving a complex solid solution phase like biotite,since extrapolations of experimental and/or thermody-namic data obtained on simple systems to "rock" sys-

DYMEK: SUBSTITUTIONS IN BIOTITE FROM WEST GREENLAND 881

tems require certain assumptions about how cation sub-stitutions proceed (Holdaway and Lee, 1977; Holdaway,1980). Consequently, it is important to examine in detailpossible substitutional mechanisms for the above-men-tioned cations through a consideration of their depen-dence on host rock composition and mineral assemblage,and where metamorphic rocks are concerned, their de-pendence on metamorphic grade.

This paper presents the results of a detailed microprobestudy on biotites occurring in high-grade schists andgneisses from West Greenland, which will be used toevaluate certain features of Ti-, Al- and A-site substitu-tions. Selected groups of analyzed biotites in metamor-phic rocks from other areas are also examined in anattempt to apply the generalities deduced from the Green-land samples. Biotites from associations such as granitoidigneous rocks, ultramafic rocks, etc., are not consideredhere, but will be dealt with separately at a later date.

Summary of cation substitutions

This section outlines some general aspects of chemicalsubstitutions in biotite that are relevant to the presentstudy. It is not intended to be a comprehensive review ofthis subject, but rather to establish a framework for datapresentation and discussion that follows.

The dominant mechanism causing Al-enrichment inbiotite involves the Al-Tschermak's substitution:

(R2+;vt + (Si4*)Iv : (Alr+;vt + (Al3+)Iv, (l)

which relates phlogopite-annite to the aluminous end-member eastonite [KMg2AlSi2Al2Or0(OH)2].r previousstudies have shown that the amount of Al-Tschermaksubstitution appears to increase slightly up to sillimanitezone, and perhaps decrease at still higher grade (Tracyand Robinson, 1978). At a given P and T,it is maximizedin biotite that coexists with an Al-saturating phase such asstaurolite, sillimanite, erc.

Foster (1960a,b) showed that many biotite analysesyielded AlvI that could not be accommodated by substitu-tion (l). She suggested that this additional Al is incorpo-rated in biotite by means of a dioctohedral-trioctohedralsubstitution:2

3(Rz+;vt : 2(Al3*)vr + (n)vt, Q)

I The formula as listed is more aluminous than what istypically referred to as eastonite in the literature, but is used herefor convenience. The existence of natural end-member eastoniteis problematic. Strunz (1970) noted that the original material wasprobably a phlogopite-vermiculite intergrowth, whereas Foster(1960b) had previously drawn attention to this situation. Ananalysis by the writer (unpublished) on a sample from the "type"locality (Easton, PA; Philadelphia Academy #26550) shows it tobe nearly pure phlogopite with -4.6 wt.Va F.

2 Foster 11960a,b) actually considered (R3*)vr. Substirution (2)thus represents a simplification of her results; (Fe3+;vt is consid-ered separately here.

which could be viewed as a muscovite comoonent inbiotite, and results in the formation of octahedral vacan-cies. Rutherford (1973) indicated that substitution (2)occurs only to a limited extent (10-l2Vo\ in the Mg-freesystem, although observations on natural biotites suggestthat dioctohedral substitution is maximized in a biotitecoexisting with muscovite, a situation not realized inRutherford's experiments.

Ti-substitution in biotite is problematic. Kunitz (1936)suggested that Tia+ replaces Sia* on tetrahedral sites, andsome recent studies of biotites from kimberlites andmantle xenoliths, where )(Si+ Al) < 4, provide somesupport for this idea. When this tetrahedral Si-Al "defi-ciency" was first recognized, Fel* was assigned to fill thetetrahedral sites (e.9., Dawson and Smith, 1977;Delaneyet al.,1980). However, Farmer and Boettcher (1981) haveshown that (Fe3*)Iu is most plausibly associated withonly those biotites that display reverse pleochroism (c/.,Faye and Hogarth, 1969), which also have >(Si+Al+Ti)( 4. Farmer and Boettcher (1981) conclude that Tia+ mayenter tetrahedral sites in those biotites with normal ple-ochroism, and XSi+Al) < 4.

Engel and Engel (1960) suggested that Ti3* replacesAl3+ on octahedral sites. Studies of the optical spectra ofbiotite, however, reveal absorption bands interpreted toindicate Fe2*-Ti4* intervalence charge transfer, and areambiguous in regard to Ti3+ (Faye, 1968; Burns andVaughan, 1975). ln contrast, Evans and Raftery (1980),based on an X-ray photoelectron diffraction study, con-cluded that Ti is trivalent in the biotites which theyinvestigated (assemblages unspecified). It is unclear whatimplications this result has for the present study, wherenearly all biotites coexist with a Tia+ oxide (ilmenite orrutile). The position taken here is that Ti occurs as(Tio*)ul, which is the prevailing opinion of most workers.

Tia+-substitution on Vl-fold sites is complicated by itshigh charge and small cation radius (0.6054) compared toMgt* (O.ZZA) and Fe2+(0.78A) (ionic radii from Shannon,1976). Several substitutional mechanisms have been sug-gested for Ti4+ in biotite. The first involves a coupled VI-and IV-fold relationship:

(R2*)vt + 2(Si4*)rv = (Ti4*)vI + 2(Al3+)rv. (3)

Substitution (3) can be viewed as a Ti-Tschermak'scomponent (analogous to R2+TiAl2O6 in pyroxene) lead-ing to the theoretical end-member Ti-eastonite [KMgzTiSiAl3Or0(OH)zl, and was favored by Czamanske andWones (1973), Robert (1976a,b), Guidotti et al. (1977) andTracy (1978). A second coupled substitution would be:

(Al'*)ut + (Si4+)rv = (Ti'*)vI + (Al3+)Iv. @)

Direct substitution of Ti for Al on octahedral sites mavoccur in two wavs:

2(Al3*)vr : (Ti4*)vI + (R2+)vr, and (5)

2(Rz+;vt : (Ti4*)vr + (!)vr. (6)

882 DYMEK: SUBSTITUTIONS IN BIOTITE FROM WEST GREENLAND

Substitution (5) can be viewed as a Ti-spinel component,whereas (6) results in the formation of octahedral vacan-cies; the latter was favored by Dahl (1969), Forbes andFlower (1974), and Dymek and Albee (1977). The final Ti-substitution involves a dehydrogenation reaction:

(R2*)vt + 2(oH)- : (Ti4+)vI + z(o2) + H2, Q)

leading to anhydrous end-member Ti-oxybiotite [KMgzTisi3Alor2l.

Clearly, there is no consensus as to the nature of Ti-substitution(s) in biotite, and recently Bohlen et al.(1980), and Abrecht and Hewitt (1980, l98l) have indicat-ed that combinations of (3), (6) and (7) are necessary toexplain their data. Whatever the case, observations onnatural biotite indicate that Ti-content is maximizedwhere biotite coexists with a Ti-saturating oxide phase,and that Ti-content increases with increasing metamor-phic grade (e.g., Kwak, 1968; Guidotti et al.,1977).

In a limited number of cases, (Fe'*)tt is assigned fromstructural formula calculations (c/., Foster, l960a,b; Cza-manske and Wones, 1973). However, as mentionedabove, the presence oftetrahedral Fel* may be revealedby reverse pleochroism, and both synthetic and naturalferri-annite [KFe3+Si3Fe3*Oro(OH)2] show this fearure(Wones, 1963a; Miyano and Miyano, 1982).

Fe3+ in octahedral coordination may proxy for Al3+ viaa Ferri-Tschermak's substitution:

(Rt*)tt + (si4+)Iv : lFet*)ut + (Al3+)Iv, (8)

yielding the end-member ferri-eastonite with composition[KMg2Fe3+Si2Al2Or0(OH)2]. This substitution should bemaximized where biotite coexists with an Fel+-saturatingphase such as magnetite or hematite.

An additional, important mechanism relevant to Fel*in biotite involves in situ oxidation of Fe2* as representedby the following reaction:

Fe2* + (OH)- : Fe3* + 02- + VzH2. (9)

Many aspects of the resulting oxy-annite end-member[KFe2+Fe;+Si3AlOr2] were discussed by Eugster andWones (1962) and Wones and Eugster (1965), and Wones(1963b) showed that Reaction (9) was entirely reversibleupon hydrogenation, and that Fe2*/Fe3* in biotite wastherefore strongly dependent upon fi1,.

Chemical variations on the A site in biotite may arisefrom at least two sources. Substitution of H3O* for K-could explain a low abundance of interlayer cationsand/or high water contents found in some mica analyses.Alternatively, a low abundance of interlayer cationscould result from the following substitution:

(K*)A + (Alr+;tv : (n)" + (Si4*)Iv 00)

which relates phlogopite-annite to a talc component[!A(Mg,Fe2*)3Si4Oro(OH)2], and would represent a po-tassic analogue to the edenite substitution in amphibole.Similarly, K-Na substitution would relate phlogopite to

its Na-counterpart, which was synthesized by Carman(1974\, and recognized recently as a new mineral (wones-ite) in schists from Vermont (Spear et al.,1978,1981). Inaddition, the Na-counterpart to eastonite has also beenidentified (preiswerkite; Keusen and Peters, 1980).

Normalization procedures

Electron microanalysis cannot distinguish betweenFe2* and Fe3*, and H2O is likewise not determined withthe microprobe. Consequently, calculations of formulaproportions from analyses of a hydrous, Fer*-bearingmineral such as biotite requires certain assumptions.

Two normalization schemes are commonly applied.The first can be expressed by the relationship:

Total Cations - (K+Na+Ca+Ba) : 7.9, (A)

which assumes complete occupancy of octahedral andtetrahedral sites (c/., Ludington and Munoz, 1975). Such"cation-based" normalization procedures are unaffectedby the valence state of iron and variations in H2O-content, and can be exceedingly useful since they permita straightforward estimation of Fe3*-contents in manysilicates and oxides (e.g., pyroxene, garnet, spinel, ilmen-ite, and probably amphibole). In biotite, assuming a fullcomplement of 10 oxygen and 2 (OH,F,CI) per formulaunit, Fe3+ would be equal to the difference between 22.0and the calculated total positive charge. However, ifoctahedral vacancies are present [e.g., substitutions (2)and (6)1, formula proportions calculated from (A) will beoverestimated in direct proportion to the number ofvacancies, and a corresponding calculated charge excesswould result, which, among other things, makes Fe3*-estimates impossible. Alternatively, a charge excess maynot be indicative of vacancies, but may be compensatedby an oxygen excess related to the occurrence of oxy-components [e.g.. substitution (7)3].

The second scheme, which assumes that all iron occursas Fe2*, involves normalization to 11 oxygen atomsexpressed by the following relationship:

Total Cation Charge : 22.0. (B)

This procedure is unaffected by vacancies, but causescation abundances to be overestimated in direct propor-tion to the number of Fe3* cations that are present, whichremains an unknown quantity. However, if there is anoxygen excess (1.e., oxycomponents), total charge f22.0, and calculated formula proportions are underesti-mated.

Some investigators (e.g., Czamanske and Wones, 1973)have used a modified form of (B) by applying a fixedFe2*/Fe3* ratio to their biotite data based on partial wetchemical analyses. In the general case where information

3 Because the initial charge calculation is made with all iron as

Fe2*, a charge excess at this point cannot be due to oxyannitesubstitution.

DYMEK: SUBSTITUTIONS /N BIOTITE FROM WEST GREENLAND 883

of this type is lacking, normalization of biotite micro-probe analyses involves a tradeoffbetween assumption ofno vacancies and no Fe3*. In a subsequent section ofthispaper, an iterative normalization procedure is presentedthat bypasses these assumptions, and permits an Fe3*-estimate to be made.

Samples studied

The results reported in this paper comprise part of an ongoinginvestigation of the mineralogy and petrology of the Archaean(>3000 Ma) Malene supracrustals, a series of sedimentary andvolcanic rocks from the Godthrib District, West Greenland,which were subjected to intense regional metamorphism at ca.2800 Ma (c/., Dymek, 1978). Biotite in samples from twolocalities at diferent grades of metamorphism are considered.The first locality (Lzng{;64'52'N, 52"12'W) was metamorphosedunder granulite grade conditions (Sill-Kf zone; -7.5 kbar,750'C). Three lithologies are distinguished:a a. potassiumfeldspar gneisses (Bio + Kf + Plag f Qtz + Garn + Crd + Sill +Ilm); b. Orthopyroxene gneisses (Bio + Plag + Qtz + Opx +Garn t Crd + Sill + Ilm/Rut * Mag); c. Pyribolite (Bio + plag +Hbl + Opx -| Cpx + Mag + 11.;.

The biotite content of these samples ranges from -l to 20volume percent. Many of the samples contain biotite grains withditrerent colors and textures, indicating at least two episodes ofmineral growth. Only primary biotite, which is brown and Ti-rich(up to 6 wt.VoTiO), is considered here.

Most samples contain a Ti-saturating phase, which is typicallyilmenite, with rutile restricted to the most Mg-rich assemblages.Magneti te, representing Fe3+-saturation, is common rnpyribolites. Quartz is abundant (>5 vol%) in all samples exceptpyribolites, in which it is typically absent. Sillimanite occurs inabout half of the potassium feldspar gneisses and in a feworthopyroxene gneisses. In all diagrams presented herein,sillimanite-bearing potassium feldspar gneisses are distinguishedfrom sillimanite-free types. Plagioclase occurs in all samples.

The second locality (Itivdlinguaq : Itiv; 64"13'N, 5l'31'W)was metamorphosed at amphibolite grade conditions (Stl-Sillzone, -5.5 kbar, 575"C). Three lithologies are distinguished herealso: a. Muscovite schist (Bio + Musc + Plag + Qtz + Sill + Stl+ Ilm); b. Semipelitic schist (Bio + Plag + Qtz + Crd + Garn +Oamph + Stl + Sill * Ilm or Rut); c. Amphibolite (Bio + plag +Qtz + Hbl t Cumm + Oamph + Garn + Ilm). The biotitecontent of these samples ranges from -l to 25 volume percent.As in samples from Lang@, there are two generations of biotite atItiv, but only primary biotite is considered. These are green, withTiO2-contents in the range 0.7-2.f wt.Vo.

All samples studied from Itiv contain quartz, plagioclase, anda Ti-saturating phase (either ilmenite or rutile). Sillimaniteoccurs in muscovite schists and in a few of the semipeliticschists.

Although regional correlation of lithologies is imprecise in the

a In this paper the following abbreviations are used: Bio :biotite; Kf : potassium feldspar; Plag = plagioclase; Qtz :quartz; Garn : garnet; Crd = cordierite;Sill : sillimanite; Ilm :ilmenite; Rut : rutile; Mag = mugn"rite; Opx : orthopyroxene;Hbl = hornblende; Cpx = clinopyroxene; Musc = muscovite;Stl : staurolite; Oamph = orthoamphibole; Cumm : cumming-tonite.

