American Mineralogist, Volume 69, pages 128-138,1984
Experimental hydrogen isotopestudies:hydrogen isotopeexchange
betweenamphibole and water
CoI-rN M. Gnnsela
Department of Geology
Edinburgh University
Edinburgh, Scotland EHg 3JW
Russell
S. HenIraoN
Department of Geological Sciences
Southern Methodist University
Dallas, Texas 75275
eNo SrNaoNM. F. Sneppenp
C.N.R.S.-C.R.P.G.B.P. 20
France
54501 Vandouevres-Les-Nancy,
Abstract
Equilibrium hydrogen isotope fractionation factors (ai,in-H,o)were determined experimentally for various amphiboles (tremolite, pargasite, ferroan pargasitic hornblende,
actinolite, and arfvedsonite)over the temperaturerange350to 950"C.D-H fractionation in
the system tremolite-water is independent of temperature from 350 to 650"C and is
: -21.7t2i above 650'C the relationship
describedby the expression1000ln a1","--11,6
: -31.00(106/72)+ 14.90.Similarly, D-H fractionationin the
becomes1000ln dierem-Huo
systemferroan pargasitichornblende-water is independentoftemperature up to 805'C and
is given by the expression1000ln of;upHro= -23.1-+2.5. Above 850'C the fractionation
factor becomes slightly more positive with increasing temperature. At 700'C and 850'C
hornblende-water hydrogen isotope fractionation factors are very similar for both the
pargasite and ferroan pargasitic hornblende. Estimated D-H fractionation for actinolitewater at 400'C is 1000ln dct-n2o - -29, whereas for arfvedsonite-water it is 1000 ln
ai.r-xro - -52. Amphibole-water hydrogen isotopefractionationsare not simply related to
the composition of the octahedral cation site; the A-site cation (if present) may also
influence the fractionation behavior of hydrogen in amphiboles.
The water content of ferroan pargasitichornblende(in equilibrium with water) decreased
from 2.1 wt.Vo at 350"C to 1.2 wt.Vo at 950'C, probably because of oxy-hornblende
reactions.
Activation energiescalculatedfor hydrogendiffusion in amphibolesfall in the range 16to
25 kcal/mole. A consequenceof the low activation energiesis that amphibolesin volcanic
rocks may readily quench in their high-temperature hydrogen isotopic compositions,
whereas amphiboles in slowly-cooled metamorphic and hydrothermal environments may
continue to exchangehydrogen with a coexisting fluid down to temperaturesmuch lower
than those of initial amphibole crystallization.
Introduction
Studies of combined hydrogen and oxygen isotope
variations in rocks and minerals are widely used to
estimate the isotopic compositions and origins of the
hydrous fluids associated with a variety of igneous,
metamorphic, and hydrothermal processes(see reviews
0003-004X/84/0102-0128$02.00
by Taylor, 1974;Sheppard,1977).The hydrogenisotopic
compositions of hydrous minerals are particularly sensitive indicators of the isotopic composition of the last fluid
with which they have equilibrated, becausemost hydrous
minerals contain relatively small amounts of hydrogen
relative to the interacting fluid, except at very small
water-to-rock ratios. In order to calculate the isotopic
r2E
r29
GRAHAM ET AL.: HYDROGEN ISOTOPE EXCHANGE IN AMPHIBOLE-H2O
composition of a fluid phasewhich has subsequentlybeen
removed from the original mineral-water system, it is
necessaryto know the isotopic fractionation factors between hydrous minerals and water as a function of
temperatureand composition.
In this study we report the results of an experimental
investigation of hydrogen isotope exchangeand fractionation between various amphiboles and water at high
pressureand over a range oftemperatures representative
of hydrothermdl and low-grade metamorphic to igneous
environments.Amphiboles are a common constituent of a
wide range of crustal and some upper mantle rocks.
Therefore, the results of this study provide important new
data for the determinaton of the stable isotopic compositions and originsofhydrous fluids in a variety ofgeological environments.
There is already a small amount of published data on
the fractionation of hydrogen isotopes between amphiboles and water. Suzuoki and Epstein (1976)measured
the D-H fractionation between a "hornblende" and water over the temperature range 400 to 750'C. Their
chemical data suggeststhat the "hornblende" was very
low in aluminium, and was probably an actinolite. The
equilibrium D-H fractionation for this "hornblende"water system was found to vary with temperature above
450"C according to the following relationship: 1000 ln
cPnur,'-n,o= -23.9(1061111
+ l.S. These authors also
proposed a general relationship between hydrogen isotope fractionation factors and the octahedral-cationcomposition of micas and hornblendes such that lfi)0 ln
ai,i*H,o : -22.4G0617'1 + 1ZXo, - 4Xue - 68XF") +
28.2, where X is the mole fraction of the respective
cation. This relationship was derived largely from experimental data for micas where OH-group bonding occurs
exclusively with octahedrally-coordinatedcations. However, preliminary experimentaldata for the system aluminous hornblende-water by Graham and Sheppard (1978)
suggestedthat hydrogen isotope fractionation in the amphibole-water system was more complex than in the
mica-water system and indicated that the Suzuoki-Epstein relationship was not totally applicable to the amphiboles.
In this study we have examined the influence of amphibole composition on hydrogen isotope fractionation factors by direct measurementof equilibrium D-H fractionation for several different amphiboles, including pure
tremolite, two hornblendes of difering Fe, Mg, and Al
contents, actinolite, and the sodic amphibole arfvedsonite. In addition, we have derived quantitative data on the
kinetics of hydrogen isotope exchange between amphibole and water, as well as on the difusion of hydrogen in
amphiboles. This information permits both quantitative
assessmentof the extent of continuedisotope exchange
between amphiboles and fluid in cooling igneous, metamorphic, and hydrothermal environments and estimation
of "closure" temperatures for the cessation of isotopic
exchange.
Starting materials
Minerals. With two exceptions (see below) all amphibole starting materials were mineral separatesof >DVo
purity ground to a mean grain size of 30-60 ,rm. Powders
were not sized, but the fine-grainedfraction was removed
by repeatedelutriation in distilled water. The isotopic and
chemical compositions of the various amphibolesused in
the study are given in Tables I and2. Tremolite #338 is
close to a pure, stoichiometric hydroxy-tremolite. Actinolite #6906 is a typical greenschist-faciesactinolite from a
tholeiitic metadolerite (e.9., Graham, 1974),and appears
to have proportions of Mg and Fe comparable to the
"hornblende" used in the experiments of Suzuoki and
Epstein(1976;Table 3). The two hornblendes,according
to the classificationofLeake (1978),are a ferroan pargasitic hornblende #322 (from a garnet amphibolite) and a
pargasite #6099 (from an amphibole-plagioclase--clinopyroxene cumulate nodule in a basaltic andesite). The
two hornblendesdiffer in their proportion of Fe and Mg,
in Ti contents, and in tetrahedral and octahedral Al
contents. The sodic amphibole #516 is an arfvedsonite
from a peralkalinegranite which has subequalamounts of
H2O and F. Other amphibolescontain only trace amounts
of F.
