American Mineralogist, Volume 58, pages 323-331, 1973

Equilibriain Systems
Gypsum-Anhydrite
GaSO*-H,Oand Ga@n-NaCl-H,O
CnlnrBs W. BrouNr
Department ol Geology, Uniaersity of Georgia, Athens, Georgia 30601

AND

FnaNx W. DrcrsoN
Department of Geology, Stanford Uniuersity, Stanford, California 94305

Abstract

The equilibria of gypsum and anhydrite with solution and vapor were studied by the use of
solubilities, which were in part gained from downward extrapolation of solubilities measured at
high temperatures to avoid kinetic difficulties encountered below 70"C. The 4-phase invariant
equilibrium of gypsum, anhydrite, simple H,O solution, and vapor is set at 56"C + 3"C and
124 -r 9 torr. The pressure change required to increase the univariant equilibrium temperature
of gypsum, anhydrite and tLO solution l"C is about 78 bars, which agreeswell with values
calculated from thermochemistry. The 3-phase univariant equilibrium of gypsum, anhydrite,
and vapor rangesfrom 124 torr at 56"C to 13.6 torr at 20'C. The effect of NaCl concentration
on the 4-phase univariant equilibrium of gypsum, anhydrite, NaCl-H,O solution, and vapor,
is to lower the equilibrium temperature from that of the invariant point, 56"C for zero NaCl
concentration.to: 48"C. 2 molal NaCl: 36'C, 4 molal NaCl; and 20'C, 6 molal NaCl' The
uncertainty in the temperatures in NaCl solutions is -t- 4"C.

Introduction (L* represents NaCl-HzO solution),wasdetermined

function of NaCl concentration.
Gypsum (CaSO+'2H2O)and anhydrite(CaSO+) as a
occur as minerals in diverse environments.Know- PreviousStudies
ledge of the temperature and pressure conditions
G + A + L -tV Equilibrium in sYstem
under which the two minerals coexist stably with
CaSOrHeO. The four-phaseequilibrium G + A +
aqueoussolutionsis essentialto understandingthe
V has received much attention from previous
genesisof specific mineral associations.This paper L +
most of whom used methods based on
presentsresults of studieson equilibrium relations workers,
solubility measurements. Controversyexistsas to the
of gypsumand anhydritein the presenceof saturated
exactconditionsprevailingat the point. Zen (1965)
aqueoussolutionsand vapor.
and Hardie (1967) include summary statementsof
A schematicrepresentationof phaserelationships
the topic, useful to readerswho wish to pursue the
in the systemCaSOr-HzOis presentedin Figure 1.
subject. The temperaturesreported to hold at the
The phasesstable in the temperature and pressure
invariant point range from 63.5oC (Hofr. et al.,
rangesof this study,15 to 70'C and 0.01 to 1000 (1903) to a value of 38'C (Toriumi and Hara,
bars, respectively,are designatedby capital letters, 1 e 3 8 .
)
as follows: G, gypsum;A, anhydrite;L, saturated Most workers have acceptedPosnjak's (1938)
solution; and V, vapor. Figure 1 shows that four solubility basedvalue of 42"C. More recent solu-
curvesrepresentingunivariant three-phaseequilibria bility studiesby D'Ans et al. (1955), Bock (1961)
intersect at a point representingthe invariant four- Marshall and Slusher (1966) and Power et al.
phaseequilibrium G + A + L + V. The conditions (1966) also supporta transitiontemperatureof 40
at the invariant point and alongthe univariant curves to 42"C. The 4O'C value calculatedfrom thermo-
of G * A + L and G + A + Vhave been deter- chemical data by Kelley et al. (1941) must be
mined as part of this study. regarded as fortuitous. Zen (1965) and Hardie
In addition, the effect of dissolvedNaCl on the (1967) in reviewing the thermochemical data
univariant four-phaseequilibrium G + A + L* + V showedthat therewas a largeuncertainty(-+22"C)
323
t24 C. W. BLOUNT AND F. W. DICKSON
SCHEMATIC
PHASERELATIONS
IN THE
SYSTEMCoS0o-HrO
T+
Ftc. 1. Schematicphaserelationsin the systemCaSO-HgO
projected onto the temperature-pressure plane. Four three-
phase univariant equilibria are represented by curves that
meet at the four-phase invariat point, 1.P. Phase designa-
tions are: Gypsum, G; anhydrite, A; H,o solution, L;
and vapor, V.
inherentin the calculations.Zen (1965) also called
attention to the significanceof the sluggishreaction
of anhydrite with simple aqueous and NaCl-HpO
solutions,which below about 70'C is so slow that
no experimenterhas successfullyreversedsolubility
determinations.Although somedifficultieshave been
met in reactinggypsumwith solutions(Hulett, 1901;
Hara et al., 1934), the determinationshave been
reversed in adequatedetail over a wide range of
temperature.Determinationsof the G * A + L + V
invariant point from plots of the apparentsolubility
of anhydrite and the true solubility of g5rpsumyield
only estimatesof minimum temperatures.
Hardie (1967) put forth a temperatureof 58oC,
based on a novel method. which revived interest
in the matter. D'Ans (1968) objectedto Hardie's
determination,regardingit as too high, and re-stated
his preferencefor Posnjak'stemperatureof 42oC.
The major points raised by D'Ans revolve about
kinetic considerations,such as the possibilitythat
metastablephases had appeared unheeded during
Hardie'sexperiments. D'Ans raisedinterestingques-
tions, but his death has prevented clarification of
his arguments.Cruft and Chao (1970) state their
preferencefor the 4042'C transition temperature.
They failed, however, to present any evidence to
support this preference.Hardie (1967) determined
the equliibrium relations of gypsum and anhydrite
by a method that compensatedfor the slow reaction
of anhydritewith solution.He reactedmixtures of
gypsumand anhydrite with solutionsof Na2SOaand
H2SO4of various concentrationat severaldifferent
temperatures.By noting which phase, gypsum or
anhydrite,tended to increaseduring long reaction
times,he was able to delineatethe stabilityfieldsof
G + L and A * L. Extrapolatinghis data to zero
dissolvedsubstanceconcentrationpermittedhim to
estimatethe temperatureat the fotrr-phasepoint to
be 58' ! 2"C. Hardie's work was done carefully;
he examinedthe solid phasesat regular intervals
to make sureonly gypsumand anhydritewere pres-
ent, and he successfullyreversedthe reactions.
G + A + L* + V equilibria in system CaSOa-
NaCl-HrO. Dissolvedsalts decreasethe tempera-
ture at which G + A + L* + V coexist.Figure 3
presentsa summaryof previotrssignificantstudies.
Seawaterbrinesat variousconcentrations wereused
in the study by Toriumi and Hara (1938). Their
solubility-basedresultscomparevery well with those
obtainedby D'Ans (1955) in NaCl solutionson a
moles chlorine per kilogram of water basis.Posnjak
(1940) reportedan equilibriumtemperatureof 30oC
in seawatersolutions4.8 timesnormalsalinitywhich
is equivalentto a 3 moles NaCl/kg HzO (3 molal
NaCl) solution.This is about 4oC higher than the
resultsof Toriumi and Hara or D'Ans. MacDonald
(1953) calculatedthe changein equilibrium tem-
perature with NaCl concentration using thermo-
chemicaldata. Bock (1961), Power et al. (1966)
and Marshall and Slusher (1966) evaluatedequi-
librium temperaturesfrom solubility data in NaCl
solutionswith results that agreedwith Posnjak and
MacDonald.Zen's (1965) study cast doubt on the
anhydrite solubilitiesbelow 70'C, and suggestedthat
the equilibrium temperature must be higher than
the solubility studiesindicate.
Hardie (1967) evaluatedthe effect of solutions
of different ionic strengths (activities of H2O) by
the method describedearlier. Hardie's results are
illustrated in Figure 3 in which his HzO activities
are related to NaCl concentrationusing vapor pres-
sure data from the International Critical Tables
(1929). Hardie'stemperaturesare distinctlyhigher
than other workers over most of the concentration
range. While anhydrite-solution reactions may be
too sluggishfor equilibrium to be obtained in HzO
or NaCl-H2O solutions, reactions in NazSOn and
H2Oa solutions do proceed rapidly enough.
G + A -f L equilibrium in system CaSO',HyO.
Manikhin and Kryukov (1968) measuredgypsum
 GY PSUM-AN HY DRITE EQU ILIB RIA 325

