American Mineralogist, Volume 67, pages 956-974, 1982
The thermodynamic properties of topaz solid solutionsand somepetrologic applications
Manr D. BenroNr
Department of the Geophysical Sciences
The University of Chicago
Chicago, Illinois 60637
Abstract
Reversedexperiments were performed on the reaction andalusite + water : hydroxyltopaz component(in solid solution) in the intervals 600-900'C and 650-2000bars using
conventional cold seal apparatus.The results allow discrimination among various activity
models for topaz solid solutions and calculation of the thermodynamic properties of
(hypothetical) hydroxyl-topaz. A proton-avoidance activity model agrees well with the
experimental data and gives reasonableextrapolations to lower temperatures. Combined
thermodynamic data for hydroxyl-topaz (OHT), fluor-topaz (FT) (from Barton et al.,
1982).and the solid solution are:
GFm.:ga,rJ
S!q.1J/K-mol
Co J/K-mol
a
oHT
-2693251
112.04
- 0.08737T
504.413
+
*ot l(l - 2Xour)
5
- 5869.53r-0
135833012
FT
-2910660
105.40
47r.414- 0.08165r+
zXFr - |
12694707-2- 5495.547-0s
Calculationsbased on these data are in good agreementwith observationof natural
assemblages.In the presenceof topaz, muscovite may be stable on the solidi of some
granitesto pressuresas low as 1-2 kbkar. Calcic plagioclaseis incompatiblewith topaz,
especially in the presence of potassium feldspar. Hydroxyl-topaz contents of topaz
coexistingwith other aluminousphasesincreaseswith increasingpressureand decreasing
temperature. With decreasing temperature at constant composition topaz removes HF
from a water-rich fluid.
which limits its usefulness,while topaz contains
(Al2O3and SiOz)nearly alwayspresent
Topaz (AI2SiO4(F,OH)) occurs as an accessory components
mineral in some granites and associatedhydrother- in associatedphases.
Topaz is a solid solution between fluor-topaz
mally altered rocks as well as in pegmatites,rhyo(Al2SiO4F2)
and (hypothetical) hydroxyl-topaz (Al2
lites, and other aluminous rocks. As one of the
Natural topaz compositionsrangefrom
principal carriers of fluorine (along with fluorite, SiO4(OH)2).
pure
in rhyolites, to about Al2
nearly
fluor-topaz,
apatite,amphiboles,and micas),topaz offers a key
deposits,although
in
hydrothermal
to understandingthe role offluorine and, to a lesser SiO4F1.4OHs.6,
(1972a)
topaz with
has
Rosenberg
synthesized
extent, water in the genesis of these rocks. In
greater
Othhydroxyl-topaz
component.
than
507o
contrastto the micasand the amphiboles,topaz has
natural
toexchange,
er
than
fluoride-hydroxyl
a simplecompositionand thus is an easierphaseto
pazes
variation.
Iron
exhibit
little
compositional
deal with experimentally and theoretically. Fluoprobably substitutingfor aluminum,
rite, while chemicallysimple,is commonly the only and chromium,
(<0.05 wt.Vo) in some
calcium-rich mineral in high-fluorine environments occur in small amounts
(Ribbe
and
Rosenberg,l97l;Deer
rhyolitictopazes
(1973a)synthe1962).
Glyuk
and
Anfilogov
rCurrentaddress:Geophysical
Laboratory,CarnegieInstitu- et al.,
green
topaz
which
they
thought
containeda
D.C. sized
tionof Washington,
2801UptonStreet,N.W.,Washington,
amount
of
iron,
and
Beus
and Dikov
considerable
20008.
Introduction
0003-004)v82l09t
0-0956$02.00
956
BARTON: TOPAZ SOLID SOLUTIONS
(1966)synthesizedan aluminum-ironfluoro-silicate
which they thoughtwas related totopaz. Rosenberg
(1972a)synthesizedtopazes in silica-deficientassemblagesinvolving the probableexchangeSi4* +
02- : Al3* + F-. This substitutionis very minor, if
presentat all, in natural topazes.
Ribbe (1980) discussesthe crystal structure of
topaz, which was determined independently by
Pauling (1928) and Alston and West (1928) and
refined by Ribbe and Gibbs (1971). Briefly, the
structureis basedon alternatingclose-packedanion
layers of composition F2O and O and contains
crankshaft chains of edge-sharingAlOaF2 octahedra. Isolated SiO4 tetrahedraconnect the octahedral chains.Ideal topaz has the spacegroup pbnm
in which all the fluorine-hydroxyl sites are equivalent. Fluorine-hydroxyl ordering was proposed by
Akizuki et al. (1979) to explain the anomalous
optical propertiesobservedin sometopazes.Structure refinements of OH-bearing topaz by Parise
(1980)discussedby Ribbe (1980)confirmed that in
low-temperaturetopaz the hydroxyl anion may be
orderedonto as few as one of the eight positionsin
the topaz unit cell.
In all topazes the crystallographic and optical
properties vary significantly with composition.
Thesevariationsform the basisof severalmethods
of compositionaldetermination(Ribbe and Rosenbery,l97l).
While there are abundant data in the literature on
the natural occurrencesand chemographyoftopaz
(seethe Applicationssectionbelow), relativelylittle
experimental work has been done. Freidel and
Sarasinin 1887first synthesizedtopaz (Morey and
Ingerson, 1937). Most work since then has also
dealt with synthesis,intentional (e.g. Michel-Levy
and Wyart, 1946;Coes, 1955)or unintentional(e.g.
Althaus, 1966; Yoder and Eugster, 1955).Topaz
figures prominently in experiments on the "system" granite-HF (Glyuk and Anfilogov, 1973a)and
in studiesof beryllium mineral stabilities(Beusand
Dikov, 1967).In neither of thesestudies,however,
was equilibrium demonstrated(e.g. by reversal of
reactions),nor were the topaz compositionsdetermined.Rosenberg(1972a,1978)has done extensive
experimentson topaz in the system Al2O3-SiO2H2O-F2O-r in which he has determinedthe compositions.Unfortunately,most were synthesisexperiments and as such are inappropriate for extracting
thermodynamicdata. A few experimentsof a quasireversed nature in Rosenberg (1972a) did yield
results consistent with the direct svnthesis data.
Synthesisexperiments(Rosenberg,1978,personal
communication)give the equilibrium constantas a
function of temperature for fluoride-hydroxyl exchange between topaz and aqueous vapor in the
presenceof quartz.
In order to quantitatively understand the petrologic role of topaz, phase equilibrium and/or calorimetric data are needed.Barton et al. (1982)presented calorimetric data for fluor-topaz. ln this study
hydroxyl-buffering experimentsprovide data which
allow calculation of the thermodynamic properties
of topaz solid solutions. In turn, calculationsare
made with these data to help interpret the conditions of formationof somegreisens,pegmatites,and
relatedrocks.
Experimental methodsand results
The experimentalportion of this study consistsof
hydoxyl-buffering experiments which permit estimation of an entropy and a standard enthalpy of
formation for hydroxyl-topaz and also allow evaluation of alternative activitv models for the solid
solution.
Starting materials
A reliable method for the determination of topaz
compositionsis critical to the experimentson topaz
solid solutions.Ribbe and Rosenberefl9Tl) developed several X-ray and optical determinative
curvesfor topaz. In this study, after unsatisfactory
preliminary results using their As21method, cell
refinements were done for each topaz used for
starting material and for each run product. The
volume and b cell edge regression equations of
Ribbe and Rosenbergwere then used to determine
the compositions (the b cell edge equation was
modified slightly to split the difference between the
composition predicted by Ribbe and Rosenbergfor
the Topaz Mountain topaz and the wet chemical
analysis done by Penfield and Minor, 1894). As
expectedthe two regressionequationsgave similar
results. Crystallographicdata were collectedusing
powder diffractometer methods (Ni-filtered CuKo
radiation with a corundum internal standard c :
4.75929(B)4, c : 12.99168(49)A).
The program of
Burnham (1962)was used to refine the data. Except
for the starting materials the same set of peaks was
used in all the refinements (hkl
114,133,
232,143,153,006,303
,320,223).
Akizuki et al. (1979)suggestthat the orderingthat
they postulate to explain the anomalous optical
properties may also affect the X-ray determinative
958
BARTON: TOPAZ SOLID SOLUTIONS
clase, muscovite, and orthoclase (Olsen, 1971).
Brown gemmy crystals were obtained from the
Field Museum (FMNH #13130). Inclusion-free
fragments
of andalusitefrom the Black Hills, South
Comp.
Dakotawere alsoused(specimencollectedby P. B.
u <R-r xOHT
a (A)
b (A)
Material2
s (A)
Moore).
(TM)
4.6475(3) 8.1891(4) 8.3920(4) 342.81(2) 0.0r4
Topaz
Electron microprobe analysis was done on an
4 . 6 4 6 5 ( 4 ) 8 . 7 8 9 7 ( 1 3 )8 , 3 9 1 8 ( 8 ) 3 4 2 . 7 r ( 5 ) 0 . 0 1 4
automatedARL-EMXinstrument at the University of
(Ol,I)
4 . 6 s 4 4 ( 6 ) 8 . 8 0 9 4 ( 9 ) 8 . 3 9 0 0 ( 7 ) 3 4 4 . 0 r( 4 ) 0 . r 0 8
lopaz
Chicagousing a ZAF correction program written by
4 . 6 5 2 7 ( 4 ) 8 . 8 0 8 8 ( 1 1 )8 . 3 9 0 0 ( 6 ) 3 4 3 . 8 6 ( 4 ) 0 . r 0 5
I. M. Steele of the University of Chicago. Beam
0
.
1
8
1
3
4
4
.
8
6
(
2
)
(BR)
4.6594(3) 8.8248(5) 8.387r(4)
Topaz
4,6596(3) 8.8252(8) 8.3885(s) 344.9s(3) o.r83
currents were 1.5 p,A at acceleratingvoltages of 15
(o?)
4 . 6 6 s 0 ( 1 0 )8 . 8 4 0 8 ( 1 8 )8 . 3 8 4 9 ( 9 ) 3 4 5 . A ] . ( 7 ) 0 . 2 5 7
Topaz
kV.
Andalusite servedas the standardfor aluminum
4 , 6 6 3 0 ( 5 ) 8 . 8 3 7 r ( r 6 )8 . 3 8 r s ( 1 6 )3 4 5 . 3 1( 6 ) O . 2 4 0
and silicon. The water and fluorine analysesfor the
(SD) 7 . 7 9 2 8 ( r 4 ) 7 . 8 9 8 0 ( r 8 )5 . 5 5 1 9 ( 1 9 )3 4 1 .8 3( 1 2 )
Andalusite
Topaz Mountain topazare from Penfield and Minor
(1894)and Barton et al. (1982).The Topaz Moun1 Errots
(1d)
given in patentheses
_
taintopaz was used as the microprobe standardfor
=
=
=
=
Pleto'
ouro
omi, BR
Brewer uine, OP
TM
oM
Tpaz llountain,
SD = South Dakota
the fluorine analysesin the other samplesof topaz.
