American Mineralogist, Volume 72, pages 312-318, 1987
Solid-solution thermodynamicsin CaCO.-MnCO.
CrrmsropHnn ClponrANcor* AlnxlNou
NlvnorsrY*
Departments of Chemistry and Geology, Arizona State University, Tempe, Arizona 85287, U.S.A.
Ansrrucr
Thermodynamic mixing data are reported for the calcite-rhodochrositesolid-solution
series.Enthalpiesof solution in an aqueoussolvent saturatedin NaCl and containing 0.745
M }lCI were obtained for CaCOr-MnCO, samples prepared at 600"C. The enthalpy of
mixing is, in kcaVmol,
AIl-. : (D(l - X)1r.493+ 0.099 + (13.17+ 0.29)X - (73.19+ 0.91))c
+ (60.27+ 0.63)Fl,
where X is the mole fraction of MnCOr. Isothermal phase-equilibrium relations between
at 700'C gave the excessfree energyof mixing acrossthe
COr, MnO, and (Mn"Ca,
")CO,
JOrnas
AG'rto*./RT: (4(l - ].4/.rJ20 a 0.068 + (0.538 + 0.012)Xl.
Positive excessfree energiesand complex enthalpies (both endo- and exothermic deviations) imply negative excessentropies of mixing. Asymmetric enthalpy behavior is also
inferred from phase-equilibrium data. This behavior is seenin materials without the longrangeMn-Ca ordering characteristicof kutnahorite.
INrnonucrroN
The join CaCO.-MnCO, is a bounding binary in the
rock-forming carbonatetetrahedron, but carbonate solid
solutions with other than minor or trace amounts of Mn
are relatively rare. However, Mn-bearing calcitesare found
in and around ore bodies, especially epithermal Pb-Zn
and Ag deposits (Tsusue, 1967). Additionally, many of
the natural occrurenceshave been describedfrom regionally metamorphosed manganiferousterranes (e.g.,Zak and
Povandra,198l; Wenk and Maurizo, 1978).In theselatter instances,Mn-bearing calcitesare often associatedwith
other minerals in which Mn is a major component, commonly in solid solution with Ca components (Peters et
al., 1978; Winter et al., 198l). Geochemicalinterpretations of manganiferousassemblagesand certain ore deposits will be facilitated given knowledge of the thermodynamics of mixing for Ca and Mn. This study provides
data pertinent to Ca-Mn mixing in rhombohedral carbonate phases.
Beneaththe oceanfloor, Mn-bearing calciteshave been
found as precipitateson foraminifera tests (Boyle, 1983).
Authigenic Mn-bearing calcites in subsurface deep-sea
sedimentsfrom the Panama Basin have also recently been
discovered (Pendersonand Price. 1982). The latter au-
thors have reported severalother submarine occurrences
and have noted that the kutnahorite composition is commonly found. Kutnahorite is the compound having the
dolomite structure in the calcite-rhodochrosite system.
The prevalence of the kutnahorite composition in submarine environments is somewhat surprising. At temperatures of the ocean floor, the development of dolomite-type order within Mn-bearing calcitesis kinetically
unlikely, and actual ordered phasesare undocumented.
Understanding of these low-temperature occurrencesis
hindered by a sparsethermodynamic data base for solid
solutions in this system.
Previous data on the thermodynamics of mixing in this
system can be obtained only from phase-equilibria studies (e.g.,Goldsmith and Graf, 1957; Boynton, l97l; de
Capitani and Peters,l98l). This study reports solution
calorimetric data for a seriesof six supersolvussynthetic
solid solutionsacrossthe calcite-rhodochrositebinary join.
New high-temperature phase-equilibria experiments are
also presentedto further constrain the thermodynamics
of mixing.
ExpnmprnNTAL
METHoDS
Phase-equilibriumexperimentsinvolving
(Ca,Mn)CO, decomposition
* Present addresses:(Capobianco) Department of Earth SciThe purpose of these experiments was to obtain the activity
ences,University of Cambridge, Cambridge, England. (Navrotsky) Department of Geological and Geophysical Sciences, coefrcient of MnCO, in the carbonatesolid solution at 70fC as
Princeton University, Princeton, New Jersey08544, U.S.A.
a function of composition. This was done by finding the satu0003404x/87l030€312$02.00
3r2
313
CAPOBIANCO AND NAVROTSKY: THERMODYNAMICS OF CaCO.-MnCO.