Godthab District, the rocks from Langp are believed to be highergrade equivalents of those from Itiv, and the followingcomparisons can be made: muscovite schist + sillimanitic Kfgneiss; semipelitic schist + orthopyroxene gneiss; amphibolite+ pyribolite.

In summary, the approach adopted here is that by studying alarge number of samples of difering bulk compositions, a morethorough understanding of the range of possible biotitecompositions at a given metamorphic grade will be achieved.Thus, it should be possible to discriminate between grade-dependent and grade-independent substitutions. This contrastswith the approach that focuses on a specific rock compositionand examines biotites in only limiting (or nearly so) assemblages.In the latter kinds of rocks, the extent of all cation substitutionsis fixed by P, T, and volatile activities, and compositionalvariation at a given metamorphic grade is precluded by thechoice of samples.

Results

G e ne ral c o mp o s itional c harac t e ris tic s

A representative analysis from each sample studied islisted in Tables l-6, where they are grouped by lithologictype. A description ofanalytical techniques is provided inthe Appendix.

Individual biotite analyses are illustrated in Figure I interms of relative amounts of MgO, FeOl+MnO, andAl2O3. All analyses are enriched in Al2O3 compared tophlogopite-annite, and span a large range of Mg-value[=Mg/(Mg+Fe+Mn)]. The much greater range of Mg-values found in the granulite-grade samples from Lang/compared to those in amphibolite-grade samples from Itivis probably related to the rock compositions studied.Although there is no apparent difference in the overallrelative concentration of Al2O3 in biotite from the twolocalities, the Al-content is related to the nature of therock type in which it occurs, reflecting a trend towardsAl-saturation. In samples from Langfl, the Al-contentincreases in the order pyribolite -+ orthopyroxene gneiss-+ Kf-gneiss. In the latter group, sillimanite-bearing sam-ples are clearly distinguished from sillimanite-free sam-ples. In samples from Itiv, the Al-content increases in theorder amphibolite --+ semipelitic schist -+ muscoviteschist. The most Al-rich biotite in semipelitic schistsoccurs in sillimanite-bearing samples.

With the exception of the above-mentioned Al-enrich-ment, and high TiO2-contents (up to -6 wt.Vo), thecompositions of the biotites are relatively simple. Theamounts of Na2O (<0.7 wt.%), MnO (<0.4 wt.Vo), andZnO (<0.2 wt.%o) are low; CaO was generally not detect-ed, whereas Cl was near the limits of detection in mostcases; F-contents are also low, although values as high as-l wt.Vo were found. Routine analyses were not per-formed for Cr2O3 and BaO, but reconnaissance workindicates low concentrations (<0.2 and <0.4 wt.Vo each).In summary, the analyses presented here appear similarto ones reported from high-grade metamorphic rocks byother investigators.


Table | . Microprobe analyses of biotite in potassium feldspar and sillimanite-potassium feldspar gneisses, Langp.

u t Z l . \ . o .

s i 0 2A l ^ 0 .

T i 0 2

Mso

Fe0,

Mn0

ZnO

C a 0

Naro

Kzo

F

c t

l o t a l

37 .201 8 . 5 4

2 . 2 7

16 .37

1 1 . 0 2

0 . 0 9

0 . 00

0 . 0 2

0 . 0 7

9 . 7 7

0 . 2 7

0 . 0 3

9 6 . t 5

36 . 80

r 8 . 8 1

13 .93

1 4 . 1 2

0 . 0 7

0 . 1 3

0 . 0 0

0 . 00

1 0 . 0 3

0 . 2 1

0 . 0 r

36. \5

r 8 . r 5

3 . 1 1

1 3 . 1 2

1 5 . 3 1

0 . 0 1

0 . 00

0 . 00

0 . 0 4

r 0 . r 3

0 . 2 3

0 , 0 1

36 .7116 .42

3 . \ 2

1 1 . 0 4

1 7 . 7 2

0 . 09

0 . r 5

0 . 00

0 . 0 4o ( q

0 . 3 4

0 . 4 8

36.28

16 .57

1 0 . 8 7

r 8 . 5 r

0 . r 8

0 . l l

0 . 0 0

0 . 0 0q R (

0 . 50

0 . 0 r

5 0 . 2 t

r 6 . r 8

3 . 80

1 2 . 1 2

16 .96

0 . 0 7

0 . 2 0

0 . 00

0 . 0 2

0 . 4 3

0 . 0 |

36.77t c 7 0

4 . 8 5

13 .78

r 5 . 5 0

0 . t 7

0 . 00

0 . 00

0 . 0 3

y . o l

0 . r 00 . 0 1

96 .27 96 .53 q q A q 95 . 38 96 .57

l v s i 2 . 7 7 6 2 . 7 3 2A l t . 2 2 \ 1 . 2 6 8

FORHULA PROPORTIONS**

2 .755 2 .697 2 .851 2 .7791 . 2 \ 5 1 . 3 0 3 r . 1 4 9 1 . 2 2 1

2 . 7 9 9 2 . 7 \ 0 2 . 7 7 7 2 . 7 2 2l . 2 o l 1 . 2 6 0 1 . 2 2 3 1 . 2 7 8

v l A l 0 .39 \ 0 .325T i 0 . 1 2 6 0 . 1 2 \f t s 1 .796 1 .768Fe 0. 578 0 . 558M n 0 . 0 0 5 0 . 0 0 5Zn 0. 000 0 . 000E 3 .000 2 . 890

Ca 0 . 002 0 . 002l l a 0 . 0 1 0 0 . 0 1 0K 0 . 9 r 7 0 . 9 0 3x 0 . 9 2 9 0 . 9 1 4

2 . 7 5 8 2 . 7 r 01 2 L 2 r ' q n

0 . 420 0 . 3430 . 1 2 7 0 . 1 2 5f . ) ) o t . > z a

0 . 885 0 .8700 . 0 0 4 0 . 0 0 40 . 0 0 7 0 . 0 0 73 . 000 2 .878

0 . 0 0 0 0 . 0 0 00 .000 0 .0000 .959 O .9420 .959 0 .9 \2

0 .373 0 .2810 . r 8 0 0 . r 7 5r . 478 | . 4 \ 70 . 9 5 8 0 . 9 4 80 . 0 0 r 0 . 0 0 10 .000 0 .0003.ooo z:gB0 .000 0 .0000 . 0 0 5 0 . 0 0 50 .977 0 .9560 . 9 8 3 0 . 9 6 2

o .355 0 .2450 . 2 0 0 0 . r 9 5r . 2 7 9 1 . 2 4 6r . l 5 1 1 . 1 2 20 .006 0 .0050 .009 0 .0093noo T.Ot0 .000 0 .0000 . 006 0 . 0060 .959 0 .935o.e6t o:5IT

2 .807 2 .7 \21 . r 9 3 1 . 2 5 8

0 . 3 r 8 0 . 2 1 80 .212 0 .2081 . 2 5 4 1 . 2 2 5r . r 9 8 1 . r 7 00 . 0 1 2 0 . 0 1 20 .006 0 .0053.ooo t.-83T

0 .000 0 .0000 .000 0 .000o .972 o .9500 .972 0 .950

0 . 2 7 20 .221I ? q q

1 . 0960 . 0050 . 0 1 I3 . 000

0 . 0 0 00 . 0 0 3n o q q

0 .962

0 . r 8 20 . 2 1 6| .365| . 0720 . 0040 . 0 1 12.852

0 . 0000 . 0030 . 9380 . 9 4 1

0 . 1 8 30 .276

0.9790 . 0 1 10 .0003 . 0 0 0

0 .0000 . 0 0 4o .9260 . 9 3 0

0 . t 0 lo '2701 .5210 .9500 . 0 1 I0 .000T:6'0 . 0 0 00 . 0040 . 9 0 80 . 9 r 2

f , 6A-24c etz-ptag-Kf- l i77-carn-I fn (22.352t) 2. 6A-24d Qtz-Plag-xf-Sif l - -Garn-I fn (22.3g0) 3. &A-36h-7 QXz'I PLag-Kf-si f7-Kf-Garn-rLm (22.477) 4. &A-29 Qtz-Plag-Kf-I ln/Mt (22.572) 5. cA-27 Qtz-Pfag-Kf-I7n (22.521) o.

6A-25e Qtz-Pfag-Kf-I7n (22.475) 7. acA-36b Qtz-Pfag-Kf-I fn-opx (?) (22,447)

* Minus oxggen for P, cJ ** Based on 7 cations (J.st e).um) and fi oxggens (2nit cofuw)

f Nunbers jn parentheses cotrestrDnd to caLcuJated totaL positive charge for the 7 cation nomLization

Ti-substitution

Ti- and Allv-contents of the biotite are illustrated inFigure 2, with formula proportions based on the lloxygen normalization (B). Biotite in the amphibolite andgranulite grade samples have a similar range of Allv, butthe latter contain substantially higher Ti.

If all TivI and Alrv were balanced by substitution (3),then the data points should lie along the lines labeled l/2.Any data falling below this line could be interpreted asresulting from the combined effects of the Al- and Ti-Tschermak's substitutions [(1) and (3)].

All but three of the analyses of biotite from Itiv liebelow the Vzline, which is a necessary but not sumcientcondition for supporting the above interpretation. How-ever, the majority of the Langl analyses contain Ti that isimpossible to balance by Alrv, thereby requiring anadditional (or different) Ti-substitutional mechanism. Ifone were to attribute this "extra" Ti to an overestimationof cation abundances due to an unknown amount of Fe3+

in each analysis, the situation would not change dramati-cally. For example, recalculating the most Fe- and Ti-richanalyses with all iron as Fe3* increases Allv by a maxi-mum of 0.12 and decreases Ti by 0.02.

Figure 3a demonstrates that a remarkably good correla-tion exists between calculated total positive charge andTi-content in biotite from Lang/, with formula propor-tions based on the 7 cation normalization scheme (A)'Note that the data points scatter about the line labeledCharge Excess : 2 Ti, which was not constructed to fitthe data, but is merely a line drawn from zero Ti-contentwith a slope of 2.5 This relationship between chargeexcess and Ti-content suggests that substitutions (3), (4)and (5) play a subordinate role for Ti-substitution in thesebiotites, since they involve the same number of cationswith the same net charge. However, the data in Figure 3aare consistent with either the Ti-vacancy (6) or Ti-oxy (7)

5 The fact that the charge excess correlates with 2Ti stronglysuggests the presence ofTia+ rather than Ti3*'

DYMEK: SUBSTITUTIONS IN BIOTITE FROM WEST GREENLAND

Table 2. Microprobe analyses of biotite in orthopyroxene gneisses, Lang/.

885

r,it Z l . 2 . ? 4 . 5 . 7 . 8 .

s i 0 2

A1203

Cr20: ,

T i 0 2

ttS 0

Fe0T

l.tn0

Z n 0

Ca0

Ba0

Naro

Kzo

c l

Tota I

37 .82I 5 . 8 9

n . a .

1 . 3 9

1 8 . 9 5

1 0 . 3 1

0 . 0 0

0 . 0 1

0 . 0 0

n . a .

0 . 0 0

l 0 . l t

0 . 3 2

0 . 0 1

3 7 . 5 7

1 7 . 1 \

0 . 0 0

1 . 4 2

1 9 . 2 9

9 . I 3

0 . 0 2

0 . 0 0

0 . 0 2

0 . 2 1

0 . 0 3

9 . 3 0

n . a .

38 . 50r 6 . 4 5

n , a ,

2 0 . 2 5

7 . 6 \

0 . r 0

0 . 1 6

0 . 0 0

0 . 0 9

1 0 . 0 0

0 . \ 2

0 . 0 r

3 7 . 9 11 5 . 3 0

2 . 5 \

1 7 9 l1 1 . 2 7

0 . 0 3

0 . 000 . 0 1

o . 2 2

9 . \ 9

0 . 3 i

0 . 0 l

3 9 . 0 4I ) . O O

n . a .

2 0 . 7 |

0 . 0 0

0 . 0 9

0. 00

n . a .

0 . 2 5

9 . l 3

0 . 2 8

0 . 0 4

3 7 . 9 1

t o . ) t

0 . 0 4

2 . 7 7

1 8 . 4 3

r 0 . 3 9

0 . 0 4

0 . 0 0

0 . 0 0

0 . 1 8

0 . 0 5

9 . 5 0

n , a .

n . a ,

3 7 . 3 5

1 6 . 7 0

0 . 0 5

3 . 0 4

1 7 . 5 t l

I 0 . 5 r {

0 . 0 2

0 . 0 0

0 . 0 0

0 . 3 8

o . 2 2

9 . \ 7

37 021 7 . 3 0

3 . 3 9l o - 2 5

r 0 . 9 5

0 . 0 5

0 . 0 0

0 . 0 0

n . a ,

0 . 0 2

1 0 . 2 9

o . 2 8

0 . 0 3

95.68 9 4 . 1 5 9 5 . 0 3 95 .77 o 4 0 n 95.q5

I V s i 2 . 7 6 7 2 . 7 5 6A l 1 . 2 3 3 | 2 \ \

A t 0 . 2 2 \ o . z o 7

T i 0 . 0 7 7 0 . 0 7 6l. '19 2 . 068 2. 050F e 0 . 6 3 1 0 . 5 2 8l 4 n 0 . 0 0 0 0 . 0 0 0Z n 0 . 0 0 1 0 0 0 1L 3 . 0 0 0 2 . 9 7 2

C a 0 . 0 0 0 0 . 0 0 0B aN a 0 0 0 0 0 . 0 0 0K 0 . 9 4 4 0 . 9 q 0I o.-tm 6-5q0

FoRt'tuLA PRoPoRT I oNs**

) 1Aa 7 1qo , A l t , 1oA

1 . 2 1 7 1 . 2 \ 1 r . 1 8 8 r . 2 0 4

0 2 0 0 0 . r 5 2 0 . 1 4 3 0 . 1 1 9

0 . 1 4 5 0 . r 4 5 0 . 1 4 9 0 . 1 4 81 . 9 6 1 1 . 9 \ 3 2 . 2 2 \ 2 , 2 1 20 . 6 9 2 0 . 6 8 6 0 . \ 7 9 O . \ 7 60 . 0 0 2 0 . 0 0 2 0 . 0 0 0 0 . 0 0 00 . 0 0 0 0 . 0 0 0 0 . 0 0 5 0 . 0 0 5? n n n , o 2 Q 2 n ^ n , o ( n

0 .00 r 0 .001 0 .000 0 .000

0 . 0 3 r 0 . 0 3 1 0 . 0 3 5 0 . 0 3 50 . 8 8 9 0 . 8 8 1 0 . 8 3 9 0 . 8 3 4a:9ZT 0:913 o.B7[ d:-86J=