S Y S T E M C o S O 4 -H z O
0 Gypsum+ Anhydrite
+ SoturotedSolution
0 GyPsum
GT)SUM
\

A Bock(1961)
E Zen( 1965)
V BlockI Woters(1968)
o(\I 0 Monik|inI l&yukou(t968)
- o BlountI Dickson(Present)
c'l I bor
Y

E
o
.ol \,
t:,"
,%
o
a
sl ?"
o
(J
\
Anhydrite
Melcher (l9lO) "oh
Hollet.ol.(1926)
Fortridge I White(1929)
Hiilil937)
Posnjok ( 1938)
B o o t hA B i d w e l(l1 9 5 0 )
PowerI Fobuss(1964)
0 Monikhin I Kryukou (1968)
-- l t.Dickson ef.ol.(1963)
B l o u n It D i c k s o(n1 9 6 9 )

50 roo r50
n oC
T E M P E R A T U Ri E
I bar
Frc. 2. The logarithm of gypsum and anhydrite solubility in molality in systemCaSOa-HzOplotted versustemperatureat
Table 2;
or the vapor pressureof the solution, 500 bars, and 1000bars. Solid lines representsolubilities calculatedfrom equationsof
of the
dashedlines representextrapolatedportions of the calculatedanhydrite solubility curves.The line connectingthe intersections
.uru., i, the projection onto the composition-temperature plane of curve separating G * L from A * L in Figure 1'
isobaric