3
equa!1on of Ribbe and
from the b cell
edge regtession
calculated
The precision of the fluorine analyseson the other
agree with the
sllghtIy
to bettel
Roseflberg (1971) modified
of Barton e! af-' (1982) for ToPaz Mountalf,
topazeswas approximately +l07o of the amount
weE chemlcal analysis
to Paz
present. As a result the fluorine and hydroxyl
contents for these samples are taken to be those
obtainedfrom the b cell edgedeterminative curve of
curves. The optically anomaloustopaz studied by Ribbe and Rosenberg. Although other elements
Akizuki et al. camefrom Ouro Preto, Minas Gerais, were searchedfor none was found. Table I shows
Brazil where the mineral associations(Olsen, 1971) the chemicaland crystallographicdata.
suggest a low temperature of formation. Sector
growth ordering, such as Akizuki er a/. suggest,is
Hy drory l- buffering exp erime nt s
more likely at low temperature; and the magnitude
Hydroxyl-buffering experiments offer informaof such ordering at the temperatures of the experimentsin this study is probably very small. Howev- tion on the stability of topaz solid solutions and in
er, indiscriminateapplicationof the results of this particular on the stability of hydroxyl-topaz. These
study should not be made to anomaloustopazes or experimentsare basedon the reaction:
to equilibria below about 400'C until this problem
(1)
Al2SiO5 + H2O: AlzSiOn(OH)i'
has been resolved.
hydroxyl-topaz
andalusite
Four topazeswere used in the hydroxyl-buffering
experiments. Nearly end member fluor-topaz crys- where the hydroxyl-topaz is in solid solution with
tals from Topaz Mountain, Thomas Range, Juab fluor-topaz. With data on the equilibrium content of
County, Utah were used both for this study and for hydroxyl-topaz at a given temperatureand pressure
the calorimetric measurements of Barton et al. and an activity model for fluorine-hydroxyl mixing,
(1982).This topaz occurs in lithophysae in rhyolites one can calculate the thermodynamic properties of
where it is associatedwith qvartz, sanidine,fluor- hydroxyl-topaz given thermodynamic data on anite, beryl, hematite,bixbyite, pseudobrookite,and dalusite and water. Figure 1 illustrates the reaction
spessartine(Ream, 1979).Lieht blue, gemmy topaz showing the experimental assemblageas the three
from Omi, Tanakami Yama, Japan which had no phase field of topaz, aluminum silicate, and waterassociatedminerals,(University of Chicagocollec- rich vapor. The reaction was reversed by doing
tion #2231) was coarsely crushed and inclusion-free matched sets of experiments at the same temperapieces were selected.White, fine-grainedmassive tures and pressuresbut startingwith topazeswhich
topaz from the Brewer mine, ChesterfieldCounty, had initially higher or lower hydroxyl contents than
South Carolina was taken from the collections of the final, equilibrium concentrations.
Theseexperimentsare hydroxyl-buffering experithe Field Museum of Natural History (FMNH
#18049).This topaz occurswith quartz, pyrite, and mentsbecauseit is only the hydroxyl-componentof
gold (Fries, 1942).Topaz from Ouro Preto, Minas the solution which is buffered, unlike the convenGerais, Btazil occurs with kaolinite, quartz, eu- tional fluorine-hydroxyl buffering experiments
Table l. Chemical and crystallographic data for topaz starting
materials
cell
ooo
Parametersl
BARTON: TOPAZ SOLID SOLUTIONS
(Munoz and Eugster, 1969) where the chemical
potentialsof both HF and water are fixed externally
to the mineral of interest. As a result only data for
the hydroxyl end-membermay be directly extracted
from this type of experiment.Attempts to do conventional fluorine-hydroxyl buffering experiments
failed for lack of a suitable buffer stable at the high
temperaturesand HF fugcities required by topaz.
Starting mixes of 60% topaz,35Voandalusite,and
5%oquartz were prepared from the materials listed
in Table 1. Optically pure, natural quartzwas added
to insure that the solid assemblageremainedsilicasaturated.A silica-deficientassemblagewould be
undesirablebecauseRosenberg(1972a)has shown
that, at least in synthesisexperiments, topaz may
containexcessaluminaand fluorine in such assemblages, which affects the X-ray determinative
curves. The mixtures were ground in an agate
morter until the average grain size, determined by
optical examination,was in the range l-3 p. Even
though reaction rates at the lower temperatures
were quite low the materialswere not ground more
finely to avoid introducingsubstantialsurfaceenergy and strain energy effects.
Approximately 10 mg of starting mix and 2 mg of
distilled water were sealedin gold capsules.Up to
five capsuleswere run simultaneouslyin conventional cold-sealbombs with internal chromel-alumel thermocouples.Pressureswere monitored by
Bourdon-tubegauges.The temperatureand pressure measurementsare believed to be accurate to
within +5o C and +40 bars. Run times variedfrom 5
daysat 900' C to 6 monthsat 600' C. Daily monitoring of the runs assured that the pressures and
temperaturesnever varied by more than the estimated accuracy and were usually kept within a
degreeor two and 20 bars of the nominal temperatures and pressures.
At the end of a run the bombs were quenchedat
pressureby cooling with an air blast. Using this
method temperaturesfell to less than 200. C in the
first 2 minutes.Capsuleswere examinedfor signsof
leakage on removal from the bomb. Although the
capsulescould not be weighedto check for leakage
becausethey tended to weld themselves to one
anotherduring the courseof the experiment,it was
generally possible to tell if any had leaked. On
opening,those that had not leaked generallyemitted a spurt of pressurizedwater; also, those that
had leaked almost always reactedto give coarsely
crystalline,thin plates of alumina which were very
distinctfrom the very fine-grainedpowder resulting
from the successfulexperiment.
9s9
At2si05
F.O_,
HF
HrO
Fig. l. Composition plot showing the limiting buffering
assemblagefor topaz solid solutions on the join AI2SiO5-H2OF2O-1 in the Al2Or-SiO2-H2O-F2O-rsystem. Note that the
compositions to the F2O-1 side of the line between HF and
Al2SiO5+ 5F2O-1 are physically unattainable. The tie lines are
schematic.The abbreviationsare: v : vapor, tpz: topaz, and :
andalusite.
Even though many of the experiments were done
in the sillimanite stability field, andalusitewas used
in all the experiments to eliminate the need to take
the ordering state of sillimanite into account. A
powder diffraction scan was made for each charge
from 19 to 35" 20 and examined for the presenceof
extraneouspeaks (in particular sillimanite,mullite,
and corundum) and the peaks of the starting phases.
In none of the experimentswas there any evidence
of sillimanite or mullite. Only those experiments
which were rejected for leakagehad any extraneous
peaksor did not show peaksoftopaz and andalusite
(the quartz peak was quite small in most casesand
was not observedin some otherwise normal runs
and in all runs checked quartz was optically determined to be present). After the first scan a portion
of the chargewas used for a cell refinement. Direction of reaction was determinedby comparisonof
the composition of the final topaz with the composition of the initial topaz (both compositions determined by application of the slightly modified b cell
edgeregressionequationof Ribbe and Rosenberg,
l97l).The experimentalresultsare listed in Table 2
alongwith calculatedcompositions.
Thermodynamicsand discussion
The thermodynamic properties of topaz solid
solutionscan be derived from the experimental data
presentedhere, and the standard state properties of
fluor-topaz from Barton et al. (1982).
BARTON: TOPAZ SOLID SOLUTIONS
In order to derive the thermodynamic properties
ofhydroxyl-topaz and to be able to apply any ofthe
thermodynamicresultsto most naturaltopazes,it is
necessaryto develop a mixing model for topaz solid
solutions. Given this model the equilibrium constant data collected from the hydroxyl-buffering
experimentsmay be used to estimate the properties
of (hypothetical) hydrox yl-topaz.
The heat capacity and entropy of hydroxyl-topaz
can be estimated from the values for fluor-topaz
and data on other fluoride-hydroxide pairs. The
entropy of hydroxyl-topaz was estimated using six
fluoride-hydroxide pairs (Robie et al., 1978)and the
data for fluor-topaz (Barton et aI., 1982)in exchangereactions of the type:
2XOH + Al2SiO4Fz:2XF + AI2SiO4(OH)2 Q)
where AS" was assumed to be 0 and a volume
correctionwas applied(Helgesonet al., 1978).The
estimated entropy of hydroxyl-topaz is 112.8 JIKmol or 7Vogreaterthan the entropy offluor-topaz.
The heat capacity of hydroxyl-topaz is also arbitrarily taken to be7% greater than the heat capacity
of fluor-topaz, although this is probably a slight
overestimate.
In all calculations the fugacity of water is assumed to equal the fugacity of pure water at the
pressureand temperature of the experiment' Other
possiblevapor speciesinclude HF and a variety of
fluoride and hydroxide complexes of Si and Al.
Calculations based on the thermodynamic data for
fluor-topaz suggests that the fugacity of HF
amountsto at most a few bars. While thermodynamic data are not available to evaluate the concentrations of the other speciesat the conditions of the
experiments they are unlikely to be more than
perhaps an order of magnitude greater than the
concentration of HF. They then should have no
more than a small effect on the equilibrium constant
and henceon the derivedfree energies.Considering
that the relative uncertainty in the mole fraction of
hydroxyl-topaz is llVo or greater, changes in the
fugacity of water of this degree would not add
greatly to the uncertainty in the derived parameters.
The simplestmixing models for topaz are those
that involve ideal mixing of the fluorine and hydroxyl on the fluorine sites in the topaz structure. In
ideal orthorhombic topaz (Pbnm) all eight fluorine
sites in the unit cell are equivalent. If the fluorine
and hydroxyl mix ideally on all these sites the
activities of fluor-topaz and hydroxyl-topaz are
given by the expressions:
qn
aorr:r:
:
*Fr
(3a)
*ogr
(3U;
for the formula units Al2SiO aF2and AI2SiO4(OH)2.
If the fluorine sites are not crystallographically
equivalent (as in topaz of lower symmetry) the
activity relations would be the same in the absence
of inter-site fluorine-hydroxyl partitioning. Any activity model must render the results from the hydroxyl-buffering experiments internally consistent.
The equilibrium condition for reaction (1) is:
AG,,:-Rrlnffi.lor*
(4)
Two related tests of internal consistencycan be
made on the activity models. The first is to check
that experiments at different pressures but at the
same temperature give the same value for the
standardstatefree energy.The secondtest is to plot
all the free energiesversus temperature to seeif the
results at the various temperaturesare consistent
with one another given reasonable values for heat
capacitiesand the entropiesof the phases.Values
for AG"/RT for the ideal activity model are given in
the secondto the last column of Table 2. Thesedata
satisfythe above criteria for internal consistency;a
curve that gives reasonablevalues for the entropy
and heat capacity of reaction can be fit through the
brackets.
