JOURNAl. OF GEOPHYSICALRESEARCH
VOL. 74, NO. 25, NOVEMBER15, 1969
Oxygen IsotopeEquilibriumbetweenMuscoviteand Water
JAMES R. 0'NEIL
U.S. Geological Survey, Menlo Park, Cali[o.rnia 94025
HUGH P. TAYLOR,JR.
Division o• Geological Sciences
California Institute o.• Technology, Pasadena, California 91109
Oxygen isotopes have been equilibrated between muscovite and aqueous alkali chloride
solution and between paragonite and alkali chloride solution in the temperature range of
400ø-650øC at I and 1.5 kb fluid pressure. Isotopic equilibrium was inferred from the fact
that compatible fractionation factors were obtained using 3 different chemical reactions to
produce the mica: (1) muscovite or paragonite was prepared by reacting natural kaolinire
with 2-3 molal KC1 or NaC1 solutions; (2) muscovite was crystallized in pure water from a
gel; and (3) synthetic paragonite was reacted with 2-3 molal KC1 solution, producingmuscovite by an alkali ion exchangereaction. The I M modification of the mica was made in
every experiment. In several casesthe extent of oxygen exchangewas traced by running com-
panion equilibrationsin solutionsof unusuallylow O's/ox6ratio. No isotopicfractionationwas
discerniblebetween muscovite and paragonite in the temperature range studied. Per mille fracrionations between muscovite and water are given by the expression 10s In a • 2.38(106T-2)
-- 3.89. These data can be combined with the results of other laboratory equilibration studies
to establish a set of calibrated oxygen isotope geothermometers.Analogous to the alkali
feldsparsystemspreviouslyreported,the direct relationshipbetweencation and oxygenisotope
exchangesuggeststhat some type of solution-redeposition
mechanismoperatedduringmuscovite-paragonitetransformationsin aqueoussolutions.Also, the extensiveoxygenisotope
exchange(with the solution)that accompanies
the formationof muscovitefrom kaolinite
impliesa breakdownof the kaolinitcstructure.This notiondoesnot concurwith hypotheses
based on rate studiesand X-ray measurementsthat the unaltered kaolinitc structureis partially inherited by the mica.
the sequenceof commonrock-formingminerals
with respectto the tendency to concentrate0 3 .
in nature. It is formedunder a wide variety of A set of coexistingmineralscommonlyexhibits
decreasing
O•/O • ratiosin the sequence:
quartz,
geologicconditionsand is abundantin finealkali feldspar, sodic plagioclase,calcic plagiograined sedimentsand many metamorphic
rocks, as well as in some aluminousigneous clase,muscovite,hornblende,biotite, and magrocks.Extensivelaboratory investigationshave netite. Other analysesof muscovitehave been
elucidated the nature of its various stacking reportedby Gatlick and Epstein [1967], Taylor
INTRODUCTION
Muscovite is the most commonmica occurring
polymorphsand their stability fields [Yoder
and Eugster,1955; Velde,1965, 1966]. Oxygen
isotopeanalysesof muscoviteswere first reported by Taylor and Epstein [1962a] and
Taylor et al. [1963], who showedthat, in general, muscoviteoccupies
a specificpositionin
x Publication authorizedby the Director, U.S.
GeologicalSurvey, Publicationsof the Division of
GeologicalSciences,California Institute of Technology, Contribution No. 1479.
Copyright ¸
1969 by the American Geophysical Union.
[1967, 1968], Taylor and Coleman[1968], and
Shieh and Taylor [1969]. These authorsindicate that oxygen isotope fractionationsbetweenquartz and muscovitein metamorphosed
rocks generally decreasewith an increasing
grade of metamorphismas judged on petrologicaland geologicalgrounds.Measuredfractionations range from about 2.0%øin igneous
rocks and high-grade metamorphic rocks to
5.0%øin low-graderocks.
The laboratory study of the equilibrium
oxygenisotopepropertiesof muscovite
reported
6012
OXYGEN
ISOTOPE EQUILIBRIUM
6013
1. Formation oJ I M mica [tom natural
•n this paper is a direct outgrowthof a similar
study on the feldspars [O'Neil and Taylor, kaolinire.
1967]. In addition to obtaininginformationon 0.75 Al,(OH)sSi4Olo-]- Na +
the nature of the mineralogicalreactionsinkaolinite
volvingmicas,these studies(in conjunction
NaAI,(OH)•A1Si•O,0-[- H + -[- 1.5
paragonite
with thoseof Clayton, O'Neil, and Mayeda on
the quartz-watersystem)establish
a calibrated
quartz-muscoviteoxygenisotopethermometer. Natural kaolinite from the McNamee mine in
This pair will probablyprove to be one of the Georgia,with a potassiumcontentof lessthan
most usefulamongcoexistingsilicatesbecause 0.1% wasusedin this procedure.The solutions
of the ubiquityof the mineralsand the moderatelyhightemperaturesensitivityof the isotopic fractionations.