One portion of the tremolite starting material contained
a small amount of minute calcite inclusions. Reaction of
this calcite with the tremolite during some of the hightemperatureexperimentsproduced detectableamountsof
diopside (tremolite * calcite --+ diopside * forsterite *
CO2 + H2O), enhanced isotope exchange rates, and
causedapparentdecreasesin water content ofthe tremolite run products. Results of the high-temperature
(>400"C) tremolite-water isotope exchangeexperiments
have therefore been omitted from the discussionof water
contents and analysis of kinetic data which follows.
However, equilibrium fractionation factors calculated
from theseexperimentsare not measurablydifferent from
those calculated from experiments using a calcite-free
tremolite starting material.
Table l.
Isotopic compositions of starting materials
(a)
I
ll
ill
IY
Y
YI
vtl
+ 5 1. 4
-2.8
-5f.E
-96.0
-t55.9
-6.0
-t20.2
l|ircral5
ite l33E
trol
actircl it. t6906
hornbladc
f322
lprnblendc160gg
arfvcdslte
t5l6
(b)
+2.\
-56.0
-
(c)
-54.7
-96.7
-160.6
6D(o,/@)
-105.t
-55.6
-62.8 (fcrren
Pargasltic
-26.E(pargeslte)
-9f.5
lprnblende)
130
GRAHAM ET AL.: HYDROGEN ISOTOPE EXCHANGE IN AMPHIBOLE-H2O
The actinolite #6%6 used in the preliminary experiments on the systemactinolite-H2O contains significantly
more than the stoichiometric amount of water (Table 2),
suggestingthe presenceof a small amount of a hydrousmineral contaminant. Small amounts of chlorite intergrowths within actinolite grains were confirmed by XRD
analysis. Further work on the actinolite-water system is
currently underway with a pure starting material and will
be reported elsewhereupon completion.
Waters. Each amphibole starting material was exchanged at various temperatures between 350 and 950"
with two or more waters of differing initial hydrogen
isotopic composition (Table I, I-VII). Periodically, mineral powders and aliquots of the various starting waters
(Table I, a-e) were sealedinto capsules.Small variations
in isotopic composition between successivealiquots of
water reflect isotopic fractionation arising from evaporation of standard waters with time. In order to insure
internal consistency, each time an individual group of
mineral-waterruns was prepared, a set of starting waters
and experimental run products were then analyzed after
each group of the exchangeexperimentswas completed.
Experimental and analytical methods
Experimental
and analyticaltechniques
The experimental
techniquesand analyticalmethodsare essentiallythose describedby Grahamet al. (1980),Isotopeexchangeruns were madeusing5 mm O.D. x 0.5 mm wall
typicallycontaining180thicknessx 2.5cm lengthAu capsules
250mgamphibolepowderand8-12mg water.Runsweremade
Table 2. Chemical analysesof amphibole starting materials and run products
si02
trercl ite
#338
actinol ite*
#6906
57.71
53.00
2.55
0.05
t0.52
Al203
n.d
Tl02
n.d
Feo(total)
0.14
arfvedsonite
#516
hornblende hornblende hornblende hornblende
{6099
#322
#322/650
t322/850
110.63
t+3.69
\4.2\
43.51
0.28
12.35
t2.5\
I t,o)
12.25
| .48
j4.67
1. 7 6
fr.f5
0 .5 0
t\.32
v' ))
0.50
I l { .4 0
14.19
4 9 .9 3
Fe2O3
r0.r4
-tat
Fe0
25.5\
0 .f 7
1 4 .t o
l1n0
n. d
0.t8
0 .6 4
f,fg0
24.09
Cao
0.05
| .39
8.47
I 3. 45
|5.40
t2.87
Na20
0.02
0.26
K20
n. d
0.06
F
0.03
n.d
Hzo
2. l8
97.61
N u m b e ro f
si
-L:!!n
99.04
8.00
7.71
0 .2 9
llvl
0 .l 5
0 .0 1
tl
Fe3+
Na
0 .r 5
| .28
o.o2
4.98
v. z>
il .08
|0.83
11.79
2.5O
rr . 2 4
il.05
o.47
0.52
u.)5
0 .5 l
0.07
t.73
0.07
2.17
0.07
l .78
0 .0 7
r.87
96.70
98.52
97.87
97.77
6 .t 9
r.8l
6 .5 0
o. o/
| .50
t.33
r.48
0 .4 1
0.69
o.74
nAo
0.20
0.07
0 .0 5
0.07
0 .4 8
2. 0 8
7.91
0.05
0 .r 8
3 .3 8
0 .0 9
0 .3 0
o.>a
1.42
| .48
| .82
1.30
o,o2
0.03
0.03
0. 0 3
3.56
2.Ol
0.27
0.24
3.20
t.93
2.38
1.77
z.q>
2.00
| .82
r.78
0.01
0.08
2.60
0.74
0 .5 1
u ,o )
0.01
0.09
0.10
0 .l 0
0.01
0.3t
0.55
0 .5 4
0 .r 0
0.03
0.03
0 .0 3
0.03
2.02
o.95
t.76
2 .t 5
1. 7 9
L87
K
OH
o.27
10.74
il.13
2.23
t.2l
lln
It9
t0.40
ions on basis of 24 (0, 0H, F)
A tl V
Fe2+
|.53
r.09
o' 90
|or.og
4.21
|1.87
o.27
Analyses of grein rcunts of starting mineral poders by electron microprobe. H2Ocontents
determined by radlo-frequency inductlon heating under vacuum, and manmetric neisurqent of
hydrogen gas yields.
H o r n b l e n d e a n a l y s e s * 3 2 2 / 6 5 0 a n d # 3 2 2 1 8 5 0a r e r u n p r o d u c t s o f 6 5 0 ' c
and 850oC hornblende 1322-Hfi *chanqe qperiments.