While the ideal mixing model fits the experimental data within uncertainty, the ideal model is not
satisfactoryfor extrapolation of the results to much
lower temperaturesor to high water pressures.The
reasonfor this is that the model allows the hydroxyl-topaz content of a topaz to reach unity. This
would imply, given the experimentalresults, that
andalusite plus water is unstable relative to pure
hydroxyl-topazat a few hundreddegreescelsius,a
conclusion which clearly conflicts with natural asNatural topazeshave at most about 30%
semblages.
hydroxyl-topaz component even when coexisting
with andalusite,pyrophyllite, or kaolinite. Rosenbere 0972a) did synthesizetopazesat low temperatures having more than 50Vohydroxyl-topaz component,but he startedwith highly metastablematerials and his final assemblage(quartz -f topaz *
pyrophyllite) was mestastableover most of the
temperature interval of his experiments. Parise
(1980)argues that due to the proximity of pairs of
fluorine sites in topaz occupation of one site in a
pair by hydroxyl would preclude the occupation of
the other by hydroxyl. To have hydroxyls on adja-
BARTON: TOPAZ SOLID SOLUTIONS
961
cent sites would require the hydrogen atoms to be Substituting (9) into (5) and simplifying one arrives
too close to one another. This leads to a "proton- at the expressions for the activities of hydroxylavoidance" restriction on the occupation of adja- topaz and fluor-topaz in topaz solid solution.
cent sites limiting the mole fraction of hydroxyltopazto lessthan 0.5
(l0a)
Given this restriction on the distribution of hydroxyl within the topaz structure an alternative
(10b)
"ideal" activity model can be derived which gives
an equally good fit to the experimental data, but Theseproportionalities may be converted to equalialso gives very reasonableresults when used to ties by consideringwhat the activities must be at
extrapolate the experimental results to lower tem- boundary conditions. In the limit of low hydroxyl
peratures.
content in the solid solution the activity of hydroxIdeal mixing models are derived through the yl-topaz using the proton-avoidance model must
equal that of the ideal model. This results because
Boltzmannrelation (Kerrick and Darken, 1975):
at low hydroxyl contents the probability of a pair of
(5)
^S:ftlno
sitesboth beingoccupiedby hydroxyls is very small
which relates the entropy of a state S to the and therefore the dominant contribution to the
multiplicity of that state O, where k is the Boltz- entropy of mixing in both models comes from the
mann constant.In this problem the activities of the F-OH combinations. Similarly, for pure fluor-tocomponents are related to the entropy of mixing paz the activity of fluor-topaz must be unity. The
through the relation:
correct activity expressionsfor the "proton-avoidance" model are:
:
si
s,
Rln a;
(6)
*ot'
(11a)
aosr:
whereSris the entropyof the pure phaseand S; is the
0 - zx,"*r)
partial molar entropy of the component in the
(1lb)
arr:(2Xrr-l)
solution. The entropy that needsto be determined
(5)
in equation is the entropy of mixing becausethe
Somecommentsare in order on these activities.
entropy of mixing and the activity are related Clearly they are not completely correct representathroughthe expression:
tions of the activitiesbecausethey give valuesof co
and 0 for dssl ilnd ap1, r€Sp€ctively, at X6s1 :
iq
R l n a'; : T ( l - X J * S m i x
0.5.
These"unreasonable" answersresult from the
e
)
o*i
"unreasonable" restriction that no two hydroxyls
where X; is the mole fraction of component i. ,f), the may occupy both sitesin any adjacentpair of sites.
multiplicity for mixing in topaz solutions with the As a consequenceof this on the molecular level it
becomesvery difficult to add another hydroxyl as
proton-avoidancerestriction, is given by:
X6s1 approaches0.5 (alternatively,it is very easy
N!
: 0.5 it is
gy _ 22Nxont
(8) to substitute a fluorine). At Xonr
(2NXoHr) !(N( | - 2xoHi)t
impossibleto substituteany more hydroxyl because
it would violate the proton-avoidance principle.
where N is the number of "molecules" of Al2SiO4 This is what requires the activity of hydroxyl-topaz
(F,OH)2and Xeql is the mole fraction of hydroxyl- to be infinite and the activity of fluor-topaz to be
topaz. This multiplicity is composedof two parts. zero at this composition.A completely equivalent
The first, the power of 2, is the numberof ways that formulation can be done using the components
pairs of sitesof which one memberis hydroxyl may Al2SiO4F2and Al2SiOaF(OH).If this is done there
be arranged. The second term accounts for the is a zero-point (configurational)entropy for Al2SiOa
number of ways that the pairs of sites may be F(OH) of Rln2 which arisesfrom the disorder of the
arranged(F-F vs. F-OH, rememberingthat OH- fluorinesand hydroxyls on the pairs of sites. Note
OH combinationsare not allowed). Substituting(8) that this is equal to the entropy of mixing for Xesl
into (5) and letting N be Avogadro's number:
: 0.5 in equation(9).
These mixing relations are very similar to those
:
Smix k ln O : Rl2Xo1rln2 + 2XoH,rln2Xo1-l
that would be derived for the Ca-Mg pyroxenes if
+ (l - 2Xonr)ln(l - Xonr)l I
(9) Ca is consideredto be restrictedto the M2 site. The
BARTON: TOPAZ SOLID SOLUTIONS
962
activity of wollastonite would be infinite and the
activity of enstatite would be zero at diopside
composition(Grover, 1980).The only differencein
the two casesis that in the topaz model there is an
additional entropic contribution due to the two
ways that pairs of F-OH can be arranged(F-OH vs.
OH-F). In both the pyroxenesand topazit is clear
that this extreme case is unreasonable,the true
solution being one where there is intra-crystalline
partitioningbetweenthe two sites; someCa on Ml
or OH-OH pairs.
Grover and Orville (1969)have discussedintracrystallinepartitioningfor caseslike this, but appli-
Table 2. Results for the hydroxyl-buffering experiments
Runl
T
oc
P
Time
bars
days
191
r91
125
125
125
K- 10b(BR)
K- 10c(OM)
K- l4a (BR)
K-14b(OM)
K- 14e (TM)
620
620
700
700
700
2000
2000
1000
1000
1000
K - 1 5 a( B R )
K- r5b (BR)
K- 15c(oM)
K- 15d(rM)
K- 16a(q"1)
750
750
750
750
750
b
XOHT -lnK - V(P-1)/RT
o
idea12 p-avoid3
8.8183(18) 0.149
8.8120(26)0.119
8.8r40(30) 0. r28
8.8070(10) 0.095
8.7967(r4) 0.045
r0.72
11.18
r0.71
rr.30
I2.79
10.37
10.98
10.4r
11.09
12.'10
2000
2000
2000
2000
1000
2A 8 . 8 1 5 5 ( 1 6 ) 0 . 1 3 5
28 8 . 8 r 7 7 ( 3 9 ) O . L 4 6
2a 8 . 8 0 7 6( 1 0 ) 0 . 0 9 8
28 8 . 7 9 8 4 ( 1 2 ) 0 . 0 s 3
27 8 . 8 0 5 9 ( r O ) 0 . 0 8 9
11.13
).1.07
11.86
r3.09
11.50
10.90
Io.72
tr.64
L2.98
r1.30
K- 16b(01..1) 7 5 0
K- 16c(TM) 7 5 0
K- 16d(rM) 7 5 0
K- Ua(O{)
800
K - r 7 b ( r M ) 800
1000
1000
1000
1000
1000
27
27
t4
t4
8.8049(15)
8.7969(6)
8.7 973(9)
8 . 8 0 3 4 (1 0 )
8.7959(8)
0,085
0.046
0.048
0.077
O.042
11 . 5 9
12.82
L2.73
It.94
13.01
I 1.40
t2,72
12.63
Lt.67
12.96
K- 18a(TM)
K- 18c(OM)
K- r8d(BR)
K-l8e(BR)
K- I 9a (TM)
700
700
700
700
800
2000
2000
2000
2000
r000
r01
r01
10r
101
37
8.8002(16)
8,8098(17)
8,8183(15)
8.8188(13)
8.7986(8)
0.062
0.r08
0.149
0.15r
0.054
12.67
r 1.56
1o.92
r0.90
12.53
t2.54
tL.32
r0.57
r0.54
12.42
K- I9c (oM)
K- 20a(TM)
K- 20b(TM)
K-20c (OM)
K- 20d (oM)
800
850
850
850
850
1000
650
650
650
650
37
31
31
31
31
L8027 (LL)
8.7955(1r)
8 . 7 9 6 6( 1 0 )
8.803r(9)
8.7995(11)
O.074
0.040
0.045
0.076
0.059
11,91
12.79
12.57
rL.52
t2.04
1I.75
t2.73
r2.4A
1 1 .3 6
11.91
K-21a(OlD
K-21b(TM)
K-22a(BR)
K-22b(01.0
K-22c(TM)
750 1000
750 1000
750 2000
750 2000
750 2000
o)
8.8037(8)
8.7966(r1)
8.8138(18)
8.8055(6)
8.7989(r4)
0,079
0.045
0.127
0.088
0.056
11,73
12.85
I 1 .3 4
t2,o7
13.16
r1.56
12.76
11 . 0 5
r1.88
t2.87
K-23a(n{)
K- 23b (ol..r)
K- 23c(TM)
K- 23d(OM)
K- 24a(rM)
850
850
850
850
900
650
650
650
650
650
8,7979(13)
8,7985(8)
8.7952(10)
8.8020(14)
8.794s(10)
0.051
0.054
0.038
0.071
0.03s
12,32
t2.2I
12.9t
11,56
r3.09
t2.21
t2.to
t2,83
r1.51
13.o2
K- 24b(oM)
K-25a(OM)
K- 25b(O.,r)
K- 25c(BR)
K- 25d(BR)
900
600
600
600
600
650
1000
1000
1000
1000
5 8.8018(12)
96 8 . 8 0 9 3 ( 1 4 )
96 8 . 8 0 9 8 ( 1 r )
96 8 . 8 2 2 9 ( L 4 )
96 8 . 8 2 3 3 ( 1 4 )
0.070
0.106
0.r08
0.171
0.173
11.71
r0.89
r0.85
9.94
9.92
11.56
10.65
10.61
9.52
9.49
r1,08
r0.50
r0.4r
10.80
1 0 .1 0
9.99
K-26a(61) 6000 2000
K-26d(BR) 600 2000
K-26e(BR) 600 2000
?7
65
63
63
63
44
5
49
49
49
8.8L25(2O) O.r2t
8.8212(15) 0.163
8.8228(10) 0,170
TM = Topaz Mountaln; OM= Orni; BR = Brewer Mlne; eee Table 1.
2
-, 2 1 =
X 9 H 1/ f 6 r g
K
K = X 6 H T/ [ f s r 9 ( 1 - 2 x 6 1 1 1 ) ]
cationhere of a more complexmodel is not justified
for the following reasons. The experimental data
can be fit equally well with either the ideal model or
the proton-avoidancemodel becausethe two models give similar activities for fluorine-rich topazes
such as those stable at the high temperaturesof the
experiments.Even for the lowest temperatureexperiments (600'C, containing the most hydroxylrich topaz) the error due to the uncertainty in the
composition is larger than the calculated difference
betweenthe two activity models.Therefore,on the
basis of experiment, this is no way to choose
betweenthe alternativemodels. However, calculations based on the proton-avoidancemodel agree
whereasthosebased
well with naturalassemblages,
on the ideal modet do not. Therefore, given the
present degree of resolution in the experimental
data, the proton-avoidancemodel, althoughan approximation, provides a satisfactorybasis for extrapolatingthe results of the experiments.
Table 2 gives values for AG"/RZ for reaction (1)
calculated using the proton-avoidance model.
These values are plotted versus llT in Figure 2.
From thesevalues,the estimatesof the heat capacity and entropy of hydroxyl+opaz and thermodynamic data on andalusiteand water (Haas et al.,
1981: Burnham et al., 1969) the standard state
thermodynamic properties of hydroxyl-topaz may
be calculated. Numerous curves could be fit
through the data within the experimental uncertainty, but the one chosenwas picked becauseit gives
an entropy very close to the estimatedentropy and
violatesthe fewest brackets.A least squaresfit was
not done becausedata of this type do not belongto
normal distributions.The standardstate thermodynamic data for hydroxyl-topaz are given in Table 3.