and anhydrite solubilities at high pressuresfrom obtained a value of 7l t I.9 batsf"C using the
0 to 50'C. Their results are shown in Figure 2. thermochemical data of Kelley et aL (1941'), 46"C
The pressureeffect on'the equilibrium has been as the equilibration temperature,and more recent
calculatedto be 83 bars/'C (Marsal, 1952) and data on the molar volumes of gypsumand anhydrite'
85.4 bars/'C (MacDonald, 1953). Zen (1965) G + A -f V equilibrium in systemCaSOa-HrO'
 326 C. W. BLOUNT AND F. W. DICKSON
^ 9 Y 9 T E MC o S O 4 -N o c t - H 2 o
Gypsum-
Anhydrife-Soturoted Solutio;Eouitibrio
I BAR
-y Hordiefl967)
-O- BlountI Dickson
(Present)
o40 \i
q
UJ
E \
f30
F solubilitiesin the kineticallyhinderedregion below
(r + Toflumie Horo(fgiJ-
UJ
=.v
UJ
F Anhydrite and gypsum solubilities. The solubil-
l0 ---
MorshollI Slusher(t966)
+ Foweret ol (1966)
6 in H2O solutionsfrom 100oto 275'C and 1 to 1000
rr,lotes
ru3oct
o, tr, *n ,lO
Frc. 3. The effect of NaCl or Cl concentration and
activity of H,O on the equilibrium temperature of the
gypsum, anhydrite, solution, and vapor assemblage,at I bar
total pressure in the system CaSO.-NaCI-H"O. H,O vapor
pressure ranges from 124 torr at 56"C and zero NaCl con-
centration, to 13.6 torr at 20.C and 6 molal NaCl.
The equilibrium of gypsum and anhydrite with vapor
cannot be measured directly because of the excessive
times required to equilibrate the solids with the
vapor phase.The positionof the curve as evaluated
from Hardie's temperatureversusionic strengthdata,
combined with data on the vapor pressure of
NaCl-HzO solutions as a function of temperature
and ionic strength,is presentedin Figure 7.
Summary ol preoious work. Controversy exists
as to the temperatureat the 4-phaseinvariant point,
as well as to the quantitative effect on lowering the
temperaturecreatedby dissolvedsalts.Hardie'swork
contrastswith the studiesof most previousworkers,
who have reported lower temperatures.The efiect
of pressureon the univariant equilibrium of gypsum,
anhydrite and saturatedsimple HzO solutionshas
not been established experimentally,but it has been
calculatedfrom thermochemical data. The tempera-
tures and H2O pressuresof the anhydrite,gypsum,
and vapor equilibrium have been estimated from
Hardie's (1967) data. Some additional work is
clearly neededto clarify gypsum-anhydriterelations.
PresentWork
We haveusedthe solubilitymethodof determining
equilibrium conditions of phase assemblages. The
method is simple in principle; it consistsof experi-
mentallydeterminingthe temperatures and pressures
at whch gypsumand anhydriteshow the samesolu-
bilitiesin a particularsolutiontype.We usedsolubil-
ity data from the literature and unpublishedsolu-
bilities of our own, measuredfor this purpose.The
solubilitieswere castinto empiricalequations,which
relatedgypsumand anhydritesolubilitiesto tempera-
ture, pressure,and NaCl concentration.Anhydrite
70oC were estimatedby downward extrapolating
higher temperaturedata.
ities of anhydrite and gypsum were measuredby
the use of the hydrothermalsolution equipment,
previouslyusedto measurethe solubilityof anhydrite
bars (Dicksonet al., 1963) and in NaCl-HzO solu-
tionsfrom 100to 450'C and I to 1000bars (Blount
and Dickson, 1969). One new determination of
anhydritesolubility was made at 80oC-1000 bars;
efforts to measurethe solubility at lower tempera-
tures failed. Gypsum solubilities were determined
from 40 to 83'C at 1, 500, and 1000 barsin simple
HzO solutions(Table 1).
The procedures followed in the anhydrite and
gypsumsolubilitydeterminations werethoseof Dick-
son el al. (7963). Well-crystallizedartificial anhy-
drite and gypsum, and carefully purified H2O, free
of COz, were used.Gypsumequilibratedwith solu-
tion within one day, in contrastto anhydrite,which
required three days to attain equilibrium at 8O'C.
Selectedsolubilitieswere reversiblydetermined.The
maximum estimateduncertainty in the values of
Table 1 is +2.5 percent of the solubility value,
i2"C and i4 bars.
Gypsum and anhydrite solubilities are illustrated
in Figure 2. There is good agreementbetween the
1 bar gypsum solubilitiesobtained during this study
and those obtained by Hullet and Allen (1902),
Mglcher(1910),Posnjak(1938),Denman(1961),
Marshall and Slusher (1966). The I bar gypsum
solubilitiesobtainedby Bock (1961), Zen (7965),
Block and Waters (1968) are significantlyhigher.
The reasons for this disagreementare unknown.
However,Hullet (1901), Hullet and Allen (1902),
Tanakaet al. (1931) and Hara et al. (1934) dis-
cuss high apparent solubilities for gypsum when it
is either poorly crystalline or very fine grained.
Hullet (1901) pointed out that gypsum was very
soft and easilyground to a fine powder when stirred.
Becauseof the possibility that the high gypsum
 GY PSUM-AN HY D RIT E EQU I LI B RI A 327