EXPERIMENTAL
PROCEDURES
All hydrothermal experimentswere carried
out in sealedgoldcapsules
usingcoldseal
bombs.
Temperatureswere monitored with ChromelAlumel thermocouples and were constant
throughoutthe run to ---+3øC.The fluid pressures were maintained
at 1000 or 1500 bars.
The bombswere quenchedto room temperature
in a few minutesby a combinationof cold air
blasts and immersion in cold water. When large
coldsealbombs were used, the quench times
were of the order of 5-10 min, but the differencesin quenchtimeshad no discernible
effects.
After quench,the solidswere removedfrom the
goldcapsules,
washedin water and acetone,and
dried in an •ven. The oxygen was liberated
from the solidsby reaction with fluorine gas at
500øC [Baertschi and Silverman, 1951; Taylor
and Epstein, 1962a] and then converted to
carbon dioxide for analysison the mass spectrometer.This analyticalprocedurehasan error
of --+0.1-0.2%o
for a singleanalysis.The starting
solutionswere analyzed directly for 0 • content
by the BrF• method of O'Neil and Epstein
[1966]. The post-reaction solutions were not
analyzed, but becausethis oxygen constituted
about 93% of the oxygenin the system,their
isotopiccompositionswere calculatedwith sufficient accuracy by material balance.
The oxygenisotopeexchangerates of natural
micas were found to be prohibitively slow in
pure water, and other methods (simple recrystallizationin waters of different isotopic
composition)had to be developed.The following procedureswere ultimately used in obtaining the oxygen isotope fractionation factors
betweenmuscovite (and paragonire) and water.
were 2-3
molal
alkali
chloride.
No
reaction
occurred after 1176 hours at 290øC. However,
at higher temperatures, reaction times were
extremely rapid. In one case, complete trans-
formation (determinedby X-ray diffraction)
was observedafter only 15 minutesat 500øC.
The nature of this reaction will be discussed in
a later section.Paragonireor muscovite(the
1 M modificationin all cases) was produced
with equal facility by using either sodium
chlorideor potassiumchloridesolutions.Most
of the runs were made with startingmaterials
that were alreadyvery near isotopicequilibrium. The 3-values of the kaolinire, Water-7,
and Water-2 are +4.0, +4.9, and + 1.2, respec-
tively. Thesematerialsofferthe advantageof
beingclosein isotopiccomposition
to the working standardof the massspectrometer.
Four
runs were made in Water-4, which is about
40%0lighterin 0•8/O•6ratio than Water-7,and
five runs were made in a water of intermediate
composition(Water-3). Grossoxygenisotope
redistributionaccompanied
all thesetransformations,and at a given temperaturethe resultant mica-water fractionations were the same
in all three solutions. The observed isotopic
fractionationscould be a consequence
of (a)
equilibriumisotopedistribution,(b) kinetic
isotope
fractionation
effectsassociated
with the
chemicalreaction,or (c) incomplete
exchange.
2. Crystallization
of muscovitegel. A gel
of muscovitecomposition
was crystallizedin
purewater.The temperature
rangeoverwhich
this procedure
can be usedis very smalland
onlyoneexperiment
wasmade(at 600øC).
3. Alkali ion (and 0 TM)exchangereaction.
NaAla(OH)aAlSisOx0
4- K +-•
paragonitc
KAl•(OH)•A1SisOx0-4- Na +
muscovite
O'NEIL
6014
TABLE
AND
Oxygen Isotope Fractionetlon Data for ProceduresI and 2
1.
Description*
Sample No.
TAYLOR
Temp., øC
Time, hours
Wateri
l0 sIn a
Procedure I
366
56
355
349
351
358
317
58
61A
6lB
73
346
352
354
139
142
341
342
347
Kaol. --2M
Kaol. --2M
Kaol. --3M
Kaol. --2M
Kaol. --2M
Kaol. --2M
Kaol. --2M
Kaol. --3M
Kaol. --3M
Kaol. --3M
Kaol. --2M
Kaol. --2M
Kaol. --2M
Kaol. --2M
Kaol. --3M
Kaol. --3M
Kaol. --2M
Kaol. --2M
Kaol. --2M
NaC1
KC1
NaC1
KC1
NaC1
KC1
KC1
NaC1
650
630
610
600
600
600
550
510
2
15
13
90
90
21
25
71
7
3
7
7
7
4
2
3
- 1.26
-0.94
NaC1 (grain exteriors)
NaC1 (grain interiors)
510
510
180
180
4
4
KC1
510
KC1
NaC1
KC1
NaC1
KC1
KC1
KC1
I(C1
500
500
500
420
420
400
400
400
Procedure
106
* Kaol.
Musc. gel q- HaO
= kaolinire. Musc.