Feo and Feeoe contents of hornblende
#322 and #322/850 deterninea by wet Lhaical nethods (analyst rl J, Saunders). FeOand
Fe203 contents and stoichlmetric
formula of arfvedsonite #516 calculated assuming that
X(cations) = 15. A separate H70 deternination on hornblende #322 bv a modified Penfield
T u b e m e t h o d ( a n a l y s t t t . J . S a u i d e r s ) g a v e w t . t H 2 O- 2 . 1 ! .
n.d - not detected. *number
of ions In actinolite
#6905 fomula calculated on-basis of 23(0) because of high HrO content
resulting frm inclusions of chlorite intergrwn with actinollte gralns (see tqt)i
GRAHAM ET AL,: HYDROGEN ISOTOPE EXCHANGE IN AMPHIBOLE-H2O
in Nimonic-105cold-sealpressurevesselsand internally-heated,
gas pressure vessels using an argon pressure medium. After
quenching,the capsuleswere pierced under vacuum, the water
extracted by heating at -150'C for one hour, and this water
converted to hydrogen for mass spectrometric analysis by passage over uranium metal at 750'C. All hydrogen yields were
measuredmanometrically to check for internal consistencywith
theoreticalvaluesand to monitor variations in the water contents
of exchangedamphiboleswith temperature. Exchangedmineral
powders were examinedoptically and by XRD analysisto check
for absence of mineral alteration. In addition, the two hornblende run products were reanalyzed by electron microprobe
and wet chemical techniques (Table 2) to document the exact
extents of possiblechemical changeduring the isotope exchange
experiments.
The water extracted from the capsules was converted to
hydrogen gas by reaction with U at 750'C (Bigeleisen et al.,
1952).The D/H ratio of the hydrogen gas was determined on a
VG-Micromass602 B mass spectrometer.Values for the relative
concentration of deuterium are given in Tables I and 3 in the
familiar 6 notation relative to the Standard Mean Ocean Water
(SMOW) standard as defined by Craig (1961)to a precision of
about + l%0, where
--'---' / (D/H)samole
\
- l I x 103.
= I'6D(%o)sample
'
\(D/H)standard
l
Using this notation, the equilibrium factor a'is related to 6D by
the expression
ai,io-szo =
(D/H)mineral
(D/H)water
I + (DDmineraVl000)
I + (6Dwater/1000)
Analytical precision and interpolation of fractionation factors
The !l%o analytical precision for the measurementof both
mineral and water D/H ratios meansthat calculatedfractionation
factors should be reproducible to +2Voo.Complete isotope exchangeat a given temperature is demonstratedeither by attainment of the same fractionation in experiments run at the same
temperature with starting waters of different isotopic composi-
.D
Fig. 1. Interpolation method for determining equilibrium
hydrogenisotopefractionation factors (c") from partial exchange
experimentsin the system amphibole-HzO (after Northrop and
Clayton, 1966;Suzuoki and Epstein, 1976).(a) tremolite #338HzO; (b) hornblende #322-H2O. See text and Table 3.
l3l
tion, or by demonstrationthat the fractionation does not change
in experimentsrun with waters of the sameisotopic composition
for diferent lengths of time (where dr + a).For most experiments, isotope exchangein the amphibole-water system was so
slow that equilibrium was not attained, and thus the interpolation
method of Northrop and Clayton (1966),as modified by Suzuoki
and Epstein (1976),was used to calculate equilibrium fractionation factors (a'), applying the relationship
(ai - l) = (a" - 1) - A(ar- aJ
to two or more experiments of the same duration run with
starting waters of different isotopic composition (where ai and ar
are the initial and final fractionation factors respectively, and A
is a constant related to the extent of isotopic exchange). The
value of o" is obtained from the intercept of 103(o"- l) in plots
- aJ; seefor example,Figure l. In this
of 103(ai- 1) vs. 103(as
study c" was calculatedalgebraicallyby a least-squaresfitting of
the isotope exchange data for three or more experiments at a
particular temperature run with starting waters of different
isotopic composition. Excellent internal consistency of data is
usually observedusing this partial-exchangeinterpolation method (see for example Fig. I, Table 3; also Graham et al., l9EO).
Results
The experimentally measured hydrogen isotope fractionations between the various amphibolesand water are
given in Table 3, together with the interpolated equilibrium fractionation factors. Exchange rates in the system
amphibole-H2O were sufficiently slow that complete exchange (f = l.fi), Table 3) was attained only in hightemperature experiments for the finest-grained starting
material (hornblende, #332) within 3 hours at 950'C and
14 days at 600'C.
Hydrogen exchange between charges and
pressure medium
Hydrogen difusion through capsule walls and D-H
exchange between charge and pressure medium during
high P-T gas-media experiments may cause significant
changes in the bulk 6D value of experimental charges.
However, the use of thick-walled (0.5 mm) gold capsules
minimizes the extent of such changes at temperatures
<650"C (Graham et al., 1980). In this study results of
exchangeexperimentsinvolving tremolite #338 and hornblende #6099 up to 850'C show no evidenceof significant
changesin the bulk DDofcharges (Table 3). By contrast,
mass balance considerations for the system hornblende
#33LH2O indicate significant increases in the bulk 6D
values for chargesin runs at temperaturesabove 700oC,
and a similar effect is more emphaticallydemonstratedfor
the runs at 550'C in the system arfvedsonite #516-HzO.
Iron-bearing amphiboles, such as the ferroan pargasitic
hornblende #322, are capable of undergoing "oxy-hornblende" reactions during exchangewith water at conditions of high P and T which involve the oxidation of Fe2+
to Fe3+ and liberation of hydrogen, according to the
reaction:
2Fe2* + 2oH- = 2Fe3* 2O2- + H2,
t32
GRAHAM ET AL.: HYDROGEN ISOTOPE EXCHANGE IN AMPHIBOLE_H2O
Table 3. Results ofhydrogen isotope exchangeexperimentsfor the system amphibole-H2O
RunT("c)p(kbars)
r,"l'Jio5r,l?llli;ll r"??"j'"[;:l]"rr " r l 3 . t l i l ; : 1 n " ,t o 3 ( o t - o r )r 0 3 r n o , l 0 3 | n o "
lo9 k2
T r e m o ti t e # 3 3 8 - H 2 q
l7
?5
31
20
27
33
2\
30
37
5tr
57*
5t*
56h
76*
78*
50*
55*
- 5 3 . 5 6 (| I l c ) +
| 2 9 .0 0 9
-9.53(tvc)
| 2 9 .0 0 5
+
5 5 .8 2 ( v q )
t29.oot
36.288 - 5 3 . 5 8 ( l i l q )
-9.62(tvc)
36.288
36.288 + 6 5 .8 2( v c)
-53.58(lllc)
r 8 .t 4 4
-9.63(lvc)
r 8 .r 8 0
1 8 . 1 9 5 + 6 5 .8 2 ( v c )
o .4 3 8 5 - 6 0 . 5 0 ( i lt d )
o .4 3 8 6 + 5 5 .7 0( v d )
0.2328 - 6 0 . 5 0 ( i l t d )
0.2328 + 5 5 . 7 0 ( v d )
o .|8 3 5 0 + 1 6 . 8 2 ( v i l )
0 . | 8 3 5 0 - 5 6 . 4 3 ( tI l e )
0 . i l i 0 - 5 0 . 5 0 (i l r d )
0 . l l r 0 + 5 5 .7 0( v d)
350'C
350'C
350"C
450'c
lr5o'c
450'c
550'c
550'C
550"C
550.c
650'C
750'C
750'C
800.c
800'c
850.C
850'C
-88.4
-r08.5
- 132.1
- g z .z
- 1 0 95.