Tlsr-B l The Solubility of Gypsum in H,O Solution as a Tesrp 2. Solubilities of Gypsum and Anhydrite ln
Function of Temperature and Pressure NaCl-H"O Solutions at Constant NaCl Concentration ln
Molality
Pre6 sure Tenperature cypsu Solubllity
ln 1n 1n hpirlcal Equatlons*

Bars oc MoIs CaS04/Kg of H20

(! 4 bars) e20c) @ .57.)
2 0.0 h n=-4.355 + 0.840 x lo-J P+ 0 OIO5t-L 700 x IO--t'

+ 0.584 x to-6t3
24 50 . 0 1 5 5u
+IO 50 .0228 u Anhydrlte 0 0 h n = - 2 . 8 7 + 1 . 2 2 0 x l 0 - 3 p - O o 2 3 7 t - O 0 0 2 8x 1 O - 3 P t
+)l 50 .0226 u
+ o 0lo x to-4.2 * o.tto * to-7Pt2.
995 50 . 0 3 3 7u
985 50 . 0 3 3 9u 2.O = -1,979 - O O2I5t + 0.5955* to-4c2 '0.0E95 x I0-6t3.
986 50 .0348u = - 2 . 4 4 7 - 0 . 0 1 4 3 5 t + 0 . 3 5 1 5 x l 0 - 4 t 2 - 0 0 3 o e zx L 0 - 6 t 3 '
4.O
468 60 . 0 2 1 8s
498 6 0 .5 .0223 s 60 = -2.926 - o.olo43t + o.:+:5 x lo'4t2 - 0.o367 x 10-6t3
24 6I .0L47 s * D - solubtltty tn @Is CsSO/kg of H2O.
140 79 . 0 1 5 2- p - pressure ln bare.
It. ) . 0 1 3 9e ! - tflPeretule it'c.
3 79 . 0 1 3 8s
7 9. 5 . 0 2 0 3u
79 .0200u
990 7 7. 5 .030 u numberof other workers (Melcher, 1910; Partridge
r0l0 . 0 3 0 1u and White, 1.929;Hill, 1937; Posnjak,1938; Mar-
565 4l .0264 u
s48 40 .0260u shall er ol. 1964).
902
902
67
68
.030 u
.031 u
High results such as those by HalI et ql. (1'926)
904 67 .030 u may representsolubilitiesof poorly crystallineanhy-
892 . 0 3 0 7u
drite, or somehemihydratemay have been present.
OT
515 67 . 0 2 2 8s
500
502
67
o/
. 0 2 2 4s
.0228 I
Resultssuch as thoseby Booth and Bidwell (1950)
992 6 6. 5 .033 may representfailure to obtain saturation or other
I002 5 6. 5 .034 u
999 . 0 3 3 5u experimentalproblems.Table 3 displaysgypsumand
998
100r
83
83.5
. 0 3 0 9s
. 0 3 1 4s
anhydrite solubilities calculated for regular inter-
1002 83 . 0 3 1 3s vals of temperatureand Pressure.
1004 83 . 0 3 0 7I
Resultson the G + A + L + V equilibrium in
u - solubillty
s = solubllity
approached fron undersaturation.
approached froE suPersaturatlon.
system CaSOt-HrO. The curves representinggyp-
sum and anhydrite solubilities as a function of tem-
perature at 1 bar total pressure(Fig. 2) intersectat
solubilities are anomalous,they were not used in 56'C. Consideringthat the calculatedsolubilitiesfor
our calculationof an empirical equation for gypsum anhydriteand gypsumare within 15 percent and tz
solubility as a function of temperatureand pressure, percent of the equilibrium value respectively, the
Table 2. The closenessof the fit of solubilitiescalcu- uncertaintyin the transitiontemperatureis i3oC.
lated from this equation is evident in Figure 2. The This value is within the experimental error of
values calculated from the empirical equation are -r
Hardie'svalue,58o zoc, and it is in closeragree-
within L2 Wrcent of most of the experimentaldata ment with van't Hoff et al.'s value of 63.5'C than
used to derive the equation. with any other number reported earlier. The vapor
There is considerablescatter in anhydrite solu- pressureat the invariant point, neglectingthe negli-
bilities obtained by difierent workers especially at
higher temperatures.The data of Dickson el a/. Tlnre 3. Solubilities of Gypsum and Anhydrite in HaO
(1963) and Blount and Dickson (1969) were car- Solution at Regular Intervals of Temperature and Pressure*
ried out reversibly and were repeatable.This data
was processedto obtain the empirical equationgiven Pressure Phsse Solublllty In Molality
Tenperature in
oc
in bara
in Table 2 from which solubilities were calculated
40 50 60 70 80
to give the graph lines in Figure 2' Logarithms of
I Gypsun .0154 .0153 .0148 .0143 ,0137
anhydrite solubility show almost a linear change 1 Anhydrite .0175 .0138 ,0108 .0086
with temperaturefrom 80oC to 260"C. In general .0218 .0205
5OO Gypsun .0240 .0233 .0226
most experimentalpoints are within f5 percent of 500 Anhydrite .0238 .0187 .0L47
the value predicted by this equation. Above 70oC 1000 Gypsu '0358 .0352 .0344 .0330 .0310
solubilitiespredictedby this equationare quite close 1000 Anhydrite .0415 .0324 .0253
to many of the experimental determinationsby a * calculated fron eDplrical equations of Table 2.-
 328 C. W. BLOUNT AND F. W. DICKSON