0.3
360
304
120
635
635
1176
1176
523
3
7
7
7
3
3
4
7
4
-0.30•
-0
-0
81
82
-o
35•
-0
30
-o
60•
+o
05
(+2
o4)•
-0
-0
02
21
-o
53•
+o
q,o
+1
+1
+1
+1
05
95
07
36
21
51
2
600
114
7
-0.78
= muscovite.
I •Water-2= q.1.16, •Water-3-- --11.73, •Water-4= --35.96, •Water-7= q.4.93.
$ Not used in calculatingthe least-squares
expression(seetext). Parenthesesindicatedemonstrably
nonequilibriumvalue.
Reactant
materials
were micas made from
in 0•8/0 TMratio than Water-2. Even though
kaolinitc,as in procedure(1), and 2-3 molal these runs indicate that isotopicexchangewas
alkali chloride solutions.These exchangereac-
only 65 to 85% complete, the runs made in
tions are directly analogousto thoseobserved Water-2 (and Water-7) started suft%ientlyclose
in alkali feldsparsystems[O'Neil and Taylor, to the equilibrium values that the data points
1967]. Alkali contentsof muscovitefrom one
can be used to provide equilibrium informa-
run weredetermined
by emission
spectrographic tion (see below).
analysis.The results(K > 5%; Na < 1%)
indicated extensive but incomplete alkali ex-
changein 67 hoursat 500øC.It is interesting
EXPERIMENTAL
RESULTS
The oxygen isotope fractionations and descriptionsof the experimentsin which they were
at a markedly slower rate than the kaolinitc measuredare given in Tables 1 and 2. The 350reactionsof procedure(1), and they alsooccur 650øC temperature range of these experiments
more slowlythan the analogousexchangereac- is fixed by muscovitebreakdownat high temtions in alkali feldspars.Most of the mica ex- peratures and slow exchangerates at low temchangerunsalsoweremadewith materialsthat peratures. For the kaolinRe reactions,apparent
were alreadyvery near isotopicequilibrium.In
equilibrium values were obtained using four
these runs, the starting mica was preparedin different waters (not at each temperature,howWater-2 (or in one case with Water-7) and ever), and thus the equilibrium distribution of
then exchangedwith the appropriate alkali isotopeswas approachedfrom: (a) both sides
chloride solution also made with Water-2 (or
of the equilibrium curve, and (b) a variety of
Water-7). Five companionruns were made in 'distances' away from equilibrium. The fracWater-4, which is approximately35%0lighter tionat.ionsare givenin terms of 1000In a where
to note that thesemica-exchangereactionsoccur
OXYGEN
ISOTOPE EQUILIBRIUM
6015
anomalieshave been observedin previous laboratory calibrationsof oxygenisotopegeothermometers;consideringthe number of stepsin
a is essentiallythe equilibrium constantfor the the proceduresand the precisionentailed,these
isotope exchangereaction between muscovite anomaliesare to be expected.
and water (written so that a single atom is
The arrows shown in Figure I indicate the
exchanged):
directionin whichthe value of 10• In a changed
during each run. The runs made in the low 0 TM
• KA12(A1Si30,d')(O•'H),.
q- H20•-•
waters (3 and 4) have arrows pointing downward since the silicate material becomespro-• KAla(A1SiaOx0xs)
(OxsH),.
qgressivelylighter. The oppositeis true for runs
Excluding the stripping experiment datum made in the heavier Water-7. Runs in Water-2
(6lB), whichis discussed
below,Table 1 shows18 are so closeto equilibrium at the start that the
fractionations
for procedure(1) measured
at var- direction of the arrow dependson the temperaious temperatures.Fourteen of these fraetiona- ture of the run. Only one run was made in
tions are possibleequilibrium values,and when Water-2, and for this case the arrow points
they are plotted as in a versusT-• in Figure 1, downward. Note that the silicate material
the data points lie on a straight line given by
changedby almost40%0in the four runs made
the expression:10• In a = 2.43(10•T-•) -- 4.02. in Water-4. Dependingon the exact mechanism
A straight line relationshipis justified by theo- o.fthe kaolinire-micareaction,the two-direction
retical argumentsof Urey [1947] and Bigeleisen approachoutlined here may or may not be a
and Mayer [1947] andby previousexperimental valid test of equilibrium.
work. The four points in Figure I that do not
Data points from procedures(2) and (3) are
lie on the least-squaresline representruns that
plotted in Figure 2 and are comparedwith the
appeared normal in every respectother than least squaresline from Figure 1. The runs made
the final isotopiccompositionof the mica. These in the light Water-4 are not shown in this
points deviate only about 0.5%0from the least- figurebecausethey clearlyrepresentnonequilibsquaresline; this amount is considerablyout- rium. It is seen that the points from proceside the presumed analytical error, however, dures (2) and (3) are in grossagreementwith
and the points are therefore rejected.Similar the data from procedure(1). In particular, the
(0'8/0 ' ') muscovite
(0•8/0 •6) water
TABLE 2. OxygenIsotope Fractionation Data for Procedure3
Sample
No.
