HEATCAPACITY OF ANHYDROUS NiClz
Sept. 5, 1952 [CONTRIBUTION FROM THE
CHEMICAL LABORATORY OF THE UNIVERSITY
OF
4443 CALIFORNIA]
The Heat Capacity of Anhydrous NiC1, from 15 t o 300°K. The Antiferromagnetic Anomaly near 52°K. Entropy and Free Energy1 BY R. H. BUSEYAND W. F. GIAUQUE RECEIVED APRIL4, 1952 The heat capacity of anhydrous nickelous chloride has been measured from 15 to 300°K. The anomalous region associated with the transition from the antiferromagnetic to the paramagnetic state has been investigated in detail. A maximum was found in the heat capacity near 52'K. which corresponds to the maximum magnetic susceptibility within the accuracy of the magnetic measurements. Thermodynamic properties are given to 300'K. and these in conjunction with the work of Coughlin from 298 to 1336'K. on a sample of the same material complete the thermodynamic description over the range 15 to 1336'K. The entropy was found to be 23.33 cal. deg.-l mole-' a t 298.16'K.
The low temperature heat capacity of anhydrous nickel chloride reported here was measured t o provide values of the entropy and free energy functions in connection with a study of the equilibrium reduction of NiClz by hydrogen. It is also of interest because of the heat effect accompanying the gradual transition from the antiferromagnetic to the paramagnetic state characteristic of higher temperatures. Since impurities might have a serious effect on the equilibrium measurements, and also on the antiferromagnetic behavior of the substance, high purity material was considered t o be of exceptional importance. The preparation of pure anhydrous NiClz was by far the most laborious and time consuming part of the research. Preparation of SiClz.-The nickelous chloride was prepared by a method which was essentially the same as that used by Baxter and Hilton2 in their work on the atomic weight of nickel. Pyrex glass was used in place of platinum vessels, which were considered unnecessary since the NiCly was finally doubly distilled. Although the starting material contained only 0.01 mole per cent. of cobalt, on a nickel basis, a preliminary purification, using Ilinsky-Knorre's3 a-nitroso-&naphthol method, for the separation of small amounts of cobalt was carried out. The procedure followed was that of Mellor and Thompson.' The product was given a preliminary drying a t 400" in a stream of anhydrous hydrogen chloride. The sublimation was carried out in a clear quartz tube about 3.7 cm. i.d. and 60 cm. long. The exit end of the quartz tube was reduced to about 0.8 cm. i.d. and pointed downward to end in a 1mm. bubbler immersed in concentrated sulfuric acid. The hydrogen chloride gas moved through the tube a t a rate of about 3 cc. per min. measured a t room temperature. Three tube furnaces, each of which was 10 cm. in length, were used in series. The first and third furnaces were operated a t 900-1000°. The intermediate furnace was held a t 400-450'. A quartz boat containing the nickel chloride to be sublimed was centered in the first furnace. Condensation occurred in the region of the middle furnace. At first it was thought that a single furnace around the boat would suffice; however, colloidal SiClz, produced by sudden condensation, was carried out of the tube. The second furnace, a t 400", was added to increase the particle size. This was partially successful but some Sic12 was still carried out in colloidal form. The addition of the third furnace, a t 1000°, caused essentially complete condensation in the 400' range. The function of the third furnace is as follows: I t had been noticed in preliminary experiments that heating with a h n i e caused a good deal of convectioii within the tube. In the final arrangement the gas had considerable opportunity (1) This work was supported in part by the Office of Naval Research,
U. S. Navy. (2) G. P. Baxter and P. A. Hilton, THISJ O U R N A L , 46, 694 (1923). (3) M. Ilinsky and G . von Knorre, Bcr.. 18, 699, 2728 (1885); 20, 283 (1887). (4) J. W. Mellor and H. V. Thompson, "A Treatise o n Quantitative Inorganic Analysis," Charles Griffin and Co., London, 1938, pp. 420421.