The following (conservative) method was used for
the estimation of errors. First, the error in the
entropy for hypothetical hydroxyl-topaz was assumedto be i8.0 J/K-mol, which is the combination of the error in the fluor-topaz entropy and the
F-OH exchangecorrection. Second, the error in
the standard free energy formation of hydroxyltopaz (for the proton-avoidancemodel) was estimated from the uncertainty in the experimental
brackets (on the order of 8 kJ due to +0.025
uncertainty in Xqqt and the width of the bracketsthe former contribution predominates) and the uncertaintiesin the free energiesfor the other phases.
The standard state free energy for hydroxyl-topaz
calculated for the ideal solution model would be
about 5500 J more positive; within error of the
BARTON: TOPAZ SOLID SOLUTIONS
]!""'l
Toc
V
-
|
|
'
'r
|
|
2RTlnb4-
I
- 5.0728x 10-6?3- 4012.56T0.s
+ 0.031835472
- 1705077-t+ 72.8278T1nT
- 0.239(1- t\
\
A
AG"
RT
: 133697
+ ffi6.9467
-firr Xonr
AtrSiOu+H.O=AtrSiO4(OH);s
_
V
963
v
(r2)
v
This expression can not be directly compared with
the resultsof Rosenberg(1978)on fluoride-hydroxI
yl exchange in topaz. However, the predicted fu.t
|
|
.
gacity ratios differ greatly from Rosenberg's. His
equation predicts that with increasing temperature
looo/TK
at constant composition the coexisting fluid beFig. 2. -AG'/RT versus llT plot for the hydroxyl-topaz
bufering experiments (values calculated using the protoncomespoorer in HF. This is contrary to the results
avoidance model; see text). Arrows indicate the direction of
predictedhere and also conflicts with other experichange.The uncertainty is approximately +0.4 at the highments on fluoride-hydroxyl exchange in silicates
temperatureend decreasingto +0.15 at the low-temperature
end. The line is also calculatedfrom the proton-avoidancemodel (e.g. the mica exchangeexperimentsof Munoz and
Ludington, 1974,1977).There doesnot seemto be a
and representsthe preferred fit.
simple numerical reasonfor this discrepancysuch
as the use of different formula units. However, the
proton-avoidancevalue (there is no special signifi- discrepancymay be within the experimental errors
cance to this agreement).Parenthesesenclosethe of the two studies.The method used by Rosenberg
value for the standard free energy of formation of (1978)involved determinationof the HF content of
hydroxyl-topaz (Table 3) to emphasize that the the fluid by mass balance (Rosenberg,personal
value is model dependent.
communication).Becauseonly a very smallfraction
From the activity model and from thermodynam- of the fluorine is partitioned into the liquid, small
ic data on hydrogenfluoride, water, and the topaz errors in the determination of the composition of
end members, the fugacity ratio of the gaseous the topaz could lead to large errors in the relative
speciescoexistingis related to topaz composition amount of HF and hencelarge error in the equilibrithrough the expression(? in K; P in bars):
um constant.The presenceofother fluoride species
^ 2000 bors
^ l O O ob o r s
- ^ 650 bors
>\\
a
I
Table 3. Thermodvnamic orooerties of some phases in the AlzOr-SiO2-H2O-F2O-1 system
co*
"298K,1 bar
J/moI
Fluor-topaz
Hydroxyl-topazf
HF (ideal
HrO (ideal
gas)
gas)
qo
-298K
Volume
c
J / K'nol
J/K.nOt
-244280
(484)
(0.73)
-29L0660
(4400)
(-26q??S1 l
q1
Source
J /bar
c
S
_6754.36
5.153
4
4
4
4
5
5
5
227
-0.20790
2.28751
105.40
(0.ls)
47L.4r4
-0.08165
r.26947
-5485.54
5.153
5
rL2.04
(8.00)
504,4r3
-o.08737
( 12000)
l. 35833
-5869.53
5.392
1I11
-27 5400
(828)
L 73 . 7 8
(0.04)
30.352
-0.00302
-
-10. 3889
2 4 7 9.
2233
-2286II
(100)
r88, 73
(0.08)
l0 .438
0 .02596
-0. 13108
2 4 7 9.
4444
7?
6.6875I
0 .3L404
-o .44689
2 g g. L g
o
Eaao." given here are for the Gibbs free
ooao=a+bT+cx106T-2
energies
of
formation.
+dx10-5t2+eT-0.5
t The cibbs free energy for
hydroxyl-topaz
-t ^
This
study,
z Robie
et
al.
(1978).
is
3 Sr.11
for
the
and
prophe!
proton-avoidance
(Lg7I).
4
model.
Ibas
et
al,
(1981).
5 Barron
9l el.
(1982)
BARTON: TOPAZ SOLID SOLUTIONS
in solution, such as suggestedby the low-temperature data of Roberson and Barnes (1978). would
result in unusuallyhigh apparentHF contents.The
pressure effect on log fHylfi1,o at constant topaz
composition(6.241Tin units of kbar-') is negligible
given the uncertainty in the thermodynamic data.
The synthesisexperimentsof Rosenberg(1972a)
are roughly compatible with the results obtained
here finding that the buffered content of hydroxyltopazin topaz solid solutionincreaseswith decreasing temperature.His results show hydroxyl contents which are systematically higher than those
calculatedfrom the results of this study. This is
consistentwith the fact that Rosenberg'smullite
and pyrophyllite buffering assemblageswere metastable compared to the andalusite * topaz (*
quartz) assemblageused here. Interestingly, the
shapeof Rosenberg'sT-X results agreeswith calculations using the ideal activity model suggestingthat
the topaz produced in his study was highly disordered.This resultsperhapsas a consequenceofthe
synthesisnature of his experiments.
Petrologicapplications
es. Hence the compositionsof the topazesand the
are governedin large part by
resultingassemblages
externally derived fluids. This section emphasizes
equilibrium calculations on topaz-bearingassemblagesand what these equilibria can reveal about
closedand open systemtopaz parageneses.
Topaz occurs in a wide variety of granitic and
related hydrothermally altered rocks. Topaz-bearing rhyolites and hypabyssalequivalentsare quite
common in some environments(Burt et al., 1980:'
Christiansen,1981; Eadington and Nashar, 1978;
Kovalenko et al., 1972). Closely related topazbearing granites are also common as are topazbearingpegmatites(reviewed in Bailey, 1977).Gradational with igneousrocks containing "primary"
topaz, and often developed within them, are topazbearingmetasomaticrocks. Theseincludegreisens,
porphyry molybdenum or tin mineralization,various types of vein deposits, and some aluminous
skarns(Thomas,1979; Hemley et al., I 980; Wallace
et al., 1978:Vlasov, 1967;Smirnov, 1977).Detailed
discussionof all the associationsof topaz in these
rocks is beyond the scope of this study, but a
numberof generalizationscan be made concerning
the petrology of topaz.
Topaz compositionsin nature may be controlled
in one of two ways: by local equilibriawith coexistTopazequilibria in the AlzOrSiOrHzO-FzO-r
ing solids and vapor, or by externally imposed
system
vapor compositions.In the caseof local equilibrium
the topaz in turn buffers the vapor composition.
The composition of topaz can be buffered by
Most parageneses
resultfrom metasomaticprocess- reactionwith many phasesin the AI2O3-SiO2-H2OFzO-t system(Figure3). The simplestreactionsare
divariant fluoride-hydroxyl exchange reactions
such as representedby reaction (4). Other, heterogeneousreactions such as (1), (13)-(16)may also
control topaz compositions.Many other buffering
Al2SiO4(OH)i'+ 3SiO2: AlzSiaOro(OHli'
pyrophyllite
quartz
topaz
4Al2SiOa(OH))':
topaz
6AIO(OH)" + Al2Si4Oro(OH)i'
pyrophYllite
diaspore
FzO-r+
cor
At20!
H.o
Fig. 3. Projection of the compositionsof some of the phasesin
the AIzO:-SiO2-H2O-F2O-1 system (parallel to the join H2OFzO-r). Abbreviations (used throughout this paper) arei tpz:
topaz, and = andalusite,cor = corundum, di : diaspore, ka :
kaolinite, kya : kyanite, py = pyrophyllite, qz : quartz, rs =
alkali-freeralstonite, sil : sillimanite, v = vapor, zu = chlorinefree zunyite.
2Al2SiO4(OH)i'+HrO :
topaz
Al2si2o5(oH)i'+ 2Alo(oH)"'
diaspore
kaolinite
(13)
(J4)
(15)
AI2SiO4(OH)T+ SiO2 + H2O : Al2Si2Os(OH)'4'
qtrartz
kaolinite
topaz
(16)
reactionscould be written involving the mineralsin
BARTON: TOPAZ SOLID SOLUTIONS
33
ooo,
3
4
Pkb
a
2
I
200 300 400 500 600 700 800 900
T"C
Fig. 4. Calculated isopleths of hydroxyl-topaz in topaz solid
solution as a function of pressure and temperature for reaction
(l). Abbreviations are the sameas for Fig. 3. The uncertainty in
composition varies as a function of the temperature and the
pressureand isopleths are model dependent;but is on the order
t35', except for the highest temperatures(>750'C) at low
pressures(<800 bars).
Figure 3, but they would not represent known
assemblages.
Figure 4 showsisoplethsof hydroxyltopazbufferedby an aluminumsilicateand a waterrich vapor. These isopleths are calculated from
equation(3) assumingthat the vapor is pure water.
The hydroxyl-topaz isoplethsare similar in shape
to normal dehydrationcurves as expectedfrom the
fact that they representthe dehydrationreactionof
hydroxyl-topazdiluted by fluor-topaz to aluminum
silicate + water.
Topaz compositionswill be bufferedby reactions
(13)-(16)at temperaturesbelow 400'C as a consequenceof the instability of andalusiteand kyanite
relative to pyrophyllite- and then kaolinite-bearing
assemblages.Figure 5 somewhat schematically
shows the buffering of topaz by these reactionsat
500 bars. Calculation of the displacementof the
reaction:
Al2si4or0(oH)z: Al2sio5 + 3sio2 + H2O (17)
pyrophyllite
andalusite quartz
with the addition of fluorine shows that at the
fluorine content where andalusite reacts to give
topazthe fluorinecontentof the pyrophyllite is only
about a tenth of a weight percent and that the
temperatureof (17)has been raisedby less than 2".
For this calculation the F-OH exchangebetween
pyrophylliteand vapor was assumedto be the same
as Munoz and Ludington (1977) found for muscovite-vapor exchange. With the first addition of
fluorine pyrophyllite is slightly stabilized. With
subsequentreaction of andalusiteto give topaz at a
higher fluorine concentration the stability of pyrophyllite is governedby reaction (13) and decreases
in temperature with increasing fluorine content. If,
with increasing fluorine content reaction (13)
moved to higher temperaturesin the limit of complete fluoride for hydroxyl substitution, fluor-topaz
would break down to fluor-pyrophyllite plus
quartz. This is incompatible with known topazkaolinite assemblagesand with the normally observedchemographyof the phases(F/OH in topaz
is much greaterthan F/OH in coexistingpyrophyllites). From this it would appearthat the pyrophyllite-bearingexperimentsof Rosenberg(1972a)must
be metastableand thus is consistentwith the unusually high hydroxyl contentsof his topazes.