241
360
367
356
357
233
244
326
292
236
305
306
Description*
Syn.parag. q- 3 M
Syn. parag. q- 3 M
Syn. parag. q- 3 M
Syn. parag. q- 2 M
Syn. parag. q- 2 M
Syn. parag. q- 3 M
Syn.parag. q- 3 M
Syn. musc. q- 3 M
Syn. parag. q- 3 M
Syn. parag. q- 3 M
Syn.parag. q- 2 M
Syn. parag. q- 2 M
KC1
KC1
KC1
KC1
KC1
KC1
KC1
NaC1
KCl
KC1
KC1
KC1
Temp.,
øC
Water
Time,
hours
103In {,l
600
600
600
550
550
500
500
500
420
400
350
350
2
4
4
7
4
2
2
2
2
2
2
4
95
44
160
62
62
283
312
216
1050
820
1120
1120
-0.73
(q- 11.76)
(+5.52)
-0.26
(q-13.81)
-0.46
+0.15
-0.50
q-0.53
q-0.68
+2.21
(q- 15.25)
%
Exchangeõ
82?
74
86
65
65
80?
60?
65?
60?
64
64
103In ao
-0.79
+0.18
+0.29
-0.10
q-1.03
q-1.41
q-3.67:1:
* Syn. = synthetic;musc. = muscovite;parag. = paragonire.
i Valuesin parentheses
are demonstrablynon-equilibriumand are not shownin Figure2. ResultNo. 233
is anomalousand not comparableto resultNo. 326; startingmaterialsare not the samepreparationsand
differ by 0.6•ooin Ots/O• ratio.
$ Not usedin calculatingleast-squares
expression
of Figure3 (seetext).
õ Questionmark indicatesestimated values.
6016
O'NEIL
AND TAYLOR
T (øC,)
(•50
600
550
500
I
I
I
I
4OO
55O
i
i
the extrapolationtechniquemore difficult. In
fact, Savin and Epstein [1968] have suggested
that hydroxyloxygenin claymineralsexchanges
at a considerably
faster rate with aqueoussolutions than doesaluminosilicate
oxygen,and also
that the hydroxyloxygenis isotopicallydifferent from the aluminosilieateoxygen.
Ianalytical
error
=2.4 3 ('106T-2)_ 4.0 2
t%t
•e/te
In the absenceof true companion runs, the
degreeof exchangehad to be estimatedin most
cases.This was doneby noting that all the runs
in light water exchangedto the extent of 64 to
86% and that, as expected,runs of the longest
duration and highesttemperaturesresultedin
--1.0
•.o
4.5
Althoughit may not be strictly legitimateto
apply the partial-exchangetechniqueto the
mica-exchangeruns for the reasonsoutlined
above, those runs made in Water-2 started so
closeto equilibriumthat the final valuesof 108
In a are very insensitiveto the per cent of
isotopicexchange.The measuredvaluesof 108
In a for procedure(3) and the calculatedequilibrium values 10• in a,eare given in Table 2.
I
I
2.0
2.5
to 6 T-2
Fig. 1. Plot of 1000 In a for muscovite-water
against 106T-2, where T is the absolute temperature in øK, for data obtained by procedure 1. A
least-squares line has been drawn through fourteen of the data points, as discussedin the text.
Arrowsindicate the direction of approachto equi-
+2.0
kaolinite--*-mus
line
-•
t03 Ina = 2.43 ({06T-2)-4
+•.5
librium.
muscovitegel datum is in exact agreement.The
te
directionof the arrowshas the samesignificance
as in Figure 1. Note that in generalthe arrows
point in the direction compatiblewith the interpretation that the exchange reactions were
incomplete.
The paragonitc-muscovite
reactionsare closer
et
to true exchangereactionsthan are the kaolinitcmuscovitereactions.
The incompletely
exchanged
runs in the analogous alkali-feldspar system
[O'Neil and Taylor, 1967] were shownto yield
equilibriumfractionationsusingthe partial ex? gel • muscovite
change extrapolation technique of Northrop
and Clayton [1966]. Thus it is instructive to
I
I
I
apply this technique to these data. However,
• .0
4.5
2.0
2.5
•o6 T-2
inasmuch as oxygen occurs in three distinct
structural positionsin micas, the various oxyFig. 2. Comparison of the least-squaresline
gens undoubtedly exchangeat different rates from Figure 1 and the raw data points from prowith the solution,thus making applicationof cedures 2 and 3.
.•••1
•
ßparago
•musc
OXYGEN
ISOTOPE EQUILIBRIUM
the highest degree of exchange.The estimated
values are given in Table 2, indicatedby question marks.
vite-water
6017
fractionation.
