American Mineralogist, Volume 71, pages68-77, 1986
A kinetic study of clinochlore and its high temperatureequivalent
forsterite+ordierite-spinel at 2 kbar water pressure
MooNsup Cnor
Department of Geology, University of Toronto
Toronto, Ontario MsS lAI, Canada
aNo J. J. Fewcnrr
Department of Geology and Division of Sciences,Erindale College
University of Toronto, Toronto, Ontario M5S 1A1, Canada
Abstract
Clinochlore and its higher temperatureequivalent assemblage,forsterite<ordierite-spinel-HrO have beensynthesizedat2kbarwater pressure,usingoxide mixtures ofclinochlore
composition as starting material. A time-temperature-transformation (TTT) diagram has
been establishedin the range 600-750'C for periods up to 250 days. Following Ostwald's
rule, the final stable phase assemblages(14 A chlorite or forsterite<ordierite-spinel) are
synthesized only through the prior formation of metastable assemblages7 A chloriteforsterite+xide mix, 7 A chlorite-forsterite-talc-oxide mix or talc-forsterite-spinel. From
the TTT diagram, the apparent activation energyfor the heterogeneousnucleation ofthe
14 A chlorite from 7 A chlorite is calculated as 6313 kcaVmole. Rate studies employing
the parabolic rate equation x : kt" indicate that the activation energy for this transformation
changesfrom 90+35 kcal/molebelow 670'C, to 177!'ol'kcal/molebetween670 and 691'C.
Such high values suggestthat the polymorphic transition of chlorite involves the rearrangement of Si-O bonds which might be the rate-determining process.The higher activation energy above 670C can be attributed to the "stable" coexistenceofforsterite and
talc with 14 A chlorite, which is more aluminous than clinochlore. The most stable Mgchlorite has the approximatecomposition(M&,nAl, ,r)(Si28eAlr,)Oro(OH)8.
Introduction
A substantialeffort has been made recently to decipher
the kinetics and mechanisms of reactions in the solidliquid systems;e.g.,the rate laws of mineral leachingand
dissolution, and growth and dissolution mechanisms(see
Lasagaand Kirkpatrick, l98l). However, relatively few
studies have been done on solid reaction systems,principally becauseofthe very slow reaction rates, metastability and lack ofappropriate techniquesto deal with nucleation problems (Greenwood, 1963).The presentstudy
hasbeenundertakento provide information on the mechanisms and kinetics involved in synthesisof clinochlore
and ofits high temperature breakdown products forsterite<ordierite-spinel. Theseassemblages
are relatedby the
reaction:
: l0MgrSiOo* MgrSi,AlnO,r.
5MgrAlrSirO,o(OHL
nHrO
clinochlore
forsterite
cordierite
(1)
+ 3MgAl,O4+ (20 - n)H,O
spinel
vapor
Reaction(l) was first investigatedby Yoder (1952) who
t Presentaddress:Departmentof Geology,StanfordUniversity,Stanford,California94305.
0003-004v86/0102{068$02.00
determined the reaction temperature as 720C at 30000
p.s.i. water pressure.He found a new polymorph, a 7 A
structure and termed it aluminous serpentine,which has
also been referred to as septechlorite,7 A chlorite and
lizardite by subsequentauthors. Yoder noted that forsteritespinel-talc was a metastable assemblagetransforming into forsterite<ordierite-spinel above the equilibrium temperature of the reaction (l). Roy and Roy
(1955) synthesizeda mixture of clinochlore and aluminous serpentineat temperaturesas low as 450"C at 20000
p.s.i. water pressure.They attributed the extreme sluggishnessofthe transition ofaluminous serpentineto clinochlore to the very large energy barrier involved in the
rearrangementof Si-O bonds. Nelson and Roy (1958)
showedthat the 7A seriesextendedbetweenthe chrysotile
and amesitecompositionsand that clinochlore had a maximum thermal stability of 710"C at 20000 p.s.i. Fawcett
and Yoder ( I 966) synthesizedthe chlorite-talc-forsterite
assemblageat 3.5 kbar and forsterite<nstatite<hlorite at
5 and l0 kbar water pressure.They concluded that at
water pressuresabove 3.5 kbar the composition of the
most stable magnesianchlorite is more aluminous than
cliniochlore, lying about halfway betweenclinochlore and
corundophyllite (Fig. l).
68
69
CHO AND FAWCETT: KINETIC STUDY OF CLINOCHLORE
si02
.
€2
-
r'
E
v,
vl
g
/o
.../.
.
Glinochlore
E
e
E
r
F
7
Forsterite
=
,t
MSO
A1203
Fig. l. Phasesencountered
in this study.The diagramis a
projection,in molepercent,from the HrO apexontothe MgG
AlrO.-SiO,plane.
Chernosky( I 974) redeterminedthe position ofreaction
(l) with reversedexperiments,using synthetic crystalline
phasesas startingmaterials (Fig. 2). He also examinedthe
variation in chlorite composition by measuring do* and
found no evidencefor compositional variability at water
pressureup to 2 kbar. His work provides the bestestimate
for the reaction coordinatesat lower pressures.
In their high pressureexperiments(up to 35 kbar), Staudigel and Schreyer(1977) also reported no shift in clinochlore composition.However, Jenkins (1980, l98l) indicated that chlorite becomesmore aluminous than the
clinochlore composition with increasing temperature at
pressuresabove 4 kbar, and that the compositional shift
in chlorite composition may be dominantly temperaturedependent.
Two major aspectsof clinochlore synthesiswith which
this study is concernedare the conversion ofthe 7 A to
14 A structure and the possibility of compositional variation in cliniochlore, which appear to be kinetically controlled processes.To investigate these kinetic problems
in clinochlore synthesis,long runs up to 249 days were
conductedat 2kbar water pressure,using oxide mixtures
of clinochlore composition as starting material.
