392
R. H. VALENTINE,G. E. BRODALE, AND W. F. GIAUQUE
Without postulating the reasons for the difference in crystal energies, it would seem possible, in view of the close similarity of structure of nickel(l1) and copper(I1) dimethylglyoximes, to evaluate this difference by comparing their heats of solution in an appropriately chosen solvent. The solvent of choice would be one in which solvation effects XTould be absent. The trend observed in the experimental data indicates that this condition has been closely approached. Comparing the standard entropy changes for both chelates in water, chloroform, benzene, and heptane, it is seen that A s t o becomes increasingly more positive. This quantity represents the entropy of fusion plus the entropy change on going from the ideal to the real solution. Since the latter term can be only zero or negative, the more positive AS2O is, the more nearly ideal the solution, and the increasing order indicates a decreasing order of solvent-solute interaction for these chelates. Furthermore, it is seen that the quantity A(AHo) = AHNi'' - AHcuo decreases going from water to chloroform solution, while A(AH0) for benzene and heptane, -3.4 and -1.8 kcal., respectively, are equal within experimental error. Regardless of any specific solute-solvent interactions in other solvents, it is difficult to imagine any significant difference in solvation for niclcel(I1) and copper(I1) dimethylglyoximes in such inert solvents as these two hydrocarbons. The fact that copper(I1) dimethylglyoxime exhibits a significantly smaller endpthermic heat of solution than the nickel chelate in both water and chloroform reflects metal-solvent interaction in keeping with the known ability of copper
Vol. 66
to exhibit a coordination number higher than four.20-23 I n this connection it is interesting to note that the authors found that addition of n-butylamine to an aqueous solution of the copper chelate permitted its quantitative extraction into chloroform, whereas a much smaller effect mas observed with the nickel chelate. An alternative explanation could be advanced in terms of a difference in ability to engage in intermolecular hydrogen bonding. There is, however, no additional evidence indicating that the copper complex would have more ability to participate in such interactions with water and chloroform than the nickel complex. It is concluded that, although the sources of the differences are not apparent, crystalline copper(I1) dimethylglyoxime is more stable by about 3 kcal. than crystalline nickel (11) dimet hylglyoxime ; also that the diff erence in hydration energies between these two compounds is more important than their difference in crystal energies in determining their relative solubilities in water. If a nickel to nickel bond does exist in the solid, it would have to be extremely weak. Acknowledgment.-The authors gratefully acknowledge the financial assistance of the U. S. Atomic Energy Commission. (20) J. Bjerrum and E. J. Nielson, Acta Chem. Scaizd., 2, 297 (1948). (21) R. M.Keefer, J . Am. Chem. SOC.,68, 2329 (1946). (22) H. B. Jonassen and J. R. Oliver, ibid., 80,2347 (1958). (23) H. A. Laitinen, E. I. Onstott, J. C. Bailar, Jr., and 8. Swann, Jr., ibid., 71,1550 (1949).
TRIFLUOROiS4ETHANE : ENTROPY, LOW TE;1IPPERATURE HEAT CAPACITY, HEATS OF FUSION AND VAPORIZA4TION,AND VAPOR PRESSURE' BYR. H. VALENTINE, G. E. BRODALE, AND W. F. GIAUQUE Low Temperature Laboratory, Departments of Chemistry and Chemical Engineering, University of California, Berkeley, Cal. Received July 31, 1061
The heat capacity of trifluoromethane has been measured from E ° K . to its boiling point. The measured values of the heat of fusion and melting point were found to be 970 cal. mole-' and 117.97"K. The heat of vaporization was 3994 cal. mole-' at the boiling point, which was determined as 190.97'K. The vapor pressure of the liquid has been. measured over the range 145 to 191OX. The vapor pressure of solid and liquid has been represented by an equation combining the (FO Hoo)/T function with an expression for gas imperfection. A table of calculated values of vapor pressure is given for both solid and liquid. The triple point pressure was calculated to be 0.046 cm. The heat of sublimation a t the absolute zero, The entropy of the ideal gas was evaluated from the experimental measurements as 57.18 gbs. AH00 = 6013.5 cal. mole-'. mole-' a t 190.97OK. and 1atm. The corresponding value calculated from molecular data is 57.23 gbs. mole-' at 298.15'K., SO = 62.05 gbs. mole-', compared to 62.00 experimental. Tables of C,O, SO, (8'0 - HoO)/T, and (Ho - Ho0)/T are given for solid, liquid, and gas. It is pointed out that CHFa is so easily purified in a fractionating column, and its melting point SO constant with the fraction melted, that i t would be useful to compare it with a gas thermometer as a temperature standard.
This paper describes a low temperature calorimetric investigation of fluoroform, C€IFs, for the purpose of determining its entropy. The boiling points of the series CH4, CH&', CH2F2,CHF3, CF, rise in a very marked maximum a t CH2F2, with reasonably smooth intermediate values for CH3F and CHF3. CHzFz could form two hydrogen
bonds, whereas CH,F and CHF3 could form only one each, which would be in line with the above boiling point order. The tetrahedral molecule perchloryl fluoride, C108F, has been found t o have residual entropy2 due to exchange of atomic positions. While this effect did not seem likely to occur in the present
(1) This work was supported in part by the National Science Foundation.