-l4lr.l
- 7 8. 6
-103.9
-l{4.0
- 8 8 .0
-138.8
-78.7
-r36.1
- l 1 0 I.
- 8 5 .9
+ 2 58
.0
-13.50
-74.00
+ 3 1. 0 0
-l3.lr0
-77.00
+31.70
-il.30
-tc,t
-13\.7
H o r n b l e n d e# ? 2 2 - f l t ( f e r r o a n p a r g a s l t l c h o r n b l e n d e )
- 6 0 . 2 0 (| l s )
')a'o
129 350'C
7 9 .4 8 8
-il.56(ilta)
l 30 350'c
7 9 .4 8 8
6
5
.
0
4
(
8 400.c
5 7. 5 5 1
| |b )
lr00'C
- 8 . 5 7 ( rI r c )
-oo. J
2l
67.551
- 5 0 . 2 0 (| l a )
-\2.\
2\. t92
93 450'C
1
1
.
5
5
(
l
l
l
a
)
.AA A
t26 450'c
2 \. 1 9 2
-\.7
95 550'c
1 2 . 0 9 6 1 0 8 . 5 0l (a )
-67.8
t27 550'C
t2.096 - i l . 5 5 ( i l t a )
- 5 0 . 2 0 (t t a )
-?c
I 56 550"C
,r
I 2 .0 9 5
-20.0
157 700'c
| 2 .0 9 5 - 5 0 . 2 0 ( l l a )
| 58 700"c
t2.096 - 1 1 . 5 5 ( i l r a )
-oo' J
43* 850'C
0 .l 5 5 0 - 1 5 . 8 4 ( i l r d )
-33.\
45* 850'c
0.|550 - 5 r . 0 0 (i l d )
- ?q A
48* 850'c
0 .l 5 5 0 - 5 r . 0 0 ( i l d )
4t* 950oc
0 . I 0 8 0 - r 5 . 8 4 ( tI t d )
42* 950'C
0 . 2 1 9 5 - r 5 . 8 4 ( rI r d )
47* 950'C
0.I080 - 5 r . 0 0 ( i l d )
Hornblende
52*
64*
51*
65*
Act i nol i te
70
72
7\
q00'c
400'C
400'c
-\8.7
-t6.0
-i{5.8
+29.6
-45.9
it.a
- 4 0 .5
- 8 .0
-10.2
-t4.4
-?n q
-\.2
+29.30
-r7.40
+ 3 7. 6 0
-57,20
-20.23
+30.37
+42,50
-54.00
'r.9tI
- 2 6 . 3 )7
-r8.201
+ 1 . 7 \)
+ f8 . 5 0
-4.00
+27.80
-9.90
+33.q0
-10.40
+75.30
-r f.4o
+35.70
+35.40
-10.50
-il. f0
+35,\O
+ 4 1 t. 9
'2.70
-4.50
+q4.50
-42.571
- |5.58J
-37.941
-t8.67J
-27.171
-22.25J
-33.871
- 2 3 . 2 7|
- 2 4 . 8 J1
-24.09"1
-22.35)
-27.271
-25.92
- 2 0 . 0 0| 1
-l8.6ltl
- 2 0 . 5 4|
-i5.691
+53.87
-36.t2
+5f.60
-35.82
+lI.8l
-1\.u7
-52.63
-23.\7
-21.36
.882
.993
.8r2
.95\
.886
EEC
-6.2\
-5.\2
-5.25
-5.57
-\.91
-20.52
-21.\6
- 1 4. 5 2
- t2.5\
-9.37
-5. |5
.927
-\, t7
,7\6
- t t .l 7
.7\6
-1.72
. 8| 5
-3,72
.Er5
-1,92
.589
r.91
,692
. 8 3 0 ' --3t ..3366
.831
.\63
.\65
.5gg
.669
.900
.904
.88r
.991
.962
|.O0
1.00
| .00
1.00
1.00
1.00
1.00
l.oo
-6.97
.5.96
-6.53
-5.52
-5.\3
-5.41
-5.21
-33.381 - z r
'oo
-t\.22)
-25.331 _fq r,
- t 3 . 9 2J
.828
.828
.898
.E95
-4.04
-4.05
-3.69
-3.70
-38.84'l
-r8.54
-29.)8
|
+r0.7qJ
.5\9
.592
.6t5
-'v''t
-7a Ae
-.a aa
-23.32
-23,22
(Ave)
-24.40
(Ave)
_re 4,
lY'":
tAve'
-4.67
(pargasi te)
#6099-Hr0
700"c
700'c
850'c
850'c
-1t,4
-51.1
-27.',t61
- 2 1 . 3 9|
-s.26)
-22.841
-23.3t
- t | . 2 7lI
-22.lrl
- 2 r . 0 9|
-t4.l5l
-tt.7\f
- | .68_l
-2t.t71
lr
lr
6
6
0 . 5 3 16
0 . 5 3 15
o.{308
o 4308
- 8 5 . 5 8( t d )
+22.00(ilrd)
-85.58(td)
+ 2 2 . 0 0 |( | r d )
+15.5
-rt.5
+22.\
-44 0
-49.90(vt)
-3.90(llte)
+ 7 3 . 6 2 ( Vt )
-54.l
-68.I
- 8 8 .5
+ 5 1. 0
-11
1
+48.5
-30.5
#5906-H20
2
2
2
5t.o29
51.029
5t.029
-50.5
-qR
i
A r f v e d s o nti e / 5 1 5 - H 2 0
|
2
12
9
32
\
r9
26
350'c
350"C
350'C
450'C
q50'c
550"C
550"C
550'c
t04.940
r 0 4 .9 2 8
10 4 .9 2 8
57 . t89
31 986
3l.985
3r.985
+1.71(lVa)
+72.80(va)
-40.78(llla)
- 4 0 . 7 8 ( lI l a )
+78.8 (vc)
+ 7 2 . 8 0( V a )
- 4 2 . 0 5 ( tI t c )
+ 2 . 4 4( t v c )
- 1 2 0 .I
.Aq
A
-r36.1
-to. I
-47.O
-|34.4
-94.9
-7.3
-38.8
-73.0
-t44.1
-t21.4
-7r.8
- 3 5. 2 0
- 7 2 .t O
-3.30
-90-00
- r0 8 . 0 0
-t.20
- 3 4 .o 8 l
*s.70
- 4 5 .| 3 J|
- 2 8 .3 1' l
- t t . 2 7J
-50.>t
I
-44.28
|
-5t. t7
- 2 6 .t 3
-4r.14(?)