2 . 7 6 1 2 . 7 5 \1 . 2 3 9 1 . 2 \ 6

0 . 2 \ 6 0 . 2 3 50 . 0 0 0 0 0 0 00 . 0 7 9 0 . 0 7 82 . i l 3 2 . 1 0 80 , 5 5 1 0 . 5 5 00 . 0 0 1 0 . 0 0 10 . 0 0 0 0 . 0 0 03 . 0 0 0 2 . 9 8 2

0 . 0 0 2 0 . 0 0 20 . 0 0 7 0 . 0 0 70 . 0 0 4 0 . 0 0 40 . 8 7 2 0 . 8 7 10:38[ o-:88]

2 . 7 9 6 2 . 7 7 3t . z u \ t . z z I

0 . 2 0 5 o . r 6 9

o 1 2 3 0 . 1 2 22 . 1 9 3 2 . 1 7 \0 . { 6 4 0 . 4 6 00 . 0 0 6 0 . 0 0 60 . 0 0 9 0 . 0 0 93 . 0 0 0 2 . 9 ! 1

0 . 0 0 0 0 . 0 0 0

0 . 0 r 3 0 . 0 r 30 . 9 2 7 o . 9 1 90 . 9 3 9 0 . 9 3 r

2 . 7 7 1 2 . 7 4 71 . 2 2 9 1 . 2 5 3

0 . r 9 9 0 . r 5 30 . 0 0 2 0 . 0 0 20 . r 5 2 0 . r 5 r2 . 0 0 9 r . 9 9 10 . 6 3 5 0 . 5 3 00 . 0 0 2 0 . 0 0 20 . 0 0 0 0 . 0 0 03 . 0 0 0 2 . 9 4 0

0 . 0 0 0 0 . 0 0 00 . 0 0 5 0 . 0 0 50 . 0 0 7 0 . 0 0 70 . 8 9 5 0 . 8 8 80 . 9 0 7 0 . 9 0 0

2 . 7 7 1 2 . 7 3 \1 . 2 2 9 1 . 2 6 6

0 . 2 3 1 0 . 1 7 50 . 0 0 4 0 . 0 0 30 . 1 7 0 0 . 1 6 71 . 9 4 0 i . 9 r 40 . 5 5 4 0 . 6 1 { 50 . 0 0 r 0 , 0 0 I0 . 0 0 0 0 . 0 0 03.000 2 .9o7

0 . 0 0 0 0 . 0 0 00 . 0 1 1 0 . 0 1 10 . o 3 2 0 . 0 3 20 . 8 9 5 o . 8 8 40.939 0 .927

2 . 7 7 51 . 2 2 5

0. 3or i

0 . r 9 1r . 8 1 40 . 6 8 70 . 0 0 30 . 0 0 03. 000

0 . 0 0 0

0 . 0 0 30 . 9 8 4^

-dI?

2 . 7 2 0r 280

0 2 1 8

0 . 1 8 71 . 7 7 80 . 6 7 \0 . 0 0 30 . 0 0 02. 860

0. 000

0 . 0 0 30 . 9 5 4

1. 6A-28d Qtz-Plag-opx-Mt 1ZZ.OazT1 2. GRD-1.44.1 Qtz-pLag-Gatn-opx-Mt (22.056) 3. GGU_3586f,2 Qtz_pLag_opx_cord_Rut (22,187)4' 6A-28h Qtz-Pfag-opx-cotd-si l f -Rut (22.796) 5. cGA-28b Qtz-plag-opx-cotd-Rut (22.147) 6. GRD-7447 Qtz-plaq-Garnl0td-Opx-Si l f - I ln (22.190) 7. Grd-1439 etz-pfag-Gatn-Opx-IJn (22.2g5) b. eCU_SSAdl, l etz_pfag4ord_Opx_SiJl_Rut (22.449)* i l inus oxvgen fot F, cf ** Baseal on 7 cat ions (Lst cofum) ani l f l oxggens (2nd coj,u@) n.a. = not anafgzedf ivumbers it patentheses eotrespond to cafeufateil tota) posixive charge based q tbe 7 cati@ notnalization

substitutions, as both lead to excess positive charge usingnormalization procedure (A). The relationship betweencalculated positive charge and Ti-content in biotite fromItiv (Fig. 3b) is more complicated than that for Lang1.This apparent diference is related to the Al-content, andwill be discussed in the next section.

The total positive charge used in Figure 3 is calculatedassuming that all iron is Fe2+. If, in fact, some of the ironis Fe3*, then the positive charge is higher and the chargeexcess greater. The net result of the assumption of allferrous iron is that data points plot lower on the positivecharge axis than they otherwise would if the amount offerric iron were known beforehand. Hence, the value forpositive charge represents a minimum value.6 In Figure

6 A slight note ofcaution should be added here due to the lackof Li2O-determinations. However, the Li-abundances observedto occur in most "common" biotites are typically very low.Unless the Greenland biotites are anomalous, inclusion of Li inthe charge calculations will not change the observed relation-ships in a significant way.

3a, nearly all analyses of biotite from pyribolites, whichcontain magnetite and are therefore Fe3*-saturated, plotbelow the line. This lowered positive charge is consistentwith the presence of Fe3+ in these analyses. However,nearly all of the analyses of biotite from potassiumfeldspar gneisses plot above the line, and more than halfof the analyses from Itiv plot above the line as well (Fig.3b). Thus a majority of the biotite analyses have a chargeexcess which exceeds that accounted for bv Ti-substitu-tion.

Al-substitution

Biotite compositions are illustrated in Figures 4aand 4bin terms of AlvI and AIIV with formula proportions basedon the 7 cation normalization (A). If the substitution ofAlvr into biotite occurred only by the Al-Tschermak'scomponent (l), then the data points should fall along thelines labeled AlvI = Allv. The presence of a Ferri-Tschermak's component (8) should result in Allv exceed-ing AlvI in direct proportion to the amount of (Fe3*)vl.


Table 3. Microprobe analyses of biotite in pyribolites, Lang/.

vJt % l . 2 . i r . 7 . 8 .

s i 0 2

A l ̂ 0 -z 5T i 0 2

I'tg 0

' - - T

l'ln0

Zn0

Ca0

Na20

Kzo

F

c l

Tota I

3 8 . 8 9

1 5 . 2 3

2 . \ 8

2 0 . \ 3

0 . 0 3

0 . 0 0

0 . 0 0

0 . 2 5

9 . 8 5

0 . 5 0

0 . 0 1

5 l . o o

1 l r . 50

3 . 2 4

1 7 5 8

1 2 . 8 3

0 . 0 4

0 . r 2

0 . 0 0

0 . r 2

v . t >

0 . 0 2

39.07

1 3 9 2

3 . 4 4

1 8 . 6 2

1 0 . 7 9

0 . 0 0

0 . 0 0

0 . 0 0

o . 1 2

9 . 3 6

0 . 0 9

0 0 3

7 7 . \ \| 4 . 9 r

4 5 0

1 5 . 7 2

r 3 . 0 5

0 . 1 3

0 . 0 2

0 0 0

0 0 2

9 . 7 70 . 5 5

0 . 0 l

q c 8 c

36.9t1

1 4 . 5 6

5 1 3

1 3 . 7 6

1 5 . 0 9

0 0 5

0 . 2 0

0 . 0 0

0 r 2q 7 R

0 . 0 7

0 0 5

3 6 . \ 1

1 lr. 86

5 . 2 8

1 3 . 2 t +

1 5 . 9 9

0 . 0 3

0 . 1 3

0 . 0 0

0 . 0 2

9 . 6 0

0 . 1 4

0 . 0 2

3 7 . 3 21 \ ) 9

5 . 3 \1 \ . 5 5

I l r .81

0 . 011

0 . 0 7

0 . 0 0

0 . 0 2

r 0 . r 50 . 3 1

0 . 0 2

3 7 . 8 1

1 \ . 2 3

r q q q

r r . 9 r0 . 0 7

0 . 0 8

0 . 0 0

0 . 0 i |

9 . 8 9

0 . 1 0

0 . 0 0

9 \ . 8 2 96. \6 95. \o 95.56 96 . 87 95 .67

I V s i 2 . 8 5 5 2 . 8 2 2A r L l q 5 1 . 1 7 8

A f 0 . 1 7 3 0 . 1 2 \T i o . 1 3 7 0 . 1 3 5

a ) 2 4 7 J 1 d

F e 0 . \ 5 2 0 . 4 \ 7l ' ln 0 . 002 0 .002Z n 0 . 0 0 0 0 . 0 0 0t 3 . c 0 0 2 . 9 1 8

C a 0 . 0 0 0 0 . 0 0 0N a 0 . 0 3 5 0 . 0 3 6t 0 . 9 2 3 0 . 9 1 2r 0 . 9 5 9 0 . 9 q 8

FORI4ULA PROPORTIONS "

2 . 8 2 2 2 . 7 6 9 2 8 3 2 2 . 7 6 \f . 1 7 8 r 2 3 1 1 . 1 5 8 1 2 3 6

0 . 1 4 5 0 . 0 5 9 0 . 1 4 9 0 . 0 4 80 . 2 5 5 O 2 . 5 O 0 . 2 9 0 0 . 2 8 91 7 5 6 1 - 7 3 3 1 5 7 3 1 . 5 3 50 . 8 2 3 0 . 8 0 7 0 9 5 8 0 . 9 { 4o . o o 8 0 . 0 0 8 0 . 0 0 3 0 . 0 0 30 . 0 0 1 0 . 0 c r 0 . 0 1 1 0 . 0 1 ' ltooo t ls-og 3.ooo t.€300. 000 0 . 000 0 . 000 0 . 0000 0 0 1 o . o o 3 0 . 0 1 8 o . o l 70 . 9 3 9 0 . 9 2 2 0 . 9 5 7 0 . 9 3 30 . 9 \ 2 0 . 9 2 5 o 9 7 5 0 9 5 1

2 . 7 9 8 2 . 7 7 71 . 2 0 2 1 . 2 2 3

0 0 7 0 0 . 0 1 { 00 r 8 0 0 . 1 7 91 . 9 4 8 l . 9 3 l r0 .193 o 7870 . 0 0 3 0 . 0 0 20 . 0 0 7 o . o o 73 0 0 0 2 . 9 4 8

0 . 0 0 0 0 . 0 0 00 . 0 1 7 0 . 0 r 70 . 9 r 9 0 . 9 r 20 . 9 3 6 a . 9 2 9

2 . 8 8 1 2 . 8 5 0r . r r 7 i r 5 0

0 094 0 0470 . 1 9 1 0 . r 8 92 . 0 1 9 2 . 0 2 50 666 0 6580 . 0 0 0 0 . 0 0 00 . 0 0 0 0 . 0 0 03 . 0 0 0 2 . 9 1 9

0 . 0 0 0 0 0 0 00 . 0 i 7 0 . 0 r 70 . 8 8 1 0 . 8 7 ro . 8 9 8 o . 8 8 8

2 . 7 9 6 2 . 7 3 51 . 2 0 \ 1 . 2 6 5

0 . r 4 2 0 . 0 5 ro . 3 0 5 0 . 2 9 81 . 5 1 6 l . l { 8 31 027 1 .0050 . 0 0 2 0 0 0 2o 0 0 7 o . o o 73 . 0 0 0 2 . 8 4 7

0 . 0 0 0 0 . 0 0 00 . 0 0 3 0 . 0 0 30 . 9 q r 0 . 9 2 00 94 \ 0 .923

2 .825 2 .7591 . 1 7 5 1 . 2 \ 1

o . r 0 9 o . o r 30 . 3 0 4 0 . 2 9 71 . 6 \ 2 1 . 6 0 \o . 9 3 8 0 . 9 1 60 . 0 0 3 0 . 0 0 30 . 0 0 q 0 . 0 0 43 .000 2 .837

0 .000 0 .0000 . 003 o . oo30 . 9 8 0 0 . 9 5 70 .983 0 .960

2 . 8 5 6 2 . 7 8 3| . 1 \ \ 1 . 2 1 7

0 . 1 2 3 0 . 0 r 70 . 3 2 0 O . 3 1 21 . 7 9 6 1 . 1 5 00 . 7 5 2 0 . 7 3 30 . 0 0 4 0 . 0 0 1 r0 . 0 0 4 0 . 0 0 43 . O t O 2 . 8 2 1

0. 000 0 . 0000 . 0 0 5 0 . 0 0 60.953 0 .9290 . 9 5 9 0 . 9 3 \

f - oGA-23b pfag-Hbf-opx-cpx-I1n/Mt (22,260 ) 2. 6A-38b Qt2-Plag-Hb1-Cpx-I ln/Mt = Noralnor metadiot i te (22.164) 3. eA-35b

plag-Hbj-opx-Cpx-r ln/MX (22-255) 4, dlA-24e Qtz-Pfag-HbL-opx-Cpx-I ln/Mt (22.421) 5. OGA-25d Qtz-Plag-HbL4Px-lLn/Mt

(22.547) 6. 6A-35k etz-pfag-Hb1-Opx-Cpx-ILn/Mt (22.492) ?. &A-37 Qtz-P:ag-Hbl-Opx-Cpx-ILn/Mt = Nordnor netadiot i te

( 2 2 . 5 2 6 ) 8 . d : A - 3 4 a Q t z - P l a 7 - O p x - I 1 n / M r ( 2 2 . 5 7 7 ) .

,Mjnus oxggen for F,Cf **Baset l on ? cat ions ( ls i cojunn) and f l oxggens (2nd caiunn)

lNun.ters in parentheecs cortespon.) xo caicufated total posi t ive charge for the 7 cat ion normafizat ion

lit t

Table 4. Microprobe analyses of biotite in muscovite schists,Itivdlinguaq.

FORI.IULA PROPORT I ONS**

1 . 2 .

All biotite analyses from Itiv and the majority of analysesfrom Lang/, however, plot above the AlvI : AlIv line,indicating the presence of AlvI not balanced by AlIv. If(Fe3+)vt from (8) were added to AlvI, the (R3+)vI excesswould be even larger (c/., Foster, 1960a, b).

The relationships shown in Figure 4 can be correlatedwith those in Figure 3 to the extent that most of theanalyses which contain excess or "uncompensated" Alvlalso yield a calculated positive charge that lies above thelines in Figure 3. Although it could be argued that thefeatures discussed in Figure 4 are an artifact of the 7cation normalization (A), when the data are portrayed interms of the I I oxygen normalization (B), a similar AlvLexcess remalns.