Teeue1. Datafrom70eC (Ca,Mn)CO.decomposition
runs
Duration
(h)
Run
3
10
13
14
16
17
19
22
28
138
288
146
96
102
98
147
173
122
Pco"
(psi)
7 250
1 18 0 0
12420
10900
1520
8 960
I 080
I 700
10550
%'
(A)
6.2660(7)
6.0335(13)
5.9752(62)
6.1166(36)
6.3597(6)
6.2220(9\
6.2234(14)
6.2000(1s)
6.1295(19)
f)
46.395(7)
47.182(14\
47.375(59)
46.905(38)
46.106(6)
46.538(9)
46.534(14)
46.657(17)
46.857(27\
TneLe2. Unit-celloarametersand compositionsfrommicroprobe
analysisfor (Ca,Mn)CO.
pnases
201
40.58
40.57
40.s6
40.s6
40.56
40 55
40.s4
40.56
40.s6
- Rhombohedral
unit-celllength.
-' Rhombohedral
cell angle.
f Two-thetaanglefor (200)peak in MnO, CuKa.
ration concentration of MnCO, in the solid solution coexisting
with a pure MnO phaseand a CO. fluid of known fugacity.
Experimentswere conducted in a cold-sealhydrothermal apparatususing CO, asa pressuremedium. Run pressurewas monitored using a Bourdon-tube gaugewith 20-psi graduations up
to 30000 psi. The run temperature was monitored from the
external thermocouple well on a standard 1.25 in. (3.18 cm)
cold-sealvesselusing a chromel-alumelthermocouplewhich was
periodically calibratedagainstthe melting point of KBr at 686t.
Two experimental chargeswere included in each equilibration.
Each charge of 3-5 mg of starting material (consisting of an
undersaturated carbonate plus MnO and an oversaturated carbonate plus MnO) was wrapped in silver foil and held at run
conditions for durations between96 and228 h. The MnO phase
was obtained by decarbonation of previously recrystallized
MnCOr. The carbonatesolid solutions used for starting material
were obtained by techniques discussedbelow. The location of
the saturation surface in compositional space was approached
from either side, and therefore these runs were consideredreversed when the X-ray films from each run produced experimentally indistinguishablepatterns.
The oxygen fugacity ofthe CO, fluid was not bufferedexcept
by the pressurevesselitself. Al1 successfulruns contained only
MnO as the oxide phase,which sets an upper limit on logffir)
at 700'C of -14. Under these conditions, the CO/CO, ratio is
so low that the CO, pressurecan be accuratelyapproximated by
the total pressrue. Huebner (1969) found the decomposition
temperature of pure MnCO, to be insensitive to fo, variation.
No attempt to control/o, was justified for this study.
The equilibrium run conditions are reported in Table 1. The
pressurefor a given run is accurateto + 50 psi. The temperature
was maintained constant at 700(+2)'C. The composition obtained from X-ray data assumingVegard'slaw is reported at the
levelof precisionof the lattice-parameterrefinements(i.e., +0.001
mol fraction). The error bars on the coefrcients of the derived
activity-coefrcient equation were obtained using a Monte Carlo
method of propagation of errors from the uncertainties given
above (seeAnderson, 1976).
Synthesis and charactedzation of samples for
solution calorimehy
Samplesalong the calcite-rhodochrositejoin were initially precipitated from concentratedaqueoussolutions of the chlorides
by adding excessKrCOr. The precipitate was rinsed, dried, and
subsequentlyrecrystallizedunder CO, pressure(8 00G15 000 psi)
at 600"C in crimped Pt capsules.Not all samplesresulting from
Xuno.