Preliminary results reporting aspectsof this work were
given by Cho and Fawcett (1982).
600
850
HzD
m0
750
TEMPEBATUBE,
G"
Fig.2. Dissociation curvesfor the reaction clinochlore : forsterite + cordierite * spinel + vapour. Solid curve is for the
reaction involving anhydrous cordierite and short-dashedone
for the reactioninvolving I .8 wt.0/0HrO in cordierite (Chernosky,
1974).Dashedline is the dissociationcurve determinedby Yoder
(1952). The dots indicate the experimentalconditions ernployed
in this study.
associated measuring and recording circuitry were calibrated
againstthe melting point ofNaCl (800.4'C). Uncertaintiesin run
temperaturesare a combination of the thermal gradient over the
capsule length (+2"6, Cermignani, 1979), uncertainty in thermocouple calibration (+2'C) and the variance ofthe day-to-day
fluctuations in recordedtemperature(+0.5 to t3'C; I standard
deviation), and total 14.5 to +7oC. Pressureswere measuredon
a 50000p.s.i., I 1-inch-diameterHeisegauge(manufacturer'scertified accuracy+50 p.s.i. at all pressures)and are believed to be
accurateto + 100 p.s.i.
The starting material was an oxide mix of MgO, AlrO, and
SiO, in the stoichiometric proportion of clinochlore composition
(5MgO' AlrO, . 3SiOr). The following reagentswere usedin preparation of the oxide mixture: Magnesium carbonate-FisherCertified Reagent(lot 762372) heated in air to produce MgO; aluminum hydroxide-FisherCertified Reagent(lot 791077) heated
in air to produce tr-alumina; silica glass-SpecpureJMC 425 fired
overnight at 900€ to drive offabsorbed water. SEM observation
revealed that the grain size of the oxide mix was less than I 5 pm
and averaged about 3 pm.
Starting materials were weighed into the annealed gold tubes
(2.5 cm in length) with excesswater, which was approximately
equal to 25 vrt.o/oof the charge. In most casesone capsule was
run per experiment. By pre-heating the furnaces it was possible
Experimental methods
to reduce the warm-up time to about 30 minutes which served
AIl runswereconducted
at 2 kbar(29000p.s.i.)waterpressure to minimize the kinetic efect ofwarm-up on the final run product.
using conventionalhydrotherrnaltechniques.Details are de- Reported "run duration" times do not include the warm-up time.
scribedby McOnieet al. (1975)and Cho (1982).Temperatures Quench time (to room temperature)was less than 3 minutes.
of theexperiments
weremonitoredby internallyfittedchromelThe products of each experiment were examined with a bin(Zuinch O.D.),and measured
alumelthermocouples
by eithera ocular and petrographic microscope, and by X-ray powder difKAYE 8000 multichannnelrecorderor a Leedsand Northrup fractometry. Due to the extremely fine-grained nature of most of
potentiometer
(model8690-2)usingzerodegrees
centigrade
as the run products, identification was based exclusively on X-ray
a referencetemperature.Each internal thermocoupleand the diffraction (XRD) methods.A Philips X-ray diffractometer,em-
70
CHO AND FAWCETT: KINETIC STUDY OF CLINOCHLORE
TC-F0-SP
TC-F0-SP-CoBD
a
TR
^-
EEE
.(J
,oo
E
ro-coRD-sP
f;
rAcHr-Fo-rc-0.r.
rll
c,
f
F
c,
ut
g
=
rl
',r
r5o
o
t Er
t"-.\
ilAcHt
tll
F
tA cHL- t0 - o.ilr.
"\
8 s i'..:..
LOGTIME( hours)
Fig. 3. A TTT diagram for the synthesisofclinochlore and forsterite + cordierite + spinel. Long-dashedcurve is the calculated
line for the completeconversion of 7 A chlorite to 14 A chlorite and short-dashedcurve for 75Voconversion.Someruns that merely
confirm the boundariesas shown, are not plotted for clarity. Abbreviations: Chl (chlorite), Fo (forsterite),Tc (talc), Sp (spinel),Cord
(cordierite) and O.M. (oxide mix).
ployingNi-filtered, Cu-rKa
radiation,wasusedto examinesmear
mountson glassslides.
Selectedsamplesof low temperature(below694'C)assemblageswereexaminedby electronmicroprobeto determinethe
chemicalcompositionof the run products.Fragmentsof run
productsweremountedwith epoxyon a glassslideandanalyzed
on an ARL instrumentwith an energy-dispersive,
solid state
potentialof 20 kV and a beamdidetector,usingaccelerating
ameterof approximately
2 pm.
Results
Mineral phasesand their compositions encounteredin
this study are plotted in Figure l. Temperature and mn
duration were varied to determine their effects on the
kinetics ofsynthesis. In order to investigate synthesisof
the clinochlore composition far below and abovethe equilibrium temperature of reaction (l), the temperaturesof
individual experiments varied from 600 to 750.C in l0
and 25 degreeincrements, and the run duration from l0
minutes to 249 days. The results of the experiments are
presentedin Table l.
The time-temperaturrtransformation diagram
A TTT diagram (Putnis and McConnell, 1980) displaying the time dependent transformations at different
temperatureshas been establishedfrom the results ofthe
various hydrothermal experiments for reaction (l) (Fig.
3).
Most boundariesdelineatingindividual fields of the assemblagesare well constrained by the synthesisexperiments and appear to be curvilinear. They show negative
slopesexceptthe one for the disappearanceof 7 A chlorite,
which shows a C-type curve typical of many TTT diagrams(Putnis and McConnell, 1980).Someruns alaparticular temperature and duration were duplicated and the
were found to be reproducible. For example,
assemblages
three 20 day runs were made at 689"C (runs 7, 30 and 3 l,
Table l): Runs 30 and 3l representthe duplicates from
the two capsulesloaded togetherin a pressurevessel.All
three run products contain the same mineral assemblage
of 14 and 7 A chlorite * talc * forsterite, although the
modal abundancesof individual phasesare slighty variable (seeTable 1 for further examples).