-34.941
* indicates experiment conducted in internally-heated gas vessel.
indicates composition of starting water (see Table l);
O+
(measured) fractionation
(neasured) fractionation
factor; qf = final
factor.
factor;
ai - initial
oe = equilibrium fractionation
Values of
for experiments in which less than l00Z exchange occurred.
oe interpolated
f = fractional
approach to equilibriun
=
F;+t
k2 = rate
constant
Because of hydrogen-bearing impurities in the argon
pressuremedium, the bomb walls exert a bufering influence on charges during experiments (Graham et al.,
1980). At high temperatures (750-950"C) the imposed
oxygen fugacity is between Ni-NiO and Fe2O3-Fe3O4,
but at lower temperatures it is closer to Ni-NiO. This
bufering effect provides the driving force for oxy-horn-
for
second order exchange reaction
(see Graham, l98l).
blende reactions which may be demonstrated by the
increased Fe3+/Fe2*, decreased HzO-content, and
changein color from green to brownish-greenof the Ferich hornblende #322 after exchangewith water at 850'C
(Table 2, column 7), and which must account for the
significant change in the bulk 6D value of the charges.
The changetoward more positive 6D valuesis in the same
GRAHAM ET AL.: HYDROGEN ISOTOPE EXCHANGE IN AMPHIBOLE-H2O
r33
direction as that observedby Graham er a/. (1980),who pressuremedium, and demonstratedby the high-temperadiscussedpossible diffusion mechanismsin such isotope ture increase in Fe3+/Fe2+(Table 2). We note that, in
general, those experiments which show the greatest inexchangeexperiments.
When significant changeoccurs in the bulk 6D value of crease in 6D value of the charge at any temperature
the amphibole-H2O charges, it is possible that LH
contain the hornblendeswith the lowest water contents,
exchangebetween chargeand pressuremedium occurs at indicating a relationship between the extent of the oxya much greater rate than that between amphibole and hornblende reaction and the extent of D-H exchange
water. If this situation occurs, and mineral-water isotope betweenchargeand pressuremedium. Infrared studiesof
exchangeis incomplete, one is not justif,ed in using the our run products, which are currently in progress, may
partial-exchangeinterpolation method of Northrop and provide new insight on the behavior of the amphibole
Clayton (1966)to derive equilibrium fractionation factors. OH-group as a function of temperature.
The water content of arfvedsonite decreases signifiTherefore, runs in which significant increasesin the bulk
6D value ofthe systemshave occurred are not considered cantly over the same 550-650"C temperature range; its
in this study unless 1007aexchangewas documented. In cause is again attributed to an oxy-hornblende reaction.
casesof total exchange,the measuredfinal fractionation The arfvedsonite contains a substantial amount of fluofactor (c") is independent (within analytical error) of the rine, but significant F-OH exchangewith fluid is unlikely
increase in DD value of the charge (Table 3).
as amphiboles strongly concentrate fluorine relative to
coexisting hydrous fluid.
Water contents of amphiboles at high
temperatures
Re lat io nship s b etw een fr ac tio natio n fac t or s ( a" )
Hydrogen yields from the amphibole run-products of
and temperature
exchange reactions were used to calculate amphibole
The relationships between the equilibrium hydrogen
water contents at various run temperatures. Radio-frequency induction heating under vacuum is consideredto
isotope fractionation factors (Table 3) and temperature
(plotted as tO6/l) for the various amphibolesare plotted
produce very accurateand reproduciblewater analysesof
in Figure 3. Also shown in Figure 3 are the experimental
amphiboles (Table 2; see also Boettcher and O'Neil,
data for the system "actinolite"-HzO studiedby Suzuoki
l9E0). Amphibole water contents are plotted against run
and Epstein (1976).
temperaturein Figure 2.
Tremolite-H2O. The fractionation factor is indepenThe water content of hornblende #322 decreasesprogressively with increasingtemperaturedown to 1.2 wt.Vo dent of temperature (within analytical error) over the
at 950"C, although there is considerable scatter in the temperaturerange 35G-650'C,and may be representedby
data. This decreaseis explained by the oxy-hornblende the expression
reaction outlined above, induced by the increase in
1000 ln dteremolit+Hzo: -21.7-+2.
oxygen fugacity with increasing temperature resulting
from the butrering effect imposed by bomb-walls and Above 650'C, the fractionation changes rapidly with
WATERCONTENTSOF AMPHIBOLES
wt96
Hp
o8
O^O
o6t
o
o
a homblcn&322
r homblcn&6009
Fig. 2. Variation of water contents of amphibole run products of amphibole-H2O exchangeexperiments (determined by radiofrequency induction heating)with run temperature.Initial water contents of amphibole starting materials are plotted on wt.7o axis.
134
GRAHAM ET AL.: HYDROGEN ISOTOPE EXCHANGE IN AMPHIBOLE-H.O
temperature, and fits the linear relationship
the two high-temperature data points is given by the
relationship
: -31.00(106/f1 + tl.lO,
lfi)0 ln dieremorit+H:o
1000Inafiupn,o: -31.0(106/f1+ t.t.
where T is absolute temperature ('K).
Hornblende-H2o. Experiments were conducted in the
system hornblende-H2O with two diferent starting compositions to assess the influence of Fe-Mg ratio on
fractionation factors, and to provide LH fractionation
data directly applicable to the origin of waters in amphiboles in both igneous and metamorphic environments.
Hydrogen isotope fractionation in the system hornblende-H2O is also independent of temperature over the
temperature range of approximately 350-850"C (Fig. 3).