for convection during which it passed through a variety of temperatures and evidently had increased opportunity for particle growth. The hydrogen chloride gas stream was replaced by nitrogen before the sublimed material was removed in a dry-box. A residue of approximately 3 mg., which was probably silica, remained in the boat from a 30-g. sample of NiCL in the first sublimation. All of the material was sublimed twice and there was about 1 mg. of residue after the second sublimation of a 35-g. sample. Approximately 550 g. of anhydrous crystalline nickelous chloride was prepared. The salt was orange to red in color depending on the crystal size. All subsequent handling of the nickel chloride was carried out in a dry-box. Analysis of Nickelous Chloride.-The nickelous chloride was analyzed for chloride and nickel content. The chloride was determined by weighing silver chloride dried to constant weight a t 325'. The nickel was determined by the dimethylglyoxime method. All weights were corrected to vacuum. The density of NiC12 was taken as 3.54 g./ml. as determined by Baxter and Hilton.z The density of nickel dimethylglyoxime is not known but was estimated as 2.5 g./ ml. The atomic weights were taken as Xi = 58.69, C1 = 35.457, Ag = 107.880. The molecular weight of nickel dimethylglyoxime, KiCsHlpNaOz,was taken as 288.914. The chloride analysis is considerably more accurate than that of the nickel. Two analyses for chloride gave 54.716 and 54.714% compared t o the theoretical value of 54.716%. The amount of AgCl weighed was about 10 g. in each case. Two analyses for nickel gave 45.28 and 45.30% compared to the theoretical 45.284%. The amounts of nickel dimethylglyoxime weighed were about 2.5 g. Calorimetric Apparatus and Procedure.-The calorimetric arrangement was of the type ordinarily used in this Laboratory. I t was similar to one described by Giauque and Egan' except that the calorimeter was made of copper instead of gold and it did not have the features relating particularly to condensed gases. The gold resistance thermometer was calibrated in terms of a standard thermocouple with the laboratory designation UT-26 originally compared to a helium gas thermometer. Both the thermocouple and resistance thermometer were compared directly with the triDle and boiling Doints of both hvdroeen and nitroeen. d&ng the present work, by condensing tlhese substances ir; the insulating vacuum space. Hz, t.p. 13.92', b.p. 20.36'K.; S?,t.p. 63.15', b.p. 77.34'K.; 0°C. was takeii as equal to 273.16"K. One innovation which will be of importance in many future cases concerns the way in which the thermocouple was attached. Previously a small thermocouple well was permanently attached t o the calorimeter. Rose's alloy was used in the well to provide thermal contact. It is often tlificult to detach the fragile tlirrrnocouple in close quarters with such an arrangement, especially since the heat applied to melt the alloy passes rapidly into the calorimeter. However, there is a more important reason for having an alternative method of attachment. When hydrated or other easily decomposed substances are investigated the melting of a fusible alloy without heating the substance above its decomposition temperature presents a serious problem. One method is to cool the sample to a low temperature and then melt the fusible alloy so rapidly that the substance will (5) W. P. Giauque and C J. Egan, J. Chcm. Phrs., 6 , 45 (1937).
W. F. GIAUQUE not heat above the allowable teitiperaturc but such treatment is not desirable especially with a fragile calorimeter. A detachable therrtiorouplc well arrangement was nutic i r i the form of a holloiv brass plug with an outside cliainetcr
1'8".
Tav.
60.44 60. 94 61.43
63.15 57.02 59.60 62.16
83.48 88.75 94.07 99.49 104.95 110.45 115.84 121.35 126.77 132.28 138.02 143.38 148.64 154.08 159.62 165.05 170.36 176.03 L'@
ATapprox.
Scries 14.14 16.03 18.34 20.47 22.72 23. 90 26.80 29.56 32.57 38.51 41.92 45.35 47.85 49.27 49.93 50. 58 51.20 55.36 56.22 57.28 5 7 ,HB 58.35 5x. 9u 69 42 59.Y8 I
Tav.