gibly small effect that the dissolved CaSOa exerts I

on the vapor pressureof the solution,is 124 f 9 torr

.@ N o C l 2 M o l e s/ K g H . 0
.08 GYPsum
(torr = mm of Hg). .o7
I M o d e i nq n d S r o t . . ( t 9 5 6 )
Hardie and we used different methods to get - .06 l B o c k{ 1 9 6 1 )
I Zcn (1965)
around the difficulty presentedby the slow reaction :<CD o5 | . Morlhollond Slush.r (t9661
rate o,f anhydrite at low temperatures.The con- 04 I a Blocl ond l{olrrr (t968)

cordance in the equilibrium temperaturesobtained o

o o3
by two different approachesadds strong support for
the general correctnessof a temperaturein the 50 v D ' A n ! G ro l ( 1 9 5 5 )
o M o d g i no n d S r o l 3 . { 1 9 5 6 }
to 60'C range, rather than in the 38 to 46"C nnge aoo A Soct (1961)
o + M o r t h o l le l o l { 1 9 6 4 )
of early workers. o
O Blounl ond oictton (1969)
Results on G + A + Lx -t V equilibria in e Elock ond Wotlrr (t968)

system CaSOa-NaCl-HrO. The equilibrium tem-

peratures were determined as a function of NaCl ?5 50 75 t00 t25 t50 t75
concentration at 1 bar pressure by essentiallythe toc
same procedure used in the system CaSO+-HBO. Ftc. 4. The effect of temperature on the solubilities of
The anhydrite solubility data in NaCl-HzO solu- gypsum and anhydrite in 2 molal NaCl solution at I bar
tions of Blount and Dickson (1969) and the 70oC total pressure.
data of D'Ans et aL (1955) were processedto
derive empirical equations that relate anhydrite Glew and Hames (1970) are noticeablyhigherthan
solubility to temperature at several NaCl concen- those used to derive the equations. If anhydrite
trations,at 1 bar (Table 2). solubility were higher than the curves suggest,the
Since many anhydrite solubility determinations gypsum-anhydritetransition temperaturewould also
were at conditions sornewhatdifferent from those be higher. The higher solubilities may indicate the
used for deriving these equations,it was necessary presenceof hemihydrate or ,a fine grained poorly
to apply small correctionsto the experimentaldata crystalline anhydrite in those experiments.
by graphical interpolation and extrapolation. Since Gypsum solubilities in NaCl-HzO solutions show
the solubility of anhydrite has been well measured considerable variation. Since also many determi-
over a range of NaCl, temperature and pressure nations are at conditions somewhatdifferent than
conditions, this procedure did not introduce signif- those desired for comparisonpurposes,it was nec-
icant uncertainties into the solubility values. The essaryto graphicallyplot this data to obtain interpo
empirical equations were derived from graphical lated values.In this analysis,considerableattention
plots of the function a,m/ndtT in which Az is the
change in molal concentrationfor the temperature
interval AT and m' is the mean molal concentration NoCl4 Moles/ Kg H.0
in the interval. The curves obtained were closely Gypsum
approximated by a quadratic equation which was o I l4ogdinond Srotr6 (t956)
l g o c k( 1 9 6 1 )
integrated using several data points to obtain the - I Zln {t965)
. M o . l h o l lo n d S l u s h a r( 1 9 6 6 )
final equation. The equationsreproducedthe solu- Yo I Bloct ond tyot.6il96g)
X Gtcw ond gome (t97O)
bility values used in their derivation within i2
percent over a temperature range from 70o to oo
250"C. Solubilities based on these equations are = A n h y dirf e
v D'Ar! rl ol {1955)
illustrated by curves in Figures 4, 5, and 6. a Boct (1961)

Solubilities reported for anhydrite in the kinet- oo

a + M o r l h o l l! t o l ( 1 9 6 4
O Elounl ond Dickton (1969)
o
ically hindered region (temperaturesbelow 70"C) o - Blocl ond Wot.r3 (t968)
Q Glew ond liom.!(t97O)
are significantly lower than values calculated from
the empirical equation. The inclusion of this data
would have produced a sharp inflection in the curves
t oc
of the function a,m/rrd aT. Anhydrite solubilities
Frc. 5. The effect of temperature on the solubilities of
above 100oCreportedby Zdanovsky(1949),Mar- gypsum and anhydrite in 4 molal NaCl solution at I bar
shall el al. (L964), Block and Waters(1968) and total pressure.