Calculated total positive charge is plotted versus Alvrin Figure 5. No trend is apparent in the data from Lang/,although analyses from the different lithologies fall intodistinct fields, once again reflecting the trend towards Al-saturation in the sillimanite-bearing samples. The datafrom Itiv, with the exception of biotite analyses in horn-blende-bearing samples, appear to define a trend in whichpositive charge excess is proportion al to 312 Alvl. This isexactly the relationship that is expected if the incorpo-

s i 0 2 3 6 . 9 5

A r 2 0 3 r 8 , 7 5

T i 0 2 1 . 4 5

t1g0 I 3 . 46

Fe0, 15 .61

H n o 0 . 2 5

zno 0 .0 t1

C a O 0 . 0 0

N a r 0 O . 2 7

K z O 9 . 1 2

F 0 . 2 6

c r 0 . 0 6

s i 2 . 7 6 6 2 . 7 3 3A l r . 2 3 \ 1 . 2 5 7

A t 0 . 4 2 0 0 . 3 6 9T i 0 . 0 8 2 0 . 0 8 1l 4 g 1 . 5 0 2 1 . 4 8 5Fe 0 .977 o -956t 1 h 0 . 0 1 5 0 . 0 1 6Z n 0 . 0 0 2 0 . 0 0 2t 3 . 0 0 0 2 . 9 I 9

C a 0 . 0 0 0 0 . 0 0 0N a 0 . 0 3 9 0 . 0 3 9K 0 . 8 7 1 0 . 8 6 rr 0 . 9 1 0 0 . 8 9 9

2 .75 \ 2 .7O21 . 2 4 6 1 . 2 9 8

0 . 543 0 . 4580 .o17 0 .o7 61 . 3 0 3 1 . 2 7 91 . 9 5 3 r . 0 3 30.o22 0.0220 . 0 0 r 0 . 0 0 13 .Tob '2 .86T '0 .000 0 .0000 .028 0 .0280.942 0.9250 . 9 7 0 0 . 9 5 3

J O . Z U

t 9 . 9 5

i l . 4 9

t o . ) )

o ' 3 t l

0 , 0 2

0 . 0 0

0 . 1 9

9 . 7 1

0 . 9 5

0 . o 2

I V

V I

T o t a l 9 5 . t l 9 6 . 3 1

7. OcA-48Q Qtz-P7ag-Musc-sxl- lJn ( 22.422+ )

2. GGU-9 L5 g 7 Qtl-P 7ag -Mus c-sx f- s i I f - r 7n-G ahh ( 2 2 . 2 59 )

* I4inus oxggen for F, CL ** Baseal on 7 catiohs (7€t

cofumn) and 77 oxggens (2ni1 co7utu)

t lvunDers in parentheses cortesPond to caiculated totaj

positive chatge for the 7 cati@ notnafization

DYMEK: SUBSTITUTIONS 1N BI}TITE FR)M WEST GREENLAND

Table 5. Microprobe analyses of biotite in semipelitic schists, Itivdlinguaq.

887

8 9 . t 0

s i 0 2

A t z a 3

T i 0 2

i lso

Fe0T

iln0

Z n 0

Ca0

N a 2 0

K z o

c l

Tota I

3 9 3 8

, 1 . 7 \

0 7 l

r 9 0 8

1 0 , 2 3

0 0 5

0 0 0

0 0 0

o 6 j

8 5 2

0 7 t

0 . 0 3

9 5 8 r

3 8 7 5

t 7 6 8

0 9 l

! 8 , 6 8

t 0 l 8

0 0 r

0 0 0

0 0 0

0 6 8

8 3 0

0 3 5

o 0 2

9 5 4 l

3 7 6 7

l 5 9 o

1 0 4

| 5 . 9 9

l 2 9 o

0 . 0 2

0 0 0

0 0 0

0 6 4

8 . 2 6

o 4 2

0 . 0 4

9\ 59

) 7 4 9

1 8 . 0 4

1 1 6

1 4 9 r

1 5 6 5

0 . 0 5

o 0 2

0 0 0

0 4 5

8 3 1

0 r 1

0 . 0 2

9 6 . z l

t 2 2

1 \ , 6 1

1 6 0 2

0 , 0 0

0 0 0

0 0 0

0 3 8

9 1 8

o 4 2

o 0 2

9 6 \ 5

3 8 . 1 6

1 7 E 9

1 2 5

r 5 0 4

1 4 5 t

0 0 0

o 0 5

0 , 0 0

0 4 8

0 4 0

0 0 1

9 6 . 1 7

3 7 4 3

r 8 5 4

1 2 9

r 4 8 8

1 5 , 1 5

0 0 0

0 t 0

0 0 0

c 5 r

8 . 1 { 5

o 1 7

0 0 2

9 6 4 7

r 8 9 01 3 0

l 4 r t 9

1 5 1 9

0 . 0 0

c , r 5

0 0 0

o \ 7

9 C 0

0 3 4

0 . o z

9 6 7 8

3 7 C O

1 7 7 3

t 3 2

1 \ 7 4

1 5 . 7 |

0 0 0

0 0 0

0 0 0

0 c 5

8 . 5 8

0 1 0

0 0 r

9 5 . 0 0

3 6 9 1

r 8 . 4 3

r 4 8

1 1 3 5

1 7 2 7

o 0 2

0 . 0 0

0 . 0 0

0 5 0

8 9 6

0 . r 6

0 0 l

97 -02

ronruLn pnopoRttols"*

1 , 0 0 5 0 0 0 53 000 2 962

2 7 4 5 2 7 2 11 255 | 279

0 1 6 1 0 3 2 30 083 c 0821 4 8 0 r . 4 5 7| 0 7 4 r c 6 50 0 0 1 0 0 0 10 . 0 0 0 0 . 0 0 01 040 2 9)9

0 0c0 0 0000 0 7 2 0 . 0 7 1o 850 o 842o.9rz o.-rl\

2 696 2 77) 2 7581 joq 1 227 1 242

0 335 0.323 0 299o 072 0.07\ 0 073r 589 1 629 1 621o 9 3 4 0 . 9 7 4 o 9 6 90 000 0 000 0 0c0o oog o ooo o ooo,stE Tidd ,:16,

0 0 0 0 0 . 0 0 0 0 0 0 0o 067 c 065 0 0540 , 8 8 2 0 . 8 r 2 o 8 0 7ilelt d:W 6fr,

0 , 0 0 53 0 0 0

0 000o o 1 30 193r 865

2 7 5 J 2 7 1 4 2 . 1 0 3 2 8 1 81 247 1 286 1 297 I 182

0 l f 5 0 3 r 5 0 2 9 8 0 3 1 60 064 0 068 0 068 0 0591 6 3 3 r 6 1 9 r 5 1 3 | 6 5 60.962 0 996 0.992 0 8950 004 0 000 0 000 0 0000 , 0 0 1 0 . 0 0 0 0 . 0 0 0 0 . 0 0 12 . 9 7 9 3 . 0 0 0 2 . 8 7 t 3 . 0 0 0

0 000 0-000 0 000 0 0000 . 0 6 6 0 0 9 5 0 0 5 5 0 . 0 6 90 . 7 8 0 0 . 8 7 1 0 . 8 6 7 0 . 8 0 60 . 8 4 6 0 . 9 2 6 0 . 9 2 2 0 . 8 7 5

2 796 2 790 2.7841 2 0 4 1 2 1 0 1 . 2 1 6

0 300 0 266 0 2570 049 0 058 0 0582 009 1 876 1 872o 614 o 799 o 7980 . 0 0 1 0 0 0 r 0 . 0 0 10 000 0.000 0 0002 973 3 000 2 986

0 000 0 000 0 0000 0 9 5 0 0 9 2 0 . 0 9 2o . 7 6 4 0 . 7 8 o 0 . 7 7 9o.€5e il3tZ ilT-

2 8011 . 1 9 3

0 ,298o 0 3 82 . o z g0 . 5 r 00 0030 . 0 0 0

0 . 0 0 00 0 9 3o 7 7 50:867

7 5 2248

3590716 l l932000

2 7 3 7

0 3 3 60 0 7 r1 5 2 20 9270 000

720280

374o736 0 39U3000

s i 2 8 r 5A l r t 8 4

A l 0 3 1 2T i o o j 8H 9 2 0 3 5F e 0 6 1 2H n 0 0 0 1Zn 0 000t J . o o o

Ca 0 000Na 0 093K o 7 1 7t LSld

2 806I r 9 4

0 l r 60 . 0 5 02 0 1 7o 6 1 70 0 0 10 0003 000

0 0000 096o . 7 6 7o l-62

2 7621 . 2 3 8

o ' ) 2 80 0 5 41 5 3 8o-9650 00(0 0 0 13 000

0 0000 066o . 7 8 3L-BIE

| 2 o B

o J330 069r 6 4 1o 8880 0000 003

0 , 0 0 00 . 0 6 80 7980 856

000 0 000

0 . 0 0 83 000

0 0000 0 5 80 . 8 9 0a 951

o 7 3 0 0 7 20 . 7 8 8o 8 5 1

1 6A-48h Qtz-P1ag-Sxt -c rd-@nph- r ln t22 075 ) 2 . 64-4A9 e tz -p taq-s t l -C td-Oanph- r ln (22 .0A4) 3 . @A-48k Qtz -P laq-s t l -c td -can- tunph-Rut - r ln (22 -044)6A-48f -2 Qtz -P lag-St l -Gar -@nph- r ln (22 .207)

Qxz-P1 ag- St 1 -c ar-oanph-I ]n ( 2 2 - 7 19 )

4 .6A-48e Qtz-Pfa9-Crd-car - rLn (22 057) 5 . 64-48x-3 ,1 exz-p lag-St t -car -s i l j - r ln 122-C92) G.

.7 - . r t ! -1 ! j Qxz-P laq-s tL -car rLn (22-120) B . 6A-4at - t e tz -p tas-car -sa j t - r t tu (22- te7) s . eA-4a j

IO. 64-48 t -3 .2 exz-p7ag-s tL -c rd-carn- rLn (22 j94)

'M inus orsgen fa r F , c f * * Eased on 7 ca t ions (1s t @jum) and 7J oxVgens (2nd co tunn)t ilunbers jn italics corresPoDd to total psitive charqe calucuJated tar the 7 cation namaltzation

ration of large amounts of AlvI in biotite occurs by meansof the vacancy-forming dioctohedral substitution (2). Inad4ition, the relationship between positive charge andAlvI in the Itiv data (Fig. 5b) provides an explanation forthe apparent lack of correlation between positive chargeand Ti in these analyses (Fig. 3b). Thus, the relationshipsbetween calculated positive charge, Ti and AlvI reflectthe fact that Ti-substitution dominates charge excess inthe granulite grade biotite (Lang{; Figs. 3a and 5a),whereas Alvlsubstitution dominates the charge excess inthe amphibolite grade biotite (Itiv; Figs. 3b and 5b).

A geometric consideration

In principal, it should be possible to devise a graphicalmethod for evaluating the problem of Ti- and Al-substitu-tion that is independent of normalization. Figure 6 is arepresentation of mica compositions that uses molecularpercentages of the oxides as its basis. The cornersS (:SiO2+2K2O+2Na2O +2BaO+2CaO), pm (:FeO1+MgO+MnO+ZnO-TiO2) and A (:2Al2O3+2Cr2O3+2TIO2-2K2O-2Na2O-2BaO-2CaO) were selected sothat the compositional fields for all possible dioctahedraland trioctahedral micas are clearly separated as shown.In addition, since phlogopite, eastonite and Ti-eastoniteare colinear, substitution of a Ti-Tschermak's componentis readily distinguishable from other Ti-substitutions.Thus, only substitution of a Ti-oxycomponent (7) or the

vacancy components [(2) and (6)] can shift data points offthe phlogopite-eastonite join towards the S corner.

The analyses of the Greenland biotites, recalculated inS-FM-A coordinates, are illustrated in Figure 7 (which isan expanded portion of the trioctahedral mica field shownin Figure 6), where it can be seen that nearly all analysesare displaced from the phlogopite-eastonite join. Al-though one could argue that the positions of the datapoints are related solely to substitution of a muscovitecomponent, this does not appear to be the case. Inparticular, the analyses of biotite from pyribolites, whichare the most Ti-rich but have the lowest Alvl-contents(cf., Fig. 4a), extend furthest from the phlogopite-easton-ite join. These relationships indicate that a Ti-Tscher-mak's component is unimportant here, but this diagramdoes not allow for a distinction between Ti-vacancv andTi-oxy substitutions.

A-site substitution

The K-contents of biotite from Lang@ and Itiv rangefrom 0.82-0. 98 and 0. 73-0. 95 cations/formula respective-ly, whereas Na-contents range up to 0.04 and 0.11 ca-tions/formula. The use of normalization (A) or (B) doesnot change these values significantly.

These data are summarized in Figure 8, which illus-trates (Na+K+Ba+Ca)A and AlIv based on the l1 oxy-gen normalization (B), and shows that all analyses lie

E88 DYMEK: ST]BSTITTJTIONS IN BIOTITE FROM WEST GREENLAND

Table 6. Microprobe analyses of biotite from amphibolites, Itivdlinguaq.

\ J t z t . 2 . ? 6 .i { .

s i 0 2

A l ̂ 0 .z )T i 0 2

tlg0

Fe0,

Mn0

z n u

Ca0

Narc

Kzo

F

c l

T o t a l

3 7 . 7 6r 7 h q

1 . 3 3

I 4 . 9 6

r 5 . 1 0

0 . 00

0 . 0 r

0 . 0 0

o . 3 6

8 . 82

0 . 2 \

0 . 0 2

3 8 . 0 41 1 . 3 \

t . J o

1 5 . r 8

t 4 . q 0

0 . 0 0

0 . 0 9

0 . 0 0

o . 2 9

8 . 7 2

0 . 2 0

0 . 0 r

3 7 . \ 3

1 1 . 1 6

I . 5 4

r 4 . r 91 5 . 1 9

0 . 0 1

0 . 0 0

0 . 0 0

0 . 2 9

8 . 3 9

0 . 3 l {

0 . 0 3

38.391 6 . 5 8

1 . 3 9

r 4 . 0 r

0 . 06

0 . 00

0 . 0 0

o 2 6

9 . 0 0

0 - l 3

0 . 06

36 . \ \1 6 . 6 4

t . 7 3

1 r . 8 3

19 .82

0 . r 6

0 . 00

0 . 0 0

0 , l 5q ? ?

