Thermochimica Actu, 139 (1989) 67-81 67

Elsevier Science Publishers B.V., Amsterdam - Printed in The Netherlands

HEAT CAPACITIES AND ENTROPIES FROM 8 TO 1000 K

OF LANGBEINITE (K,Mg,(SO,.,),), ANHYDRITE (CaSO,)
AND OF GYPSUM (CaSO, 2H,O) TO 325 K * l

RICHARD A. ROBIE, SUSAN RUSSELL-ROBINSON and BRUCE S. HEMINGWAY

Mail Stop 959, U.S. Geological Survey, Reston, VA 20292 (U.S.A.)
(Received 4 April 1988)

ABSTRACT

Heat capacities of K,Mg,(SO,), (langbeinite) and CaSO, (anhydrite) were measured

from approximately 8 to 1000 K by combined adiabatic shield calorimetry (8-365 K) and
differential scanning calorimetry (350-1000 K). Heat capacities were also measured on
natural crystals of gypsum (CaSO,.2H,O) between 8.1 and 323.5 K. The molar entropies at
298.15 K, Sz(298.15 K), are 378.8f0.6, 107.4kO.2 and 193.8f0.3 J K-’ mol-’ for
langbeinite, anhydrite and gypsum, respectively. The heat capacity in J K-r mol-’ of
langbeinite can be represented by the equation
C;,,(K,Mg,(SO,),, T) = 535.9+0.11011T-1.0200x106/T2 -4.909x10-?*
-4040.2/T”.5
between 300 and 1000 K with an average deviation of + 0.4%. For anhydrite the heat capacity
between 300 and 1000 K is given by
C&,,(CaS04, T) = 372.8-0.1574T+1.695x106/T2 +7.993x10m5T2 -4330.8/T’.’
with an average deviation of + 0.4%.
Combining our heat-capacity and entropy data with the solution calorimetric results of
Kelley et al. (U.S. Bur. Mines Tech. Paper, 625, 1941) yields an equilibrium temperature for
the reaction gypsum --, anhydrite-t 2 water of 314.7 K (41.5 o C).
Our observations are in agreement with the conclusions of Speer and Salje (Phys. Chem.
Miner., 13 (1986) 17); we see no evidence in our heat capacity measurements for the
transformation of cubic langbeinite (P2,3) to a low temperature orthorhombic (P2,2,2,)
form as is seen in the isostructural Co, Zn, Ca, Mn and Cd langbeinites.
Although Bond (Bell Sys. Tech. J., 22 (1943) 145) reported that langbeinite was piezoelec-
tric at room temperature, we found no evidence in our Ci measurements for a Curie
temperature above which langbeinite would no longer be piezoelectric.

INTRODUCTION

The sulfate minerals langbeinite, anhydrite and gypsum are major compo-
nents of many commercial salt deposits. Langbeinite (K,Mg,(SO,),) is a

* Dedicated to Professor Edgar F. Westrum, Jr., on the occasion of his 70th birthday and in
honor of his contributions to calorimetry and thermal analysis.

0040-6031/89/$03.50 0 1989 Elsevier Science Publishers B.V.

68

major source of potash for agricultural use. In 1982, for example, 16 million
tonnes of mixed K,Mg,(SO,) ,-KC1 ore were mined near Carlsbad, NM.
The major coexisting phase in these potash-rich salt deposits is anhydrite,
CaSO, .
One of the locations selected by the Waste Isolation Pilot Plant (WIPP)
which is designed to demonstrate the safe storage of high-level radioactive
waste is sited in these same Permian salt deposits near Carlsbad, NM.
Engineering calculations of the heat buildup in the surrounding rocks as a
consequence of this radioactive waste storage require data for the heat
capacities of the several mineral phases which comprise the salt deposits.
Accurate heat capacity data exist for two of the principal phases which make
up the salt deposits, sylvite (KCl), and halite (NaCl). There are, however, no
Ci data available for langbeinite, while those for anhydrite were obtained on
a somewhat impure mineral sample, and those for gypsum are over 55 years
old and extend upwards in temperature only to 302 K. Modern heat
capacity data on these sulfate minerals would therefore be of considerable
engineering use for nuclear waste repository modeling.
At room temperature langbeinite is cubic (point group 23 (tetratoidal),
space group P2,3 (198)) and is reported by Bond [3] to be piezoelectric. One
would suspect that at some higher temperature (i.e. the Curie temperature)
langbeinite would transform to the normal (non-piezoelectric) state and that
this transformation would be accompanied by an anomaly in the heat
capacity such as that exhibited, for example, by quartz or BaTiO,.
The heat capacity of CaSO, was previously measured by Kelley et al. [l]
at 23 temperatures between 53.5 and 295.7 K using as a sample anhydrite
crystals from Arden, NV which contained some gypsum. Kelley [4] also
derived values for the heat capacity at higher temperatures based upon the
heat-content measurements of Lastschenko and Kompanski [5] between 293
and 1393 K. Lastschenko and Kompanski’s sample was also a mineral
sample and contained 0.23% SiO, and 1.23% H,O.
The heat capacity of gypsum was previously determined by Latimer et al.
[6] between 18.7 and 302.1 K. Heat capacity data are apparently not
available for langbeinite.