6.3756(9)
6.3238(e)
6.2771(5)
6.1628(23)
6.1060(6)
s.9446(6)
5.9010(9)
46.077(1
0)
46.21
9(9)
46.359(4)
46.745(3)
46.e39(6)
47.545(6)
47.764(10)
0.0000(2)
0.1184(70)
0.2023(60)
0.4308(1
20)
60)
0.55s0(1
0.91s4(20)
1.000(2)
this procedure produced crystals large enough for analysis by
microprobe. Microprobe analysesfor sampleswith large enough
grain size are reported in Table 2 (basedon at least l0 analyses
from separate $ains) along with the refined lattice parameters
of samplesusedto constructa working curve. The lattice parametersfor all carbonatematerialsin this study were obtained from
refined X-ray powder patterns using an internal standard of Si
with Guinier cameratechniques.Severalprevious workers have
reported an essentiallylinear relation betweencomposition and
unit-cell length (Goldsmith and Graf, 1957; de Capitani and
Peters, 1981; Fubin and Stone, 1983). Our data confirm these
resultsand justifi' the Vegard'slaw assumptionfor compositional estimates.
Calorimetric techniques
After preliminary experimentswith severalaqueousand molten salt solvents at 80-350'C, an aqueous solvent saturated in
NaCl and containing 0.745 M HCI at room temperature was
adopted. The high-ionic-strength acid brine has a low vapor
pressureand a low CO, solubility and rapidly dissolvescarbonates reproducibly at 85'C. Heats of solution were obtained in
this solvent using a Tian-Calvet heat-flow calorimeter operating
at 85'C. An advantageof this solvent is the small magnitude of
the heats of solution. To permit unequivocal interpretation of
the enthalpiesofmixing obtained from the enthalpiesofsolution
in this chemically complex solvent, severalprecautionsweretak'
en. (1) The heat of mixing at a given composition along the
CaCOr-MnCO, join was found by comparing the measuredheat
of solution of the solid solution with the measuredheat of solution of a mechanicalmixture of endmember carbonatesat the
same composition as the solid solution. (2) The volume of solvent and the mass of all samples dissolved were held strictly
constant for all runs (20.0-mg sampleswere dissolved in 20 mL
of solvent). This procedure made it unnecessaryto correct for
the work done in CO, evolution or to unravel the complex chemical interactions within the solvent. The thermochemical cycle
used with this technique is therefore very simple. For a given
composition X,
(Ca,-rMn")COr,"n dissolved aqueouscomponents + CO2c*)
(l - X)CaCOr,", + (X)MnCOr,,a dissolved aqueouscomponents + CO2c*)
Since the right sides of both reactions are the same, even with
regardto the number ofmoles ofreactants and products and the
concentrationsof dissolved components,the diference between
the heats of solution gives the enthalpy of mixing at X.
Prior to dissolution, sampleswere encapsulatedin thin-walled
pyrex tubes to prevent prereaction of the sampleswith the corrosive solvent. After thermal equilibrium was attained in the
CAPOBIANCO AND NAVROTSKY: THERMODYNAMICS OF CaCO,-MnCO,
314
Trele 3. Activityfor (Ca,Mn)CO.solutionsat
700qCand902.6bars
Run
16
e
17
19
22
14
28
10
13
X""co"
o,
In yuo"
0.035
0.233
0.327
0.324
0.374
0.551
0.523
o.727
0.851
E
1.1682
0.8425
0.7207
0.7436
0.6683
0.4022
0.4204
0.2082
0.1045
.t.
9t
h
6
0.
0.00
0.20
0.t0
mol frution
M!@g
vs. compositionfor MnCO' comFig. l. Activity coefficient
calorimeter,runs were initiated by crushingthe pyrex sample
curve is shown,and experitube.The dissolutionwascompletedwithin 30 s, but the result- ponentat 700'C.A least-squares
phaseequilibrium.
points
are
from
decomposition
mental
ing calorimetricpeaklastedbetweenI and 2 h because
of the
long time constantof the large-volumecalorimeterused.The
total heateffectmeasured
wasaboutI cal.Furtherdetailsofthe
(1986).
experimental
methodarefoundin Capobianco
to the following equation, in which differencesin compressibilitiesand thermal expansivitiesof the solid phasExpnnrlrnNTAl RESuLTSAND
es have been neglected:
THERMODYNAMICANALYSIS
R?irn
a'"'o'::'i;::!'P.o.XP,
Carbonatedecompositionexperiments
The experimental data reported in Table I pertain to
the equilibrium three-phaseassemblagein the following
chemical equilibrium:
- Pr. (3)
Equation 3 was used to calculate values of ln "y."-, activity coefficientsreported in Table 3. Positive deviations
MnCO,(xl): MnO(xl) + CO,(fluid).