It would be fruitless to
attempt to identify the sourceof the problem
consideringthe number of complicatedsteps
At 350øC the extrapolated value of companion runs 305 and 306 is +3.67, corresponding to an apparent extent of exchangeo.f 64%
in 1120 hours. This extrapolated value is incompatible with the feldspar-water values of
+3.93 [O'Neil and Taylor, 1967, Table 1, average of runs 307 and 308] at the sametemperature becauseisotopicanalysesof coexistingfeldspar and muscovite in naturally occurring
involved in obtaining a single data point. It is
muscovite fractionation (see Table 3). The
350øCextrapolatedvalue is thereforejudgedto
be anomalousand was not used in calculating
the final least-squares
expressionfor the musco-
sion calculated from these selected data is
probable,however,that the extrapolationtechniquecannotbe usedfor micasexcept,perhaps,
as an approximationwhen the O18/O
• extrapolation is very small (thus ruling out the
350øC case).
The fourteen'good'kaolinire-muscovite
points,
the muscovite-gelpoint, and the extrapolated
values for the paragonire-muscoviteruns are
assemblages
precludesucha smallalkalifeldspar- plotted in Figure 3. The least-squares
expres-
l0s In a-
2.38(106T-2) -- 3.89
This expressionis the best representationof
TABLE 3. OxygenIsotope 'Temperatures'of Various Alkali Feldspar-MuscovitePairs from Igneous
and Metamorphic Rocks
Feldspar*
$ Muscovite*
$ Temperature,øCt
8.1 (An25)
7.6
1400 4- 500
10.7 (KF)
9.6
655 4- 75
9.4 (KF)
8.6
1000 -!- 200
8.4 (KF)
7.3
655 4- 75
Igneous Rocks
1. H-l, 2-mica granite,
Dixville Notch, New Hampshire
2. H-11, 2-mica granite,
Lake Sunapee,New Hampshire
3. T-404, 2-mica granite,
Fitchburg, Massachusetts
4.
Elberton
Granite
New Comolli Quarry, Georgia
5. Pala pegmatite aplite
Pala, California
6. Trondhjemite (quartz diorite) contact
zone of Sawtooth stock, Santa Rosa
Mtus., Nevada
7.6
12.1
13.4
14.2
(An0)
(An,5)
(Anx•)
(An,•)
6.3
11.2
12.7
13.1
530
670
950
520
4- 50
4-4-100
4-4-200
4-4-75
Metamorphic Rocks
7. EV 13 Chester Dome gneiss
Gassetts,Vermont
8. EV 18, pod in garnet-zoneschist
West Marlboro, Vermont
9. LA 10q schist, Mr. Grant,
Vermont (chloritoid-kyanite zone)
10. Pelona albite-chlorite-actinolite greenstone
schist,San Gabriel Mtns., California
8.2 (KF)
6.7
450 4- 40
9.7 (An,o)
8.2
385 4- 30
13.9 (Ans)
12.0
315 4- 20
12.1 (Aris)
10.2
315 4- 20
Note:Samples
1 and2 arefromTaylor[1968];samples
3, 7, and8 arefromTaylor(unpublished
data).'
sample4 is from Taylorand Epstein[1962a];sample5 is quotedin Epsteinand Taylor [1967];samples6
are from Shiehand Taylor [1969];sample9 is from Taylor et al. [1963];and sample10 is from Taylor and
Coleman [1968].
* Valuesof $ for feldsparand muscoviteare reportedrelativeto standardmeanoceanwater (SMOW).
Symbolsin parentheses
indicateK feldspar(KF) or the anorthitecontentof theplagioclase
(An0)to (An2•).
t Errors shownfor the calculatedtemperaturesare basedon the assumptionthat valuesof 1000 In a
are accurateto 4-0.1. Note that the sensitivityof the feldspar-muscovite
geothermometeris very poor at
high temperature.
6018
O'NEIL
AND
T (øC)
650
600
550
500
400
logicallyvery 'reasonable'temperaturesthat are
350
obtainedin a variety of metamorphicrocktypes
with the quartz-muscovite geothermometer
[Epstein and Taylor, 1967]. These are, of
course,only plausibility argumentsand do not
constituteabsoluteproof of equilibrium.There
is little reasonto expectthat the variouspolymorphs of muscovitewould have measurably
differentisotopicproperties,becausethe essential bond linkagesinvolving the oxygenatoms
(Si-O-Si and Si-O-A1) are not changed.That
'103
Inc•=2.38(.106T-2)-5.89
c: +o.5
-
is, the calibration of the 'metastable' I M mica
-
//
_
e
e koolin•fe---muscovite
ß paragonire--- muscovite
? gel--'- muscovite
1.0
'•5
20
25
t06T-2
Fig. 3. Plot of 1000 In • versus 106T-•' for the
selected
muscovite-water
fractionations
from
all
three procedures and the least-squaresline constructed
from
TAYLOR
these data.
the equilibrium muscovite-waterfractionations
over the temperaturerange studied.