.l
.09
NoCl Moles/ Kg H.o
o8 Gypsum
o7
oN
Bocr (1961)
W
T
Z!n {1965)
o n6
Morrholl ond Slush!.
:z
.o4
o
03
o
= prer.nt resultsat 500 and 1000 bars. Their anhy-
A n h y di rf e
v D'An! et ql (1955)
os
@
(n A Eock {1961)
o + M o r r h o l l€ l o l ( 1 9 6 4 )
(J
O Blounl ond Dicklon {19691
25 75
t oc
Frc. 6. The effect of temperature on the solubilities of
gypsum and anhydrite in 6 molal NaCl solution at I bar
total pressure.
was paid to what information was availableregard-
ing experimentalprocedures,especiallyconsidering
the tendency of gypsum to undergo grinding under
many experimental situations as discussed pre-
viously. Somevaluesin the literature may be anom-
alouslyhigh due to the presenceof poorly crystalline
or very finely pulverized material. In this effort
we attemptedto obtain an internally consistentset
of data and reject valueswhich appearedeither low
or high. The gypsum solubility curves illustrated in
Figures 4, 5, and 6 are in fair agreementwith the
resultsby MadginandSwales(1956), Bock (1961)
Block and Waters (1968), and someof the results
of Marshalland Slusher(1966). The resultsby Zen
(1965) and by Marshall and Slusher(1966) (in
6 molal NaCl) are significantlyhigher. The uncer-
tainty regarding gypsum solubilities are probably
around i3 percent.
The temperaturesevaluated by the above pro-
cedurewere:48o -+ 4oC,2molalNaCl; 36o +- 4"C,
4 molal NaCl; and 20" ! 4"C, 6 molal NaCl. The
uncertainty in each temperaturewas estimatedfol-
lowing the procedure used for that in H2O only.
Figure 3 showsall data plotted astemperatureversus
NaCl concentration. Our curve agrees well with
Hardie's, but it falls above curves of the other
workers, except at high NaCl concentration,where
all curves converge. The agreement of our and
Hardie's work supports the validity of a higher
temperature position for the correct curve.
Results on G * A + L equilibrium in system
CaSO;-H2O. The 3-phaseunivariant equilibrium of
GY PSUM-A N HY DRITE EQU ILIBRI A
gypsum, anhydrite and saturatedI{rO solution was
ivaluated from solubilitiesmeasuredas functions of
(r956)
temperatureand pressure.Figure 2 showssolubility
versustemperatureplots for 1, 500, and 1000 bars
for gypsum and anhydrite. Solubilities reported by
.
11966l
Ir Manikhin and Kryukov (1968) for gypsum agree
+l
tl with previous results at 1 bar but are higher than
1i
drite solubilities are scatteredabove and below the
calculatedsolubility curves shown in Figure 2. It is
doubtful that they achievedequilibrium during their
anhydrite solubility measurements.Their gypsum
solubilities also seem to be anomalous' The equi-
125 t75
librium temperature obtained from the curves in
-F 3oC
Figure Z ari 56 +- 3oC, 62 t 3'C and 69
for 1, 500, and 1000 b'arsrespectively.The uncer-
tainty in the slope,dp/dt, is probably not in excess
o,f -+1oC for 100O bars. Our value of 78 t 7
bars/oC is not far from valuescalculatedby Marsal
(1952) and MacDonald (1953), which were 83
bars/oC and 85.4 bars/oC respectively. 7nn (1'965)
recalculated the pressure coefficient using more
recentmolar volume data and obtained 70.9 '+ l'9
barsfoC which agreeswith our recalculationsusing
Hardie's more recent revised thermochemicalequa-
tions (Hardie, 1967) and Clark's (1966) molar
volume data. In view of the tendencyfor quantities
calculated from thermochemical considerationsto
show relatively large uncertainties,the agreementof
the thermochemicallycalculatedpressurecoefficients
with the experimentalvalue is excellent.
Resultson G * A + V equilibrium in system
CaSOrHeO. The temperaturesand HzO pressures
at which gypsum and anhydrite coexist in the pres-
enceof vapor cannot be determineddirectly because
of kinetic hindrancesto attainment of equilibrium'
They can be estimatedfrom the temperatureversus
NaCl concentration data presented in Figure 3,
togetherwith data on vapor pressuresof NaCl-HzO
solutions of different NaCl concentrationsand tem-
peratures(I nternationnlCritical Tables,1933) . Two
effectshave been neglected:the tiny lowering of the
vapor pressureof the solutions resulting from dis-
soived CaSOe;and the small positive effect of the
confining pressureof 1 bar on the vapor pressure
of the solutions.
The approachis basedon the well-known princi-
ple of physical chemistrythat a multi-phasesystem
in equilibrium at a given temperatureand pressure
remains at equilibrium if one of the phasesis re-
moved. other circumstancesbeing held constant'