At temperatureshigher than720"C, talc + forsterite +
spinel occurs as a metastablephase assemblagein short
duration runs, crystallizing earlier than the stableassemblage,forsterite * cordierite * spinel. The proportion of
metastabletalc increasesgradually with time in runs of
short duration. However, it decreasesrapidly and finally
becomeszero as the more stable cordierite appearsand
becomesmore abundant with increasingrun duration.
Below 740t, 7 A chlorite-forsterite-talc or 7 A chlorite-forsterite assemblages
have beenidentified in shorter
runs. The occurrenceof non-aluminous phaseswith chlorite requires additional material that is more aluminous
7l
CHO AND FAWCETT: KINETIC STUDY OF CLINOCHLORE
Table l.
Run b.
Tenperatures
59?.3!?.0, n-56
598.117.7, n=106
6 2 4 . 5 ! 5 . 1 ,n - 1 6
624.9!5.5, n=34
6 2 7 . 1 ! 5 . 0 ,F l 4
623.7!7.0, n-37
3{
33
126
129
549.1i5,2, n=9
648.914.9. n-lg
650.5!6,7, n-19
550.9i6.7' n=25
9d
20d
5rd
70d
td
60
53
95
670.111.5' n--2
559.915.8' n-5
669.4J5.4' n-15
6 7 2 , 3 ! 5 . 1 ,n = 1 0
39
105
670.9!6.4. ^=32
570.715.8, F56
108
670.715.8, n-56
62
674.0!6,2, n-57
112
96
680
575.616.1, n=3
5 8 0 . 1 1 5 . 4 ,n - l {
682.4!5.4. n-zo
102
58I.815.5' n=41
110
6€I.l!5.3,
50d
122d
9d
20d
r50d
1d
9d
20d
n-51
52d
91
58I.515.4, n=71
72d
79
689.3!5.0' n=10
690
590
689.?14.5, n-6
589.4J4.7, n=l8
l0n
lh
3h
1d
ChI
d
rntenslty
= Percent
(Chlorite),
(Spinel),
Sp
-
12d
t 55d
20d
52d
122
Abbrevlatlons:
**xchl
hratlon
89
90
124
46
131
36
12
1l
*Ir
ln
(day)
ralio
of
ad
Fo
(001)
tEansfoination
Plducts
Run No.
7A sI
7 & t4A chl
zA chl
7 & 1{A cbl
TErtAsl
7 a r4A chl
0.10
0.14
0,30
t4
20
43
7A Chl-tr. Fo
?s14Afrt
14 & 7A chl
14A Chl
o.21
0.51
0.65
30
a7
94
7A or-tr.
b
7A Chl-tr. Fo
Tdl4Asl
1l e ?A Chl-tr.
ft-tr.Fo
lt r 7A chl
14 E 7A sl-t!.
Tc-t!,Fo
14 r ?A chl-tr.
Tc-tr.Fo
144 sl-tr.Tc-t!.Fo
7A frl-Frtr.Tc
7A chr4c-Fo
14 & 7A sl-Tc-Fo
14 & ?A cht-tr.
&-tr.Fo
1l E 7A Chl-tr.
&-tr.Fo
14 6 7A Chl-tr.
b-tr.Fo
144 ChI-tr,Tc-tr.Fo
7A chl-Fo
7A cht-Fo-rr. t
7A chl-Fetr.
?c
?A frt-tc-Fo
7 r l4A chl-lc-Fo
(rorslerite),
(CoEdierlte),
6!d
t!,
Clystalline
0,r8
0.50
7I
0.{3
0,54
5t
71
n.62
89
0,56
94
0.38
0.50
5a
7I
0.57
8I
o.62
89
0.55
94
(ntnute),
(hour),
(trace)
and
(002)
of
dlorite
7
5 8 9 . 5 1 1 . 8 , n = 1t 9
20d
589.1!1.5, n-20
689.111.5' n-20
590,11{,8, n=29
20d
20d
28d
l9
27
589.9i4.5, n=34
589,5!5.3, n=56
589,5i5.3, n--56
36d
12d
72d
r11
692.815.8, n-73
109d
29
28A
9I
105
r03
119
593.5i5.5, n-84
593.216.3' n=150
708.514.8' n=10
7O4.2!3.1, n=ZO
702.6!6.4, F42
701.315.6, n=29
r40d
249d
l0n
20d
104
7O2.1!6.2, n=7I
s2d
67
80
53
70
55
716,2!4.5. n-2
717.5
713.814.8' n-5
7 1 7 . 1 1 4 . 5 ,n = 8
7 1 5 . 8 1 4 . 9 ,n = 1 2
716.9!5.2' n-l5
712.1!5.2. n-53
llh
ld
100
84
72
719.615.4, n=3
729.4!4.5. n=5
724.016.1| a=4
722.1!4,9' n-9
7 2 1 . 1 1 4 . 8 'n = 1 5
125.7!6.0' ^=25
75
92
77
118
58
83
55
52
742.315.3' n=10
752.515.8' n=10
713
740
74!.1!6.4t n-2
744.1!6.4, n-2
742.7!7.1, n-8
7 4 0 . 3 ! 5 . 2 ,n - 1 8
98
93
order
of
chlolite
plFolpha.
plForphs,
than clinochlore composition. However, no phasesother
than chlorite, forsterite and talc appearedin XRD analyses.This may be attributed to a very small amount of
the assumed aluminous phase(s),or to the amorphous
nature of the assumedaluminous material. The latter may
be the casesinceabundant isotropic aggregate,which appear to be partially-reacted oxide mixture, are detected
by the petrographicmicroscopein shorterruns. The amorphous oxide mixture occurs as aggregates,composed
of extremely fine grains of less than 0.5 pm size and the
very fine grain size ofthis material precludesexact identification.