(l) from ideality are indicated acrossthe entire join as seen
in Figure l. The least-squaresregressioncurve for these
X-ray data given in Table I indicate that the MnO phase
data is also plotted and was obtained by considerationof
remains essentiallyunchangedfor these isothermal runs
the function g(-X),defined as
regardlessof the pressureof equilibration. Becauseof this,
- X*n or)'.
(4)
g(n : h 7r,-rl(l
information pertaining to the activity coefficient of MnCO,
is readily available from these P,T,XM,cq points. The
If dX) doesnot vary with composition, the solid solution
standardstate for Reaction I is defined as 700'C and the
is a regular solution. However, our data indicate more
pressureat which pure MnCO, is in equilibrium with its
complicated mixing behavior for the calcite-rhodochropure oxide constituents. This pressure is 902.6 bars as
site system. The mathematical form chosento represent
calculated from data given in Robie et al. (1984) for g(D
in the least-squaresregressionis a polynomial with
MnCO, and CO, fugacitiesof Holloway (1977) as 902.6
an appendedhyperbolic term:
bars.
The activity of the MnCO, component in the carbonate
(5)
is derived from these polybaric, isothermal experiments
d,n:|n,u - x"".o,),.
u =**.
by simplifuing the following generalexpression:
The,4 parameter,with no subscript, allows the fitted acf,,
coefficientcurve to be displacedalong the ordinate
tivity
Vrn o, dP + RToln aPfl,cos
* |
Irflpnco,
Jp"
axis (seeFig. l) and results from errors in standard thermodynamic data, in CO, fugacities, and in measuredP
and Z values. If one does not include this term and attempts, instead,to usea higher-degreepolynomial to representthe data, experimentally unwaranted curvature is
: pf.p"o
*
v*"o dp + plg,
introduced into the activity-composition relationship. The
f'"'
equation representedin Figure I is given by
.^',1:@P*
* [",'ff) ar.
(2)
lD ?vncor: (0.0597+ 0.041l)
+ (2.1968 + 0.0640X1- xr,"or)',
- (1.0773r 0.0241x1- xn""o,)'.
(6)
The PZintegrals rangefrom the standard state (Po,Zo)to
the experimental equilibration conditions (P,,[). Since The constant (l) on the right-hand side of Equation 5
the oxide and fluid are pure and there is no excessvolume correspondsto a free-energydiscrepancyof I 15 callmol
in the carbonate solid solution, Equation 2 is simplified between the calculated standard state and the 700'C ex-
315
CAPOBIANCO AND NAVROTSKY: THERMODYNAMICS OF CaCO"-MnCO"
0.t
0.7
d
06
U
'!
E
U
0.1
x
s
p
h
d
0 .I
.00
020
0.40
F
mol frrction
l!{nCOg
Fig. 2. Excessfree energyas a functionof compositionat
700"Cfor the calcite-rhodochrosite
binarv.
Fig.3. carorimetric.ffi;TTtn
for600'csampres
fit to the data.
and least-squares
periments.This value is within the errors associatedwith
the calculation of the equilibrium standard-state pressure. The mixing relationship is contained only within
the last two terms of Equation 6. The leading term is an
adjustment to the standard state and is not used below
in expressionsfor the activity coefrcient or excessfree
energy.The errors quoted are standard deviations ofthe
means for the coefficientsobtained by regressingEquation 5 fifty times using experimental data (P,I,X) points
that have been randomized within their known uncertainties. However, the Monte Carlo statisticsare not used
for the values of A,Ao,l,, which are basedon the regression of the nominal P,T,X data.
One can apply the Gibbs-Duhem relation to Equation
6 to compute the activity of the CaCO, component.
Equation 7 gives the result.
ln ?c,co,: (0.5809 + 0.0735XXr,.oJ2
+ (1.0773+ 0.024lXXr,.o,)r. (7)
These last two equations may be combined to give the
excessfree energyof mixing at 700"C using Equation 8
AG.-/RT: X,ln "tr(n + Xrln y(X).
side of the binary. A solvus is anticipated at lower temperatures,but for the 700'C experiments, the convexity
of the total free energy of mixing implies a continuous
solid solution consistentwith known phasebehavior.