CP
Series A 272.60 16.82 280.41 16.92 218.40 17.01 296.84 17.12 305.90 17.21 315.87 17.28 325.62 17.40 336.36 17.49 Series B i9.91 8.499 84.56 8.919 Series C 61.45 6.204 65.03 6.583 69.21 7.130 73.56 7.659 78.27 8.201
C'P
8.799 9.367 9,897 10.42 10.89 11.54 11.X 12.16 12.52 12.86 13.17 13.46 13.i 2 13.97 14.21 14.42 14.62 1'1.81
(!.ti:; . tis ,
ti4
.A 1 ,
s;
. Ji
--
.)J
X -., . d.2 52 .61 .Cil ,
50
49 , ,48
0.417 ,521 ,668 .822 1.000 1.105 1.376 1.668 2,025 2.884 3,471 4 1% 4 71; 5,108 .i 32(j .
C'P
14.99 15.16 15.32 15.47 15.62 15.75 15 90 16.03 16.14 16.27 16.36 16.48 16.56 16.67 16.76 16.85 16.94
la,..
L,
0.94 1.74 1.35 2.31 1.87 3.11 2.75 2.78 3.23 3.14 3.50 3,2,2 1.72
re., 181.61 187.07 192 85 198.5% 204.05 210.05 216 40 222.94 229.28 235.42 241.76 248.13 254.41 260.89 268.00 275.23 283.07
20 57 22.35 24 93 28.07 32.19 38.24 42.45 46.61 56.70 59.31 62.05 I
I
Series F 1.50 1.90 3.19 3.04 4.93 3.80 4.25 3.93 2.65 2.46 2.94 Series G
0.839 0.9iti 1.203 1.515 1.9'78 2.831 3.535 4.410 6.683 5.9:31 fj . 2 2 2
2 1 . $3; 1 (i;; :i ;j!l 21, l t i
0 904
27.01 i31.89 37.22 43. i 5 47.54
3.45
1 .4ijO 1 . 949 2 684 3 . TOO
-a>,,
,511 . ( !-1
2.3;
3 803
51.45
3 . X4l.i 5.903 5.954 5,992 6 . 037 t;. l l ( i ti. 177 6.351
3 1 , ti2
0 .1.1 1-1 . 1-1
5.55rJ 5 8O$l 3 581 .i (i9fj
5
i"I
2.87 Series E 2.65 5.716 2.45 5 963 2.63 0 2-43
31.77 61.91 52.04 52.18 52. 32 5%.45 52.59 52.96 53.56 54, I 5 (54. 7:3 33 . i 9 58.22 til. ti2
4.M 5.34
6 . 2:3 2.45
.14 I14 . 1*3 , 1:;
. I:( .14
. 59 .59
1.1:;1
4 . B3ti
5 ;;w 5 !I76 !i,05:j (i . 144
0.241 t i , 367 6,498 f i ,738 U ,422
5.878 5.617 5,527
VOl. 74
of 0 . 5 cin. and a length of 2.5 cm. The plug had a total This fitted into a hollow brass cylindcr, with taper of 4'. iuatching taper, which was permanently soldered to the calorimeter. The thermocouple junction was left pertnanently soldered in the well within the plug which could I)c pulled and held tightly in place by means of a screw and lock ivasher. A suitable dismantling tool was xieccssary t o force the joirit apart when the calorimeter as disassembletl. 111 the prcserit case there rvas no reason for iiot heating the anhydrous nickel chloride; however, it seemed desirable t o try out the detachable therrnocouple well during an esperiment which would not be more than temporarily inconvenienced in case the method failed. The performance was erit irely satisfactory. The Heat Capacity Data.-The observed heat capacity data arc given in Table I in the order in which they were ~neasurerl. The sample weighed 292.614 g. in z~ucuoand the molecular scight of YiClz WAS taken as 129.604. One defined calorie wits taken as 4.1840 absolute joules. Series D , 15, F and G were designed to give information about thc ittioinaly in the heat capacity curve which has a ruaximuin near 52°K. The shape of the curve is characteristic of a cooperative phenomenon and indicates that spcrial care should be taken with respect to slow attainment of equilibrium. I n order t o test this point the sample was cooled from 77°K. to liquid hydrogen temperatures as rapidly as possible before series D. The calorimetric arrangement is such as to prevent very rapid cooling and the rate was 0.7" rnin.-' a t 60°K. and 0.4" rnin.-l a t 40°K. Rleasuremcnts of series E were made immediately after those of series D by rerooling the sample from 64.5 to 55.7"K. at the average rate of 0.07" mixi.-]. Series F was obtained after the sample had been cooled slowlg from 60 t o -WOK. The cooling rate was 0.07" miti.-' a t 60 , 0.09" miti.-' at 50", and 0.05' min.-' at 40°K. Series G was measured for the express purpose of determining the low temperature part of the curve after slow cooling with suficiently small intervals to determine the shape of the anomalous curve. The cooling rate was 0.12" rnin.-l a t GO", 0.13' min.-l a t both 50 and 40'K.