SAMPLE PREPARATION

Coarse crystals of K, Mg, (SO,) 3 were prepared by crystallization from a

melt of an equimolar mixture of K,SO, and MgSO,. Because of its very fine
grain size and hygroscopic nature, reagent grade MgSO, (Fisher M-65) was
first heated at 1150 K for 2 h in a 100 cm3 platinum crucible immedately
prior to weighing. The appropriate number of moles of reagent K,SO,
(Fisher P-304) which had been heated at 400 K before weighing was added
 69

to the crucible containing the MgSO,, mixed well, and placed in an electric
furnace at 1225 + 10 K in which it was held for 4 h. The furnace controller
was then turned off and the crucible and its contents allowed to cool to
room temperature overnight. This procedure yielded crystals which mea-
sured 2 mm on edge.
The sample was freed from the crucible, crushed to pass through a 10
mesh screen (1.68 mm), washed with cold distilled water to remove possible
excess K,SO,, and vacuum dried. It was then sieved and the material
passing a 35 mesh screen was discarded. The sample was heated in an oven
at 385 K for 1 h before loading in the calorimeter.
An X-ray diffractometer tracing using Cu K, showed only those lines for
langbeinite given in the National Bureau of Standards (NBS) pattern (Swan-
son et al. [7]). The unit cell edge of the calorimetric sample was 0.99196 +
0.00015 nm and was obtained using NBS Standard Reference Material 640
silicon (a = 0.543088 nm, Hubbard et al. [S]) as an internal standard, Cu
K,, radiation (h = 0.1540598 nm), and 12 lines for which 29 2 49”. Our
value agrees with that of Swanson et al. [7] to within f 0.014%. The molar
volume VG (298.15 K) of langbeinite calculated from our cell edge and using
Cohen and Taylor’s [9] value for Avogradro’s number (6.022137(4) X 1O23
mol-‘) is 146.95 + 0.04 cm3.
Anhydrous CaSO, was prepared using the procedure described by Manross
[lo]. Reagent grade CaCl, .2H,O (Fisher C-79) was heated in a resistance
furnace in a 100 cm3 platinum crucible to 1123 K and held for 3 h. It was
removed from the furnace while still molten and quenched in air. Reagent
grade K,SO, (Fisher P-304) was then added to the crucible. A typical
charge was 40 g of CaCl 2 (anhydrous) and 8.5 g of K,SO,. The crucible was
returned to the furnace and held at 1150 K for 3 h. The furnace controller
was then reset to 950 K and the crucible and its contents cooled over
approximately 1 h to 950 K and then held at 950 K for 16 h. The crucible
was removed from the furnace and cooled to room temperature. The salt
pellet was broken free of the crucible, crushed coarsely in an alumina mortar
and washed with copious quantities of hot water to remove the residual
CaCl, and KCl.
Tabular crystals of CaSO, up to 1.2 cm in length were grown in this
fashion. For the calorimetric measurements the material was sieved and the
crystals passing a 35 mesh sieve (0.42 mm) were discarded. Spectrographic
analysis indicated the presence of 0.09% K,O in the CaSO, sample.
Unit cell parameters for our sample were measured using NBS Standard
Reference Material 640 (silicon) as an internal standard. Our values are
a = (0.69893 + 0.00038), b = (0.69964 + 0.00030) and c = (0.62409 &-
0.00028) nm and agree to within 0.04% with those of McMurdie et al. [ll].
Our data lead to a molar volume Vz(298.15 K) of 45.96 + 0.03 cm3.
McMurdie et al.‘s [ll] cell parameters yield a value for Vz(298.15 K) =
46.008 * 0.006 cm3.
70

The gypsum sample was part of a large, 15 cm long, glass-clear crystal of

selenite from Brazil. Semi-quantitative spectroscopic analysis indicated the
presence of 300 p.p.m. strontium, i.e. 0.35 wt.% SrO, as the only significant
impurity.