Calorimetric experiments
The enthalpy of solution data for samplesprepared at
600"C are shown in Table 4. The enthalpies of solution
of mechanical mixtures do not vary linearly with composition, a fact suggestingcomplex chemical interaction
within the solvent. However, this complexity is of no
consequenceto the enthalpies of mixing obtained as the
differencebetween enthalpiesof solution for mechanical
mixtures and solid solutions. The enthalpy of mixing in
CaCOr-MnCO, is shown in Figure 3. Note the highly
asymmetric behavior with both positive and negativedeviations from ideality.
The least-squarescurye representedin Figure 3 was
obtained by regressionofa function \(-Y), defined as
I(X): aa^./(r- n(n.
(8)
(10)
This transformation is advantageousbecausethe resulting curve may be fit without constraining it to passthrough
any nonexperimentalpoints (i.e.,al X : 0 or X: I Afl- :
AG75I-/RT: (X)(l - W.r20 + 0.068 + (0.538
0, but tr(X) + 0). This enthalpy-interaction parameter is
+ 0.012)4,
(9)
plotted in Figure 4. Experimental values for the endwhere X is the mole fraction of MnCOr. A plot of AGpoints on this plot have been approximated from the acat 700"Cis shown in Figure 2. The values are everywhere tivity-coefficient data derived from the phase-equilibripositive with a maximum of 670 callmol on the Mn-rich um experiments. The excessfree energiesat 600'C for
One obtains for the calcite-rhodochrositesystem
TeeLe4. Calorimetricdata for (Ca, Mn)CO. solutions (kcal/mol)
Xunco"
0.1124
0.30s7
0.4508
ss
mm
2.281(941
2.442(70,
2.877(78)
2.987(31)
Enthalpies
of solution
4.666(61)
3.657(42)
3.175(49)
3.798(99)
0.161(117)
0.111(84)
Enthalpiesof mixing
-0.482(6s)
-0.868(117)
0.6278
0.8376
0.9154
5:t42(26)
4.670(87)
s.650(77)
s.469(121)
-0.472(90)
-0.182(144)
Note.' ss = solid solutions; mm = mechanicalmixtures; values in parentheses represent standard deviation in
last decimal place.
316
CAPOBIANCO AND NAVROTSKY: THERMODYNAMICS OF CaCO3-MnCO'
o
d
q
o0
e
b,
A
E
x
o
d
F1
mol
freuo!
Mn@g
Fig. 4. Enthalpyinteractionparametervs. compositionfor
calcite-rhodochrosite
binary. Asterisksare extrapolatedfrom
phase-equilibrium
data;squares
arecalorimetricdata.
Fig.5. Entharpy
Hill ,"-o'.,withreasr-
"rJ;:J;H
of mixing obrainedfrom the temsquarescurveand enthalpy
peraturedependence
functions.Note the
of excess-free-energy
similarshapeof the functions.
MnCO, in CaCO, at infinite dilution and for CaCO, in
MnCO, at infinite dilution (RI ln 7) are l94l and 2875
callmol, respectively.Thesevalues correspondto I(X) at
X : 0 and X : l, respectively,if at these extreme compositions there is no excessentropy. The other assumption in this approximation is that we may use the 700"C
data to represententhalpies at 600'C; this is not unreasonableif the first assumption is true. The auxiliary data,
constrainingthe endpoints, do not produce any curvature
inconsistentwith the calorimetric data plotted in Figure
4. However, the positive value for the MnCO, endpoint
causesvery slight endothermic behavior for AI1"' very
near pure MnCOr. The least-squarescurve shown in Figure 4 is a cubic polynomial (in mole fraction MnCOr)
and results in the enthalpy of mixing given by
AH-.: (4(1 - nl.493l
0 . 0 9 9+ ( 1 3 . 1 7+ 0 . 2 9 ) X
- ( 7 3 . 1 9+ 0 . 9 1 ) X ,+ ( 6 0 . 2 7+ 0 . 6 3 ) ) 3 1 . ( l l )
(Ca,Mn) ordering analogousto that in the ordered lowtemperature phase, kutnahorite. However, greater symmetry with respect to the 50:50 composition might be
expectedif theseresults were wholly attributable to kutnahorite formation. Also, the temperature of formation
of the solid solution (600"C) is well above that at which
kutnahorite disorders,and the X-ray patterns ofthe synthetic materials show no dolomite-type reflections. It is
possible that the asymmetry is due to the superposition
of more than one energetically distinct intracrystalline
process.