Figure 5 also illustratesother reactionsinvolving
topaz, including compositional limits for silica undersaturatedconditions.Note that at temperatures
lower than those of reaction (18) the maximum
Pr,=Proro,=
5OObors
T'C
rez
(31
'p'([I)
,p,(!1,)
:iti
'* ([l)
ro{.'i,'
OH
OH+F
Fig. 5. Schematictemperature versus OHI(OH + F) diagram
showing some, but not all, of the reactions which limit topaz
compositionsin the system AI2O3-SiO2-H2O-F2O-1at 500 bars
total pressure. The lines represent the compositions of the
various phases coexisting with the phases in the parentheses
(i.e., tpz(py,qz) is the projected composition of topaz in
equilibrium with pyrophyllite and quartz). The abbreviationsare
the sameas for Fig. 3.
BARTON: TOPAZ SOLID SOLUTIONS
(2 - r)AlzSiaOls(OH*F2-J :
pyrophyllite
(x - r)Al2SiOs+ 3(2 - r)SiO2
quartz
andalusite
(18)
(2 - x)AI2SiO4(F2-.OHJ+ (x - r)H2O
topaz
hydroxyl-contentfor topaz is found in silica-undersaturatedassemblages;
the topazesin silica-saturated assemblagescontain more fluorine. Implicit in
the constructionof Figure 5 is the assumptionthat
F/OH in topaz ) kaolinite > diaspore> pyrophyllite. This sequenceis arbitrary except for the position of topaz which always has F/OH much greater
than that ofthe other phases(cf.Deer et al., 1962).
Topaz-bearingassemblagesin the Al2O3-SiO2H2O-F2O-r systemare quite common. Andalusitetopaz-quartz has been reported by Sykes and
Moody (1978)and Hildebrand (1961),and might be
expectedin regionsof advancedargillic alterationin
hydrothermal ore deposits (Hemley et al., 1980).
Kempe (1967)describessillimanite (kyanite) + topaz + quartz assemblagesfrom Tanzania, and
Sheridanet al. (1968)describetopaz * sillimanite+
quartz from the Front Range in Colorado. Topaz,
pyrophyllite, andalusiteand quartz have been describedin vein assemblages
from hydrothermalore
deposits(Meyer and Hemley, 1967;Hemley et al.,
1980)and pyrophyllite developedby alteration of
pre-existing topaz-andalusite-quartz rock is describedby Sykesand Moody (1978).Olsen (1971)
and Sainsbury (1964)specifically describe kaolinite-topaz-quartzassociationsand suchassociations
are common in supergeneor late hypogenealteration of topaz-bearinggranitesand greisens.Silicaundersaturated "apo-carbonate" greisens often
have coexistingtopaz and diaspore(Vlasov, 1967).
All theseassemblages
are consistentwith the model
for the low-temperaturephase relations of topaz
representedby Figure 5. Although calculatedtopaz
compositionsat low temperaturesare suspectdue
to the uncertainties in the thermodynamic data,
such calculationsshould give a reasonablequalitative understandingof topaz behavior, becauseof
the large activity coefficients of hydroxyl-topaz
nearXesl : 0.5.
Few topazesfrom well-documentedbufferingassemblagesin this system have been analyzed.
Those that have are generally consistentwith the
present results. Topaz from Ouro Preto, Minas
Gerais,Brazil is found with kaoliniteand quartzand
contains 25 mole Vo hydroxyl-topaz (Table 1),
roughly consistent with Figure 5. Thomas (1979)
reports that fluid inclusion filling temperaturesfor
topaz from Ouro Preto are in the range 260-280' C.
The topaz from the Hillsborough pyrophyllite deposit alsocontainsabout 25 moleVohydroxyl-topaz
component (Sykes and Moody, 1978) consistent
with the composition expected near reaction (18)
(Figs. 4 and 5). It is interesting to note that the
pyrophyllite compositioncalculatedfor this assemblage has a fluor-pyrophyllite component of but a
few percent, in reasonable agreement with the
compositionreported by Sykes and Moody.
Kempe (1967) provides optical and crystallographicdata for severaltopazesfrom sillimanite +
topaz -f qtartz + kyanite assemblagesfrom Tanzania. Estimatedcompositionsfrom Kempe's datafall
in the rangeof ll-16% hydroxyl-topaz component.
However, Rosenberg(1972a)reports that at least
some of this topaz has as little as 5Vohydroxyltopaz component.If these are stable assemblages
then Ps,s ( P,o, kf. Fie. 4). Topaz in a sillimanite
+ qtJartzassemblagefrom Colorado (Sheridan et
al., 1968)contains abofi llVo hydroxyl-topaz component(Rosenberg,1972a)which would be consistent with formation under vapor-saturatedconditions above about I kbar and 675"C.
Without independent pressure or temperature
estimatesthe topaz composition from a buffered
assemblagecan at most give the temperatureas a
function of pressure(or vice versa). Only for univariant reactionscan the temperatureand pressure
both be estimated from the composition of the
topaz. The uncertaintiesin derived pressuresand
temperatureswould be highly correlated even for
the univariant reactions such as (18) and hence it
would be difficult to tightly constrain the pressure
or the temperaturein the absenceof someindependent information.
For fluorine-rich, silica-undersaturatedportions
of this systemtopaz compositionsmay be buffered
by reactions involving alkali-free ralstonite
(Al(F,OH)3) and chlorine-free zunyite (Alr3Sis
(Fig. 3). Rosenberg(1972a)
Ozo(OH,F)re(F,OH))
observedthesephasesin his synthesisexperiments
and suggestedthat natural ralstonite-topazassemblagesmight occur in places like the St. Peter's
Dome area, Colorado, where both minerals have
beenreported. Zunyite plus pyrophyllite or kaolinite is equivalentto topaz plus vapor and is found in
some ore deposits characterizedby intense acid
leaching(suchas Red Mtn., Colorado;Butte, Mon-
BARTON: TOPAZ SOLID SOLUTIONS
tana; Meyer and Hemley, 1967).This suggeststhat
zunyite + topaz + pyrophyllite (or kaolinite) +
vapor may be a stable assemblageunder some
conditions.Zunyite + quartz + kaolinite (or pyrophyllite) would be incompatiblewith the commonly
observedassemblagetopaz -l kaolinite + (inferred)
vapor suggestingthat zunyite, like pyrophyllite,
could have a limited temperaturerange of stability
at least under high silica activities.
Topazequilibria in more complex systems
Although many topaz-bearing rocks can be described in terms of the four-component system
discussedabove most topaz-bearingassemblages
include a variety of other minerals requiring the
addition of other components, notably alkalies.
Burt (1979, 1980,1981),Beus and Dikov (1967),
Glyuk and Anfilogov (1973a), Kalyuzhnaya and
Kalyuzhni (1963), Kogarko (1966), Kupriyanova
(1970),and othershavediscussedthe paragenesis
of
topaz under a variety of conditions on the basisof
field observationsand synthesisexperiments.
Under alkali-rich and/or silica-poor conditions
topaz decomposes to alkali-aluminum fluorides
such as cryolite (Na:AlFo) or chiolite (NasAlrFr+)
(Burt, 1979; Kogarko, 1966). This explains the
absenceof topaz in syenitesand the reaction relations observedbetweentopaz and thesefluoridesin
very alkalic granitesand granitic pegmatites.
Other reactionsinvolving topazinclude reactions
with beryllium minerals (Burt, 1975a)where with
increasingfluoride activity the aluminousberyllium
mineralsdecomposeto give topaz plus phenakiteor
bertrandite.Reactionsof this type have been proposedas natural fluorine buffers for somepegmatite
and greisenassemblages(Burt, 1975b).For example, beryl * phenakite + topaz + quartz and beryl
+ bertrandite+ topaz * quartz assemblages
occur
in the pegmatitesand greisensin the Mt. Antero
region, Chaffee Co., Colorado (Adams, 1954).
Theseassemblages
buffer the HF fugacitiesby the
heterogeneous
reactions:
+ 4HF : 3BezSiO+
2Be3Al2Si6Or8
phenakite
beryl
If the topaz composition and a temperature or
pressurewere known for one of theseassemblages,
and the activity of water could be assumedto be
nearly equal to unity, the temperatureor pressure,
and HF fugacity could be calculated. Analogous
reactionscan be written for the decompositionof
aluminousphasesof other minor componentswith
increasingactivity of HF to less aluminousphases
and topaz. However, such reactions are probably
not common fluorine buffers becausethe bulk of the
fluorine buffering will be performed by the major
componentcompoundsin the rock-i.e. the alkali
aluminosilicates.
Of the alkali aluminosilicatesthe most important
here are the feldspars and the micas. Numerous
fluorine buffering reactions can be written in terms
of reactionsinvolving the feldspars,micas, aluminum silicates,and one or more fluorides.Fluorite is
the only one of the common alkali or alkaline earth
halidesstablewith topaz. Fluorination reactions for
calcium-bearingfeldsparsproduce fluorite plus an
aluminousphase + quartz. They extract the anorthite componentin a manneranalogousto metamorphic reactions involving (clino)zoisite or calcite.
Three of the most common buffering reactions
involving anorthite componentare:
CaAlzSizOi'+2HF :
anorthite
Al2SiOs + SiO2 + CaF2 +
quartz
fluorite
andalusite
sillimantite
HzO (20a)
CaAlzSizOE* 4HF :
anorthite
Al2sio4Fi'
topaz
+
sio2 + caF2 +
qtartz
fluorite
2II2O (20b)
CaAl2Si2O$'+ KAISi2OS' + 2HF :
K-feldspar
anorthite
KA12[A1Si3Oro](OH)i" + CaFz +
fluorite
muscovite
2SiOz(20c)
qoartz
Reaction (20a) is the buffering reaction for fluorine in some metamorphicrocks (Bohlen and Es+2AI2SiO4FT+ 2H2O + 7SiO2
(19a) sene,1978)and representsthe AFSQ fluorinebuffer
(Munoz and Eugster, 1969).This reaction reprequartz
topaz
sentsthe low-fluorine complement of reaction (20b)
4Be3Al2Si6Or8
+ 8HF : 3Be+SizOz(OH)z
which is the buffering reaction for anorthite compoberyl
bertrandite
nent (in feldspars)in some greisensand rhyolites.
+ 4AI2SiO4FT+ H2O + 14SiO2
(leb) Figure 6 illustrates the effect of fluorination on the
quartz
topaz
anorthitecontent of plagioclasein the presenceof
968
BARTON: TOPAZ SOLID SOLUTIONS
\""i2.
-3.5
t
I t{tr
log +
,Hp
_q.
lltrl
r.o 0.30.2 0.t o.o5
Xnn
o.ol
r0o.30.2 0.1 o.05
trrl
I
o.ol
I
I
o-t-2
log o4n
Fig. 6. Reactions (l), (20a), (20b), (20c), (zta), and (2lb)
plotted for the log of the activity of anorthite versuslogfi1pfg,6
at 500 bars and 400 and 600" C. Xo, is the mole fraction of
anorthite at the temperature of interest (calculated from the
solutionmodel of Newton et al., l9E0).Abbreviationsare ksp :
K-feldspar, mu = muscovite, a1 = anorthite (in plagioclase),and
as for Fig. 3.
common granitic phases. These calculations are
based on the thermodynamicdata for topaz (this
study) and the data for the other phases(Haas er
al., l98l; Robie et al., 1978).Under most conditions at HF fugacitieshigh enoughto stabilizetopaz
the activity of anorthitemust be fairly low, especially in the presenceof potassiumfeldspar. Figure 6
was calculated assumingpure K-feldspar, if one
considersthe most sodic composition (coexisting
with albite)the shift in the curve is <0.1 log units to
higherHF.