Because the reactions utilized
in this work
involved chemicalchanges,they are not simple
isotope exchangereactionsand the equilibrium
nature of the fractionationscan only be strongly
inferred, not proved. It was mentioned in the
discussionof the first procedurethat the measured fractionationscould conceivablybe a consequenceof kinetic isotopeeffects.li is the crux
of the equilibriuminterpretation that the oxygen
isotope [ractionations obtained with the three
proceduresare consiste• with one another. I•
would be highly [ortuitous i[ kinetic or other
nonequilibriumprocesses
operatedin directions
so as to producesuchsystematicisotopicresults
•rom three so di#erent procedures.
Additional weight is given the equilibrium
interpretationof thesedata by (1) the linear
relation between the per mille fractionations
and T-" shownin Figure 1, and (2) the petro-
producedin theseexperimentsprobablyapplies
equally well to 2 M micas, the most common
form occurringin nature.
The alkali feldspar-muscovitefractionationsin
igneous and metamorphic rocks are relatively
small (0.6 to 1.9%o;see Table 3). This implies
that the feldspar-muscoviteisotopic geothermometeris not very temperaturesensitive,and
analytical errors in the measuredO•8/OTMratios
result in a large uncertainty in calculatedtemperaturesof formation of natural mineral pairs.
Nevertheless,it is usefulto calculatecrystallization temperatures for those rocks on which
isotope measurementshave been made on coexisting feldspar and muscovite, particularly
becausethe feldspar-H20 calibration of O'Neil
and Taylor [1967] and the muscovite-H20
calibration of the present work were both done
usingcation-exchange
techniquesin alkali chloride solutions.These calculated 'temperatures'
are presentedin Table 3. Note that the 'temperatures'quotedin Table 3 for variousigneous
and metamorphic rocks are geologicallyvery
reasonable,particularly consideringthe uncertainties
involved
and
the
fact
that
some of
these pairs probably did not form in oxygen
isotopicequilibrium.
The experimentsbearing on the oxygenisotope relationshipsbetweenmuscoviteand paragonire are 139 and 142 at 420øC and 349 and
351 at 600øC. These runs show no muscovite-
paragonitc fractionation, and also the paragonire datum of experiment 61A falls on the
same straight line as the good muscovitedata.
In addition, no observable 0'8/0 •6 differences
were found betweenmuscoviteand paragonitc
from chloritoid-kyanite zone rocks from Vermont [Taylor et al., 1963]. Hence,althoughthe
data are limited, it would appear that the
sodium and potassiummicas do not measur-
OXYGEN ISOTOPE EQUILIBRIUM
ably fractionate O"relative to one another at
equilibrium. This is consistentwith the findings
in the alkali feldspar system for which both
natural and experimentaldata indicate identical isotopic properties for the sodium and
potassium end members [O'Neil and Taylor,
1967].
MECHANISM
OF THE EXCHANGE REACTIONS
Based on rate studies and X-ray measure-
6019
would presumablysupport Velde's proposal.
The knowledgethat the oxygenlattice of the
kaolinitc structureis completelybroken down
must be reconciled with this observation. It is
well known in nature that pseudomorphous
replacementcan occur with preservation of
external forms and delicateinternal structures,
even though a complete reconstructionof the
lattice has obviouslyoccurred(e.g. limonitereplacingpyrite, silicareplacingwood,etc.).
In experiment61, kaolinitc was reacted for
ments,¾elde[1965] hassuggested
that portions
of the kaolinitc crystal structure are inherited 180 hours at 510øC with a NaC1 solution that
intact by the mica in the reaction between was approximately40%ølower in O•8/OTMratio
kaolinitc and KOIt to give muscovite.Inas- than the kaolinitc. The solid products were
much as kaolinitc and muscovite have identical
A1/Si ratios, the transformationinvolvesonly
addition of potassiumion and removal of hydrogenion and water, and the inheritanceproposalseemedvery reasonable.However,in the
reactionof procedure(1) in whichpotassiumis
suppliedin KC1 solution,the oxygenisotope
exchange accompanyingthe reaction proves
that essentiallyevery bond to oxygenin the
kaolinitc
was broken
and
reformed
in
con-
structingthe muscovitelattice. That is, all parts
of the kaolinitcstructureare completelybroken
down in the transformation.
In this sense the
then reacted with •
of the stoichiometric
amount of fluorine necessaryto liberate all the
oxygen from the sample. Inasmuch as the
fiuorination reaction is believed to take place
with the fluorine successively'stripping' off
outer layers of the oxygen-containing
solid, this
liberated oxygen (61A) comesfrom the outer
portions of the paragonitcgrains. The oxygen
isotope fractionation between paragonitc and
water calculatedfrom the analysisof this oxygen
is compatiblewith the equilibriumfractionations
measured at other temperatures. The oxygen
isotope analysis of the unreacted portions of
reactionis identical to the muscovite-paragonite the grains (6lB) was 2%0heavier than 61A;
this may indicatean unexchanged
remnant core.
ous solution.