A partial separation of the energetic contributions to
the calorimetric enthalpy of mixing can be done using the
phase-equilibriumwork of Boynton (1971). He detercurve for
mined an isothermal (90"C) excess-free-energy
the calcite-rhodochrositesystem from aqueoussolution/
crystal partition coefficients for Ca and Mn. He forced
As in the activity-coefficient fit, the errors in Equation I I the solid phasesin his solution-precipitation experiments
are estimated by standard deviations of the means of to maintain homogeneouscomposition, even at this low
regressioncoefficientsfrom separatedata sets generated temperature,by an external cycling of CO, pressure.The
through Monte Carlo error propagation.
excessfree energy(expressedas a function ofXr".o,) derived from his experimentsis representedby
DrscussroN
The asymmetry of the enthalpy of mixing is extreme
for the 600'C samples.Isothermal addition of CaCO. to
pure MnCO, causesa rapid stabilization of the energyof
the phase,whereasadding MnCO. to pure CaCO, results
in an endothermic maximum. The composition corresponding to the Ca-rich maximum is not well constrained,but falls near Xr".o, : [. Perhapsthe phasecan
begin to stabilize in energyonly when Mn is in sufficient
concentrationthat the averageCa-O octahedronis linked
to at Ieast one Mn.
The central region of the binary indicates modest energy stabilization. Sincethe phase-equilibrium data show
positive excessfree energies,the exothermic heatsof mixing imply significant excessentropies of mixing. It is
tempting to relate the observednegative excessenthalpy
and inferred negative excessentropy to the influence of
AG*/RT: (n0 - nQ.074s+ 2.6288X
- 4.rr0rx).
(r2)
However, Equation 12 does not produce a stabilization
at or near the kutnahorite composition as expectedif his
data referred to a stable equilibrium involving the ordered phase.Boynton did demonstratethat his data representedequilibrium values, but it is probable that the
equilibrium was metastable involving only disordered
carbonate solid solutions. This is useful for the present
discussion becauseit permits a calculation of the temperature differential of the excessfree energy of mixing
in a continuous carbonate solid solution without the influenceofkutnahorite.
The excessfree energyofour study showedno evidence
for kutnahorite becauseit was obtained from samples
well above the stabilitv field of kutnahorite (700"c).
CAPOBIANCO AND NAVROTSKY: THERMODYNAMICS OF CaCO.-MnCO.
3t7
0.
0.
.
Eo.
U
b
F]
E
o.
-'
ko.
I
x
F]
-'
mol
fretion
MnCOg
Fig. 6. Excess-free-energy
curvesat 100-degintervalsfrom
90 to 590'C.Note that the excessfree energyincreases
with
temperature.
Calculations
referto phases
with no Ca-Mnorderrng.
Boynton's data (90'C) did not exhibit stabilization near
kutnahorite presumably becausekinetic factors inhibited
the ordering process.Since both data setsare isothermal
determinations, one can calculate a simple temperature
differential for AG".:
metastable
calcites
mol fraction
MnCO3
ExperimenFig. 7. Phasediagramfor calcite-rhodochrosite.
tal solvusgivenby de Capitaniand Peters(1981);metastable
solvuscalculatedfrom temperature-dependent
excess-free-energyfunctionsgivenin this study.
tially long-range ordered kutnahorite is 450"C (Goldsmith and Graf, 1957),one must conclude that disorder0 3 ) ing occurs over a rather large temperature interval.
The derivative is assumed independent of temperature Analogous relations among long-range and short-range
and is given by
order and their energeticeffectshave been discussedfor
Mg,TiOo (Wechslerand Navrotsky, 1984).
=
/RT)/d7
d(AG"
lQ-3(0.0739 3.427X
To distinguish the above possibilities, calorimetric
+ 6.738X).