0 . 0 5

0 . 0 l

36.3716.26

1 t ? q

1 9 . 5 8o . l 2

0 . 00

0 - 00

0 . 1 3

9. ! {4

0 . 0 2

0 . 0 r

96 . 00 q q q ! 9 \ . 4 2 95 .97 9 5 . 1 l lq q ( q

FORI4ULA PROPORT I ONS

I V s i 2 . 8 0 2 2 . 7 8 1A l 1 . r 9 8 r . 2 1 9

Ar 0 .329 0 .296T i 0 . 0 7 4 0 . 0 7 4Ms 1 -655 l . 5 l r 3Fe 0. 9 l t I 0 934l . ln 0. 000 0. 000Z n 0 . 0 0 1 0 . 0 0 1L 3 . O O O 2 . 9 \ 7

C a 0 0 0 0 0 0 0 0N a 0 . 0 5 2 0 . 0 5 1K 0 . 8 3 5 0 . 8 2 9r 0 . 8 8 7 0 . 8 8 0

2 . 8 7 5 2 . 8 3 91 . 1 2 5 r . l 5 r

0 . 3 l t 8 0 . 2940 . 0 7 8 0 . 0 7 7| . 5 6 4 1 . 5 \ 5r . 0 0 5 0 . 9 9 40 . 0 0 4 0 . 0 0 40 . 0 0 0 0 . 0 0 03 . 0 0 0 2 . 9 1 3

0. 000 0. 0000 0 3 8 0 . 0 3 70 . 8 5 0 0 . 8 4 9a3es 0.T86

2 . 7 8 1 2 . 7 5 71 . 2 1 9 1 . 2 \ 3

0 . 2 1 9 o . 2 \ 2n n o o n n Q R

1 . 3 \ 6 ' t . 3 3 5

1 . 2 6 5 1 . 2 5 50 . 0 1 0 0 . 0 1 00 . 0 0 0 0 . 0 0 03 . 0 0 0 2 . 9 4 0

0 0 0 0 0 . 0 0 00 022 0 .0220 . 9 0 9 0 . 9 0 1o . 9 3 1 0 . 9 2 3

2 .8o7 2 .769r . 1 9 3 1 . 2 3 1

0 . 2 8 7 0 . 2 2 9o . 1 2 9 0 . 1 2 1I . 3 0 5 r . 2 8 8| . 2 7 1 1 . 2 5 30 . 0 0 8 0 . 0 0 80 . 0 0 0 0 . 0 0 03. OOO ' . 906-

0 . 0 0 0 0 . 0 0 00 . 0 r 9 0 . 0 r 90 . 9 3 0 0 . 9 1 7o . 9 [9 0 .9 i 6

2 .825 2 . . 8O1 2 .827 2 .8001 . ' t 7 5 ' t 1 9 9 | . 1 7 3 1 . 2 0 0

0 . 3 \ \ o . 3 0 7 0 . 3 5 5 0 . 3 1 , {o .076 0 015 0 .088 0 .087r . 6 8 1 1 . 6 6 7 r . 5 9 8 1 . 5 8 30 . 3 9 5 0 . 8 8 7 0 . 9 5 0 0 . 9 5 10 . 0 0 0 0 . 0 0 0 0 . 0 0 1 0 . 0 0 10 . 0 0 5 o . o o 5 0 . 0 0 0 _ o . o o 03 . 0 0 0 2 . 9 4 0 3 . 0 o 0 2 . 9 3 5

0 .000 0 .000 0 .000 0 0000 . 0 4 2 0 . 0 4 r 0 . 0 4 2 0 0 4 20 . 8 2 5 0 . 8 r 9 0 . 8 0 i j 0 . 8 0 1n.T63 o:SA b.Tr o.8'R-

V I

I . q ; A - 4 8 a b e t z - p 7 a g - G a r n - C u f f i - o i l p h - I f m ( 2 2 , 1 6 6 t ) 2 . 6 A - 4 8 u Q x z - P f a g - C u m - O a n p h - F b f - I l n ( 2 2 . 1 8 9 )

3 . 6 A - 4 8 2 Q t z - P l a g - G a f n - C u n m - H b 7 - r f m ( 2 2 . 2 0 ? ) 4 . G G U - 9 1 7 2 2 Q X z - P 7 a g - H b f - C u m - r 7 n ( 2 2 . 2 7 7 )

5 . U G A - 4 8 9 Q t z - P f a g - H b f - G a r n - I L n ( 2 2 . 7 8 9 ) 6 . O G A - 4 8 a Q t z - P f a 7 - H b 7 - I f n 2 2 . 3 0 0 )

*Minus oxggen fq F, cL **Based on 7 cat ions ( lst coLunn) and 7f oxggens (2nd cofaff i )

fNumbers i i patentheses corresponal to calcuTaxed xotaf posi t ive charge for the 7 caxion nornaf izat ion

within the field bounded by talc-phlogopite-eastonite.There is no major distinction between the two sets ofdata, although analyses from Itiv extend to lower valuesof A-site occupancy. The possibility that A-site occupan-cy increases with metamorphic grade was suggested byHoldaway (1980), and the data presented here are notinconsistent with that interpretation. In addition, the datain Figure 8 suggest substitution of up to -20%o of a talccomponent (10), and a study of biotite in talc-bearingrocks is in progress to investigate this problem further.

The "composition" of the A-site shows a definitegrade-dependent variation. As illustrated in Figure 9,granulite grade biotite (Laneil is characterized by higherIV(K+Na) compared to amphibolite grade biotite (Itiv).Analyses from the three lithologies under considerationdisplay a similar shift towards higher K-contents in thegranulite grade samples, making a bulk composition con-trol appear unlikely. If these data are in fact reflecting agrade dependent phenomenon, then they suggest a higherthermal stability for the K end-member. Alternatively,the data could be reflecting a crystal-chemical effect in

which high Ti-content precludes substantial substitutionof the smaller Na for K. Additional discussion of thispoint is presented in the following section'

Discussion

Previous studies of biotite have established the com-plex nature of cation substitutions in that mineral, but theprecise mechanisms for several of these remain uncer-tain. The data on Greenland biotites presented in theprevious section extend the observations made by otherinvestigators, and indicate that their contents of Al andTi, as well as the composition of their A site, change fromamphibolite to granulite grade, and are systematicallyrelated to rock composition and mineral assemblage. Inaddition, these new data suggest that Ti-Tschermak'ssubstitution is subordinate in these biotites, but the dataare consistent with either Ti-vacancy or Ti-oxy substitu-tion. Additional vacancy substitution is suggested by thepresence of AlvI not balanced by AlIv.

In this section, the chemical data for the Greenlandbiotites are summarized along with data from other

FeOrrMnO i fo 30 20 lO MgO

Fig. l. Composition of granulite grade biotite from Lang{ @)and amphibolite grade biotite from Itivdlinguaq (b) illustrated interms of relative concentrations of MgO, Al2O3 and FeOl +MnO. ln (b), analyses marked with the solid symbol (A) containhornblende as the dominant amphibole; these analyses are alsodistinguished in subsequent figures.

EE9

Fig. 3. Calculated total positive charge and Ti-content inbiotite from Lang4 @) and Itivdlinguaq (b), with formulaproportions based on the 7 cation normalization (A).


a

oo

6F

0E

-

J

sF

-1.00 l. l0 1.20 ., E l30 l. l0 150A t -

Fig. 2. Ti- and Allv-contents of biotite from West Greenlandwith formula proportions based on the I I oxygen normalization(B).

sources in an attempt to resolve this ambiguity of vacancyvs. oxy substitution. Thereafter, an iterative normaliza-tion procedure is presented, which utilizes the conclu-sions deduced from the cation substitutions, and permitsFe3+-estimates to be made from microprobe analyses ofbiotite. In addition, a model system is presented thatdiscusses the interrelationships among the various cationsubstitutional mechanisms, on the basis of which biotiteanalyses can be recast into selected end-members. Final-ly, the implications of the data are explored with respectto biotite crystal chemistry. The discussion begins, how-ever, with an examination of biotite in metamorphic rocksfrom selected other localities following the same generalmethods as for the Greenland biotites.

Comparison with other data

Several sets of electron microprobe analyses of meta-morphic biotite are considered in this section to deter-mine whether the observations on Ti- and Al-substitutiondescribed above have more widespread applicability.These data sets include biotite from the following associa-tions: metagraywacke and metapelite from a contactaureole, Yellowknife District (Ramsay, 1973); mafic tofelsic granulite grade gneisses from the Reading Prong,

(8) ITIVDLINGUAQ

F-M"$@G=;[|stlI o Semirel i t ic Schisl l

l^" a-pr,imrire ?

Tit(attr-t)=trz

I f IVDLINCUAO (8)O) ITIVOLIN6UAA

chor9eExcoss, z r i

O M u s c o v i t € s c h i g l

o Semlpel i t ic schislA A A m p h i b o l i i e


0

F

0Fo

z2

F

atyl

1 0 l l I 2 A I E r ' 3 1 . 4 l s

Fig. 4. Formula proportions of Alvr and AIIV in biotite fromLangl (a) and It ivdlinguaq (b) based on the 7 cationnormalization (A).

New Jersey (Dallmeyer, 1974b); staurolite, staurolite-sillimanite and sillimanite-potassium feldspar zone meta-pelites, Maine (Guidotti et al., 1975, 1977); cordierite-potassium feldspar and sillimanite-potassium feldsparzone metapelites, Bavaria (Blumel and Schreyer, 1977);and muscovite-sillimanite to sillimanite-potassium feld-spar zone metapelites, Massachusetts (Tracy, 1978).These samples span a somewhat larger range of metamor-phic grade than the ones from Greenland, and, except forthe samples of Ramsay (1973) and Dallmeyer (1974b),represent predominantly Al-saturated biotite from limit-ing assemblages.

Ti-contents and calculated total positive charge forthese data sets are illustrated in Figure 10, with formulaproportions based on normalization (A). All analysesyield a charge excess that-to a greater or lesser extentcorrelates with Ti-contenl (cf., Fig. 2). Except for thedata of Dallmeyer, calculated positive charge for all butone analysis falls above the Charge Excess = 2Ti lines.

Fig. 5. Calculated total positlve charge and Alvl-content inbiotite from Lang{ @) and Itivdlinguaq (b), with formulaproportions based on the 7 cation normalization.

Fig. 6. Ternary representation of mica compositions, with co-ordinates defined as follows:

S = SiOz + 2K2O + 2Na2O + 2BaO + 2CaO

A: 2AlzOt * 2Cr2O3 + zTioz - 2K2O - 2Na2O

- 2BaO - 2CaO

FM = FeOr + MgO + MnO + ZnO - TiOz

All compositions prefixed by "ferri" have Fe3* in lieu of Al3* inoctahedral sites. Ti-biotite is the Ti-vacancy end-member,siderophyllite is the ferrous iron analogue of eastonite, andceladonite is K(AlMgn)SirOro(OH)z; other compositions aregiven in the text. The stippled region in the trioctohedral micafield is expanded in Fig. 7.

(i l LANGA

a A ^s^ o 8 o

a aa r 4 . o " $ .

o o ^ P

A lEft.o .a o d f u - o ;

, mdg_ pa a - c _ o r p - : .

a d l

o cn f f i

trsilf,f G€bslo Kf Gneiss I

J o oprcEk II a PyriboliE I

'.r .

o Muscov i la sch is l

o Semipe l i t i c sch is^ a Amphibo l i te

cHAnoE EXCESS.t /a A tE-o 375

DYMEK : SUBSTITUTIONS 1N BIOTITE FROM WEST GREENLAND

ITIVDLINGUAO

E9l

P y r i b o l i l e

ITI VDLI NGUAO

";"i';BkeeAr+1

005 090 095 100 090K/ ( K * No)

Fig. 9. Ti us. K/(K+Na) in biotitesymbols as in previous figures.

095 I 00

from West Greenland;

Amphr bo l i l e-vFig. 7. Compositions of West Greenland biotites portrayed in

S-A-Fm coordinates, as indicated in Fig. 6. The data points alllie within the region bounded by phlogopite-eastonite-Ti-oxybiotite; the fact that the data are shifted off the phlogopite-eastonite join towards the S corner indicates the presence of a Ti-oxycomponent, a Ti-vacancy component, or a muscovitecomponent (c/., Fig. 6)

Similarly, except for Dallmeyer's data, all analyses haveAlvI > AIIV as shown in Figure ll. Thus, for therelatively Al-poor analyses of granulite grade biotite,calculated charge excess is compensated almost com-

K M S3SisA I Oto(O H)2 K( Mg2 Al )(S i2Alz) Oro(OH )2

0 ' ' o A I E l 3 o

Fig. 8. (a) Ko-Alrv variat ion diagram i l lustrat ing thesubsti tut ional relat ionships among talc, phlogopite, andeastonite. (b) Expansion of the stippled region shown in (a)illustrating A-site occupancy vs. Alrv in the Greenland biotites,with formula proportions based on the I I oxygen normalization(B). No systematic difference is apparent between the two datasets, although the analyses of biotite from Itivdlinguaq extend tolower A-site occuDancv.

pletely by Ti-substitution, whereas for the other data setsinvolving relatively Al-rich biotite, charge excess is com-pensated by both Ti and AlvI. These relationships (Figs.10 and I l) do not contradict any of the previously statedconclusions regarding Ti- and Al-substitution in theGreenland biotites, and appear to lend them strong sup-port.

Ti-vacancy vs. Ti-oxy substitution

The correlations between Ti and calculated positivecharge in biotite from West Greenland (Fig. 2) andelsewhere (Fig. l0) are ambiguous because they areconsistent with two interpretations that are not necessari-ly mutually exclusive: vacancies and oxycomponents.

T i

Fig. 10. Calculated total positive charge and Ti-contents inbiotite from the various localities cited in the text, with formulaproportions based on the 7 cation normalization (A).

z

" 1oo

S e m i p e l i l i c s c h i s l

230f

fft

892 DYMEK: SUBSTITUTIONS 1N BIOTITE FROM WEST GREENLAND

The graphical treatment of mica compositions (Fig. 7)resulted in the same ambiguity.

The data presented in this report together with observa-tions made previously by other workers strongly supportthe conclusion that the Ti-content of biotite coexistingwith a Ti-oxide phase increases with grade of metamor-phism, commonly reaching extreme values in granulitegrade rocks (>6 wt.%). A consequence of this observa-tion is that Ti increases the stability field of biotite.

It follows from this conclusion that either vacancy-formation or oxy-substitution promote an increased sta-bility for biotite. The formation of vacancies in biotitemay create a crystallographically favorable environmentfor increased Ti-substitution. An oxycomponent, on theother hand, might provide a partial explanation for theubiquitous occurrence ofa "hydrous" phase (i.e., biotite)in otherwise "anhydrous" rocks such as granulites. Stat-ed another way, Ti-substitution in biotite could involveprogressive "dehydration. "

Incorporation of Ti into biotite as an oxycomponentwould result in a distinct negative correlation between Tiand (OH)-, which is virtually impossible to test utilizingthe microprobe analyses presented here. Nevertheless,assuming that the diference between the analytical totaland l00Vo represents the H2O-content, the calculatedformula proportions of (OH)- were compared to Ti-content. Examination of the data revealed no correlationwhatsoever, but this may not be a meaningful test, due tothe large uncertainties involved.