The occurrenceof phaseswith little or no alumina, i.e.,
forsterite and talc, in addition to 7 A chlorite may imply
that the nucleation rates offorsterite and talc are greater
than that of chlorite as pointed out by Staudigel and
Schreyer(1977).
7 A chlorite persists metastably for a very long time
(over 72 days at 600'C) and finally converts to l4 A chlorite through the mixture of 7 A and 14 A chlorite. The
transformation of 7 A chlorite to 14 A chlorite will be
discussedin detail in a later section.
In summary, the present experiments suggestthat the
assemblages
chlorite-forsterite, below 670.C (Chernosky,
phases
nmber
10h
ld
4d
@uld
of
s 7A frl-tr.
Tc-tr. Fo
14 r 7A &I+-Eo
14 r 7A Cbl-Tc-Fo
1t & 7a Sr-tr.
Tc-tE. Fo
14 s ?A chl-Tc-Fo
14 e 7A Chl-tE. Fo
1{ r 7A frl-tr,
0.45
54
0.51
o.47
0.51
73
67
13
0.53
0.60
76
86
t{ & 7A 61-u,
Tc-tE . Fo
14AsI-tr.Tc-tr.Fo
1{A chl-tr.Tc-t!.Fo
?A &I-Fetr.
t
14 & 7A chI-Tc-Fo
1{ & 7A ch1-Tc-Fo
t4 & 7A Chr4c-Fr
0.55
93
0.70
0.58
t00
97
0.53
0.53
0.59
76
16
84
lt
14 e ?A chr-Tc-Fe
sP
t!. 6rd-tr,
Tc-Fr7A ChI
&-ao-? r l4A frt
lt 6 7A chr{c-Fo
Tc-8o-r4 & 7A frI-SP(?)
Tc-EeSI{ord
Eo-Co!d-SP
Cord-FrsP
Tc-? e 144 ChI-Ao
&-Fo-sp
Tc-Fo-Sp{ord
21.8d
@rd-Fo-sP
l0n
10n
lh
10h
r.5d
3d
9d
l7d
Fo-rc-7A 6t
Tc-ArSp
Ac-Fo-Sp
Tc-Fe8p
ft-Fo-Sp-lr,
Tc-Frststr.
no!
peak,
6Ed
Cord
Fo-AP{oEd
!o
@lres[Ends
llsted
real-o! ienla!1on
!
9d
20d
40d
72d
Plductg
Cryatalllne
46d
characteristic
identlfled
prefer
rr*n
of
strongest
tbel.r
phaees
reflecltons
Duration
30
3l
The
h
in
TenFratures
73
0.35
(TaIc),
&
n
(cont.)
Table l.
Run data
one
although
be esllnated
of
the
Eellably
decreaslng
lntenslty
relativ€
proPortlons
bcauEe
of
of
of
lhe
the
ef f ect.
temperature
reasurenents.
1975)and talc-forsterittsspinel(Yoder, 1952;Segnit,1963)
described by previous investigators, are metastable and
crystallizeearlier than the final stableassemblagebut persist for long periods of time. Most stable phase assemblagesare synthesizedonly through an intermediate assemblage,providing a good demonstration of Ostwald's
rule ofthe prior formation of unstablephases.The gradual
transformation of several metastablephase assemblages
to final stable assemblagescan be traced on the timetemperature-transformationdiagram. The recognition of
usinga TTT diagram shows
metastablephaseassemblages
that reactionsare extremely sluggishboth at low temperatures and near the reaction temperature. This provides
further reinforcementof the needto demonstratereaction
reversalsin phaseequilibrium studies.Therefore,the importance of this diagram cannot be overemphasizedin
understandingsolid statereactions.Unfortunately, at the
present time very few TTT diagrams are available for
metamorphic reactions.
Rate stu.dies
Rate studies have been carried out for the transformation of 7 A chlorite to 14 A chlorite below the equilibrium temperatureof reaction ( I ), usingproducts of runs
72
CHO AND FAWCETT: KINETIC STUDY OF CLINOCHLORE
508
x
TIME,days
Fig. 4. Variation of d and X- as a function of run duration (in days). Mean experimental temperaturesare 626, 650, 670, 68 I,
and 691"C. Length ofsolid bar representsprecision ofmeasurement and dotted lines connect duplicate runs.
that yielded 7 A and 14 A chlorite assemblages.Quantitative determination of the extent of transformation was
determined by measuring the peak heights from basal
reflectionsof chlorite on powder X-ray diffraction charts.
The relative intensities (1,) of (001) and (002) basal reflections from a 7 A and 14 A chlorite mixture were used
for quantitative analysesof 14 A chlorite, becauseboth 7
A and 14 A chlorites will orient to the same degreeto
eliminate the effect of the preferred orientation (Mossman
etal.,1967). All sampleswere scannedthree to eight times
at lolminute to determine the ratios of the peak heights
(Table l). In calculating the percent reaction (X"n), it is
assumedthat the intensity ratio of (001y(002) for pure
14 A chlorite is 0.7 which is the maximum value obtained
in this study as shown in Figure 4. The ratio increases
with run duration and reachesa steady state value of0.7
in experimentsofabout 150 days. The precisionofthe
method was estimated to be better than + l0o/oX.ny
Figure 4 is the plot of { versus run duration. Mean
experimental temperatureswerc 626, 650, 670, 68 l, and
691"C.Duplicate runs are also plotted, which show a fluctuation of up to l5o/oX"n. As a small degreeof reaction
cannot be determined by XRD, the initial stageof transformation at each specific temperature is difficult to define. Large uncertainties in X"n, also make it difficult to
distinguish experimentally between the various solid state
reaction models (Sharpet al., 1966).However, the curves
connecting experimental data points appear to be parabolic in shape,with rapid initial reaction.