For the Fe-rich composition, D-H fractionation between
ferroan pargasitic hornblende #322 and water iS given by
the expression
Experiments were also conducted to determine the
hydrogen isotope fractionation between pargasite #6099
and water at 700"C and 850"C, in order to assess the
influence of Fe: Mg ratio on the fractionation factor. The
pargasite#609 has a significantly larger Mg: Fe ratio and
Ti-content, but has an otherwise comparablecomposition
to the ferroan pargasitic hornblende #322. At 850"C,
exchangewas incomplete after 12 hours. Measured fractionation factors are slightly more positive than those for
the ferroan hornblende #322,btt the difference is small
(<5%o) at both temperatures. The small difference in
fractionation factors between these two hornblendesis in
the same sense as that predicted from the studies of
1000ln af;urnro: -23.1+2.5.
Suzuoki and Epstein (1976),i.e., deuterium is enriched in
This result confirms and extends the preliminary data of the more Mg-rich phase.
Graham and Sheppard (1978) for the same hornblende
Actinolite-H2O. Three experimentswere conducted to
starting material. Above 850'C the fractionation becomes estimatethe approximate magnitudeof DH fractionation
slightly more positive with increasing temperature (Fig. betweenactinolite and water at 400"C.Theseexperiments
3). At present there is insufrcient experimentaldata to fix were designedto provide data to permit comparisonwith
the slope of the fractionation curve with confidencein the the experimental data of Suzuoki and Epstein (1976)on
temperature range 850-1000'C. For lack of data to the an amphibole of probable "actinolitic" composition, and
contrary, we have assumed that the high-temperature to aid in assessingthe overall influence of composition on
behavior in the system Fe-rich hornblende-H2Ois similar mineral-water fractionation factor. A knowledge of the
to that in the system actinolite-H2O, thus the line through hydrogen isotope fractionation factor between actinolite
Trc)
550
1fi
92t - ffrl
Fig. 3. Experimentally determined relationships between hydrogen isotope fractionation factor (103in a' amphibole-HzO) and
temperature(plotted as 106/T2)for the system amphibole-H2O. Date for "actinolite"-H2O after Suzuoki and Epstein (1976).Error
bars indicate analytical precision.
GRAHAM ET AL.: HYDROGEN ISOTOPE EXCHANGE IN AMPHIBOLE-H2O
and water at typical greenschist-faciesconditions is essential to interpreting stable isotope studies of hydration
and metamorphismof basaltic oceanic crust.
As shown in Figure 3, the experimental data (Table 3)
indicate that the equilibrium fractionation factor in the
system actinolite #6m6-H2O at 400"C is approximated
by the expression
1000ln dict-H:o = -29.
However, this result is uncertain because of the minor
chlorite contamination of the actinolite. Hydrogen isotope fractionation between chlorite and water was not
experimentally studied. The 6D value of the natural
chlorite coexisting with actinolite #6906 (6D : -55.6%0,
Table l) in the parent greenschist-faciesmetadolorite is
-48%o (Tui et al., 1973').Thus, the equilibrium fractionation factor for pure actinolite is unlikely to be more
positive than -Z9%ooand could likely be slightly more
negative.A more detailed study of the systemsactinoliteH2O and chlorite-H2Ois in progress.
Arfvedsonite-H2O. Fractionation factors in this system
are subject to large uncertainty becauseof the problems
ofchangingbulk 6D valuesofchargesencounteredin the
isotope exchange experiments, possibly even down to
temperaturesas low as 450'C (Table 3). The fractionation
at 350"Cis approximately 3Woomore negativethan for the
calciferous amphiboles at this temperature.
Discussion
The hydrogen isotope fractionation curves for different
minerals, even within the same structural group, may be
rather complex (Graham et al., 1980)and the amphiboles
are no exception (Fig. 3). The simple relationships between the mineral-water fractionation factor and the
octahedral cation composition proposed by Suzuoki and
Epstein (1976)for the micas do not apply to the amphiboles. The fractionation factors for both tremolite and
hornblende are similar and independent of temperature
over the low-temperature range, despite their chemical
and structural diversity. The two compositionally distinct
hornblendeshave comparablefractionations at high temperature. Our data for tremolite are significantly different
from those of Suzuoki and Epstein (1976)for "actinolite"
at temperaturesbelow 650"C,but the fractionation factors
for both tremolite and "actinolite" may be represented
by a similar curve at temperatures above 650"C. The
fractionation factor for D-H exchangebetween actinolite
and water at 4fi)'C measured in our reconnaissance
experimentsis intermediate between the tremolite-water
fractionation factor (this study) and the "actinolite"water fractionation factor of Suzuoki and Epstein at the
same temperature.
Uncertainty regarding the detailed chemistry of the
"actinolite" used in the hydrogen isotope exchangeexperimentsof Suzuoki and Epstein (1976)createsdfficulty
in quantitatively comparing their results with our frac-
t35
tionation curves for tremolite and hornblende' Graham
and Sheppard(1978)showed that the predicted fractionation curve for hornblende #322 using Suzuoki and
Epstein's (1976) octahedral cation composition model
differs radically from the experimentally determined
curve (Fig. 3) for this amphibole. The cation sites in the
amphibolesmost likely to influence behavior of the OHgroup are the octahedral (M1, M3) sites and the A site.
The importance of the latter interaction was demonstrated by the infrared study of Rowbotham and Farmer (1973)
on the GH stretching frequency in the tremolite-richterite series. These authors observed that substitution of
both Na and K into the vacant A site of tremolite causeda
change in the stretching frequency of the O-H bond.
Thus, the diference between the "actinolite" and tremolite fractionation curves (Fig. 3) may reflect to a large
extent their different proportions of Fe and Mg. This
diference may be largely compensatedin our hornblende
samplesby their high Na (A) and Alvl contents' However, our measured actinolite-H2O fractionation factor at
400'C (1000 ln a!"s-g,6 : -29) is significantly more
positive than that for the Suzuoki and Epstein (1976)
"actinolite" (1000 ln olact',-Hzo: -39) at the same
temperature, despite the apparently comparable Fe/Mg
ratio of the two amphiboles. We note that the "actinolite" apparently lacks octahedral aluminum (see Table 3
of Suzuoki and Epstein, 1976; cf. Table 2, this study)'
whereasits A-site Na content is undetermined.
Despite these complexities, the similarity of the hydrogen isotope fractionation curves for the chemically-diverse calciferous amphiboles we have studied (Fig. 3)'
and the insensitivity ofisotope fractionation factors over
a wide range of temperature, suggeststhat our fractionation curves may be applied to a potentially wide rangeof
calciferousamphibolesin igneousand metamorphicrocks
in stable isotope studies offluid-rock interations.