EXPERIMENTAL METHODS AND RESULTS

Our heat capacity measurements were made by the intermittent heating

method under quasi-adiabatic conditions using the calorimeter-cryostat and
automatic data-acquisition system described by Robie [12]. Temperatures
were determined with a Mince model S 1059-l platinum resistance ther-
mometer (R, = 100.02 Q) that had been calibrated by the Temperature
Measurements Section of the National Bureau of Standards on IPTS-68
between 13.8 and 505 K, and on a provisional scale used in this laboratory
between 4.2 and 13.8 K [12]. The samples were sealed in the calorimeter
under purified helium gas at a pressure of 5 kPa (approximately 4 x lop5
mol of He). The langbeinite sample had a mass of 35.713 g, that of anhydrite
was 33.748 g and the mass of the gypsum sample was 30.014 g.

B 50 188 15-a 200 250 300 350

T~?llpW-Cit"~~. K
Fig. 1. Experimental heat capacities of synthetic anhydrite (CaSO,) crystals (A) and natural
gypsum crystals from Brazil (0).
 71

TABLE 1
Experimental heat capacities (J mol-’ K-l) of K,Mg,(SO,), measured with an adiabatic
calorimeter; formula weight 414.9794 g mol-’

Temp. Heat Temp. Heat Temp. Heat

(K) capacity (K) capacity (K) capacity
Series 1 124.09 187.2 144.12 208.2
302.22 321.2 128.53 192.0 148.28 212.1
306.73 323.4 132.92 196.8 152.48 216.0
311.29 326.3 137.28 201.4 156.66 220.1
315.87 328.2 160.83 223.8
320.31 330.8 Series 5 164.99 227.6
324.75 333.5 8.24 0.836 169.13 231.1
329.18 335.9 8.89 1.117 173.26 234.5
333.62 337.1 9.78 1.580 177.39 238.0
338.05 339.9 10.85 2.197 181.50 241.8
12.11 2.969
Series 3 13.55 4.109 Series 7
338.77 339.9 15.11 5.580 191.93 250.3
344.09 342.6 16.84 7.397 196.83 253.7
349.51 344.8 18.75 9.761 201.77 257.5
354.91 347.2 20.88 12.90 206.77 261.4
360.18 349.5 23.24 16.78 211.56 264.7
365.43 351.4 25.88 21.67 216.43 268.2
370.66 353.6 28.84 27.76 221.35 271.8
32.17 35.34 226.35 275.1
Series 4 35.91 44.38 231.46 278.8
51.76 84.60 40.11 54.96 236.69 282.0
56.41 93.74 44.77 67.70 242.03 285.4
66.04 110.1 49.91 81.31 247.45 288.7
71.49 118.6 54.89 90.64 252.94 292.0
76.67 126.6 66.43 110.6 258.47 295.1
81.67 133.9 68.81 114.4 264.08 298.4
86.53 140.9 71.12 117.9 269.76 302.0
91.39 147.7 73.40 121.4 275.43 305.6
96.30 154.2 281.08 308.8
101.12 160.4 Series 6 286.72 312.4
105.84 166.2 131.24 195.2 292.33 316.1
110.49 171.7 135.40 199.4 297.93 319.2
115.08 177.1 139.72 203.8 303.52 321.8
119.61 182.3

Our experimental heat capacities, measured with the adiabatic calorime-

ter, are listed in their chronological order of measurement in Tables 1, 2 and
3 for synthetic langbeinite, synthetic anhydrite and natural gypsum, respec-
tively. The data for anhydrite and gypsum are shown graphically in Fig. 1.
The Ci data were extrapolated graphically to 0 K using a plot of C: ,/T
against T2. Above 25 K the data were smoothed by splines.
72

TABLE 2
Experimental heat capacities (J mol-’ K-‘) of CaSO, measured with an adiabatic calorime-
ter; formula weight: 136.138 g mol-’

Temp. Heat Temp. Heat Temp. Heat

(K) capacity (K) capacity (K) capacity
Series 1 43.66 12.46 167.27 71.08
304.26 102.1 48.69 15.80 172.63 72.62
309.87 103.7 53.63 19.23 177.97 74.08
315.58 104.2 57.79 22.10 183.31 75.68
321.29 105.1 61.40 24.52 188.63 77.11
326.98 106.1 64.74 26.67 193.96 78.48
332.64 107.0 67.91 29.09 199.28 79.96
338.28 108.0
Series 4 Series 5
Series 2 56.21 21.06 204.40 81.17
343.71 109.1 59.44 23.27 209.91 82.40
349.08 110.1 62.42 25.27 215.75 83.87
354.73 110.9 65.60 27.34 221.67 85.32
360.42 111.5 69.18 29.61 227.34 86.48
366.04 112.6 72.67 31.78 233.12 87.85
76.11 33.85 239.02 89.22
Series 3 79.49 35.77 245.02 90.45
8.65 0.0299 82.80 37.64 251.16 91.65
9.89 0.0925 87.53 40.20 257.42 93.00
10.87 0.1247 93.62 43.33 263.70 94.24
12.03 0.1700 99.64 46.26 269.98 95.59
13.46 0.2734 105.66 48.99 276.24 96.74
14.95 0.4122 111.60 51.58 282.49 97.99
16.63 0.6012 117.45 54.03 288.74 99.74
18.50 0.8760 123.22 56.26 294.97 100.9
20.56 1.273 128.91 58.39 301.19 101.7
22.85 1.827 134.52 60.47
25.41 2.612 140.08 62.45 Series 6
28.27 3.693 145.59 64.33 301.50 101.5
31.48 5.164 151.06 66.13 306.30 102.7
35.08 7.095 156.49 67.79 311.19 103.7
39.14 9.344 161.89 69.42