Evaluation of analyses for which direct H2O determina-tions are available provide an alternative approach to thisproblem. Stevens (1946) observed that many wet chemi-cal analyses of mica minerals were deficient in (OH)-,and was the first to suggest that this deficiency is compen-sated by a Ti-oxycomponent. Foster (1964) likewise not-ed that calculated (OH)- contents of many trioctohedralmicas were less than the theoretical 2.0/formula, but alsoobserved that 02- contents showed a correspondingexcess. She concluded that an oxyannite component (9)could accomodate some, but not all of this deficiency.

Dodge er al. (1969) pointed out that the (OH)- deficien-cies discussed by Stevens (1946) might be due to incorrectanalyses of water content. Consequently, they took par-ticular care in carrying out H2O-determinations in theirstudy of biotite in granitic rocks from the Sierra Nevadabatholith. Despite this, all but one of their analysesyielded (OH+F+CI) < 2.0, and they also attributed thisdeficiency to an oxyannite component. An examinationof their analyses by the writer reveals no correlationbetween Ti and (OH)-.

Forbes (1972) also studied variation in the hydroxylcontent of biotite, and showed that a remarkable correla-tion exists (f : -0.96) between >(OH+F+CI) and thedifference in charge calculated for cations in octahe-dral and tetrahedral sites [i.e., Acharge : ) VI"5..g" -

) IV"h".g"l. The correlation between hydroxyl contentand interlayer site charge was very poor (r2 : -0.22),

indicating that <5Vo of the observed (OH)- variationcould be ascribed to interlayer cations. In order to explainhis observations, Forbes postulated the existence of"tetrasilicic phlogopite" [KMg:Si+Orr(OH)], which isrelated to phlogopite by the substitution: Sia+ : Al3+ +H+. Although Forbes was able to synthesize this com-pound, he determined that it was unstable with respect toother phases; its role in the compositional variation ofnaturally occurring biotites is unclear.

Experimental studies lend support to both vacancy-and oxy-substitution. For example, Forbes and Flower(1974) synthesized a Ti-vacancy biotite in the Mg-system,whereas Abrecht and Hewitt (1980, l98l) producedbiotites with various amounts of Ti-vacancy, Ti-oxy andTi-Tschermak's substitutions. The latter authors wereunable to find evidence for a vacancy substitution in theFe-system.

Hazen and Burnham (1973) considered Ti-substitutionin biotite and noted that the cation configuration(R2*Ti!)vISiIv yields an ideal -t2 charge contribution tonearest oxygen atoms. These same authors also noted (p.898) that their structural refinements of Pike's Peakannite indicated the presence of octahedral vacanciesapproximately equal in number to the amount of Ti.

Takeda and Ross (1975) reported a crystallographicstudy ofoxybiotite in a rhyodacite ash flow tufffrom theValles Mountains, New Mexico, and concluded thatsubequal amounts of an oxyannite component and a Ti-oxycomponent are present in this sample. Bohlen et a/.(1980) described a biotite in orthogneiss from the AuSable Forks area, New York, and concluded that anoxycomponent represented the dominant mechanism forTi-substitution in their sample.

Studies of the infrared spectra of micas have revealedseveral interesting features relevant to the present discus-sion. Such spectra are, in general, characteized by threesets of absorption bands termed N (normal), I (impurity)and V (vacancy) (Vedder, 1964). N- and I-bands areproduced by stretching of (OH) oriented approximatelynormal to the mica a-b plane (Serratosa and Bradley,1958). Three separable N-bands may occur, which show afrequency shift in response to Mg-Fe2+ variation; thesehave been ascribed to cation-hydroxyl configurationssuch as Mg3*-(OH), Mg?+Fe2+-(oH), and Mg2*Fe3*-(OH) (Wilkins,1967).I-bands are associated with variousR2*R3*-(OH) configurations (Wilkins, 1967), and theirfrequencies and intensities have been correlated to varia-tions in the abundance of (Al, Fe3*)tl (Rousseaux e/ a/.,1973).

V-bands display pronounced infrared dichroism, andare related to stretching of (OH) which lies approximatelywithin the mica a-b plane and is coordinated to twocations and a vacancy. Thus V-bands in biotite are notunlike absorption bands associated with the vacant site indioctohedral micas, such as muscovite (Vedder and Mc-Donald, 1963; Vedder, 1964).

Three overlapping, but resolvable V-bands are com-

monly observed, the intensities of which vary as afunction of mica composition (Vedder, 1964) and watercontent (Rouxchet, 1970). However, the cation configura-tions that yield V-bands are somewhat uncertain. Vedder(1964) initially suggested R2*-R2*-n(OH), R2*-R3+-!(OH) and R3+-R3+-E(OH). Farmer et al. (1971) notedthat the intensity of two of the V-bands is greatest inbiotite with high Al and Fe3*, and made specific assign-ments for Fe2*-Fe3*-!(oH) and Fe2*-Al3*-!(oH).Neither Vedder (1964) nor Farmer et al. (1971) consid-ered the explicit involvement of Ti in any of the observedspectra. However, Rousseaux et al. (1973) noted that Mg-rich biotite with low contents of Al and Fe+ displayed asingle strong V-band which they assigned to R2*-Ti4*-n(OH). They extended this assignment to other biotitesbased on the characteristic frequency of this absorptionband. The important point here is the spectroscopicidentification of (OH) absorption bands associated withoctahedral vacancies in biotites with a wide range ofchemical compositions.

Vedder and Wilkins (1969) carried out a series ofdehydroxylation studies on several micas, and noted thatthe V-bands disappeared from the biotite spectra attemperatures in the range 500-800'C, whereas N- and I-bands persisted to at least 900'C. They suggested that inbiotite, as in muscovite, two (OH)- near an octahedralvacancy condensed to form H2O, which then leaves themica by diffusion. In the case of biotite, (OH)- near avacancy is expelled long before those associated with acompletely filled octahedral site (Vedder and Wilkins,r96e).

These authors also noted that in situ oxidation of Fe2*to Fe3+, with concomitant loss of H2 [oxyannite (9)]might complicate this interpretation. However, the factthat V-band intensity could not be restored by hydrogena-tion confirmed that hydroxyl loss occurred by H2Ocondensation, and not by simple hydrogen loss. In con-trast, the intensities of N- and l-bands can be nearlycompletely restored by hydrogenation, indicating thar nolarge loss of (OH) by condensation from N or I "sites"has occurred, but rather by hydrogen loss associated withoxidation of Fe2*.

In summary, observations on variations of the H2O-content in biotite do not offer a solution to the problem ofTi-vacancy vs. Ti-oxy substitution, and most investiga-tors would seem to favor an oxyannite component as theprincipal source of this variation. The results of X-raycrystallographic and experimental studies do not estab-lish a unique mechanism for Ti-substitution either (and itmay turn out ultimately that more than a single one isinvolved). The results ofinfrared spectral studies supportthe vacancy substitution, whereas the dehydroxylationexperiments may provide a possible explanation for someevidence favoring the Ti-oxycomponent: Ti initially en-ters biotite via a vacancy substitution, and later (inresponse to some thermal event?), H2O is lost by (OH)-condensation producing an oxygen "excess." This mech-

893

o

H

0 3 0

r z o r 3 0l r E

Fig. I l. Formula proportions of Alvr and Alrv in biotite fromthe various locallties cited in the text, with formula proportionsbased on the 7 cation normalization (A); symbols as in Fig. 10.

anism could also explain the low water contents in somebiotite analyses.

Collectively, the observations outlined above do notcontraindicate a Ti-vacancy substitution, and the writertentatively concludes that it provides the best explanationfor Ti-substitution in the biotites considered here. Theremaining sections of this paper explore some conse-quences of this interpretation.

An iterative normalization

If Ti-substitution in biotite involves vacancies, as sug-gested here, then the 7 cation normalization procedure(A) is incorrect and should be abandoned. A more appro-priate cation-based normalization scheme would be:

Total Cations - (K+Na+Ca+Ba) + Ti = 7.0, (C)

which allocates a vacancy for each Ti cation. However,normalization (C) does not eliminate charge excess inmany analyses, particularly in those with large amountsof AlvI.

Therefore, a fourth normalization procedure can becarried out, which takes into account vacancies associat-ed with Al-substitution by means of the dioctahedralcomponent [substitution (2)]. The Al of interest in (2) isactually excess octahedral R3 *( :AIYJ), where:

AIYJ : (Al+Cr)vI - AlIv + (K+Na+2Ca+2Ba) (l l)

so that the fourth normalization takes the following form:

Total Cations - (K+Na+Ca+Ba) + Ti + lr\ryl = 7.0,(D)

which allocates one vacancy for every two excess octahe-dral Al ions in addition to the vacancy already allocatedto Ti. Normalization of analyses according to (D) elimi-nates calculated charge excess in all cases and permits a

DYMEK: SUBSTITUTIONS 1N BIOTITE FROM WEST GREENLAND

o r o

1 5 01 1 0


Table 7. Selected examples of biotite formula proportions calculated on the basis of the iterative normalization procedure.

4 . 9 .2 .

v r A r 0 . 3 1 6F e 3 + 0 . 0 q 2T i 0 . 1 2 3M s 1 . 7 6 5F e Z + 0 . 5 2 5M n 0 . 0 0 5Z n 0 . 0 0 0r iFiT

2 . 7 7 3 2 . 7 5 61 . 2 2 7 1 . 2 \ 40 . 0 0 0 0 . 0 0 0

o .283 o. 2090 . 0 6 5 0 . r 0 30 . 0 7 3 0 . 1 2 71 . 5 3 8 1 . 2 8 20 . 8 5 5 r . r q 50 . 0 0 0 0 . 0 0 80 . 0 0 r 0 . 0 0 0

l v s i 2 . 7 2 7 2 . 7 2 \A r 1 . 2 7 3 1 . 2 1 6Fe3+ 0 000 0.000

2 . 727 2 . 800 2 .5801 . 2 7 3 1 . 2 0 0 r . 2 8 90 . 0 0 0 0 . 0 0 0 0 . 0 3 1

0 . i l 4 0 . 0 9 3 0 . 0 0 00 . 2 5 7 0 . 1 6 7 0 . 4 1 50 . 1 4 3 0 . r 3 4 0 . 2 9 21 . 9 2 1 2 . 1 9 3 1 . \ 5 30 . 4 2 1 0 . 2 7 6 0 . 5 3 70 . 0 0 2 0 . 0 0 2 0 . 0 0 20 . 0 0 r 0 . 0 0 0 0 . 0 0 7T3r7 T.866 LTol0 . 0 0 r0 . 0 3 10 . 8 7 rb.3bz

2 .590 2 .79OL 3 1 0 1 . 2 1 O0 . 000 0 . 000

o . \ 3 1 o . 3 3 30 . r 0 0 0 . 0 1 20 . 0 7 5 0 . 0 5 91 . 2 7 3 r . 6 4 06 a r o n R 7 (

0 . 0 2 r 0 . 0 0 00 . 0 0 1 0 . 0 0 32 .837 2 .931 2 .927 2 .873

0 . 0020 . 0 r 00 . 9 0 r0 . 9 1 3

0 . r 9 r0 . 1 \ 20 . 2061 . 2 t 71 . 0 2 10 . 0 1 I0 . 0067:7rE0 . 0 0 00 . 0 0 00 . 9 4 40=--9-47

0 . 0000 .027o .920d:rE7

A C aN aKT

C h a r g e

0 . 0 0 0 0 . 0 0 00 . 0 3 5 0 . 0 0 30 . 9 0 5 0 . 9 0 rb. e-Id d:J-'dT

0 .000 0 . 0000 . 0 5 8 0 . 0 5 r0 .798 o .82603z6 oin

0 . 0000 . 0 r 9o . 9 1 3n q ? ?

21 . 958 2 r . 858 21 . 7 \3 21 . 833 21 . 553 21 . 900 2 r . 988 21 . 93 \ 21 . 897

f . l a b 7 e 1 , # 7 ( 6 A - 2 4 c s i f l - K f g n e i s s ) 2 . T a b f e 1 , # 5 ( 6 A - 2 7 K f g n e i s s ) 3 . T a b f e 2#4 (6A-28h opx gneiss) 4. Table 3, H7 (6A-23b P7riboi i te) 5. Tabfe 3, H6 (oCA-

36k PgriboTite) 6. Tabfe 4, #2 (eA-48q Nuscovite schist) 7. TabJe 5, #6 (OGA-

48\-2 Senipel i t ic schist) 8. Table 6, 4L ((EA-Aqab ArnphjboLite) 9- Tabfe 6, #6(&A-48a Hornbfende anphiboTite)

* Totaf posi t ive charge cafcuTated wit l ) af f i ron as Fe2*

rough estimate for Fe3+ [excluding that associated withoxyannite (9)1.

It turns out that normalization (D) works only for Al-rich biotite, as most Al-poor analyses yield negative Alf"r.This effect most probably reflects the presence ofoctahe-dral Fe3+, the amount of which is not known prior tonormalization. Nevertheless, the "negative magnitude"of AlY"l provides a minimum estimate of Fe3+ needed tobalance AlIv. Formula proportions for a selected numberofanalyses from Tables l-6 calculated from the iterativenormalization are listed in Table 7.

Model system for mica composition

The substitutional relationships discussed previouslyfor Ti and Al (excepting oxycomponents) are summarizedin a hypothetical mica composition polyhedron illustratedin Figure 12. Although this diagram was constructed forthe Mg-Al system, Fe2* and Fe3* components can besubstituted at each composition point where appropriate.

Figure 12 distinguishes those cation substitutions [(l),(2), (3), (6)l that are accessible directly from phlogopitefrom those Ti-substitutions [(4), (5)] which require analuminous component as well. Stated another way, [(2)-(5F(6)l and [(l)-(3)-(4)] are respective linear combina-tions of each other. Some of the substitutions shown inFigure 12 are in fact redundant, and it can be shown thatmany proposed substitutions in micas are linear combina-tions of an Al-Tschermak's component (1), a dioctohedralcomponent (2), a Ti-spinel component (5), and a talccomponent (10), coupled with the simple cation exchangecomponents Fe : Mg, K = Na, Al = Fe3+ and Al : Cr.A plagioclase component (NaSi : CaAl) and a simpleBa : Ca exchange would relate all brittle micas to thesystem shown in Figure 12.

For example, the Ti-Tschermak's (3) and TiAl : AlSi(4) substitutions are linear combinations of the Al-Tscher-mak's (l) and Ti-spinel (5) components [(3) : (5)-2(l) and(4) = (5)-(l)1. Robert (1976b) suggested that the followingsubstitution represents an independent method for pro-ducing octahedral vacancies:7

2 A l l v + M g t I : 2 S i I v + l v l (12)

Substitution (12) is, in fact, a linear combination of the Al-Tschermak's (l) and dioctohedral (2) components [(12) :(2)-2(r)).