 330 C. W. BLOUNT AND F. W. DICKSON
Gypsum and anhydrite equilibratedwith NaCl-HzO
solution and vapor remain in equilibrium with the
vapor phase even if the NaCl-HzO solution is re-
moved. The vapor pressureof the series of NaCl-
H2O solutions that are in equilibrium with gypsum
and anhydrite at different temperatures can be
closely approximated from known data. Figure 7
presentsa plot of H2O pressureversustemperature
data for the reaction.
Hardie (1967) did not presenthis data in this
form; however,his H2O activity versustemperature
data on the equilibriumof gypsum,anhydrite,solu-
tion and vapor, can be converted conveniently to
vapor pressureversus temperaturedata. His data,
when plotted on a diagram such as Figure 7, yield
a curve indistinguishablefrom ours, in view of the
uncertainties.
The univariant curve representingthe G * A + V
equilibrium falls below the vapor pressurecurve for
H2O liquid and vapor at all temperaturesbelow
56'C. At 56"C, the G + A * V curve terminates
at the invariant point for G * A + L * V. At tem-
SYSTEM CoSOo-H20
- onhydrite
Gypsum - Vopor
Equilibrium
O BLOUNT ondDICKSON
o uanore's(1967)H2o
ociiviliesconverfed
(ts to trro
E for the solubility method to work, equilibrium solu-
o
F bility values must be available.Reversiblereactions
c of the solid and liquid phasesare essentialto dem-
lrJ
E
a
(t)
tu
E.
o-
0r0203040s060
T E M P E R A T U R Ei n O C
Flc. 7. The equilibrium of gypsum, anhydrite, and vapor,
at I bar total pressure, represented by a plot of vapor
pressure versus temperature. The upper terminus of the
curve is at the four-phase invariant point G + A + L + V,
located at the uppermost open circle according to Blount
and Dickson and at the uppermost solid circle according
to Hardie (1967). (torr = mm Hg).
peraturesabove56oC,the G + A + V equilibrium
can only occur metastably.
Conclusions
The temperatures of the equilibriaof gypsumand
anhydrite with solution and vapor, evaluatedby
most previousworkers from solubilities,appear to
be too low. The major causeof difficultyis the slow-
ness with which anhydrite reacts with simple H2O
or NaCl-H2Osolutionsbelow 70"C. Hardie (1967)
took advantageof the more rapid reactionof anhy-
drite with H2SO4 solutions to evaluateindirectly
higher equilibrium temperatures.By application of
a still different appro,achbasedon downwardextrap-
olation of high temperaturesolubilities to lower
temperatures, we haveevaluatedequilibriumtemper-
atureswhich are in generalagreementwith Hardie's.
We feel, therefore, that our results add strong sup-
port to the correctnessof the higher temperatures
reported by Hardie.
Our direct measurementsof the solubilities of
gypsum and anhydrite in H2O solutions enabledus
to evaluatethe pressureeffect on the G + A + L
equilibrium, which agreed well with the pressure
effect calculated frorn thermochemicaldata.
The experiencesof workers studyingthe gypsum-
anhydriteequilibria by use of the solubility approach
illustrates the strengths and weaknessesof the
method.The solubilityapproachavoidsmany of the
kinetic difficulties other methods encounter, such
as the appearanceof metastablephases.However,
onstratingthat equilibrium has been attained.The
history of the studieson gypsum and anhydrite sta-
bility relationswell illustratesthe hazardsof accept-
ing steady state or slowly changing concentration
values as equilibrium solubilities.
Acknowledgments
The authors are indebted to Professor Konrad Kraus-
kopf, Geology Department, Stanford University, who
kindly reviewed the manuscript. Mr. Warren Ervine and
Mr. Francisco Querol-Sufie, Geology Department, Stanford
University, made valuable comments. The work was sup-
ported by the National Science Foundation Grants G7405
and P862.
References
BrouNr, C. W., mrlo F. W. Drcrsox (1969) The solubility
of anhydrite (CaSO') in NaCl-H,O from 100 to 450' and
I to 1000 bars. Geochi,m.Cosmochim.Acta, 33,227-245.
 GY PSUM-AN HY D RITE EQU I LIB RI A
Brocr, hcor, eNo O. B. Wrrens, Jn. (1968) The
CaSOr-Na,SOr-NaCl-H,O system at 25' to 100'C.
l. Chem. Eng. Data 13, 336-344.
Bocr, E. (1961) On the solubility of anhydrous calcium
sulfate and gypsum in concentratedsolutions of soduim
chloride at 25'C, 30'C, 40"C, and 50'C. Can. L Chem.
39,1746-1751.
Boorn, H. S., eNo R. M. BrpwBn (1950) Solubilities of
salts in water at high temperatures../. Amer. Chem. Soc.
72,2567-2575.
Crenr, S. P., Ed. (1966) Handbook of Physical Constants.