The extrapolation of the data for anhydrite below 10 K corresponds to a

Debye temperature of 502 K which is in only fair agreement with the value
435 K which one calculates from the room temperature elastic constant data
of Schwerdtner et al. [13]. For langbeinite our extrapolation below 10 K
corresponds to an apparent Debye temperature of 282 f 10 K which is
completely out of line with the 8, calculated from the elastic constants of
Haussuhl [14] and of Liebertz and Quadflieg [15], whose data yield 8, of
500 and 505 K, respectively.
 73

TABLE 3
Experimental heat capacities (J mol-’ K-‘) of gypsum (CaSO,.2H,O) measured with
adiabatic calorimeter; formula weight 172.168 g mol-’

Temp. Heat Temp. Heat Temp. Heat

(K) capacity (K) capacity (K) capacity
Series 1 188.35 137.5 9.04 0.2052
53.16 35.77 193.34 140.3 9.98 0.3100
58.47 41.93 198.37 143.0 11.05 0.4419
63.21 46.70 203.22 145.6 12.37 0.6414
68.52 51.73 208.07 148.1 13.80 0.9336
74.74 57.76 212.93 150.5 15.35 1.327
80.72 63.28 217.82 152.9 17.07 1.868
86.64 68.51 222.75 155.3 18.99 2.659
92.56 73.64 227.75 157.6 21.10 3.712
98.43 78.41 232.84 160.0 23.44 5.092
104.20 82.95 237.99 162.3 26.10 6.929
109.86 87.20 243.17 164.7 29.06 9.348
115.44 91.37 248.37 167.0 32.39 12.45
120.94 95.35 253.63 169.2 36.16 16.30
126.38 99.20 258.96 171.4 40.40 21.07
131.76 102.9 264.29 173.6 45.18 26.37
137.09 106.5 269.61 176.0 50.52 32.61
142.36 110.0 274.92 178.1 56.41 39.66
147.59 113.5 280.21 180.2 62.57 46.09
152.78 116.7 285.49 182.4
157.93 119.9 290.74 184.9 Series 4
163.06 123.0 295.98 186.5 311.00 192.0
168.15 126.0 301.20 188.4 315.83 193.7
173.22 129.1 306.41 190.3
178.27 132.0
Series 5
Series 2 Series 3 319.05 194.7
183.30 134.7 8.09 0.1402 323.51 196.4

Smooth values for Cp9,,,, S;(T)- S:(O), [Hz(T)- Hi(O)]/T, and

-[G:(T) - KXO)I/T are listed in Tables 4, 5 and 6 for langbeinite,
anhydrite and gypsum respectively.
Our Cp9m values for anhydrite are systematically greater than those given
by Kelley et al. [3] by 1.4% at 50 K decreasing to 0.4% at 200 K and rising
again to 1.4% at 300 K. For gypsum our results are significantly lower (up to
14% at 20 K) than the values of Latimer et al. [6] up to 50 K above which
our values are consistently larger by 0.2-0.6s to 300 K.
The heat capacities of langbeinite and anhydrite were also measured
between 350 and 990 K using a differential scanning calorimeter (Perkin-
Elmer DSC-2 *)_ The sample weights for langbeinite and anhydrite were

* Reference to specific manufacturers and model numbers is made to facilitate under-

standing and does not imply endorsement by the U.S. Geological Survey.
74

TABLE 4
Standard molar thermodynamic functions for K,Mg,(S0,)3 between 0 and 370 K