Desboroughet al. (1980)have recently proposed
that (20c)could account for unusually low temperatures derived from feldspar geothermometry on
fluorite-bearing, two-mica granites from the pikes
Peak batholith. Siderophyllite component in biotite
is anotherplausiblesink for aluminain thesefluorination reactions.Becauseof these fluorination reactions,greisenizationand other phenomenarelated to the generationof relatively high HF activities
are restricted to low-calcium granitic rocks. Con-
versely,rocks with greisen-stylealteration,be they
granites,rhyolites, pegmatites,or aluminous sediments have albitic plagioclase.As Burnham (1979)
points out, the strict associationof certain types of
ore deposits with alkalic rocks may be more a
consequenceof their bulk compositionthan of any
originalenrichmentin trace ore components.Calcic
rocks will buffer the acid fluoride species at low
concentrationsprecluding the significant involvement of thesespeciesin the ore-formingprocess.
Topaz-bearingassemblagescould however arise
from calcic source magmas given the following
scenario:fluids derived from the magmamust start
with relatively low HF/H2O ratios, below the field
of topazstabilityfor the magmaticconditions.If the
fluid then cools without beingable to react with the
calcium-richhost it will move into the field of topaz
(c/. Figs. 6 andT). On encounteringa low-calcium
host it could then produce topaz.
Figure 6 also indicates that the AFSQ buffer is
only marginally stable with respect to anorthite +
fluorite + topaz + quartz, and could be metastable
given the errorsin the thermodynamicdata. Bohlen
and Essene (1978) have discussed some of the
chemographicconsequencesof this.
Two commonfluorine buffersinvolving feldspars
and related to topaz are:
KAISi3OF + Al2sio5 * H2O :
K-feldspar andalusite
KAl2[AlSi3OroXOH)i.+ SiO2
muscovite
quarlz
(2la)
+ Al2SiOaF.i :
KAlSi3CI8"
K-feldspar
topaz
KAl2[AlSi3Oro]FT+ SiO2
quartz
muscovite
(21b)
Thesereactionsrelate the three mineralscharacteristic of greisen-stylealteration-muscovite, topaz,
and quartz. Figure 7 shows the stability fields of
thesephasesrelative to reactionsinvolving ahdalusite and potassiumfeldspar.This figure was calculatedby extractingthe thermodynamicpropertiesof
muscoviteand potassiumfeldsparfrom the experiments of Chatterjeeand Johannes(1974)on (fluorine-free) reaction (Zla). The minimum slope
through their brackets was chosen in order to
minimize the difference between the extracted entropy of reaction and that determined from calorimetry (see discussionof Krupka et al., 1979).No
correction has been made for the ordering of the
BARTON: TOPAZ SOLID SOLUTIONS
potassiumfeldsparwith decreasingtemperatureor
for possibleorderingin the muscovite;thesecorrections are small comparedwith the other uncertainties. The muscoviteF-OH exchangedata of Munoz
and Ludington (1977)were used to calculate new
values for the thermodynamic properties of fluormuscovite.New/f.off;"6 ratios for the AFSQ buffer (usedby Munoz and Ludington) were calculated
from the data of Robie et al. /1978)for fluorite and
the data of Haas et al. (1981)for the other phases.
This recalculation shows that the preferred fugacity
ratios for the muscovite-vapor exchangeare approximately 0.6 log units less than the values derived by Munoz and Ludington.
Reactions(l), (21a), and (21b) are plotted in
Figure 7. The addition of fluorine to the vapor
stabilizesmuscoviteplus quartz relative to andalusite plus K-feldspar plus vapor, while at high fluorine concentrationsmuscoviteplus quartz becomes
unstable relative to topaz plus K-feldspar (reactions (2|a) and (21b)respectively).The point where
all threeof thesereactionsintersectgives the maximum temperaturefor the stability of the assemblage
muscoviteand quartz at that pressure.The region
between the andalusite-muscovite-quartz-K-feldspar reaction and the topaz-muscovite-quartz-Kfeldspar reaction is the region where topaz, muscovite, and qtartz are stable together. This field
representsmany greisens.Addition of other componentsto the systemwill affectthe absolutepositions
of the curves by small amounts while leaving the
topology the same.
Burt (1976)correctly arguesthat psp increases
with increasingpr.o for reaction (21b), however
from chemographyalone it is not possibleto determine the direction of reaction with decreasingtemperature.Although the presentcalculationsdo indicate that on cooling topaz-rkspar-fvapor produces
muscovite * qtartz in agreementwith Burt's conclusion,smallchangesin the entropiesof the F-OH
exchangereactionscould produce the opposite effect. Consequently, the agreement between the
calculationsand observation(below) gives further
support to the applicability of the thermodynamic
model.
Rosenberg(1972b)documentsan exampleof this
in the Brown Derby pegmatite (Colorado) where
early topaz + K-feldspar (*quartz) has been replacedby micas(muscovite/lithiummicas).Muscovite commonlyreplacestopaz in greisensand related deposits (e.9. Panasquiera,Kelley and Rye,
1980;and in other deposits,summarizedin Vlasov,
1967).The additionof other components(suchas Li
or Fe) will increasethe stability of mica relative to
lopaz + K-feldspar. In many cases the micas
replacingtopaz depart significantlyin composition
from muscovite; as a result, they may replace
topazesofhigher fluorine content. For instance,the
lithian muscovitesin the Brown Derby pegmatite
replace topaz with Xo11r : 0.090 (Rosenberg,
r972b).
Granitic porphyry molybdenum deposits are
characteristicallyenriched in fluorine and usually
containzonesin which topaz is a major component
(Mutschler et al., l98l). Broadly speaking,the
zoningin thesedepositsrangesfrom (deepto shallow) potassicto silicic/greisento sericiticto propylitic. In many deposits several mineralizing events
are superimposedand locally alteration may depart
from the typical sequence (Carten et ql., l98l;
Wallaceet al., 1968,1978;White et al., l98l).
Gunow et al. (1980)studied the compositionsof
micas from the alteration zones in the Henderson
(Colorado)porphyry molybdenumdeposit. The F/
(F + OH) ratios in the muscovitesare highestin the
potassic zone and decreasewith decreasingtemperaturesof origin. Theseratios vary from >0.25 to
about 0.10 over a temperaturerangefrom >500'to
<350'C (Gunow et al. (1980)based on the fluid
inclusiondata of Kamilli (1978)).The topaz varies
in compositionfrom Xottr : 0.14 to 0.04 and is not
known to be correlatedwith temperaturein its more
restrictedparagenesis.Comparisonof thesevalues
with Figure 7 shows reasonableagreement.The
Hendersonmuscovites,in some cases,contain appreciableamountsof magnesiumwhich greatly increasestheir tendencvto hold fluorine (Munoz and
Ludington, 1974). The discrepancy between the
fluorine contents found by Gunow et al. and the
values calculated here can be largely attributed to
this factor.
The greisenassemblageswhich constitute characteristicpart of porphyry-stylemolybdenum(tungsten-tin) mineralization,and occur in many other
kinds of deposits,representa fluorine-rich variety
of advancedargillic alteration (Meyer and Hemley,
1967).Fluid inclusion and other geologicevidence
indicates that these deposits probably formed at
pressuresless than I kbar and at temperatures
generally less than 500oC, although the early alteration in some casestook place at higher temperatures(Hall et al.,1974;Ivanovaet al., 1978;Kelley
and Rye, 1980;Thomas,1979:'Wallace et al., 1978
and many others). Boiling of the hydrothermal
970
BARTON: TOPAZ SOLID SOLUTIONS
f.c
- rH20
-3so
4oo
i?3
5oo
55o
Fig. 7. Log fsplfnro versus llT plot showing the calculated
positionsof somereactions in the silica-rich portion of the K2OAI2O3-SiO5H2O-F2O-rat 500 bars. Also shown are isopleths
for hydroxyl-muscovite (dotted) and hydroxyl-topaz (dashed).
Abbreviations are the same as for Fis. 3 and 6.
cutting K-feldspar alteration in the Hendersondeposit have sericitic envelopes(E. Seedorff, pers.
comm.) a sequenceconsistent with simple acidbase exchange (cf. Figure 4 in Burt, 1981) and
similar to the patterns seenin many greisens(Vlasov, 1967).
Addition of fluorine increases the stability of
muscovite + qvartz with respect to K-feldspar +
aluminum silicate + quartz (Figure 8). The maximum stability of muscovite is achieved at topaz
saturationwhere muscovite + quartz: K-feldspar
* andalusite+ topaz + vapor. Increasingthe HF
fugacity at lower temperaturesresults in complete
decompositionof the mica (cf. Figs. 7 and 5). The
topazphaserelationsin Figures6-8 are topologically consistentwith the acidity-salinity diagramspresentedby Burt (1981).
The additionof fluorine as HF (Wyllie and Tuttle,
1961; Glyuk and Anfilogov, 1973a; Kovalenko,
1977;Manning,1981)or as a salt (Anfilogovet al.,
1973;Glyuk and Anfilogov, 1973b;P. J. Wyllie,
personalcommunication)to granitic bulk composition has a dramaticeffecton the solidus,lowering it
by at leastseveraltens of degreeseven at low (a few
percent)total fluorine contents. Fluorine-free(taken from Wyllie, 1977)and schematicfluorine-bearing (buffered by topaz * andalusite) solidi for the
and for a muscovite
systemK2O-AI2O3-SiO2-H2O
granite are illustrated in Figure 8. These reactions
fluids is also a common feature (e.g. Landis and
Rye,1974 Ivanovaet al., 1978;and others).These
observationsmay be combined to interpret, very
crudely, the sequenceof alteration assemblagesin
terms of the Figure 7. The aqueousfluids produced
by crystallizationof a fluorine-rich alkalic magma
would be rich in halogen acids and alkali halides
(Burnham, 1979). On entering the surrounding
rocks thesefluids could react with the feldsparsto
produce potassiumfeldspar at the expenseof the
earlier minerals. Continued cooling, perhaps accompaniedby boiling, would drive the fluid composition into the stability field of the aluminum silicates(Montoyaand Hemley, 1975;Burnham,1979;
Burt, 1981)which, in the presenceof significant
amountsof acid fluoride species,would be repre4
sentedby topaz. Further cooling and reactionwith
the primary minerals farther out would make the
solutions more alkaline and move from the topaz P k b 3
field into the muscovitefield. Projectedonto Figure
7 this trajectory would start in the upper right-hand
corner and move to the lower left-hand corner. In
essence,the shift from the topaz field into the mica
field is governed by progressiveremoval of acid
fluoride speciesfrom the fluid by reaction with the
oL
primary minerals. If the HF/H2O fugacity ratio
450
550
600
650
750
500
700
Tol^
were not reducedby either fixation of fluorine in a
IV
solid phase or by hydrogen-alkali exchangethen
Fig. 8. Calculated positions for the muscovite + quartz
the fluid composition would remain in the topaz breakdownreaction for the fluorine-free system and for fluorine
field on cooling. This removal of fluorine from the buffered by andalusite + topaz + vapor. Schematic solidii are
fluid by formation of fluorine-bearingsolids distin- shown for melting in the fluorine-free and topaz-buffered
granite and for the K2O-AI2O3-SiO2guishesthe behavior of fluorine in these systems systemsfor a muscovite
H2O-F2O-r system. See the text for further discussion.
from that of chlorine which is not incorporated to a Abbreviations are: L = liquid, als : aluminum silicate, w =
significant degree in the solids. Topaz veinlets water, and as for Fig. 3 and 6.