Holdaway (1980) considered chemical formulae appro-priate for "averge" biotite compositions in medium- tohigh-grade pelitic schists, and was particularly concernedabout how to deal with (Ti, Al)vI. Although he noted (p.7 14) that " . . . octahedral sites for Ti-free biotite are 98Vofull and octahedral vacancies increase with Ti . . .", heremarked that the Ti-vacancy substitution (6) was proba-bly not a good assumption. By carrying out a linear leastsquares regression on 47 microprobe analyses of biotite[including the data of Guidotti et al. (1977) and Tracy(1978), which have also been considered here (c/., Figure10)1, he deduced the following complex substitutionalrelationship:

(R2*)11!, + (AP*)'Y[7 + trdso= GF*)i16o + nillz + KS5o. (13)

However, he considered (13) to represent the sum of twograde-dependent substitutions :

7 In this study, Robert also described the systematic increasein the intensity of the infrared V-band.

Fig. 12. Hypothetical mica composition polyhedron derivedfrom the substitutional components described in the text. Ingeneral, many substitutions in micas can be shown to be linearcombinations of (l), (2) and (5), so that even in this simplerepresentation, some of the substitutions are redundant.

(R'*)i{6o + (AP*)il[7: (Ti4*)l{60 + !il[, ( l4a)

(Rz*)ilL + ndso = !il! + Kfso. (l4b)

In terms of the substitutions considered here, (13) and(14) represent linear combinations of Al-Tschermak (l),dioctohedral (2), Ti-vacancy (6), and talc (10) components[(l4a) = (6) - 0.33(2) and (l4b) : 0.2s(2) - 0.5(l) -

0.5(10), such that (13) : (6) - 0.5(1) - 0.083(2) -

0.5(10)1. The additivity ofthe substitutions leading to (13)serves to illustrate why Tivl - nvt.

It was suggested to the writer in review that substitu-tions other than a talc component (10) may cause varia-tions in the occupancy of the interlayer site. Two of thesecould be:

KA + 1R2*;vt : nA + (R3*)vI, and

KA + 1R3n1vI : nA + (Ti4*)vI.

895

(15)

(16)

However, (15) and (16) are also linear combinations of theprincipal substitutions discussed here [(15) : (l) + (10)and (16) : (2)-(10H6)-(1)1.

The model system has an interesting consequence forTi substitution in micas. Specifically, the Ti-vacancysubstitution (6) is actually a linear combination of thedioctohedral (2) and Ti-spinel (5) components [(6) = (2f(5)1. Thus Ti-substitution in muscovite occurs directly by(5) whereas in biotite it is by a combination of (5) and (2).This interpretation helps to explain why the correlationbetween Ti-content and calculated positive charge is bestin biotite with low AIYI.

Using this model system, together with formula propor-tions derived from the iterative normalization method(D), and Fe3+ estimated from charge calculations, it ispossible to cast any biotite analysis into end-members.End-members calculated from the analyses listed in Table7 are presented in Table 8.

Cry s t al-c he mic al c o ns ide rations

It has been known for some time that the lateraldimensions of the octahedral and tetrahedral sheets inbiotite do not match precisely. Donnay et al. (1964)indicated that this structural misfit between the sheets isaccommodated primarily by a rotation of the tetrahedraabout [001] (see Fig. I of McCauley and Newnham, 197 l),and in phlogopite the tetrahedral rotation angle (a) isapproximately 8o. Hazen and Wones (1972) have dis-cussed how substitution of Fe2* (0.78A) for Mg2+ (0.724)causes the octahedral sheet to expand, which increases

DYMEK: SUBSTITUTIONS 1N BIOTITE FROM WEST GREENLAND

Table 8. Mica end-members calculated for analyses listed in Table 7.

E n d - f l e m b e r F o r m u l a ' I 2 3 . 4 . 5 . 6 - 7 , 8 . 9 .

( B )o ( rn3 )u ' { s ; , . ) r vo , o {oH) ,

( x )A ( r s r i B ) v | ( s i r n r ) l vo ,o1sn1 ,

1x ;A1 r9 r r "3 * ) v l { s i r n t r ) l vo ,o { oH) ,

{ r )A {n r rD )v l ( s i r n r ) l vo ,o 1q6 ; ,

1x ;A l r e re t ) v | { s i r n t r ) l vo ,o (on ) ,

{ x )A {ns r ) v | { s i r n t ) l vo ,o {o r i ) ,

( A ) T a l c

( 8 ) T i - P h l o g o p i t e

( C ) F e r r i - E a s t o n i t e

( 0 ) l { u s c o v l t e

( E ) E a s t o i t e

(F) Ph logop i te

0 . 0 0 0

3 1 . 6 1 9 . 1

\ 3 . 2 4 0 .

9 . 1 5 . 0

r 4 . 3 1 3 . 4

25.7 16 .7

0 0 0 0

l t 2 9 . 3

3 9 r 5 \ . 5

6 . 8

12-7

1 0 . 3

0 0 0 . 0

28.3 20 .9

45.5 \9 3

8 7

1 2 . 3

\ 2 1 \ 2

9 6

4 1 . 6

0 . 0

0 0

1 9 6

5 . 1

1 0 . 0

8 . 8

2 6 3

\ 2 - 1

1 3 4

6 9

1 , 2

0 . 0

33.2

4 5 - 3

1 2 3

1 . 3

Me /(Hs+F1)

lsl(n9+re2+)

Fe- lFe f

726 .5 r l

.738 5q4

937 .878

.739 .832

8zo .888

.521 .523

.6 \9 .638

.652 .65 \

986 .928

596 .553

-73o .518

5\6 .903

.507

.528

917

( A ) = t . o - ( K + N # c a + 8 a ) ;

( e ) = ( n t + c r ) v l - z x o ;

( B ) = r i ; ( c ) = ( r " 3 * ) l j r . , ( D ) = o 5 x t ( A t + c r + F e 3 + ) v r - ( e t l V - r - I " - z c " - 2 s " ) l ;

( F ) = 0 . 3 3 3 x I n 2 * - 3 x A - 2 x ( c + E ) - B l

Na, Ca, Ba grouped with K; Fe2+. Mn, Zn groupd vith Ng.

896

Fig. 13. Ti-content vs. Mg/Fe in biotite from West Greenland,with formula proportions based on the iterative cationnormalization (D); symbols as in previous figures.

both the structural misfit and a. Fe-substitution reaches alimit where the mean octahedral radius is -0.764, whichcorresponds to Fe/(Fe + Mg) - 0.67. At this point, o :0', and the tetrahedra constitute a regular hexagonalarray. Compositions with more iron are stable only ifthere is oxyannite substitution, i.e., the presence of Fe3*decreases the dimension of the octrahedral sheet andpromotes additional Fe2 | substitution.

Similarly, substitution of the smaller Tia* cation(0.6054) should promote increased Fe/Mg, or more prob-ably, high Fe/Mg favors increased Ti-substitution. Figure13 shows that in the Greenland biotites, Ti tends todecrease with increaing Mg/Fe, although there appear tobe two distinct trends in the Lang{ data, which areprobably related to diferences in Al-content. Since mostof the biotite analyses presented here are Ti-saturated,the relationships shown in Figure l3 are consistent withsome type of crystal-chemical control on Ti-substitutionby Fe/Mg. Thus the structural distortions caused byincreased Fe/Mg are partly offset by increasing Ti. Inaddition, the presence of octahedral vacancies associatedwith Ti should permit the necessary structural accommo-dations to occur more easily. However, the efects of Tion the structure of biotite must be considered in view ofthe fact that the tetrahedral sheet is not Si3Al as inphlogopite-annite, but distinctly more aluminous (c/.,Figs. 2 and 4).

ITTVDLTNGUAQ l l

1S\i r

q o e F _

o2o o25 o'o ,rr lr t luo,

o oo 045 o 50

Fig. 14. K/(K+Na) us. Fe/(Fe+Mg) in amphibolite gradebiotite from Itivdlinguaq, which shows that the most Mg-richbiotites aie also the most Na-rich.


1 0 0

Substitution of either the Al- or Ferri-Tschermak'scomponents causes expansion of the tetrahedral sheetand contraction of the octahedral sheet, with a resultingincrease in structural misfit. Although tetrahedral rota-tion occurs to compensate for this added misfit, therotation is limited by the minimum bond length associatedwith the interlayer cation, i.e., tetrahedral rotation cannotproceed beyond the point where the mean K-O bondlength decreases to 2.8A (Itewitt and Wones, 1975).These authors indicated that this limit to the rotationangle would then limit Al-substitution in K-biotite, butthey were able to synthesize a more aluminous biotite inthe Na system, in which case the smaller interlayer cationpermits a greater amount of tetrahedral rotation to occur.

The amount of Tschermak's components in the Green-land biotites, as gauged by AlIv (Figs. 2 and 4), is similarin analyses from both the amphibolite and granulite gradelocalities. However, the biotite from Itiv contains moreAlvI than that from Langl (Fig. 4), and the decrease inthe size ofthe octahedral sheet associated with this extraAlvl would necessitate additional tetrahedral rotation.The fact that the more aluminous biotite from Itiv is alsomore sodic (Fig. 9) suggests that the required extratetrahedral rotation is compensated by increased Na-substitution. In the granulite grade biotite from Lang{,which contains abundant Ti that should also increasestructural misfit, it may be unnecessary to incorporate thesmaller Na cation since contraction of the octahedra inthis case is not as great where the larger Ti is dominantover the smaller Al.

However, Fe/Mg in biotite may exert a more funda-mental control on K/Na. In samples from Langp, themost Mg-rich biotite contains the highest amount of Na,and in the analyses from ltiv, K(K + Na) increasesuniformly with increasing Fe/(Fe + Mg) (Fig. l4). There-fore, the expansion of the octahedral sheet and corre-sponding decrease in a caused by substitution of Fe forMg, which is reversed by Ti (and Al) substitution, iscounterbalanced slightly by a higher K-content in moreFe-rich biotite, since increasing K/(K + Na) also causes dto decrease.

Two representations of Ti and Alvl in the Greenlandbiotites are illustrated in Figure 15. In both diagrams,there is a broad general decrease of Ti with increasingAlvI. However, the positions of individual data points inFigure 15b, which is based on the iterative cation normal-ization procedure (D), are systematically shifted towardslower values of AlvI compared to Figure l5a, which isbased on the 7 cation normalization (A). Data points forthe I I oxygen normalization (B) plot at intermediatepositions.

The shift in Figure 15 occurs because the abundance ofeach cation is lowered using the iterative normalization:Si decreases, which increases AlIv, and Alvl decreases asa result. Therefore, the difference between Figures l5aand l5b is to some extent an "artifact" of the normaliza-tion process. Note however that analyses of Ti-rich

0,9

0 , 9

z

Mg/ Fe

DYMEK: SUBSTITUTIONS /N BIOTITE FROM WEST GREENLAND 897

biotite yield little to no Alvr in Figure l5b. In several ofthese analyses, (Si + Al) is less than 4.0, and, in general,the higher the Ti-content, the lower the sum of Al + Si.The data indicate up to 0. 15 cations of an additionalelement in tetrahedral coordination, and the most logicalcandidate is Fe3*. Czamanske and Wones (1973) havealso noted the possible occurrence of tetrahedral Fe3* inbiotite with high contents of Ti. Although high Ti inbiotite may be partly accommodated by (Fe3*)Iv, since itwould tend to decrease structural misfit, this assignmentis problematic due to the lack of reverse pleochroism inthese samples (c/., Farmer and Boettcher, 1981).

Summary and conclusions

This report has summarized in detail the principalcharacteristics of biotite in a range of rock types from twolocalities in West Greenland. Compared to amphibolitegrade biotite, those in granulite grade samples are charac-terized by higher Ti, similar AIIV, and lower Alvr; K/(K +Na) and possibly >(K + Na) are also higher in granulitegrade biotite. Chemical variation at each locality isstrongly dependent on rock composition and mineralassemblage.

Substitution of Ti, and to a lesser extent Al, onoctahedral sites in biotite is interpreted in terms ofvacancy-formation. Based on these conclusions, an itera-tive cation-based normalization procedure is describedthat eliminates charge excess associated with Ti and Al,and permits Fe3+ to be estimated from the relationship:Fe3* : 22.0 - TotalPositive Charge. Fe3* calculated bythis method ranges from 0.05-0.45 Fe1, with the largestamounts occurring in biotite from pyribolites, which aremagnetite-bearing, and the smallest amounts in musco-vite schists and sill-Kf gneisses, some of which aregraphite-bearing. In addition, formula proportions calcu-lated from the iterative normalization for biotites withhigh TiO2 (>5 wt.Vo), may indicate the occurrence of(Fe3*)Iv.

A mica composition polyhedron is presented whichshows how complex substitutional mechanisms can bereduced to combinations of a few simple types, and thatTi-substitution in muscovite and biotite lead to the sameend-member composition. Mica end-members were con-structed utilizing these simple substitutions, and calculat-ed Fe3*.

The formulation of biotite analyses into end-members,as well as the methods for estimating Fe3*, assume ananion framework with a net charge of -22.00 and nooxycomponents, although oxyannite substitution shouldnot affect any ofthe conclusions reached here. Additionalwork on Ti-rich biotite, which combines high-qualityhydrogen analyses with spectroscopic studies (infrared,optical, Mcissbauer), together with X-ray crystallographicstudies on site populations, is needed to explore furtherthe interrelationships among Ti-content, vacancies andoxycomponents, and to evaluate the possible occurrenceof tetrahedral Fe3*.

(A) tonxl lrznrrol "r"

. ^ r ' 'a

'^ '^l

l rE

0 . 3

- 0 . 2

0 . 1

Fig. 15. Ti vs. AlvI in biotite from West Greenland. In (a)

formula proportions are based on the 7 cation normalization (A),and in (b) on the iterative cation normalization (D); note thesystematic shift towards lower Alvr when the iterativenormalization is used.

Acknowledgments

The initial study of biotite in the Langp gneisses comprisedpart of a Ph.D. dissertation at the California Institute of Technol-ogy carried out under the supervision of A. L. Albee, with thefinancial support of NSF grant DES 75-03417 . Completion of themanuscript was aided by the Clark Fund of Harvard University,and NSF grants EAR 78-23412 and EAR 8l-09466. Support foradditional field work in Greenland was provided by a grant fromthe National Geographic Society.

The writer wishes to thank J. B. Thompson for discussions oncation substitutions, and for suggesting the format of Figure 6. S.Dobos and C. A. Francis read an early version of the manuscript;thorough reviews by S. W. Bailey, D. A. Hewitt and D. R.Wones resulted in clarification of several issues, and are mostappreciated. The writer also thanks V. R. McGregor. GeologicalSurvey of Greenland, for providing some of the samples used inthis study (prefixed GGU).