(14)
samples quenched from other temperatures will be reThis approximation to the temperature dependenceof quired.The variation of the calorimetric excessenthalpy
the excessfree energy of mixing in the (assumedto be) with annealing could then be used to assesspossible recompletely disordered solid solution can be compared to ordering effects[seeReederand Wenk (1983) for the case
the calorimetric data by using the Gibbs-Helmholtz of CaMg(COr)rl and to estimate the total enthalpy reequatlon:
quired to disorder kutnahorite. In either case,kutnahorite
is stable by at least 0.63 kcal/mol in enthalpy relative to
d(LG.,/ Rn/ dT : - aH* / RT,.
(15)
a disorderedsolid solution, as indicated by the diference
One can usethis equation to calculatea metastableexcess in enthalpy of mixing derived from the calorimetric and
enthalpy at 600"C. This function along with the calori- from the phase-equilibrium data.
metrically determined excessenthalpy for samples presor,vus rN MnCO3-CaCO,
A lmusuBr.E
pared at 600"C are plotted in Figure 5. Notice that the
derived excessenthalpy curve exhibits similar asymmetThe effect of the suppressionof kutnahorite ordering
ric behavior to the calorimetric data. Apparently the on the phasebehavior is seenin a calculation of a metaasymmetry is unrelated to the kutnahorite ordering pro- stablesolvus using Equation 14.The derived temperature
cess.Thecalorimetric function shows larger exothermic dependenceof the excessfree energy of mixing may be
efects in the central region. The differencebetweenthese used to calculate a mixing curve at all temperaturesbefunctions is perhapsthe enthalpic effect ofpartial order- tween 90 and 700'C. A plot ofthe excessfree energyfor
ing at 600"C. Alternatively, the additional energy stabi- 100-deg intervals is shown in Figure 6. The calculated
lization may have been acquired during quench.
excessfree energiesincreasewith increasingtemperature.
The above observationsare consistentwith the follow- Negative excessentropiesare again inferred, but now for
ing ideas about the kutnahorite order-disorder process. mixing processesthat exclude ordering due to kutnahoAt 700'C (the temperatureof phase-equilibrium studies), rite. The calculatedsolvus is shown in Figure 7 along with
no significant long- or short-rangeorder exists. Samples the experimentally determined solvus of de Capitani and
quenched from 600'C contain some short-range order, Peters (1981). The metastablesolvus closesat 355t,
either presentat T or produced during quench. They are whereasthe experimental solvus extendsto about 550'C.
not long-rangeordered. If the short-range order persists Solid solubility is generallyincreasedin this systemwhen
at 600'C, then, since the onset ofdisordering for an ini- kutnahorite-forming processesare suppressed.Carbonate
(Aci6*c/RT - AGB6"./RZ)
/ (700 - 90) = 71o*' / RT")/dT.
318
CAPOBIANCO AND NAVROTSKY: THERMODYNAMICS OF CaCO3-MnCO,
phaseswith compositions near kutnahorite, or even beneath the stablesolvus, may therefore sometimesbe representative of metastable interphase equilibrium. It has
been suggested(Tsusue, 1967, among others) that a Carich solvus should exist in this system at temperatures
too low for solid-state phase-equilibrium experiments. The
present data are insufrcient to calculate the presumed
phase behavior (analogousto the Ca-Mg system) since
we lack necessarythermodynamic information on kutnahorite. However, the metastablesolvus offers an alternative explanation for natural metamorphic samplesthat
do not exhibit immiscibility on the Mn-rich side of the
join (see Essene,1983). One should also be aware that
naturally coexisting Ca-rich and kutnahorite-composition phasesdo not necessarilyconfirm the analogy with
the Ca-Mg system,unlessthe 50:50 phasecan be shown
to contain kutnahorite ordering.
CoNcr,usroN
We have found, by two independentmethods (calorimetry and phase-equilibrium studies),that the enthalpy of
mixing within the calcite-rhodochrosite binary is very
asymmetric. The more volatile carbonate component
(MnCOr) causesa slight energydestabilization of calcite,
and the refractory component (CaCOr) produces initial
exothermic mixing effectsin rhodochrosite. The crystalchemical interpretation of this behavior awaits further
structural characleization. We have inferred the presence of significant negative excessentropies from both
the calorimetric and phase-equilibriadata. An additional
exothermic enthalpy effect observed in the calorimetric
data has been attributed to short-rangekutnahorite order
that either persists considerably above the onset of disordering for kutnahorite or is acquired during quench.