The simple model equation to best fit the experimental
data is the parabolic rate equation (Kridelbaugh, 1973;
Matthews, 1980):
X"r,r: k't'
(2)
where X", is the percent of transformation, k, the rate
constant,t is time in days, and n is the order of reaction.
Writing equation (2) in logarithmic form:
logX"n,:logk+nlogt
(3)
gives a straight line with a slope of n and an intercept of
log k in a log X"n,versus log t plot.
Figure 5 is the plot of experimental dala at each temperature showing linear-regressionlines based on equation (3). The values ofthe constantsn and log k for equation (3) are presentedin Table 2. Times for the 7 5o/oand
complete conversion of 7 A chlorite to 14 A chlorite are
calculated from the linear-regressionequations at each
temperature.They are also listed in Table 2 and plotted
in Figure 3. As can be expected, it required very long
durations for the complete conversion both at low temperature (312 days at 626C) and near the reaction temperature.The apparentinflections at 670C (Fig. 3) might
indicate a changein reaction mechanism that will be discussedlater.
The uncertainties in the above values are calculated by
weighting the standarddeviations at eachdata point @evington, 1969, Chapter 6). When the standarddeviations
(ol and o2) for each data point are different becauseof
the nonJinear scale,the averages,o : (ol * o2)/2, are
CHO AND FAWCETT: KINETIC STUDY OF CLINOCHLORE
taken throughout the study. Two different standard deviations for a few casesare listed together.
The order ofreaction decreasesgradually with increasing temperaturefrom approximately 0.9 at 626 and 650C
to 0.15 at 69I'C (Table 2). This variation in n indicates
a changein the mechanismof the reaction (Busenbergand
Plummer, 1982).
The rate constant ft) is related to an activation energy
(HJ bV the Arrhenius equation:
k: A exp (-H"/RT)
log k:
TIME,days
(4)
where A is a temperature-independentfrequency factor,
R is the gas constant and I is the absolute temperature.
Writing equation (4) in logarithmic form:
log A - H^/2.303RT
IJ
x
g
ct
(5)
Therefore,the Arrhenius plot oflog k versus l/T givesa
straight line of slope -HJ2.303RT and intercept log A.
Figure 6 is the Arrhenius plot of k derived from equation (5). There is an apparentbreak in the slope at about
670C, which indicates a changein reaction mechanism.
An activation energyof 90+35 kcal/mole (f1",) was calculated at temperaturesbelow 670'C and an activation
energy of 177!t^t3kcal/mole'?(H"r) at temperatures be012
tween 670 and 691'C. The uncertainties in H^, and Hu,
tOGTIME.daYs
are calculated by considering only the uncertainties in k
Fig. 5. A plot of log X"o,versus log time (in days) showing
and temperature, respectively.A singleline, with no change
in slope could be fitted to all five data points on Figure linear-regtessionlines at eachtemperature.Symbolsare the same
6. We prefer the interpretation shownin the figurebecause as in Figure 4. Uncertainties on X." are all the same but appear
independent evidence, discussedlater in the paper, sug- different due to the log scale.
gestsa changein chlorite composition at about 670"C.
Our interpretation of thesedata is also consistentwith the
inflection noted at this temperature for the boundary curves
shown on Figure 3.
The high value of H,r, as compared to I1., can be
related to the stableexistenceofforsterite + talc with 14
A chlorite at about 670-696C, as will be demonstrated
H",= lZlflkcal/mole
in a later section. As the 14 A chlorite in the chloriteforsterite-talc assemblageis more aluminous than the ideal clinochlore composition, H", may represent the activation energyfor the transformation of Mg-chlorite polymorphs that are more aluminous than the ideal clinochlore
composition.
An activation energy larger than 90 kcaVmole might
indicate that diffusion is probably not the rate-controlling
Hr,,=90135kcal/molc
step for the transformation of chlorite polymorphs (see
Table 3). As suggestedby Roy and Roy (1955), a number
of Si-O bonds must actually be broken and rearranged,
so that every alternatelayer ofSia+ ions in the silica layer
of 7 A chlorite can form new Si-O bonds with the linked
tetrahedrapointing in the opposite direction. Sucha rearrangement of the Si-O bonds might be the rate-deterttz
10.2 r0.4 10.6 10.8 ll.0
mining processfor the transformation from Z A to tq A
'
K
1
yt'm1
structure.
Fig. 6. The Arrhenius plot of log k versus l/I
(in degrees
2Two differentstandarddeviationsdueto the non-linearscale Kelvin). A changein the reaction mechanism is suggestedby the
arelistedtogether.
change in slope at approximately 670{.
74
CHO AND FAWCETT: KINETIC STUDY OF CLINOCHLORE
[0G TIME,hourc
t.8
2.0
2.2
2.4
2.6
2.8
3.0
r0.5
I
-
CJ
0
o
q
c,
650
lrt
E
ozo
E
ut
4
=
EI
fl.0
TIME,days
Fig. 7. A plot of log time against l/Z(in degreesKelvin) showing the bracketing data (solid dots) between 7 A chloriteforsteriteoxide mixture (open triangle) and 7 A and 14 A chlorite (open rectangle). Thick solid line represents an estimate of the phase
boundary betweenthe two assemblages.Dotted area represents"envelope ofuncertainty." *Three day runs are included in order
to reduce the range ofuncertainty. This does not affect the present interpretation becausefI" without 3 day runs is 62 kcallmole.