Temp. Heat Entropy Enthalpy Gibbs energy

T capacity ST-S,0 function function
(K) cp” (H; - %)/T -(G; - H,“)/T
(J mol-’ K-‘)
5 0.22 0.07 0.05 0.02
10 1.69 0.56 0.42 0.14
15 5.46 1.87 1.40 0.47
20 11.55 4.21 3.12 1.09
25 19.99 7.65 5.61 2.04
30 30.33 12.18 8.84 3.34
35 42.15 17.73 12.74 4.99
40 54.96 24.19 17.21 6.98
45 68.14 31.42 22.14 9.28
50 80.47 39.25 27.37 11.88
60 100.2 55.76 37.94 17.81
70 116.3 72.44 48.02 24.42
80 131.5 88.97 57.51 31.46
90 145.7 105.3 66.53 38.76
100 158.9 121.3 75.12 46.22
110 171.2 137.1 83.30 53.76
120 182.7 152.5 91.11 61.35
130 193.7 167.5 98.59 68.94
140 204.1 182.3 105.8 76.51
150 213.8 196.7 112.6 84.04
160 223.0 210.8 119.3 91.52
170 231.8 224.6 125.6 98.95
180 240.4 238.1 131.8 106.3
190 248.5 251.3 137.7 113.6
200 256.3 264.2 143.4 120.8
210 263.7 276.9 149.0 127.9
220 270.8 289.3 154.4 135.0
230 277.6 301.5 159.6 142.0
240 284.1 313.5 164.6 148.9
250 290.2 325.2 169.5 155.7
260 296.2 336.7 174.3 162.4
270 302.2 348.0 178.9 169.1
280 308.3 359.1 183.4 175.7
290 314.3 370.0 187.8 182.2
300 320.0 380.8 192.1 188.6
310 325.5 391.4 196.4 195.0
320 330.7 401.8 200.5 201.3
330 335.8 412.0 204.5 207.5
340 340.6 422.1 208.4 213.7
350 345.1 432.1 212.3 219.8
360 349.3 441.8 216.0 225.8
370 353.4 451.5 219.7 231.8
273.15 304.2 351.5 180.3 171.2
298.15 319.0 378.8 191.4 187.4
TABLE 5
Standard molar thermodynamic functions for CaSO, between 0 and 360 K

Temp. Heat Entropy Enthalpy Gibbs energy

T capacity S;-S,o function function
(K) cp” (H; - W/T -(G; - H,o)/T
(J mol-’ K-‘)
5 0.010 0.003 0.002 0.001
10 0.094 0.029 0.022 0.007
15 0.419 0.115 0.089 0.026
20 1.167 0.325 0.254 0.071
25 2.476 0.712 0.555 0.158
30 4.457 1.328 1.031 0.298
35 7.017 2.202 1.698 0.504
40 9.956 3.326 2.542 0.784
45 13.29 4.689 3.549 1.140
50 16.72 6.266 4.694 1.572
60 23.62 9.927 7.276 2.652
70 30.17 14.07 10.08 3.982
80 36.08 18.49 12.97 5.517
90 41.48 23.05 15.84 7.211
100 46.41 27.68 18.66 9.027
110 50.90 32.32 21.39 10.93
120 55.02 36.93 24.02 12.91
130 58.83 41.49 26.55 14.93
140 62.41 45.98 28.99 16.99
150 65.76 50.40 31.33 19.07
160 68.88 54.75 33.58 21.16
170 71.86 59.01 35.75 23.27
180 74.72 63.20 37.83 25.37
190 77.46 67.31 39.85 27.47
200 80.06 71.35 41.79 29.56
210 82.51 75.32 43.67 31.65
220 84.87 79.21 45.49 33.72
230 87.15 83.04 47.25 35.78
240 89.35 86.79 48.96 37.83
250 91.46 90.48 50.62 39.86
260 93.52 94.11 52.23 41.88
270 95.57 97.68 53.80 43.88
280 97.64 101.2 55.33 45.86
290 99.67 104.7 56.82 47.83
300 101.5 108.1 58.28 49.78
310 103.3 111.4 59.71 51.72
320 105.0 114.7 61.10 53.63
330 106.7 118.0 62.45 55.54
340 108.4 121.2 63.78 57.42
350 110.1 124.4 65.08 59.29
360 111.6 127.5 66.35 61.14
273.15 96.22 98.79 54.28 44.51
298.15 101.2 107.4 58.01 49.42
76

TABLE 6
Standard molar thermodynamic functions for CaSO,.2H,O between 0 and 320 K