The Anomaly in the Heat Capacity of Nickelous 1 shows the heat capacity data Chloride.-Figure between 40 and 6 5 O . K . Series G shows that the heat capacity rises rather rapidly on the low temperature side and then decreases rather abruptly from the maximum value. This is typical of curves obtained in various cases of cooperative phenomena associated with effects ranging from molecular rotation within crystals, ferromagnetism, antiferromagnetism or liquid helium 11. The magnetic susceptibility of nickelous chloridefi-8 increases with decreasing temperature until it reaches a maximum which coincides with the heat capacity anomaly within the limit of error of the magnetic data. Below the transition region, (frequently and incorrectly referred to as the "Curie point" or "Curie temperature") the susceptibility decreases slightly but is almost without teniperature coefficient. This means that magnetization a t very low temperatures may be thought of as :ti1 ordered mechanical process without an appreciable entropy change. This is consistent with the thermodynamic formula where I, H and S refer to the intensity of niagnetization, field strength and entropy, respectively. The entropy of NiClz should approach zero at very low temperatures unless some disorder of molecular
. 69 . Cis
,j,532
5.53i
(1;)
1.22 3.20 3.5s
5.ti12 5 . y"9 ti.198
(1925).
II. 11. R'olijer, Cornin P h y s . Lab. Uniu.
0.f
L e i d e n . KO. 173b
17) H. R. U'oltjer arid H. Kamerlingh Onnes, ibid., S o . 173c. ( 8 ) C . Starr. F. Bitter and A . R. K a u f m a n n , P k y s Rri* , 68, 977 (1940).
Sept. 5, 1952
4445
HEATCAPACITY OF ANHYDROUS NiC&
dimensions should be frozen in and there is no reason to suspect this. Neel,g Landaulo and others11J2 have suggested that spontaneously magnetized layers of atoms, oppositely oriented, alternate throughout the crystal and this appears to agree with all the known facts. Such an arrangement has been proved recently for the case of MnO and several other similar magnetic oxides by the very interesting neutron diffraction experiments of Shull, Strauser and Wollan. l a The calorimetric behavior of the sample indicates that the processes in the transition region are reversible to a high degree of accuracy. Several points between 50 and 57" in series D, which followed rapid cooling, do not quite fall on the curve within the TABLEI1 THE~MODYNAMXC PROPERTIES OF NiClz
7.0
6.0
5.5
6 5.0
4.5
0°C. = 273.16; mol. wt. 129.604; cal. deg.-l mole-' T 15 20 25 30 35 40 45 50 51 52 52.35 53 54 55 60 70
0.461 0.787 1.206 1 719 2.352 3.126 4.052 5 385 5 766 6.321 6.825 5.586 5.529 5.569 6.017 7 216
80
S.411
90
100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 2 50 360 270 280 290 298.16 300
CP
!I.J9:!
10.4i 11 32 12.07 12.72 13.29 13 78 14.22 14 60 14.92 15.24 15.52 15.76
1.5 (37 10.16 16 31 l(i 50 16.fi3 16.79 16.91 17.03 17.13 17.15
S O
0.168 0.344 0.563 0.826 1,140 1.503 1.923 2.413 2.523 2.640 2.683 2.757 2.860 2.982 3.464 4 479