In other experiments at lower temperatures
One couldargue that indeedthe initial con- where reactionwas lesscomplete,both kaolinite
version of kaolinitc to muscovite takes place and mica reflectionswere observed in the X-ray
with inheritanceof the kaolinitc structure,and diffraction patterns.
The strippingexperimentand electronmicrooxygenisotopeexchangesubsequently
takes
and alkali feldsparexchangereactionsin aque-
placebetweenthe micaandwater.The rapidity scope observations,in conjunctionwith the
of the isotopeexchange(essentiallyequal in oxygenisotopedata, provideinformationon the
rate to the chemical reaction) argues against
detailed nature of the reaction between kaolinitc
this.Also,O'NeilandTaylor[1967]haveshown and potassium(or sodium)ion in solution.The
that exchange
ratesare evenslowerin solutions starting systemis, of course,far removedfrom
containing
a commonion (e.g.,K + in the solu- chemicalequilibriumand the potassium-bearing
tion and in the muscovite) than in pure water.
solutionwill readily attack a kaolinitc grain.
Further insightsinto the nature of this Based on analogy with alkali-exchangephereactionare affordedby the 'stripping'experi- nomena in feldspar-H•Osystems[O'Neil and
ment 61 and by a comparison
of the starting Taylor, 1967], we concludethat the newlykaolinitcand resultantmica by electronmicros- formed muscoviteprobably nucleateson the
kaolinitcin the samegrain.
copy.The photomicrographs
in Figure4 show similarly-structured
that the muscovite(in this caseproducedin Our data showthat oxygenisotopeequilibrium
2 hoursat 630øC) is well crystallizedand of has been established between this muscovite and
the sameaveragegrain size as the kaolinitc the solution, and at sufficientlylow temperastartingmaterial.This observationsuggests
a tures (belowthe dehydrationboundaryof kaotransformationakin to pseudomorphous
replace- linitc) the muscoviteprobably progressively
ment,andwithoutthe oxygen
isotope
data,this replacesthe kaolinRecore in the mannerof a
6020
O'NEIL
AND TAYLOR
}•i{ii•i::i•iiiii•::i•i•ii•:,•i•i•i
i::!:i•i:::,::-•::;;sr3:P'
'•'-{-•}•:
":{
•:'-".::..3•
:.•:•.Z•.•:`:•:.;:•i•::•::•::•e::•::•::::::::::•:{•{•::•:::::::•:•::•::•::•::•::•{•::•::::::::::•::::•::::::•::::::•::•:
............
========================================
•' •' •............
••::i::i{i::i•i•i•:•i::i•i•i•i::i::i::ii::i::i:
•{•;:•.si::•..:2•ii;..'•...-:.{.-....'•s.e:;:iii•:•iiii:•{%i•.•:.:::
..... •.....•:
..•!•. •===========================================================================
I
I
t
Micron
Fig. 4. Photomicrograph of muscovite prepared from kaolinitc by reaction with KC1 solution at 630øC for 2 hours (above) compared with photomicrograph of the kaolinitc starting
material (opposite). Allowing for the fact that the mica flakes have adhered together to a
certain extent, note that the average grain sizes and general shapesare apparently unaffected
by the reaction even though the oxygen has been almost completely exchanged.
reactionfront sweepingthrough the grain.
Acknowledgments. We wish to thank Profes-
Oxygen isotopeequilibrium is maintained be- sor R. N. Clayton for his helpful criticism of the
Mr. Paul Yanagasawa for performing
tween the minute amount of solutionpresentat manuscript,
most of the fluorine
extractions
on the solid
the reaction boundary and the bulk solution. phases,Dr. John Hudson for help with the elecChemical
communication
with
the interior
of
tron microscope, and Mrs. Elizabeth Bingham
the grain is presumablyaccomplished
througha for emission spectrographicanalyses.
Financial support was provided by the National
network of solutionpathwaysgeneratedby the
Science Foundation, grant GA-513, and by the
crystal imperfections normally attendant on Alfred P. Sloan Foundation.
growthof mineralgrainsin the laboratory.The
observedrapid attainment of oxygen isotope
l•EFERENCES
equilibrium at the reaction front is best explained by fine-scalesolution redepositionat
this boundary, as was postulated for feldspar
exchangereactionsin aqueoussolutions[O'Neil
and Taylor, 1967].
Baertschi, P., and S. R. Silverman, The determination of relative abundancesof the oxygen isotopes in silicate rocks, Geochim. Cosmochim.
Acta, 1, 317-328, 1951.