Nucleation activation energyof la i chlorite
The activation energyfor the nucleationof I 4 A chlorite
from 7 A chlorite-forsterite-oxide mixture can be calculated from the TTT diagram.
The rate of nucleationK at any temperatureis the product of the thermodynamic factor and kinetic factor (Putnis
and McConnell, 1980):
where C' is constant. For all low temperaturesAG* = 0
(Putnis and McConnell, 1980; Putnis and Bish, 1983),
and
ln t:
II,/nRT - lnC"
(9)
where C" is constant. The activation energy for the nucleation of 14 A chlorite can be found from the slope in
K: C exp (-AG*/RZ) exp (-H^/RT)
(6) the ln t versus l/Iplot at low temperatures.It should be
noted that the activation energyfor nucleation, calculated
where C is constant, AG* is free energy of activation for using the equation (9) represents maximum value
the
bethe formation of a critical nucleus, and H^ is activation
causeAG* is assumedto be zero.
energy for a kinetic process.Writing (6) in logarithmic
The curve for the appearanceof 14 A chlorite is almost
form:
linear at temperaturesbelow 670'C (Fig. 3). Experimental
ln K: ln C - (AG*/RT) - (H^/RZ")
(7) data are plotted again in Figure 7 on a log t versus l/Z
diagram to define the curve for the first appearanceof 14
From equation (3), ln X"o,: ln K * n ln t. Therefore,
A chlorite. The uncertainties in the bracketing data bewere calculatedusingequation
n ln r: (AG*/R7) + (H,/RT) - ln C'
(g) tweenthe two assemblages
6a of Demarestand Haselton(1981).Only the uncertainties in temperatureare consideredsince the errors in run
Table2. Valuesofthe constantsn and log k for the equation,
plotted in log scale are small in long runs. An
log X"" : log k + n log /, and the calculatedtimes for the 75o/o duration
activation
energyof 6313kcal/mole was calculatedfor the
andcompleteconversion
of 7 A chloriteto 14A chloriteat each
first
appearance
of 14 A chlorite from the TTT curve,
experimentaltemperature
using n : 0.9 (from Table 2), as compared to a value of
?enlErature
Tlme for
(dayE)
@nversion
90 + 35 kcallmole from the rate study. The errorsare taken
('c)
n
I@
k
751
100t
from
the slopesof two extreme lines.
-0.2504!0.4454
625.6
0.902!0.221
226.7!58.8
3I].8i I 12.9
0.91810.096
The activation energy calculated from the TTT curve
0.338510.1524
17.1t0.2
64.5r,t.9
670.4
0.65510.I0?
0.7554!0.1782
{6.0!0.5
?1.315.3
is smaller than the activation energy obtained from the
68r.4
0.44910.041
1.201510.0648
31..710.5
60.2!2.5
590.5
0.f5810.0f5
t.544{10.0287
29.0!2.3
179.2!19.4
rate studies,implying that nucleation is not the rate-controlling mechanism. However, because of large uncer-
75
CHO AND FAWCETT: KINETIC STUDY OF CLINOCHLORE
Table 3. Activation energy scale (kcaVmole);from Fyfe et al.
( l 978)
Dehydration
and
No.
seple
Tenperature ("C)
Duracld
{day6)
hydracion
12-20
uigratio
15-30
Brchange diffuEion
of
Table 4. Representative electron microprobe analyses of two
selectedchlorites
alkali
cation8
Ehrough
olEn
structurea
of Na, K, interatitial
(vacanci')
Oxides et.t
diffu6ion
25-40
MgO
Al203
SIO2
Total
Displacigetransfornatlon
50-70
Exchange niglation
50-90
O-Si-O
bonds frm
of si,
Lattie
energie6,
29,81
16,90
26,47
31,74
17.22
28,11
77,13
A1
non-brldglng
dygens
U9
100-200
24
689.s
72
39
670.9
36
break up of O-si{
bond6
si
tainties involved, there is still a possibility that nucleation
of 14 A chlorite is a rate-controlling process.In this case,
the difference in calculated activation energiescould be
attributed to two factors.First, small amounts of forsterite
and oxide mixture coexistingwith 7 A chlorite might have
decreasedthe activation energyfor the conversion of7 A
chlorite to 14A chlorite by providing appropriatematerial
for nucleation of 14 A chlorite. Second, the activation
energy obtained from the TTT curve may represent an
transformation, since
activation energyof a heterogeneous
the charge container may have provided heterogeneous
nucleation sites and decreasedthe activation energy for
the appearanceof l4 A chlorite. Heterogeneousphenomena in nucleation and $owth of chlorites are ubiquitous
in the run products, as demonstrated in the SEM study
(Cho and Fawcett, 1986).
Chlorit e-for sterit e-t al c assembIage
This assemblagewas first identified by Fawcett and
Yoder (1966) in 3.5 kbar water pressureexperimentsand
was interpreted by them to be a stable one in which Mgchlorite becomesmore aluminous than clinochlore composition. Jenkins (1980, l98l) further indicated this shift
in composition at pressuresabove 4 kbar. At l4 kbar and
850'C he estimated the composition of the most stable
chlorite to be (Mg ssAlrrrxsi28sAlrr2)Oro(OH)r.Chernosky (1974) did not detect any compositional variability
for water pressuresup to 2 kbar and concluded that the
most stable Mg-chlorite is of clinochlore composition.
Severallonger runs, up to 249 days, were conducted at
temperaturesbetween 670'and 693'C, in order to investigate the stability or metastability of the chlorite-forsterite-talc assemblage.This assemblagepersistedin each
of those runs, even beyond the time calculated for the
completeconversion of 7 A chlorite to 14 A chlorite (Fig.