Temp. Heat Entropy Enthalpy Gibbs energy

T capacity s; - s,o function function
(K) Cp” (H; - %‘)/T -(G; - H,“)/T
(J mol-’ K-t)
5 0.040 0.016 0.011 0.005
10 0.309 0.102 0.074 0.028
15 1.23 0.370 0.280 0.090
20 3.17 0.959 0.733 0.226
25 6.15 1.96 1.50 0.460
30 10.19 3.43 2.60 0.830
35 15.07 5.36 4.02 1.34
40 20.50 7.72 5.74 1.98
45 26.21 10.46 7.69 2.77
50 32.03 13.52 9.84 3.68
60 43.26 20.37 14.48 5.89
70 53.39 27.81 19.33 8.48
80 62.71 35.56 24.18 11.38
90 71.43 43.45 28.95 14.50
100 79.63 51.40 33.61 17.79
110 87.36 59.36 38.15 21.21
120 94.71 67.28 42.56 24.72
130 101.7 75.14 46.84 28.30
140 108.5 82.93 51.01 31.92
150 114.9 90.63 55.05 35.58
160 121.2 98.25 58.99 39.26
170 127.2 105.8 62.83 42.95
180 132.9 113.2 66.56 46.65
190 138.5 120.5 70.20 50.35
200 143.9 127.8 73.75 54.04
210 149.0 134.9 77.22 57.72
220 154.0 142.0 80.59 61.39
230 158.7 148.9 83.89 65.05
240 163.2 155.8 87.10 68.68
250 167.7 162.5 90.23 72.30
260 171.9 169.2 93.29 75.90
270 176.1 175.8 96.28 79.48
280 180.2 182.2 99.21 83.03
290 184.1 188.6 102.1 86.57
300 187.9 194.9 104.9 90.07
310 191.6 201.2 107.6 93.56
320 195.3 207.3 110.3 97.02
273.15 177.4 177.8 97.21 80.60
298.15 187.2 193.8 104.4 89.43

30.203 and 30.010 mg, respectively. The calorimeter was operated at a rate
of 10 K min-’ and the sensitivity was set at 0.021 W. The results are listed
in Tables 7 and 8 for langbeinite and anhydrite, respectively. Our Ci m
 77

TABLE 7
Experimental molar heat capacities (J K-’ mol-‘) of K,Mg,(SO,), measured with a
differential scanning calorimeter

Temp. Heat Temp. Heat Temp. Heat

(K) capacity (K) capacity (K) capacity
Series 1 Series 3 815.0 451.6
338.9 339.9 567.5 412.2
348.8 344.2 577.5 413.0 Series 6
358.8 348.3 587.4 413.9 806.0 450.2
368.7 353.0 597.3 416.0 816.0 451.1
378.7 355.6 607.3 417.6 825.9 451.7
388.6 358.4 617.2 418.0 835.9 453.1
398.6 361.6 627.2 419.8 845.7 453.6
408.5 365.3 637.1 421.5 855.4 454.4
418.4 365.4 647.0 423.1 865.1 456.2
428.4 372.7 657.0 425.2 874.8 459.3
438.3 376.2 666.9 427.8 883.5 463.8
448.3 379.5 676.8 431.9
458.3 383.8 686.8 436.6 Series 7
468.2 387.2 695.7 439.4 913.6 461.2
478.1 390.3 923.3 465.2
488.0 393.7 Series 4 933.0 462.8
497.0 397.4 666.9 425.0 941.7 464.7
676.8 428.5
Series 2 686.8 428.7 Series 8
468.1 386.2 696.7 432.1 962.0 460.1
478.1 388.2 706.7 434.3 971.7 460.6
488.0 390.5 716.6 434.8 981.4 465.5
498.0 393.5 726.5 435.9 990.0 468.4
507.9 395.7 736.5 436.8
517.8 398.2 745.4 437.3 Series 9
527.8 401.0 835.9 455.9
537.7 403.8 Series 5 845.7 454.5
547.6 406.2 736.5 439.9 855.4 453.2
557.6 408.6 746.4 441.8 865.1 455.4
567.5 410.8 756.3 442.0 874.8 459.5
577.5 413.5 766.3 444.1 884.5 455.8
587.4 416.5 776.2 445.7 894.2 460.4
596.3 419.4 786.2 447.4 903.9 458.8
796.1 448.0 912.6 460.1
806.0 449.9

measurements for langbeinite form a smooth, monotonic sigmoidal curve

between 9 and 990 K with no indication of either a low temperature
transition to a less symmetric (orthorhombic) form or of a Curie point above
room temperature.
The C; data obtained with the scanning calorimeter were combined with
those measured with the low temperature adiabatic calorimeter to generate
78

TABLE 8
Experimental molar heat capacities (J K-’ mol-‘) of CaSO, measured with a differential
scanning calorimeter