Bigeleisen, J., and M. G. Mayer, Calculation of
OXYGEN
ISOTOPE EQUILIBRIUM
6021
.:
'-'
". .....
:c.
"•:"
';'::
,,.
.?i:-
:;i!!:f--"a
%
.
:.•:.-.,!;'*---'•:•:,::.
....
***-"-":'"';=-:'"%-****':"
'*==='*
......
.-:-'"":*:;
.....
'
....;.,.-,
'.;;-.***;a*"•-..--',i,i;
•:,:.:**:::4,a,-:,a,!,;,;;..--'.-:.-'**:
":""***";"
.................................
":-::i
'..' :;;...,:-'•"-'"-'"*:*"-":*
......
....
;.' ....... ,
'.......
;.:..
:*:'•.?•'•;;:,;.:"':;;,ii:i:'-;•i!ii,!;**i
::'""'2'""•';':"""
"*
:'
'•"
.... *:'*:
ß
.......
..-;..'"'"';:i
.......
:.:..,,....?:
...
..... •"'";:•
****:
ß,;..:........
- ....
-?;-•;**'-'
......
*,•----*-'*':
....
:-:?.-?;;;;;i..-...-i?6;';"':-'
-i<i:"'"
:':-:
"-;•.;'."'
'....
.---'-...
-•
.;.;•.,.:
::
"**-•
.... ...
..,,:;%'-•-.
,- ::::%..
"a•.
"*;;'-'-'".:.
I
"
................
;;i:!>*';.***
'
4 Micron
equilibrium constants for isotopic'exchangereaction, J. Chem. Phys., 15, 261-267, 1947.
Epstein, S., and H. P. Taylor, Jr., Variation of
O*s/O•6 in minerals and rocks, in Researches in
Geochemistry, vol. 2, edited by P. H. Abelson
pp. 29-62, John Wiley and Sons, New York,
':'-.
..........
I
other areas, Contrib. Mineral. Petrol., 20, 306356, 1969.
Taylor, H. P., Jr., and S. Epstein, Relationships
between O•s/Ox6ratios in coexisting minerals of
igneous and metamorphic rocks, 1, Principles
and experimental results, Bull. Geol. Soc.
1967.
Amer., 73, 461-480, 1962a.
Garlick, G. D., and S. Epstein, Oxygen isotope Taylor, H. P., Jr., and S. Epstein, Relation beratios in coexisting minerals of regionally metatween O•s/O• ratios in coexisting minerals of
morphosed rocks, Geochim. Cosmochim. Acta,
igneous and metamorphic rocks, 2, Application
31, 181-124, 1967.
to petrologic problems, Bull. Geol. Soc. Amer.,
73, 675-693, 1962b.
Northrop, D. A., and R. N. Clayton, Oxygen
isotope fractionations in wstems containing
Taylor, It. P., Jr., A. L. Albee, and S. Epstein,
O•s/o •ø ratios of coexisting minerals in three
dolomite, J. Geol., 74, 174-196, 1966.
assemblagesof kyanite-zone pelitic schist., J.
O'Neil, J. R., and S. Epstein, A method œor
Geol., 71, 513-522, 1963.
oxygen isotope analysis of milligram quantiTaylor, It. P., Jr., Oxygen isotope studies of hyties of water and some of its applications, J.
drothermal mineral deposits, in Geochemistry
Geophys. Res., 71, 4955-4961, 1966.
o/Hydrothermal Ore Deposits, edited by H. L.
O'Neil, J. R., and H. P. Taylor, Jr., The oxygen
Barnes, pp. 109-142, Holt, Rinehart and Winsisotope and cation exchange chemistry of feldton, New York, 1967.
spar, Amer. Mineral., 52, 1414-1437, 1967.
Shieh, Y. N., and H. P. Taylor, Jr., Oxygen and
Taylor, H. P., Jr., The oxygen isotope geochemishydrogen isotope studies of contact. metamortry of igneous rocks, Contrib. Mineral. and
phisin in the Santa Rosa Range, Nevada, and
Petrol., 19, 1-71, 1968.
6022
O'NEIL
AND TAYLOR
Taylor, H. P., Jr., and R. G. Coleman, Oxs/O
•6
ratios of coexisting minerals in glaucophane-
bearing metamorphic rocks, Bull. Geol. Soc.
Amer., 79, 1727-1756,1968.
Urey, H. C., Thermodynamic properties of iso-
Velde, B., Upper stability of muscovite,Amer.
Mineral.,51, 924-929, 1966.
Yoder, I-I. S., and I-I. P. Eugster, Synthetic and
natural muscovites,Geochim.Cosmochim.Acta,
8, 225-280, 1955.
topic substances,J. Chem. Soc., 562-581, 1947.
Velde, B., Experimental determination of mus-
covite polymorphstabtittles,Amer. Mineral., 50,
436-449, 1965.
(Received July 31, 1967;
revisedJune 30, 1969.)