3). The 150 day experiment at 674C helps confirm the
persistenceof the assemblageand, together with shorter
duration runs, establishestemperaturelimits for the chlorite-forsterite-talc assemblage (Fig. 3). Field boundary
configurations on the TTT diagram strongly suggestthat
this assemblageis stable over a relatively narrow temperature interval.
To investigate the stability of the chlorite-forsterite-
4.889
2.192
2.912
4.939
2 . 12 0
2.911
talc assemblage,the peak position of the doooreflection
was determined for 14 A chlorite obtained in runs longer
than 3 days (Chernosky,1974).Three to eight oscillations
were made ftom 24.5 to 27" 20 at a scanningspeedof I /4"
20/min., using quartz as an internal standard. The serpentine determinative curve (Chernosky, 1975) which is
also applicable for 14 A chlorite (Schiffman and Liou,
1980) was used to estimate the Al content of these
synthetic chlorites. The precision of this method is
+0.015'2d (CuKa), which correspondsto a compositional
10.04 in the formula,
variation of about
(Mgu-,Al-)(Si4--Al")O,o(OH)e (Caruso and Chernosky,
r979).
Severalelectron microprobe analysesof chlorite were
performed to calibrate the XRD measurementof chlorite
composition. Unfortunately, the analyses always gave
lower totals than the ideal chlorite composition by l0l5ol0,which may be accountedfor by the porous nature
of the disaggregatedchlorites (Table 4). However, a few
analysesshowing good chlorite stoichiometry proved to
be consistent with XRD results, especiallyin the 690"C
run product.
The resultsare listed in Table 5 and illustrated in Figure
8. It is evident that with increasingtemperature there is
a sudden decreasein dooospacing at about 670'C, and
that the chlorite composition variesfrom pure clinochlore
at temperaturesbelow 670'C to more aluminous chlorite
of the approximatecompositionwith x: Lll+0.04 at
temperaturesbetween 680 and 693C. This x value is
consistentwith x: l.l2 obtainedat 14 kbar by Jenkins
(1981). Therefore, the maximum stability of clinochlore
is defined not by reaction (l) but by an overall reaction
such as:
l.IIMgrAlrSirO,o(OH)r
clinochlore
: (M&.rAlt
'XSi2 sAlr rr)Oro(OH)8
Mg-chlorite
+ 0.264Mg,SiOn
forsterite
+ 0.044MgrSLOro(OH),
talc
+ 0.396H,O
vapor
(10)
76
CHO AND FAIVCETT: KINETIC STUDY OF CLINOCHLORE
t.00
o{
.E
3.570
1.031
cf
ct
ct
-t
r.06
t.09
rite with x : I . I I can be estimated from the TTT diagram.
In runs longer than 46 days at 701'C, cordierite and spinel
have formed together with the chlorite-forsterite-talc
assemblage, suggesting that the latter assemblage is not the
stable one above 701'C. This is also consistent with Chernosky's (1974) bracketed reaction temperature of
696+ 10"C. Yoder's (1952) reaction temperature which is
20 degrees higher than that ofthis study probably resulted
from the metastable persistence of the chlorite-forsteritetalc assemblage above the equilibrium reaction temperature. Thus the importance of reversals in classical synthesis experiments cannot be overemphasized in determining phase equilibria.
L12
Acknowledgments
This study representspart of M. Cho's M.Sc. thesis at the
University of Toronto. Dr. J. M. Allen and Dr. G. M. Anderson
have contributed greafly in discussionsduring the courseofthis
work. The research was supported by grants from the Natural
TEMPERATURE,'C
Sciencesand EngineeringResearchCouncil of Canada to J. J.
Fig. 8. Variation of dooo
and Al contentin termsof x value Fawcett and by the H. V. Ellsworth Fellowship in Mineralogy at
in syntheticchloritesas a functionof temperature.
Opencircles the University of Toronto (M. Cho). Reviews of a preliminary
representthe analyses
determinedby XRD and solid dots by version of this manuscript by J. M. Allen, J. V. Chernosky, D.
electronmicroprobe.
A. Hewitt, D. M. Jenkins, J. G. Liou and T. P. Loomis resulted
in significant improvements to the presentation.
1.15
600
625
650
670
690
whose reaction temperatureis approximalely 670.C at 2
kbar water pressure.Further detailed studies are required
in order to characterize the reaction governing the transformation of clinochlore to the more aluminous chlorite,
becauseFigure 8 does not preclude the possibility of a
continuous reaction involving the coupled substitution of
2Al = Mg + Si in chlorite. If the latter is true, clinochlore
will gradually changeits composition towards the more
aluminous chlorite (x : l.l 1) within a short temperature
interval ofless than 20'C at 2kbar water pressure.
The upper thermal stability of the most stableMg-chlo-
References
Bevington, P. R. (1969) Data Reduction and Error Analysis for
the Physical Sciences.McGraw-Hill Inc., New York.
Busenberg,E. and Plummer, L. N. (1982) The kinetics of dissolution of dolomite in COr-HrO systemsat 1.5 to 65"C and
0 to 1 atm P-r. American Journal of Science,282, 45-78.
Caruso, L. J. and Chernosky, J. V., Jr. (1979) The stability of
lizardite. Canadian Mineralogist, 17, 757-7 69.
Cermignani, C. (1979) Metarnorphic reactionsin the systemalbite-anorthire-nepheline-NarCOr{aCOr-HrO,
with application to the Haliburton-Bancroft alkaline rocks. Unpublished
Ph.D. thesis,Univ. of Toronto.
Chernosky,J. V., Jr. (1974) The upper stability ofclinochlore at
low pressureand the free energyof formation of Mg-cordierite.
Table5. Variationin doon
of chloriteandthecalculated
x values,
American Mineralogist, 59, 496-507.
usingthe calibrationcurveofChernosky(1975)
Chernosky,J. V., Jr. (1975) Aggegate refractive indices and unit
cell parameters of synthetic serpentine in the system MgODuration
Tenperature
x values
AlrO3-SiOr-HrO. American Mineralogist, 60, 200-208.