Kinetics of hydrogen isotope exchange
The kinetics of hydrogen isotope exchange between
amphibolesand water may be quantified by application of
reaction-kinetictheory (Graham, l98l), assumingthat
exchange reactions may be approximated by secondorder kinetics. The relevant equation is
f(l-0:Kzt'
where f is the fractional approach to equilibrium, and t is
the time of the reaction. Activation energies(O) for the
rate-determining step, the transport of hydrogen in the
hydrous mineral, may be derived from the Arrhenius
relationship
log K2 : log a - 812'303RT'
where a is a constant. Values of log K2 for amphibolewater exchange experiments are given in Table 3, and
plotted against temperature in Figure 4. Derived activation energiesare listed in Table 4. These data are com-
t36
GRAHAM ET AL.: HYDROGEN ISOTOPE EXCHANGE IN AMPHIBOLEH2O
tmlite
.
t'mbkxle (nz) r
tv*teUe (ooor)a
Table 4. Activation energies (Q) for hydrogen isotope exchangein amphibolesderived from the Arrhenius relationshipby
linear least-squaresfitting of calculated second-orderrate constants (log Kz)
Tenperature
range
('c)
tr'ml
o
lte #338
&tinolite
/6906
hornblended322
!z
o
a
bq12
(kcals/
9-a tom)
350-800
400-670
-o.72
-3.49
16.0
0.96
0.91
-5]34
24.4
0.99
J5O'550
l.6l
-5.3\
2\.\
0.91
least squares fit of kinetic data (table 3) to th. equatlon
fog K2 - a * b (@f ), where b - e/2.303R, and 12 is the correlation
c@fficient.
Kinetic data for actinol ite after crahsm (1981), ano
Lindr
based on experirental
data of Suzuoki and Epsteln (1976).
hornblende#322show alarger scatter, perhapsreflecting
the simultaneous progress to varying degrees of oxyFig. 4. Arrheniusplot of secondorder rate constants(tog
hornblende reactions during the hydrogen isotope exK) vs. reciprocaltemperaturefor hydrogenisotopeexchange
changereactions.
reactionkineticsin the systemamphibole-H2O.
Datafrom this
study(Table3), SuzuokiandEpstein(1976)andGraham(t98t).
Diffusion of hydrogen in amphiboles: closure
Linesareleast-squares
fits to data.Seetext for discussion.
temperatures
pared with kinetic data for the experiments of Suzuoki
and Epstein (1976) for the system "actinolite"-water
derivedby Graham(1981).
Activation energies for hydrogen transport in hornblende #322 and "actinolite" are comparable (24 kcall
mole H) and slightly larger than the value for tremolite (16
kcaUmole);all these values lie in the range of values (1431 kcaUmole)derived for hydrogen diffusion in a variety
of hydrous minerals (Graham, l98l). Kinetic data for
Assumptions and calculatrons. Closure temperatures
for cessation of hydrogen isotope exchange between
amphibolesand hydrous fluid or between amphibolesand
coexisting hydrous minerals in igneous or metamorphic
rocks may be calculatedfrom estimatedditrusion parameters for hydrogen diffusion in amphiboles. Methods of
calculating hydrogen diffusion coefficients from experimental hydrogen isotope exchangedata are described in
detail by Graham (1981), and the application of these
methods to the amphiboles is briefly outlined here.
H DIFFUSPN IN AMPHIBOTES
%w
Fig. 5. Arrhenius plot of difusion coefficientsvs. reciprocal temperaturefor diftrsion of hydrogen in amphiboles.Data from this
study and after Suzuoki and Epstein (1976).Arrhenius plot for hydrogen ditrusion in muscovite (cylinder model) after Graham (1981)
shownfor comparison.Dashedlines indicate extrapolation beyond experimentalrange.Error bars indicate uncertaintiesin calculated
diffusion coefficientsarising from uncertainties in mineral grain geometry estimates.See text for discussion.
GRAHAM ET AL.: HYDROGEN ISOTOPE EXCHANGE IN AMPHIBOLE-H2O
Table 5. Activation energies (O) for hydrogen diffusion in
various amphibolesderived from the Arrhenius relationship by
linear least-squaresfitting of estimateddiffusion coefficientsfor
infinite cylinder and plate models (see text, Fig. 5, Graham
(le8l)).
r37
Conclusions
(1) The bulk D/H ratio in Fe-rich amphibole-water
exchangeexperimentsmay increaseby a variable amount
as a result of hydrogen diffusion through capsule walls,
despite the use of thick-walled (0.5 mm) capsules. The
effect is probably related to the occurrence of "oxy(
k
c
r
l
t
/
ran9a
9-rtom)
hornblende" oxidation of the amphiboles.
(2) Hydrogen isotope fractionation is independent of
(.) llJ.sgr
temperature up to 650"C in the system tremolite-water
-E.0, -t,72
t7 . 0
0.95
trcrcl I tr f3!8"
350-8oo
- 5 . 2 0 - 5 .r 5
23.6
0.96
rct I nol I tr #5906+
400-670
and 800-850"C in the system hornblende-water. The
-7.62 -\,19
2 0 |.
0.94
ho.nblcnd. #322d
350-550
relationship between hydrogen isotope fractionation factor and octahedral cation composition for OH-mineral(b)@
water systemssuggestedby Suzuoki and Epstein (1976)is
6
.
7
5
3
.
7
8
t
7
.
t
tr.rcl I tc /r38'
0.94
650-850
not directly applicable to the amphiboles. A-site cations
6
.
8
0
q
.
r
6
hornbl.nd. #3224
r9.0
0.95
350-550
may exert an important influence on the O-H bond in
Llnear least-squar.s fit of estlmated dlffuslon coefficients where
amphiboles.
l o g D - a * u 1 J 9 11,, b - Q / 2 . 3 0 3 R , a n d 1 2 l s t h € c o r r e l a t l o n c o e f f i c l c n t .
(3) The temperature-independence
of hydrogen isotope
* d.ta for t.ercllte
assume prlsns of avcrage lcngth 50 um and average
hornblende and water over a large
between
fractionations
aspcct r6tlo 2.46.
temperaturerange meansthat the hydrogen isotope com+ data for actinolite dcrlved from experlrents of Suzuokl .nd Epstein
( t 9 7 6 ) , a n d a s s u m ea v e r a g e p r l s m r . d i u s o f 5 6 u m ( l o o - 2 0 0 n e s h ) ;
position of aqueous fluid in isotopic equilibrium with
(
3
e
c
Gr€ham, l98l).