Finally, we suggestthat some naturally occurring phases
of kutnahorite composition might be better termed
"pseudokutnahorite" since they may be formed without
any dolomite-type ordering.
AcxNowr,nocMENTS
This work was supported by the National Science Foundation, Solid
State Chemistry Program, Grants DMR 8106026 and DMR 8521562.
Boyle, E.A. (1983) Manganesecarbonate overgrowths on foraminiferal
tests.Geochimicaet CosmochimicaActa,47, l8l 5-1819'
Boynton, W.V. (1971) An investigationof the thermodynamicsof calciterhodochrositesolid solutions Ph.D. Thesis, Carnegie-MellonUniversity, Pittsburgh.
Capobianco,C. (1986) Thermodynamic relations among severalcarbonate solid solutions. Ph.D. Thesis, Arizona State University, Tempe,
Anzona.
de Capitani, C., and Peters,T. (1981) The solvus in the systemMnCOr
CaCOr. Contributions to Mineralogy and Petrology,76,394-400.
Essene,E.J. (1983) Solid solutions and solvi among metamorphic carbonateswith applicationsto geologicthermobarometry.Mineralogical
Societyof America Reviews in Mineralogy, 11,77-96.
Fubin, 8., and Stone,F.S. (1983) Physico-chemicalpropertiesofMnCOr
CaCO, and MnO-CaO solid solutions.Journal of the Chemical Society,
Faraday Transactions l, 79, 1215-1227.
Goldsmith, J.R., Graf, D.L. (1957) The system CaO-MnO-CO,: Solid
solution and decomposition relations. Geochimica et Cosmochimica
Acta.11.310-334.
Holloway, J.R. (1977) Fugacity and activity of molecular speciesin supercritical fluids. In D.G. Fraser,Ed Thermodynamics in geology,p.
, oston.
l 6 l - 1 8 1 . D . R e i d e lB
Huebner, J.S. (1969) Stability relations of rhodochrosite in the system
American Mineralogist, 54' 457-481.
manganese-carbon-oxygen.
Pendersen,T.F., and Price, N.B. (1982) The geochemistryofmanganese
carbonatein Panama basin sediments.Geochimica et Cosmochimica
Acta,46, 59-6E
Peters,T., Trommsdorfl V., and Sommerauer,J. (1978) Manganesepyroxenoids and carbonates:Critical phaserelations in metamorphic assemblagesfrom the Alps. Contributions to Mineralogy and Petrology,
66, 383-388.
Reeder,R.J., and Wenk, H.R (1983) Structute refinementsof some thermally disordereddolomites American Mineralogist, 68, 7 69-776.
Robie, R.A , Haselton,H.T., and Hemingway, B'S. (1984)Heat capacities
and entropiesof rhodochrosite(MnCOJ and siderite (FeCOr)between
5 and 600 K. American Mineralogist, 69, 349-357.
Tsusue, A. (1967) Magresian kutnahorite from Ryujiima mine, Japan.
American Mineralogist, 52, 1751-1763.
Wechsler,B., and Nawotsky, A. (1984) Thermodynamics and structural
chemistry of compounds in the system MgO-TiOr. Journal of Solid
StareChemistry,55, 165-180.
Wenk, H.R., and Maurizio, R. (1978) Kutnahorite, a rare Mn mineral
from Piz Cam (BergellAlps). SchweizerischeMineralogischeund PetrographischeMitteilungen, 58, 97-100.
Winter, G.A., Essene,E.J., and Peacor,D.R' (1981) Carbonateand pyroxenoids from the manganesedeposit near Bald Knob, North Carolina. American Mineralogist, 66, 278-298.
Zak, L., and Ponvondra,P. (198l) Kutnahorite from the Chvaleticepyrite
and manganesedeposit, east Bohemia. Tschermaks Mineralogische und
PbtrographischeMitteilungen, 28, 55-63'
RnrnnnNcns
Anderson,G.M. (1976) Error propagationby the Monte Carlo method in
geochemicalcalculations.Geochimicaet CosmochimicaActa, 40, 1533l 538.
Mer.n-rscnrp'r REcEll'ED Ml.v 22, 1986
MnNuscrrrr
AccEprED Novel,rnsn 20. 1986