BARTON: TOPAZ SOLID SOLUTIONS
971
rate substantial amounts of fluorine up to the point
wheretopazbecomesstable,increasingthe thermal
stabilitiesof these sheet silicatesby up to several
tens of degrees. The hydroxyl content of topaz
bufferedby andalusite,kyanite, and sillimaniteincreaseswith increasing pressure and decreasing
temperature.Low-temperature stability relations of
topaz are uncertain, but calculations based on the
thermodynamic model indicate that topaz is stable
to low temperatureswith a variety of minerals in the
Al2O3-SiO2-HzO-FzO-rsystem. The calculations
predict compositions (XoHr < 0.35) and assemblagesthat accord well with observation.
Under fluorine-rich conditions in more complex
systems topaz usurps the role of the aluminum
silicates. For fluorine contents buffered by topaz,
muscovite may be stable on the solidii of some
granitesin the pressurerange of one to two kilobars. Muscovite plus quartz becomeunstablewith
respect to topaz plus K-feldspar with increasing
fluoride activity. Presence of topaz is generally
restricted to rocks with low calcium contents or
where the calcium is restricted to fluorite. The
anorthite content of plagioclase coexisting with
topazmust be quite low under nearly all conditions.
This fact restricts the development of topaz-rich
rocks to thosewhich either have initially low calcium contentsor those in which most of the calcium
has already been incorporated in fluorite, correSummary
spondingto the generalrestriction of greisensand
Thermodynamic data for topaz solid solutions relateddepositsto low-calciumrocks and the rarity
providea meansfor understandingthe conditionsof of topaz in skarns.
formation of the wide variety of rocks which conAcknowledgments
tain topaz.In this study experimentson the reaction
R. Goldsmith, StevenD. Ludington,
Donald
M.
Burt,
Julian
andalusite+ water : hydroxyl-topaz componentin
AlexandraNavrotsky, Robert C. Newton, and Peter J. Wyllie all
topaz solid solution permit calculationof the ther- provided helpful commentsand reviews. This study was supmodynamicproperties of (hypothetical)hydroxyl- ported by National Science Foundation grants to Julian R.
topaz and evaluation of the applicability of alterna- Goldsmith(EAR78-13675)and Peter J. Wyllie (EAR76-20413).
tive activity models for the solid solution. Both While at the University of ChicagoI was supportedby a National
ideal and proton-avoidance mixing models agree Science Foundation graduate fellowship and by a McCormick
Fellowship from the University of Chicago.
with the experimentaldata, but the proton-avoidancemodel gives much more reasonableresults on
References
extrapolationto lower temperatures.
(1953)Beryllium depositsof the Mount Antero
Combined thermodynamic data on topaz solid Adams, J. W.
region, Chatree County, Colorado. U.S. Geological Survey
solutions from this study and that of Barton et al.
Bulletin 982-D,95-119.
(1982)allow calculationof many reactionsinvolving Akizuki, M., Hampar,M. S. and Zussman,H. (1979)An explanation of anomalousoptical properties of topaz. Mineralogical
topaz and some coexisting fluid compositions.
Magazine,43,237-241.
These calculations show that topaz is efficient at
Alston, N. A. and West, J. (1928) The structure of topaz.
removing HF from a coexisting vapor phase, an
Proceedings
of the Royal Society,A., 121,358-367.
efhciencywhich increaseswith decreasingtempera- Althaus, E. (1966)Stability rangeofpyrophyllite under the effect
ture. Topaz is richer in fluorine than coexisting
of acids. I. Experimental study. Contributions to Mineralogy
and Petrology,13, 3l-50.
sheetsilicates,althoughsuch silicatesmay incorpointersectwith the muscovite + quartz breakdown
reactionsto give invariant points for the respective
systems.The curves betweenthe pairs of invariant
points representthe topaz-missingreactions. Not
shown are several other melting reactions that
radiate from each of these invariant points. The
extra lines connecting the two pairs of invariant
pointsfor the potassiumsystemand for the muscovite graniteare the topaz-missingmelting reactions
for the respectivefluorine-bearingsystems.One of
the consequencesof this topology is that the minimum pressurefor the existenceof muscoviteon the
solidus is lowered by about one-half, helping to
explain the presenceof muscovite in many twomica granites which have features incompatible
with higher pressures(see Miller et al., l98l).
Another consequenceis that subsolvus(two alkalifeldspar) granites can form down to quite low
pressures.Pichavant(1981)draws similar conclusionsfor boron-rich magmasbased on melting experiments.
Topaz-bearinghypabyssal and volcanic rocks,
are known in some areas (Eadingtonand Nashar,
1978;Burt et al., 1980;Kovalenko e/ al., 1972:
Lindsey, 1979).The topazfrom theserocks tendsto
havea very low hydroxyl contentwhich is compatible with a high-temperature,low-pressure,but not
necessarilymagmatic,origin.
972
BARTON: TOPAZ SOLID SOLUTIONS
Anfilogov, V. N., Glyuk, D. S. and Trufanova, L. G. (1973)
Phase relations between granite and NaF at a water vapor
pressureof 1000 kglcmz. Geokhimiya, 44-48 (transl. GeochemistryInternational,10, 30-33, 1973).
Bailey, J. C. (1977) Fluorine in granitic rocks and melts: a
review. ChemicalGeology, 19, l-42.
Barton, M. D., Haselton, H. T., Hemingway, B. S., Kleppa,
O. J. and Robie, R. A. (1982)The thermodynamicproperties
of fluor-topaz. American Mineralogist, 67, 350-355.
Beus, A. A. and Dikov, Yu. P. (1967)The Geochemistryof
Beryllium in the Processes of Endogenetic Mineralization.
Nedra. Moscow.
Bohlen, S. R. and Essene, E. J. (1978) The significanceof
rnetamorphicfluorite in the Adirondacks. Geochimica et Cosmochimica Acta. 42. 1669-1678.
Bumham, C. W. (1962) Lattice constant refinement. Carnegie
Institution of WashingtonYearbook, 6l, 132-115.
Burnham, C. W. (1979) Hydrothermal fluids at the magmatic
stage.In H. L. Barnes, Ed., Geochemistryof Hydrothermal
Ore Deposits,2nd edition, p. 7l-76, John Wiley and Sons,
Inc., New York.
Burnham, C. W., Holloway, J. R., and Davis, N. F. (1969)
Thermodynamic properties of water to 1,000'C and 10,000
bars. Geological Society ofAmerica Special Paper 132.
Burt, D. M. (1975a)Beryllium mineral stabilities in CaO-BeOSiO2-P2OrF2O-1 and the breakdown of beryl. Economic
Geology, 70, 1279-1292.
Burt, D. M. (1975b)Natural fluorine buffers in the system BeOAl2O3-SiO2-F2O-r.(abstr.) Transactionsof the American
GeophysicalUnion, 56, 467.
Burt, D. M. (1976)Generationof high HF fugacities in miarolitic
pegmatitesand granites: the possible role of topaz. (abstr.)
Geological Society of America Abstracts with Programs, 8,
798.
Burt, D. M. (1979)Geometrical analysis of cryolite and related
mineral stabilities in the system Na2O-AI2O3-SiO2-F2OI.
(abstr.) Transactionsof the American Geophysical Union, 60,
965.
Burt, D. M. (1980) Acidity-salinity diagrams: Application to
greisenand porphyry deposits. (abstr.) Geological Society of
America Abstracts with Programs, 12,396.
Burt, D. M. (1981) Acidity-salinity diagrams-application to
greisenand porphyry deposits.Economic Geology,76,832843.
Burt, D. M., Bikun, J. V. and Christiansen,E. H. (1980)Distribution and petrogenesisof topaz rhyolites, western U.S.A.
(abstr.) Geological Society of America Abstracts with Programs,12,396.
Carten,R. B., Shannon,J. R., Ward, A. D. and Geraghty,E. P.
(1981)Controlson molybdenitedepositionat Hendersonmine,
Ernpire, Colorado. (abstr.) Geological Society of America
Abstractswith Programs,13, 423.
Chatte{ee,N. D. and Johannes,W. (1974)Thermalstabilityand
standardthermodynamic properties of synthetic 2M1-muscovite, KAI [AlSi3Oro](OH)2.
Contributionsto Mineralogyand
Petrology,48, 89-114.
Coes, L., Jr. (1955) High pressure minerals. Journal of the
American Ceramic Society, 38, 298-312.
Deer, W. A., Howie, R. A. and Zussman, J. (1962) Rockforming Minerals. Vol. l, Ortho- and Ring Silicates, John
Wiley and Sons,Inc., New York.
Desborough,G. A., Ludington, S. D. and Sharp, W. N. (1980)
Redskin granite: a rare-metal-rich Precambrianpluto, Colorado, U.S.A. MineralogicalMagazine,43, 959-966.
Eadington,P. J. and Nashar,B. (1978)Evidencefor the magmatic origin of quartz-topaz rocks from the New England Batholith, Australia. Contributions to Mineralogy and Petrology, 67,
433-438.
Fries, C., Jr. (1942)Topaz deposits near the Brewer mine,
ChesterfieldCounty, South Carolina,U.S. GeologicalSurvey
Bulletin936-C,59J8.
Glyuk, D. S. and Anfilogov, V. N. (1973a)Phaseequilibria in
the system granite-H2o-HF at a pressure of 10ffi kglcmz.
Geokhimiya, 434-438(transl. Geochemistry International, 10,
32r-32s, 1973).
Glyuk, D. S. and Anfilogov, V. N. (1973b)Phaseequilibria in
the system granite-H2O-KF at water vapor pressuresof 1000
kglcm2.Doklady AkademiaNauk S.S.S.R.,210,918-940.
Grover, J. E. (1980)Thermodynamicsof pyroxenes.In P. H.
Ribbe, Ed., Reviews in Mineralogy, Yol.7, Pyroxenes,p.
341-418.MineralogicalSocietyof America,Washington,D.C.
Grover, J. E. and Orville, P. M. (1969) The partitioning of
cations between coexisting single- and multi-site phaseswith
application to the assemblagesorthopyroxene-{linopyroxene
and orthopyroxene-olivine. Geochimica et Cosmochimica
Acta. 33. 205-226.
Gunow,A. J., Ludington,S. and Munoz, J. L. (1980)Fluorinein
micas from the Hendersonmolybdenite deposit, Colorado.
.
Economic Geology, 75, 1127-1137
Haas,J. L., Robinson,G. L., Jr. and Hemingway,B. S. (1981)
Thermodynamic tabulationsfor selectedphasesin the system
at l0l.325 kPa (l atmosphere)between
CaO-AI2O3-SiO2-H2O
273.15and 1800K.Physicaland ChemicalReferenceData, 10,
575-669.
Hall, W. E., Friedman,I. and Nash,J.T. (1974\Fluid inclusion
and light-stableisotope study of the Climax molybdenum
deposits,Colorado.EconomicGeology,69, 884-901.
Helgeson,
H. C., Delany,J. M., Nesbitt,H. W. andBird, D. K.
(1978)Summaryand critiqueof the thermodynamicproperties
of rock-formingminerals.American Journal of Science,278A, r-229.