500'C data are excluded
a data for hornblende assune prlsns of average length r0 !m and
hornblendes in closed igneous and metamorphic rock
average aspect ratio 2.19.
systems for temperatures of >350"C may be reliably
estimated even when uncertainty exists over the exact
closure temperature for cessation of hydrogen isotope
The calculation of diffirsivities involves a knowledge of
grain geometry. Amphibole starting materials in this exchange.
(4) Low activation energies (16-25 kcal/mole) for hystudy were roughly sized by removal of fine-grained
drogen
diffusion in amphiboles are comparable to those
material from crushed mineral powders. Mean prism
for hydrogen diffusion in other hydrous minerals. Closure
lengths for the hornblende #322 and tremolite #338
temperaturesfor cessationofhydrogen isotope exchange
startingpowders were in the range30-50 pm. Mean prism
between amphiboles and fluid on slow cooling are far
lengths for hornblende #ffi99 were in the range 50-70
below likely crystallization temperaturesof amphibolesin
,rm. Mean aspect ratios were 2.46, 2.19 and 1.74 for most metamorphic rocks, but rapidly cooled amphiboles
tremolite #338, hornblende#322, and hornblende#6099,
in volcanic rocks should readily quench in their high
respectively. These values are insufficiently large to
(e.g., 900-1000'C)D/H ratios.
permit a choice between geometrical models. Uncertain- t€mperature
(5) Water contents of Fe-bearing hornblendesin equities in grain size lead to uncertainties in calculated
librium with water may decrease significantly with indifusion coemcientsof less than one order of magnitude
creasingtemperature as a result ofoxy-hornblende reacfor a given choice of geometrical model. The prismatic
tions.
habit and chain structure of the amphiboles suggestthat
diffusional anisotropiesshould be parallel or perpendicuAcknowledgments
lar to prism edges, and therefore the infinite plate and
(Crank,
4.43)
5.36
and
1975,equations
cylinder models
were kindly supSamplesand amphibolemineralseparates
were applied to our experimental data. These values are plied by A. M. Grahamand P. Ehlig. We particularlywish to
based on quite small extrapolation of our diffusion data. thank J. Borthwickfor laboratoryassistancewith the stable
Adopting the cylinder model, for example, a 0.1 mm isotopeanalysis.Researchand analyticalwork at the Experidiameter hornblende prism in a lava would require sever- mentalPetrologyUnit and ElectronMicroprobeUnit at Edinthe Scottish
al days at 950"Cfor complete hydrogen isotope exchange burghUniversityand the IsotopeGeologyUnit at
by the
are
supported
Reactor
Centre
and
Research
Universities
with hydrous fluid; a 0.5 mm diameter hornblende prism
CouncilandtheScottishUniverNaturalEnvironmentResearch
would require several months for complete exchangeat
with electronmicroprobe
sities.We thankP. Hill for assistance
this temperature, whereas no detectableexchangewould analyses.This work was supported,in part, by NSF grant
occur in less than several hours. We therefore conclude EAR82.1E3EO.
that amphibolecrystals in rapidly quenchedlavas will, in
References
the absenceof extensive oxidation or alteration, preserve
their high-temperature(magmatic)hydrogenisotope com- Bigeleisen,
J., Perlman,M. L., and Prosser,H' C. (1952)
position, and they should therefore preserve useful inforConversionof hydrogenicmaterialsto hydrogenRi isotopic
.
mation on the hydrogen isotope compositions and origins
analysis.AnalyticalChemistry,24, 1356-1357
Boettcher.A. L. and O'Neil, J. R. (1980)Stableisotope'
of "magmatic" waters.
l3E
GRAHAM ET AL.: HYDROGEN ISOTOhE EXCHANGE IN AMPHIBOLE-HaO
chemical, and petrographic studies of high-pressureamphi- Northrop, D. A. and Clayton, R. N. (1966) Oxygen isotope
boles and micas: evidence for metasomatismin the mantle
fractionation in systemscontaining dolomite. Journal of Geolsource regions of alkali basalts and kimberlites. American
ogy,74,174-196.
Joumal of Sciences.280-A. 594-621.
Rowbotham, G. and Farmer, V. C. (1973)The effect of "A" site
Craig, H. (1961)Standardfor reporting concentrationsof deuteoccupancyupon the hydroxyl stretchingfrequency in clinoamrium and oxygen-l8 in natural waters. Science, 133, 1833phiboles. Contributions to Mineralogy and Petrology, 38,1471834.
149.
Crank, J. (1975)The Mathematicsof Diffusion. Oxford UniversiSheppard, S. M. F. (1977) Identification of the origin of orety Press, London.
forming solutions by the use of stable isotopes. In Volcanic
Graham, C. M. (1979) Metabasite amphiboles of the Scottish
Processesin Ore Genesis,p. 254l.Institution of Mining and
Dalradian. Contributions to Mineralogy and petrology, 47,
Metallurgy and Geological Society, London.
165-185.
Suzuoki, T. and Epstein, S. (1976)Hydrogen isotope fractionGraham, C. M. (1981) Experimental hydrogen isotope studies
ation between OH-bearing minerals and water. Geochimicaet
III: diffusion of hydrogen in hydrous minerals, and stable
CosmochimicaActa, 40, 1229-nn.
isotope exchange in metamorphic rocks. Contributions to
Taylor, H. P. (1974) The application of oxygen and hydrogen
Mineralogy and Petrology, 76, 216-228.
isotope studiesto problemsof hydrothermal alteration and ore
Graham, C. M. and Sheppard,S. M. F. (1978)Hydrogen isorope
deposition. Economic Geology, 60, 843-E83.
fractionation between aluminous hornblende and water. proTuri, B., Graham,C. M., and Sheppard,S. M. F. (1973)Carbon
gressin Experimental Petrology, 4, 152-153.Naturat Environand oxygen isotope studies of the Dalradian metabasicrocks,
ment Research Council of U.K. Publication Series D, ll,
and the role of pore fluid during regionalmetamorphism.Third
London.
European Colloquium of Geochronology, Cosmochronology,
Graham,C. M., Sheppard,S. M. F., and Heaton,T. H. E. (19E0)
and Isotope Geology (Oxford), 69.
Experimental hydrogen isotope studies-I. Systematics of
hydrogen isotope fractionation in the systems epidote-H2O,
zoisite-H2O and AIO(OH)-H2O. Geochimica et Cosmochimica Acta, 44,353-i&.
Leake, B. E. (1978) Nomenclautre of amphiboles. American
Manuscript received, October 25, 19E2;
Mineralogist, 63, 1023-1052.
acceptedfor publication, June 28, 1983.