Geol. Soc.Amer. Mem.97, 587p.
Cnurr, Eocen F., ANDPAo-CHrN Cnlo (1970) Nucleation
kinetics of the gypsum-anhydrite system.3rd Symp. on
SaIt,1969,109-118.
D'ANs, J., D. BnBorscHNErDER, H. Ercx, lNo H. E. FneuNo
(1955) Untersuchungeniiber die Calcium Sulfate. Kali.
U. Steinsalz,1, 17-38.
(1968) Der Ubergangspunkt Gips - Anhdydrite,
Kali U. Steinsalz,5,109-111.
DENMAN, WevNs L. (1961) Maximum re-use of cooling
water . . based on gypsum content and solubility. Ind.
Eng.'Chem.53, 817-822.
DrcxsoN, F. W., CnenI-EsW. BLouNT, eNo Groncr TuNsrl
(1953) Use of hydrothermal solution equipment to deter-
mine the solubility of anhydrite in water from 100'C
to 275"C and from I bar to 1000 bars pressure.Amer. l.
Sci. Xil, 6l-78.
Gr-Bw, D. N., eNo D. A. Herrps (1970) Gypsum, disodium
pentacalcium sulfate, and anhydrite solubilities in con-
centrated sodium chlorite solutions. Can, l. Chem. 48,
3731-3738.
Herr, R. E., J. A. Rorn, erto C. E. CoreueN (1926) The
solubility of calcium sulfate at boiler-water temp€ratures.
J. Amer. Chem. Soc. 4E, 927-938.
Hml, R., Y. TeNere AND K. Nexer"rune (1934) On the
calcium sulphate in sea water. I. Solubilities of dihydrate
and anhydrite in sea waters of various concentrations at
0'-200"C. Tohoku Imperial Unhs. Tech. Rep. 11,87-109.
Henore, LrwnBNce A. (1967) The gypsum-anhydrite equi-
librium at 1 atmosphere pressure. Amer. Mineral 52,
l7r-200.
Hrrr, A. E. (1937) The transition temperatureof gypsum
to anhydrite. I. Amer. Chem. Soc. 59,2242-2244.
Horr, I. H. vlN'r, E. F, AnusrnoNc, W. HrNnrcnsrN,
F. Werosnr eNp G. Jusr (1903) Gips und An\Vdrite.
Z. Phys. Chem. 45,257-306.
Hurrrr, Gponcs A. (1901) Beziehungenzwischen ober-
fliichenspannung und Ltjslichkeit. Z. Phys. Chem. 37,
385-406.
eNp L. E. Arrw (1902) The solubility of gypsum.
I. Amer. Chem. Soc.24,667--679.
International Critical Tables (1926-1933). New York,
McGraw-Hill Book Company.
33r
KBnBv, K. K., J. C. Sourrrenp, lNo C. T. ANoBnsoN
(1941) Thermodynamic properties of gypsum and its
dehydration products. tl . S' Bur' Mines Tech, Pap. 625.
MlcDoNer,o, GonooN I. F. (1963) Anhydrite-gypsum equi'
librium relations. Amer. J. Sci. 251,884-898.
MeucrN, W. M., rNo D. A. SwerEs (1956) Solubilities in
the system CaSO.NaCI-H,O at 25" and 35'. I- AppL
Chem.6.482487.
MlNlKxrN, V. L eNp P. A. Knvuxov (1968) Effect of
pressure on the solubility of sodium and calcium sulfates.
Porouye Rastl)ory Metody lkh. Izuch., 13344 [in Rus'
sianl. G. V. Bogomolov, Editor, Izd. "Nauka Tekhnika":
Mrusr, U.S.S.R.
MARSAL,DrBrnrcn (1952) Der Einfluss des Druckes auf das
System CaSO.H,o. Beit' Mineral. Pettogr. 3r 289-296.
Mlnsu,ln, W. L., ewo R. Srusnr'n (1966) Thermodynamics
of calcium sulfate dihydrate in aqueous sodium chloride
solutions,0-1 10". l. Phys. Chem. 70, 40154027 .
Mensnetl, W. J., R. Sr-usnnn, eNo E. V' JoNes (1964)
Aqueous systems at high temperature. XIV. Solubility
and thermodynamic relationships for CaSO, in NaCl-HgO
solutions from 4O' to 200"C, 0 to 4 Molal NaCl'
L Chem. Eng. Data,9, 187-191.
Mercrsn, A. C. (1910) The solubility of silver chloride,
barium sulfate and calcium sulfate at high temperatures.
L Amer. Chem. Soc.32, 50-66.
Pentnlocr, E. P., .lNo A. H. Wnrre (1929) The solubility
of calcium sulfate from 0 to 200'C' J' Amer, Chem. Soc.
51, 360-370.
Posw.rer, E. (1938) The system, CaSO.H,O. Amer. J'
Sci. 235-A' 247-272.
(1940) Deposition of calcium sulfate from sea
watet, Amer. J. 5ri. 2:3E,559-568.
Powen, W. H. lNo B. M. F$uss (1964) Thermodynamic
properties of saline water. Research and Development
Progress Report No. 104, Office of Saline Water, U' S'
Department of the Interior' 82.
lxo C. N. SerreRFrero (1965) Transient
solute concentrations and phase changesof calcium sulfate
in aqueous sodium chloride. l. Chem. Eng' Data ll,
149-154.
TeNlre, Y., K. NereuuRA, AND R. H.lu (1931) On the
calcium sulphato in sea water. I. Soc. Chem' Ind', lapan
34. 284-287.
Tonrur'rI, T. eNp R. Hene (1938) On the transition point
of calcium sulphate in water and concentrated sea water'
Tohoku Imperial Unio. Tech. Rep. 12' 17-94.
ZoANovsxY, A. B. (1949) Heterogeneous salt equilibria'
Tr. Vses.-Nauchn. Issled. Inst, Galurgii, 21, 336-358'
ZBN, E-AN (1965) Solubility measurementsin the system
CaSOr-NaCl-H'O at 35",50', and 70"C and one atmo'
sphere pressure.J' Petrology' 6' 124-l@'
Manuscript receiued,'October4, I97I; acceptedlor publi-
cation October 6' 1972.