References

Abrecht, J. and Hewitt, D. A. (1980) Ti-substitution in syntheticFe-biotites. (abstr.) Geological Society of America Abstractswith Programs, 12,377.

Abrecht, J. and Hewitt, D. A. (1981) Substitutions in syntheticTi-biotites. (abstr.) Geological Society of America Abstractswith Programs , 13, 393 .

Albee, A. L. and Ray, L. (1970) Correction factors for electron

(BJ NORMALIZATION . .0"

r.'*@6*Ssu

I I I

898

probe microanalysis of silicates, oxides, carbonates, phos-phates, and sulfates. Analytical Chemistry, 42, 1408-1414.

Bence, A. E. and Albee, A. L. (1968) Empirical correctionfactors for the electron microanalysis of silicates and oxides.Journal of Geology, 76, f82-403.

Blumel, P. and Schreyer, W. E. (1977) Phase relations in peliticand psammitic gneisses of the sillimanite-potash feldspar andcordierite-potash feldspar zones in the Moldanubicum of theLam-Bodenmais Area, Bavaria. Journal of Petrology, 18, 43 l-459.

Bohlen, S. R., Peacor, D. R., and Essene, E. J. (1980) Crystalchemistry of a metamorphic biotite and its significance inwater barometry. American Mineralogist, 65, 55-62.

Burns, R. G. and Vaughan, D. J. (1975) Polaized electronicspectra. In C. Karr, Ed., Infrared and Raman Spectroscopy ofLunar and Terrestrial Minerals, p. 39-72. Academic press,New York.

Carman, J. H. (1974) Synthetic sodium phlogopite and its twohydrates: stabilities, properties, and mineralogic implications.American Mineralogist, 59, 261-273.

Champion, D. E., Albee, A. L., and Chodos, A. A. (1975)Reproducibility and operator bias in a computer controlledsystem for quantitative electron microprobe analysis. pro-ceedings of the l0th National Conference on Electron probeAnalysis, Las Vegas, 55.

Czamanske, G. K. and Wones, D. R. (1973) Oxidation duringmagmatic diferentiation, Finnmarka Complex, Oslo Area,Norway: Part 2, The Mafic Silicates. Journal of Petrology, 14,349-380.

Dahl, O. (1969) Irregular distribution of iron and magnesiumamong coexisting biotite and garnet. Lithos, 2, 3ll-!22.

Dallmeyer, R. D. (1974a) The role of crystal structure in control-ling the partitioning of Mg and Fe between coexisting garnetand biotite. American Mineralogist, 59, 201-203.

Dallmeyer, R. D. (1974b) Metamorphic history of the northeasr-ern Reading Prong, New York and northern New Jersey.Journal of Petrology, 15,325-360.

Dawson, J. B. and Smith, J. V. (1977) The ulnro (mica-amphibole-rutile-ilmenite-diopside) suite of zenoliths in kim-berlite. Geochimica et Cosmochimica Acta. 41. i09-323.

Delaney, J. S., Smith, J. V., Carswell , D. A., and Dawson, J. B.(1980) Chemistry of micas from kimberlites and zenoliths-Il.Primary- and secondary-textured micas from peridotite zeno-liths. Geochimca et Cosmochimica Acta. 44,85'l-872.

Dodge, F. C. W. , Smith, V. C., and Mays, R. E. (t969) Biot i tesfrom granitic rocks of the central Sierra Nevada Batholith,California. Journal of Petrology, 10, 250-27 l.

Donnay, G., Donnay, J. D. H., andTakeda, H. (l9g) Trioctohe-dral oneJayer micas. II. Prediction of the structure fromcomposition and cell dimensions. Acta Crystallographica, 17,1374-138t.

Dymek, R. F. (1978) Metamorphism of the Archaean MaleneSupracrustals, Godthib District, West Greenland. In proceed-ings ofthe 1978 Archaean Geochemistry Field Conference, p.339-342. University ofToronto Press, Toronto, Canada.

Dymek, R. F. and Albee, A. L. (1977') Titanium and aluminum inbiotite from high-grade Archaean gneisses, Lang@, WestGreenland. Transactions, American Geophysical Union(EOS), 58, 525.

Engel, A. E. J. and Engel, C. (1960) Progressive metamorphismand granitization of the major paragneiss, northwest Adiron-dack Mountains, New York, Pt. 2, Mineralogy, GeologicalSociety of America Bulletin, 71, l-58.


Eugster, H. P. and Wones, D. R. (1962) Stability relations of theferruginous biotite, annite. Journal of Petrology, 3,82-125.

Evans, S. and Raftery, E. (1980) X-ray photoelectron studies oftitanium in biotite and phlogopite. Clay Minerals ,15,2W-218.

Farmer, G. L. and Boettcher, A. L. (1981) Petrologic andcrystal-chemical significance of some deep-seated phlogo-pites. American Mineralogist, 66, 1154-1163.

Farmer, V. C., Russel, J. D., McHardy, W. J., Newman,A. C. D., Ahlr ichs, J. L., and Rimsaite, J. Y. (1971) Evidencefor loss of protons and octahedral iron from oxidized biotitesand vermiculites. Mineralogical Magazine, 38, l2l-1i7.

Faye, G. H. (1968) The optical absorption spectra of iron in six-coordinate sites in chlorite, biotite, phlogopite and vivianite.Some aspects of pleochroism in the sheet silicates. CanadianMineralogist, 9, 403-425.

Faye, G. H. and Hogarth, D. D. (1969) On the origin of reversepleochroism of a phlogopite. Canadian Mineralogist, 10,25-34.

Forbes, W . C. (1972) An interpretation of the hydroxyl contentsof biotites and muscovites. Mineralogical Magazine, 38,712-720.

Forbes, W. C. and Flower, M. F. J. (1974) Phase relations oftitan-phlogopite, KzMgiTiAlzSi6O26(OH)a: A refractory phasein the upper mantle? Earth and Planetary Science Letters,22,60-66.

Foster, M. D. (1960a) Layer charge relations in the dioctahedraland trioctahedral micas. American Mineralogist, 45, 383-398.

Foster, M. D. (1960b) Interpretation of the compositions oftrioctahedral micas. U.S. Geological Survey Professional Pa-per 354-8, 1l-48.

Foster, M. D. (1964) Water content of micas and chlorites. U.S.Geological Survey Professional Paper 474-F.

Goldman, D. and Albee, A. L. (1977) Correlation of Mg/Fepartitioning between garnet and biotite with r80/r60 partition-ing between quartz and magnetite. American Journal of Sci-ence.277. 750J67.

Guidotti, C. V., Cheney, J. T., and Conatore, P. D. (1975)Interrelationship betwen Mg/Fe and octahedral Al content inbiotite. American Mineralogist, 60, 849-853.

Guidott i , C. V., Cheney, J. T., and.Guggenheim, S. (1977)Distribution of titanium between coexisting muscovite andbiotite in pelitic schists from northwestern Maine. AmericanMineralogist, 62, 438-448.

Hazen, R. M. and Burnham, C. W. (1973) The crystal structureof oneJayer phlogopite and annite. American Mineralgoist, 58,889-900.

Hazen, R. M. and Wones, D. R. (1972) The efect of cationsubstitutions on the physical properties of trioctohedral micas.American Mineralogist, 57, 103-129.

Hewitt, D. A. and Wones, D. R. (1975) Physical properties ofsome synthetic Fe-Mg-Al trioctohedral biotites. AmericanMineralogist, 60, 854-862.

Holdaway, M. J. (1980) Chemical formulae and activity mod€lsfor biotite, muscovite, and chlorite applicable to pelitic meta-morphic rocks. American Mineralogis t, 65, 7 l l-7 19.

Holdaway, M. J. and Lee, S. M. (1977) Fe-Mg cordierite stabilityin high-grade pelitic rocks based on experimental, theoreticaland natural observations. Contributions to Mineralogy andPetrology, 63, 17 5-198.

Keusen, H. R. and Peters, Tj. (1980) Preiswerkite, an Al-richtrioctohedral sodium mica from the Geisspfad ultramafic com-plex (Penninic Alp). American Mineralogist , 65 , ll34-l137 .

Kunitz, W. (1936) Beitrag zur kenntis der magmastischen Asso-

DYMEK: SUBSTITUTIONS IN BIOTITE FROM WEST GREENLAND 899

ziationex. III. Die Rolle des Titans und Zirkoniums in dengesteinbildenden Silikaten. Neues Jahrbuch fiir Mineralogie,Geofogie und Palbontolo gie, 70, 385-416.

Kwak, T. A. P. (1968) Ti in biotite and muscovite as anindication of metamorphic grade in almandine amphibolitefacies rocks from Sudbury, Ontario. Geochimica et Cosmochi-mica Acta. 32. 1222-1229.

Ludington, S. and Munoz, J. L. (1975) Application of fluor-hydroxyl exchange data to natural micas. (abstr.) GeologicalSociety of America Abstracts with Programs, T, 1179.

McCauley, J. W. and Newnham, R. E. (1971) Origin andprediction of ditrigonal distortions in micas. American Miner-alogist, 56, 1626-1638.

Miijano, T. and Miiiano, S. (1982) Ferri-annite from the DalesGorge Member iron-formations, Wittenboom area, WesternAustralia. American Mineralogist, 67, 1179-1194.

Ramsay, C. R. (1973) Controls of biotite zone mineral chemistryin Archaean metasediments near Yellowknife, Northwest Ter-ritories, Canada. Journal of Petrology, 14, 467-488.

Robert, J. L. (1976a) Phlogopite solid solutions in the systemK2O-MgO-AI2O3-SiO2-H2O. Chemical Geology, 17, 195-212.

Robert, J. L. (1976b) Titanium solubility in synthetic phlogopitesolid solutions. Chemical Geology, 17, 213-227.

Rousseaux, J. M., Laverde, C. G., Nathan, Y., and Rouxhet,P. G. (1973) Correlation between the hydroxyl stretchingbands and the chemical composition of trioctohedral micas.Proceedings of the International Clay Conference (Madrid),89-98.

Rouxhet, P. G. (1979) Hydroxyl stretching bands in micas: Aquantitative interpretation. Clay Minerals, 8, 375-388.

Rutherford, M. J. (1973) The phase relations of aluminous ironbiotites in the system KAlSi3Os-KAlSiO4-Al2O3-Fe-O-H.Journal of Petrology, 14, 159-180.

Serratosa, J. M. and Bradley, W. F. (1958) Determination of theorientation of (OH) bond axes in layer silicates by infraredabsorption. Journal of Physical Chemistry, 62, 1164-1167.

Shannon, R. D. (1976) Revised effective ionic radii and systemat-ic studies ofinteratomic distances in halides and chalogenides.Acta Crytallographica, A32, 7 5l-:7 67 .

Spear, F. S., Hazen, R. M., and Rumble, D., I I I (1978) Sodiumtrioctohedral mica from the Post Pond Volcanics, Vermont.Transactions, American Geophysical Union (EOS), 59, 1216.

Spear, F. S., Hazen, R. M., and Rumble, D. I I I (1981) Wonesite:a new rock-forming silicate from the Post Pond Volcanics,Vermont. American Mineralogist, 66, 100-105.

Stevens, R. E. (1946) A system for calculating analyses of micasand related minerals to end members. U.S. Geological SurveyBulletin 950, l0l-127.

Strunz, H. (1970) Mineralogische Tabellan. Geest und Portig,Leipzig.

Takeda, H. and Ross, M. (1975) Mica polytypism: Dissimilaritiesin the crystal structure of coexisting lM and 2M, biotite.American Mineralgoist, 60, 1030-1040.

Tracy, R. J. (1978) High grade metamorphic reactions and partialmelting in pelitic schist, west-central Massachusetts. Ameri-can Journal of Science, 278, l5O-178.

Tracy, R. J. and Robinson, P. R. (1978) Metamorphic isogradmapping in central Massachusetts and the study of changingmineral compositions in metamorphism. Geological Society ofAmerica Abstracts with Programs, 10, 89.

Vedder, W. (l9g) Correlations between infrared spectrum andchemical composition of micas. American Mineralogist, 49,7t6-:767.

Vedder, W. and McDonald, R. S. (1963) Vibrations of the OHions in muscovite. Journal of Chemical Physics, 38, 1583-1590.

Vedder, W. and Wilkins, R. W. T. (1969) Dehydroxylation andrehydroxylation, oxidation and reduction of micas. AmericanMineralogist, 54, 482-509.

Wilkins, R. W. T. (1967) The hydroxyl-stretching region of thebiotite mica spectrum. Mineralogical Magazine, 36, 325-333.

Wones, D. R. (1963a) Phase equilibria of "ferriannite," KFe]*Fe3*Si3Oro(OH)2. American Journal of Science, 261, 581-596.

Wones, D. R. (1963b) Physical properties of synthetic biotites onthe join phlogopite-annite. American Mineralogist, 48, 1300-1 3 2 1 .

Wones, D. R. and Eugster, H. P. (1965) Stability of biotite:Experiment, theory, and application. American Mineralogist,50, 1228-1272.

Manuscript received, March 3, l98l;accepted for publication, February 21, 1983.

AppendixAnalvtical techniaues

All of the data utilized in this study represent completeanalyses for up to 14 elements (Na, Mg, Al, Si, K, Ca, Ti, Cr,Mn, Fe, Zn,Ba, F, Cl). Most of the analyses were performed atCaltech on an automated 3 spectrometer rralc-5 sAE electronmicroprobe interfaced to a DEc pDp-8/E minicomputer for controland online data reduction. Additional analyses were performedat Harvard on a 3 spectrometer ARL EMx-sM microprobe usingKRIsEL automation and interfaced to a DEc pop-l l /05

minicomputer for control and onJine data reduction. X-rayintensities were converted to oxide weight percentages byemploying the methods of Bence and Albee (1968), andcorrection factors modified from Albee and Ray (1970).Operating conditions were 15 kV accelerating potential and 0.05p,A (CIT) or 0.02 p,A (HU) sample current (on brass); a nearlyidentical set of simple oxides and silicates was used as primarystandards in both laboratories. At CIT counting times rangedfrom 15-90 seconds and were controlled by the computer to yieldprecision of better than llo for oxides with abundances of >lwt.Vo, andbetter than 5%o for oxides in the range of 0.1-l.0wt.Vo.At HU, individual measurements were terminated after 30seconds counting time or 60,000 accumulated counts, whichresults in a statistical precision similar to that achieved at CIT.Based on repetitive analyses of the same two internal standardsin both laboratories, accuracy of the analyses is judged to bebetter than -2Vo (relative) for major elements and 2-5Vo(relative) for minor elements (c/., Champion et al.,1975).

Titanium, aluminum and interlayer cation substitutions … · American Mineralogist, Volume 6E,...

Documents

Transcript of Titanium, aluminum and interlayer cation substitutions … · American Mineralogist, Volume 6E,...