('C)
Run no. in days
2e(004)
d(004)
Cho, M. (1982) A kinetic study of the clinochlore composition
90
1s6
598. 1
24.914
at 2 kbar water pressure.Unpublished M.Sc. thesis, Univ. of
3.574
0.955
131
50
24.923
3.573
0.985
Toronto.
125
12
624.7
24.92L
3.573 0.981
Cho, M. and Fawcett, J. J. (1982) Mechanisms and kinetics of
123
122
625,7
24.9L1
3.573
0.972
349
548. 1
the reaction clinochlore : forsterite * cordierite + spinel +
24.926
3.572
0.993
33
20
648.9
24.919
3.573 0.976
vapour. (abstr.). Transactions of the American Geophysical
Union @OS),63,465.
43
36
650.4
24.924
3.572
0.998
126
51
650.6
Cho, M. and Fawcett, J. J. (1986) Morphologies and growth
24.910
3.574 0.955
129
71
650.9
24.920
3.573
0.979
mechanisms of synthetic Mg-chlorite and cordierite. American
603
669.9
24.944
3.570 1.036
Mineralogist, 7 l, 78-84.
669.4
24.922
3.573
0.983
't.034
39
36
670.9
Demarest,H. H., Jr. and Haselton,H. T., Jr. (1981)Erroranalysis
24.943
3.570
1.061
for bracketed phase equilibrium data. Geochimica et Cos62
150
674.0
24.953
1.065
3.560
mochimica Acta, 45, 217-224.
102
36
6 8 1. 8
24.974
3.555 1.112
97
72
Fawcett,J. J. and Yoder, H. S., Jr. (1966) Phaserelationshipsof
681.5
24.970
3.566 1.102
lt
3
689.4
24.472
3.566
1.107
chlorites in the system MgO-AlrOr-SiOr-HrO. American
720
589. s
24.967
3.566
1.094
Mineralogist,51, 353-380.
32
29
690. I
24.954
3.658 1.071
Fyfe, W. S., Price, N. J., and Thompson, A. B. (1978) Develop24
72
699.5
24.969
3.566
1.099
t.096
ments in Geochemistry, Vol. l: Fluids in the Earth's Crust.
27
72
24.971
3.566 r .104
Elsevier Scientific Publishing Company, Amsterdam.
29
140
24.976
3.565 1. I t7
29A
249
693.2
24.972
Greenwood, H. J. (1963) The synthesisand stability of antho3.566 1.107
phyllite. Journal of Petrology,4, 317-351.
CHO AND FAWCETT: KINETIC STUDY OF CLINOCHLORE
Jenkins,D. M. (1980) Crystallochemicalcharacterizationof synthetic Mg-chlorite. (abstr.)GeologicalSocietyof America Abstractswith Programs, 12,455.
Jenkins, D. M. (1981) Experimental phaserelations of hydrous
peridotites modelled in the system HrO-CaO-MgO-AlrO3SiOr. Contributions to Mineralogy and Petrology, 77, 166-
r76.
Kridelbaugh, S. J. (1973) The kinetics ofthe reaction: calcite :
quartz + wollastonite + carbon dioxide at elevated temperatures and pressures.American Journal of Science213, 757777.
Lasaga,A. C. and Kirkpatrick, R. J. (1981) Kinetics of Geochemical Processes.Reviews in Mineralogy, vol. 8. Mineralogical Society of America, Washington, D.C.
Matthews, A. (1980) Influences of kinetics and mechanism in
metamorphism: a study of albite crystallization. Geochimica
et Cosmochimica Acta, 44,387-402.
McOnie, A. W., Fawcett,J. J., and James,R. S. (1975) The
stability of intermediate chlorites of the clinochlore-daphnite
seriesat 2 kbar P"*. American Mineralogist, 60, 1047-1062.
Mossman,M. H., Freas,D. H., and Bailey,S.W. (1967)Orienting
internal standard method for clay mineral X-ray analyses.Clays
and Clay Minerals, 15,441-453.
Nelson, B. W. and Roy, R. (1958) Synthesisof the chlorites and
their structual and chemical constitutions. American Mineralogist,43, 707-725.
77
Putnis, A. and McConnell, J. D. C. (1980) Principles of Mineral
Behaviour. Elsevier, New York.
Putnis, A. and Bish, D. L. (1983) The mechanism and kinetics
ofAl,Si ordering in Mg-cordierite. American Mineralogist, 68,
60-65.
Roy, D. M. and Roy, R. (1955)Synthesisand stability of minerals
in the system MgO-AlrO3-SiOr-HtO. American Mineralogist,
40, t47-r78.
Schiffman, P. and Liou, J. G. (1980) Synthesisand stability relations of Mg-Al pumpellyite, CanAlrMgSirOr,(OH)t.Journal
of Petrology, 21, 441-474.
Segnit, R. E. (1963) Synthesisofclinochlore at high pressures.
Americian Mineralogist, 48, 1080-1089.
Sharp, J. H., Brindley, G. W., and Achar, B. N. N. (1966) Numerical data for some commonly used solid state reaction
equations.American Ceramic Society Journal, 49, 379-382.
Staudigel,H. and Schreyer,W. (1977)The upperthermal stability
of clinochlore MgrAl(AlSirO,J(OH)r, at lG-35 kb P".o. Contributions to Mineralogy and Petrology, 61, 187-198.
Yoder, H. S., Jr. (1952) The MgO-AlrO3-SiO2-H2O systemand
th€ related metamorphic facies.American Journal of Science,
Bowen vol., 569-627.
Manuscript received,December1, 1983;
acceptedforpublication, September9, 1985.