Hemley, J. J., Montoya, J. W., Marinenko, J. W. and Luce,
R. W. (1980)Equilibria in the system Al2O3-SiOrH2O and
some general implications for alteration/mineralizationprocesses.EconomicGeology,75, 210-228.
Hildebrand,F. A. (1961)Andalusite-topazgreisinnear Caguas,
East-centralPuerto Rico. U.S. GeologicalSurvey Professional
Paper 424-8,222J24.
Ivanova,G. F., Morotina, Z.M. and Naumov, V. B. (1978)
Formation of features of the mineral associations of the
Yugodzy molybdenum-tungsten deposit (Mongolia). Akademy Nauk SSSR GeologiyaRudnykh Mestorozhdeny,19,
26-36 (transl. International Geology Review, 20, 855-863,
1978).
Kalyuzhnaya,K. M. and Kalyuzhnyi,V. A. (1963)The paragenesis of accessory beryl, phenakite, and euclase in topazmorion (smoky qtrartz) pegmatites. Mineralogiskii Sbornik
(L'vov), 17,135-147.
Kamilli, R. J. (1978) The genesis of stockwork molybdenum
deposits:implications from fluid inclusion studies at the Hendersonmine. (abstr.) GeologicalSociety of America Abstracts
with Programs,10, 431.
Kelly, William C. and Rye, Robert O. (1979) Geotogic, fluid
inclusion, and stableisotope studiesof the tin-tungsten depos-
BARTON: TOPAZ SOLID SOLUTIONS
973
its of Panesqueira,Portugal. Economic Geology, 74, l72lMunoz, J. L. and Ludington, S. (1974)Fluorine-hydroxyl ex1823.
changein biotite. American Journal of Science, 274, t96-413.
Kempe,D. R. C. (1967)Some topaz-, sillimanite-,and kyanite- Munoz, J. L. and Ludington, S. (1977)Fluorine-hydroxyl exbearing rocks from Tanzania. Mineralogical Magazine, 280,
changein synthetic muscovite and its application to musco515-52r.
AmericanMineralogist,62, 304-308.
vite-biotiteassemblages.
Kerrick, D. M. and Darken, L. S. (1975)Statisticalthermody- Mutschler, F. E., Wright, E. G., Ludington, S. and Abbott,
namic models for ideal oxide and silicate solid solutions with
J. T. (1981)Granitemolybdenitesystems.EconornicGeology,
applicationto plagioclase.Geochimicaet CosmochimicaActa,
76,874-897.
39, t43l-1442.
Newton, R. C., Charlu,T. V. and Kleppa, O. J. (1980)ThermoKogarko, L. N. (1966) Physico-chemical analysis of cryolite
chemistry ofthe high structural stateplagioclases.Geochimica
paragenesis.
Geokhimiya,1300-1310.
et CosmochimicaActa, 44,933-941.
Kovalenko, N. I. (1977) The reactions between granite and
Olsen, D. R. (1971)Origin of topaz depositsnear Ouro Preto,
aqueoushydrofluoric acid in relation to the origin offluorineMinas Gerais,Brazil. EconomicGeology,6,627-631.
bearinggranites. Geokhimiya, 503-575(transl., Geochemistry Parise, J. B. (1980) Order-disorder phenomenain minerals:
International.14. 108-118.1977).
Structuralstudies.Ph.D. Dissertation,JamesCook University
Kovalenko,V. I., Fin'ko, V. I., Letnikov, F. A. and Kuzmin,
of N. Queensland.
M. I. (1972) Certain effusive and subvolcanic rocks with
Pauling,L. (1928)The crystal structure of topaz. Proceedingsof
relatively high concentrations of rare elements. International
the NationalAcademy of Sciences,U.S.A., 14, 603-606.
GeologyReview, 14, 599-608.
Pensfield,S. L. and Minor, J. C., Jr. (1894)On the chemical
Krupka, K. M., Robie, R. A. and Hemingway, B. S. (1979)
composition and related physical properties of topaz. Ameri
High-temperatureheat capacitiesof corundum, periclase, ancan Journalof Science,3rd series,47,387:396.
glass,muscovite,pyrophyllite, KAISi3OE Pichavant,M. (1981)An experimentalstudy of the effect of
orthite, CaAl2Si2O6
glass,grossular,and NaAlSi3O3glass.American Mineralogist,
boron on a water saturated haplogranite at I kbar vapour
pressure.Contributionsto Mineralogyand Petrology,76,43064,86-101.
Kupriyanova, I. L (1970)Parageneticanalysisofmineral associ439.
ations in beryllium-bearing greisens. Zapiski Vsesoyuznogo Ream,L. R. (1979)The ThomasRange,Wah Wah Mountains,
MineralogischekogoObshchestva,99, 277-285.
and vicinity, western Utah. Mineralogical Record, lO, 261Landis, G. P. and Rye, R. O. (1974)Geologic, fluid inclusion,
278.
and stable isotope studies of the Pasto Bueno tungsten-base Ribbe, P. H. (1980)Topaz. In P. H. Ribbe, Ed., Reviews in
metal ore deposit, northern Peru. Economic Geology, 69,
Mineralogy, Vol. 5. Orthosilicates, p. 215-230. Mineralogical
1025-1059.
Societyof America, Washington,D.C.
Lindsey, D. A. (1979)Preliminary report on Tertiary volcanism
Ribbe, P. H. and Gibbs, G. V. (1971)The crystal structureof
and uranium mineralizationin the Thomas Rangeand northern
topaz and its relation to physical properties. American MinerDrum Mountains,JuabCounty, Utah. U.S. GeologicalSurvey
alogist,56, 24-30.
Open-fileRept. 79-1076.
Ribbe, P. H. and Rosenberg,P. E. (1971)Optical and X-ray
Manning,D. A. C. (1981)The efect of fluorineon liquidusphase
determinativemethods for fluorine in topaz. American Minerrelations in the system Qz-Ab-Or with excess water at I kb.
alogist,56, l8l2-lE2l.
Contributionsto Mineralogyand Petrology,76,206-215.
Roberson,C. E. and Barnes, R. B. (1978)Stability of fluoride
MacKenzie, W. B. (1970) Hydrothermal alteration associated
complex with silica and its distribution in natural water
with the Urad and Henderson molybdenite deposits, Clear
systems.ChemicalGeology,2l, 25-J4.
Creek County, Colorado. Ph.D. Dissertation,University of
Robie,R. A., Hemingway,B. S. and Fisher, J. R. (1978)TherMichigan, Ann Arbor.
modynamic properties of minerals and related substancesat
298.15 K and I bar (105 Pascals)pressure and at higher
Meyer, C. and Hemley, J. J. (1967)Wall rock alteration. In
temperatures.U.S. GeologicalSurvey Bulletin 1452.
H. L. Barnes, Ed. Geochemistryof Hydrothermal Ore DeposRosenberg,P. E. (1972a)Compositional variations in synthetic
its, lst edition, p. 166-235.Holt, Reinhart, Winston, New
topaz.AmericanMineralogist,57, 168-187.
York.
Michel-Levy,A. and Wyart, J. (1946)Reproductionsynthetique Rosenberg,P. E. (1972b)Paragenesisof the topaz-bearingportion of the Brown Derby No. I pegmatite, Gunnison County,
de la topaz, de la cryolite et du corindon. Bulletin de la Soci6td
Colorado. American Mineralogist, 57, 57l-583.
Francaisede Mindralogieet de Cristallographie,69, 156-161.
Rosenberg,P. E. (1978) Fluorine-hydroxyl exchange in topaz
Miller, C. E., Stoddard, E. F., Bradfish, L. J. and Dollase,
(abstr.). Transactions of the American Geophysical Union
W. A. (1981) Composition of igneous muscovites: genetic
E O S . 5 9 .1 2 1 8 .
implications. CanadianMineralogist, 19, 25-34.
Sainsbury, C. L. (1964) Geology of the Lost River mine area,
Montoya,J. W. and Hemley, J. J. (1975)Activity-+omposition
Alaska. U.S. GeologicalSurvey Bulletin 1129.
relations and stabilities in alkali feldspar and mica alteration
Sheridan,D. M., Taylor, R. B. and Marsh, S. P. (1968)Rutile
reactions. Economic Geology, 70, 577-583.
and topaz in Precambrian gneiss Jefferson and Clear Creek
Morey, G. W. and Ingerson,E. (1937)The pneumatolyticand
Counties, Colorado. U.S. Geological Survey Circular 567.
hydrothermal alteration and synthesis of silicates. Economic
Smirnov, V.l. (1977)Ore Depositsof the U.S.S.R., Vol. III.
Geology,32,607J60.
Pitman Publishing. San Francisco.
Munoz,J. L. and Eugster,H. P. (1969)Experimentalcontrol of
Stull, D. R. and Prophet, H. (1971)JANAF Thermochemical
fluorine reactionsin hydrothermal systems.American MinerTables.2nd edition.f.{ational StandardReferenceData Series,
alogist, 54, 943-959.
NationalBureauof Standards(U.S.), 37.
974
BARTON: TOPAZ SOLID SOLUTIONS
Sykes,M. L. and Moody, J. B. (1978)Pyrophyllite and meramorphism in the Carolina slate belt. American Mineralogist,
63, 96-108.
Thomas,R. (1979)Untersuchungvon Eisnschlussen
zur thermodynamischen und physikochemischen Charakteristik lagerstattenbildender Losungen und Prozesse im magmatischen
und postmagmatischen
Berieich.Ph.D. Dissertation,Bergakademie Freiberg, Freiberg.
Vlasov, K. A. (1967) Geochemistry and Mineralogy of Rare
Elements and Cenetic Types of their Deposits, Vol. Iil,
Genetic Types of Rare Element Deposits. Nauka, Moscow
(transl. Israel Programfor Scientific Translations, Jerusalem).
Wallace,S. R., Muncaster,N. K., Jonson,D. C., MacKenzie,
W. B., Bookstrom,A. A. and Surface,V. F. (1968)Multiple
intrusion and mineralization at Climax, Colorado. In J. D.
Ridge, Ed., Ore Deposits of the United States, 1933-1967
(Graton-Salesvolume), 605-640, American Institute of Mining and MetallurgicalEngineers.
Wallace,S. R., MacKenzie,W. B., Blair, R. G. and Muncaster,
N. K. (1978)Ceologyof the Urad and HendersonMolybdenite
Deposits,Clear Creek County, Colorado,with a sectionon a
comparisonof these depositswith those at Climax, Colorado.
EconomicGeology,73, 325-368.
White, W. H., Bookstrom, A. A., Kamilli, R. J., Ganster,
M. W., Smith, R. P., Ranta, D. E. and Steininger, R. C.
(1981)Characterand origin of Climax-typemolybdenumdeposits.EconomicGeology,Seventy-fifthAnniversaryvolume,
270-316.
Wyllie, P. J. (1977)Crustal anatexis:an experimentalreview.
Tectophysics,
43,4lJL
Wyllie, P. J. and Tuttle, O. F. (1961)Experimentalinvestigation
of silicate systems containing two volatile components. Part
IL The effects of NH3 and HF, in addition to H2O on the
melting temperaturesof albite and granite. American Journal
of Science.259, 128-143.
Yoder, H. S. and Eugster, H. P. (1955)Syntheticand natural
muscovites.Geochimicaet CosmochimicaActa, 8,225-280.
Manuscriptreceived,February24, 1982;
acceptedfor publication,June 10, 1982