Temp. Heat Temp. Heat Temp. Heat

(K) capacity (K) capacity (K) capacity
Series 1 587.4 136.4 746.4 144.6
338.9 107.3 596.3 137.8 756.3 144.9
348.8 108.7 766.3 145.3
358.8 110.2 Series 3 776.2 145.7
368.7 111.4 567.5 132.3 786.2 146.8
378.7 112.7 577.5 132.7 796.1 147.3
388.6 114.2 587.4 134.0 806.0 148.0
398.6 115.6 597.3 134.9 815.0 148.5
408.5 116.8 607.3 135.7
418.4 118.1 617.2 136.4 Series 6
428.4 119.2 627.2 137.1 806.0 147.9
438.3 120.6 637.1 137.7 816.0 148.3
448.3 121.6 647.0 138.3 825.9 149.0
458.3 122.5 657.0 138.5 835.9 149.8
468.2 123.7 666.9 139.2 845.9 149.8
478.1 124.8 676.8 140.9 855.4 150.1
488.0 126.0 686.8 140.5 865.1 151.4
497.0 126.7 695.7 140.9 874.8 152.3
883.5 152.4
Series 2 Series 4
468.1 124.8 666.9 138.3 Series 7
478.1 125.5 676.8 139.6 913.6 155.3
488.0 126.2 686.8 140.3 923.3 155.7
498.0 127.0 696.7 140.9 933.0 156.3
507.9 127.8 706.7 141.2 941.7 157.3
517.8 128.5 716.6 141.8
527.8 129.5 726.5 142.3 Series 8
537.7 130.6 736.5 143.0 962.0 156.9
547.6 131.6 145.4 143.6 971.7 158.0
557.6 133.0 981.4 159.5
567.5 133.8 Series 5 990.0 158.5
577.5 135.4 736.5 144.2
-

the following equations in J K-’ mol-’ valid from 300 to 1000 K

C&(K,Mg,(SO,),, T) = 535.9 + O.llOllT- 1.0200 x 106/T2

- 4.909 x 10-5T2 - 4040.2/T”.5

C;_,(CaSO,, T) = 372.8 - O.l574T+ 1.695 x 106/T2

+7.993 X lo-'T2- 4330.8/T0.5

The average deviation of each data set from the fitted equation was 0.4%.
 79

THE GYPSUM-ANHYDRITE REACTION

The dehydration of gypsum to form anhydrite apparently occurs readily

in nature. Replacement of anhydrite crystals by gypsum is also common. In
several localities bedded deposits of alternating layers of gypsum and
anhydrite exist. There have been a number of studies of the equilibrium
temperature of the reaction
CaSO, - 2H 2O = CaSO, + 2H,O (1)
gypsum anhydrite water
Three of the more recent studies are Zen [16], Hardie [17] and Blount and
Dickson [18]. The last two papers take issue with the previously accepted
value of 314 K (42°C) for the equilibrium temperatures of reaction (1)
based principally on the work of Posnjak [19] and opt for a higher tempera-
ture of 330 & 2 K (57 o C).
Our heat capacity data for anhydrite and gypsum between 300 and 325 K
were fitted to linear equations yielding
Cp9,(anhydrite) = 49.9 + 0.172T (2)
and
C;, ,(gypsum) = 76.9 + 0.37T (3)
For water we used the value 75.345 J K-’ mol-’ from the CODATA tables
(Garvin et al., [20]). For reaction (1) we get
AC; = 123.69 - 0.198T (300-325 K) (4)
Using Kelley et al.‘s [3] value for A,H”(298.15 K) for this reaction, 16 862 +
84 J, we can write
A,H”(T) = - 11216 + 123.69T - 0.099T2 (300-325 K) (5)
Combining our entropy data with the equation for AC’; we get
A,S” = 123.69 - 0.198T + 123.69 X In T - 715.9 (6)
and
A,G” = - 11216 + 0.099T2 - 123.69 X TX In T + 715.9T (7)
From this expression A,G” is 0 at 314.7 K (41.5OC). The uncertainties in
A,G” arising from the f 84 J uncertainty in A,H” correspond to f 1.4 K at
this temperature, whereas the f 0.39 J K-i uncertainty in A,S0(298.15 K)
yields an uncertainty off 2.1 K at the equilibrium temperature. Thus, in the
extreme case the calorimetric uncertainties result in a possible spread
of f 3.5 K in the calculated equilibrium temperature. Blount and Dickson’s
[18] assertion that “the 40 o C value calculated from thermochemical data by
Kelley et al. [4] must be regarded as fortuitous” is unwarranted. The solution
calorimetric results of Kelley et al. [4] for A,H”(298.15 K) when combined
80

with the present more accurate entropies support Posnjak’s [19] value for the
equilibrium temperature ( - 41” C). It is well known from the work of
Weiser et al. [21] that gypsum crystals in the absence of an aqueous solution
can persist metastabily up to - 333 K.
The pressure effect on this equilibrium can be obtained as follows. From
the X-ray unit cell parameters (Morris et al. [22]) and using 6.022137(4) x
1O23 mol-’ for Avogadro’s number, we calculate 74.58 f 0.02 cm3 for the
molar volume of gypsum. Using our value for the molar volume of anhydrite
45.96 ) 0.03 cm3 and 18.167 + 0.005 cm3 for vz(H,O, 314.7 K) we get
A,V”(314.7 K) = 7.714 + 0.029 cm3 or 0.7714 J bar-‘. From our equation for
A$“ we calculate A,S”(314.7 K) = 56.89 K + 0.39 J K-i, and thus dp/dT
= (56.89 + 0.39)/(0.7714 -t 0.003) = 73.7 f 0.5 bar K-‘.

CONCLUSIONS

Combination of our entropy and heat capacity data for gypsum and
anhydrite with the enthalpy of reaction measured by Kelley et al. [l] leads to
314.7 + 3.5 K for the temperature at which gypsum and anhydrite are in
equilibrium in aqueous solution. This temperature is in excellent agreement
with that determined by Posnjak [19] from solubility measurements and does
not support the higher temperatures obtained by Hardie [17] and Blount
and Dickson [18]. Our low temperature Ci m measurements on K,Mg,-
(SO,),, down to 9 K, show no suggestion of a transition and therefore
support Speer and Salje [2], who concluded from their structural investiga-
tions that the cubic to orthorhombic phase transition exhibited by Ca, Cd,
Co, Mn and Zn “langbeinites” would not occur in K,Mg,(SO,),. However,
the excessively high Cr& of langbeinite below 20 K might possibly be
indicative of a transition occurring at a temperature below 9 K. We believe
that this is most unlikely.

ACKNOWLEDGMENTS

We wish to thank our U.S. Geological Survey colleagues G.E. Ericksen

for the gift of the gypsum crystal and D.E. Voigt and J.S. Huebner for the
use of and help with their X-ray diffractometer and unit-cell refinement
computer program.

REFERENCES

1 K.K. Kelley, J.C. Southard and CT. Anderson, U.S. Bur. Mines Tech. Paper, 625, 1941.
2 D. Speer and E. Salje, Phys. Chem. Miner., 13 (1986) 17.
 81

3 W.L. Bond, Bell Sys. Tech. J., 22 (1943) 145.

4 K.K. Kelley, U.S. Bur. Mines Bull., 584 (1960).
5 P.N. Lastschenko and D.I. Kompanski, J. Russ. Phys. Chem. Sot., 60 (1928) 579.
6 W.M. Latimer, J.F.G. Hicks, Jr. and P.W. Schutz, J. Chem. Phys., 1 (1933) 620.
7 H.E. Swanson, H.F. McMurdie, M.C. Morris and E.H. Evans, NBS Monograph 25, 6
(1968) 40.
8 C.R. Hubbard, H.E. Swanson and F.A. Mauer, J. Appl. Crystallogr., 8 (1975) 25.
9 E.R. Cohen and B.N. Taylor, J. Res. NBS, 92 (1987) 85.
10 N.S. Manross, Liebigs Ann. Chim., 82 (1852) 348.
11 H.F. McMurdie, M.C. Morris, E.H. Evans, B. Paretzkin and W. Wong-Ng, Powder
Diffraction, 1 (1986) 265.
12 R.A. Robie, Calorimetry, in G.C. Ulmer and H.L. Barnes (Eds.), Hydrothermal Experi-
mental Techniques, Wiley-Interscience, New York, 1987, p. 389.
13 W.M. Schwerdtner, J.C.-M. Tau and P.B. Hertz, Can. J. Earth Sci., 2 (1965) 673.
14 S. Haussuhl, Acta Crystallogr., 18 (1965) 839.
15 J. Liebertz and P. Quadflieg, J. Cryst. Growth, 6 (1969) 109.
16 E-an Zen, J. Petrology, 6 (1965) 124.
17 L.A. Hardie, Am. Mineral., 52 (1967) 171.
18 C.W. Blount and F.W. Dickson, Am. Mineral., 58 (1973) 323.
19 E. Posnjak, Am. .?. Sci. A, 35 (1938) 247.
20 D. Garvin, V.B. Parker and H.J. White, Jr., CODATA Thermodynamic Tables, Hemi-
sphere Publishing, Washington, 1987.
21 H.B. Weiser, W.O. Milligan and W.C. Eckholm, J. Am. Chem. Sot., 58 (1936) 1261.
22 M.C. Morris, H.F. McMurdie, E.H. Evans, B. Paretzkin, C.R. Hubbard and S.J. Carmel,
NBS Monograph 25, 17 (1980) 16.

Note added in proof

Professor Juliana Boerio-Goates (Brigham Young University) has independently measured

the heat capacity of K,Mg,(SO,),. She has found that, by using very low heating rates and
small temperature rises, she was able to show the existence of three small transitions in C;
between 50 and 70 K.
Our measurements on K,Mg,(SO,), in this temperature range (see Table 1, series 4 and
5) were made with heating rates of 0.006 to 0.025 K s-i and temperature rises of between 2.3
and 5.6 K, and it is quite possible that our much larger temperature rises caused us to miss
these transitions.