(2) J. K. Koehler and W. F. Giauque, J . Am. Chem. SOC.,80, 2658 (1958).
THERMODYNAMIC PROPERTIES OF TRIFLUOROMETHANE
March, 1962
case, especially in view of the low freezing point, there was at least a possibility of some disorder, such as the 2% detected in one crystalline form of carbonyl ~ h l o r i d e . ~We did not, of course, consider that the energy change involved would correspond to bhe breaking of a possibly weak hydrogen bond but rather the difference in energy between a hydrogen bond made in either of two very similar situations. However, the measurements have shown that the crystal approaches zero entropy a t limiting low temperatures. Calorimetric Apparatus and Temperature Scale.-The meamrements were made in Gold Calorimeter V.4 A gold resistance thermometer-heater was used for precision in temperature measurements and standard copper-constantan thermocouple No. 102 was used as a temperature reference. The thermocouple was checked a t the triple and boiling points of hydrogen, and the tr ple point of nitro en, by condensing these substances within the calorimeter. The thermocouple was in agreement at the triple point (63.15"K.) of nitrogen. and reauired a correction of about -0.08" at the Lripre point (13.\4'K.) and -0.06 a t the boiling point (20.36"K.) of hydrogen. 0°C. was taken as 273.15"K. One defined calorie-was taken as 4.1840 absolute joules. Sample of Trifluoromethane.-The trifluoromethane was a 3-llb. lot of 98,% pure material obtained from the hlatheson Co. It was distilled in a vacuum jacketed, silvered, low t,emperature helices filled fractionating column2 at a reflux ratio of about 200:l. The distillation pressure was about 55 cm. A 191.557-g. (in vacuo) sample was selected from the middle fraction and condensed into the calorimeter. The heat capacity curve below the melting point, while rising rapidly, gave no indication of premelting and we infer that the sample contained less than one part in one hundred thousand on a molal basis of liquid soluble-solid insoluble impurity. This probably is to be expected in the case of such a low boiling liquid with properties considerably different from those of likely impurities.
H:eat Capacity Measurements.-The heat capacity measurement's are given in Table I. The TABLE I HEATCAPACITY OF TRIFLUOROMETHANE 0°C = 273.15' K., mol. wt. CHFP = 70.018, 2.73582 moles in the calorimeter; heat capacity in gibbs mole-' cal. deg.-l mole-' TW, OK.
CIllSSl.
T,", OK.
Crneas.
Tav , OK.
=
defined
Crneas~
SerieR I 15.37 1.91 67.79 11.54 102.10 14.56 16.76 2.40 73.97 12.07 105.63 11.99 19.10 3.10 80.44 12.61 108.24 15.33 22.00 3.92 86.80 13.12 Series 111 24.28 4.65 92.57 13.64 122.73 20.25 28.13 5.70 97.85 14.11 129.82 20.10 30.39 6.24 102.73 14.60 137.93 20.0% 33.29 6.90 107.27 15 IS 146.32 20.00 36.25 7.54 111.29 15.78 154.60 20.03 39.32 8.07 115.02" 16.21 163.29 20.08 42.62 8.65 Series I1 171.95 20.20 46 57 9.24 80.97 12.59 1P0.82 20.37 50.99 9.81 89.01 13.27 189.33 20.59 56.00 10.42 93.33 13 69 61.72 11.18 97.87 14.13 Resistance thermometer temporarily strained. Heat capac,ity calculated from thermocouple.
393
under saturation pressure, was made on the results for liquid fluoroform in computing the smoothed values of Cp which are given later. The Melting h i n t and Heat of Fusion of Trifluoromethane.-The melting point mas observed 115 a function of the fraction melted. There was essentially no change. Trifluoromethane appears to be so easily purified that it would make a good temperature reference if it were compared directly with a gas thermometer. Other desirable characteristics in this connection are that it is non-flammable, non-toxic, and can be readily condensed into suitable pressure vessels for storage or transportation. The melting point observations are given in Table 11. The resistance thermometer gave very high precision, demonstrating the constancy of the melting point to O.O0lo, however, the resistance thermometer was calibrated in terms of the standard thermocouple and the temperature given is =kO.O5O in an absolute sense. TABLE I1 MELTINGPOINT OF TRIFLUOROMETHAXH 0°C. = 273.15" K. Time, min.
% Melted
T, OK., resistance thermometer
0 Heat added 39 3 59 Heat added 71 12 119 Heat added 214 49 354 49 1394 49 Accepted value
117.970 117.971 117,971 117.971 117.971 117.97 rt 0.05"K.
The heat of fusion was determined in the usual manner of starting heat input below the melting point and ending above it, with appropriate corrections for the JC, dT. The data are given in Table 111. TABLE I11 HEATOF FUSION OF TRIFLUOROMETHANE Cal. mole-', triple point = 117.97" K. TI.OK.
Tz, OK.
Total heat input (cor.)
116 806 117 149 117.079
121.472 122 990 120.884
1061.7 1083 2 1043 0
JCp d T
AH
89.1 973 115 8 967 74 3 969 Av. 970 A 3
Heat of Vaporization of Triflu0rornethane.The method used for measuring the heat of vaporization was essentially that described earlier,6 except that the 5-1. bulb which accepted the gas at constant pressure was not thermostated or calibrated. The amount of fluoroform was determined later by condensation in a stainless steel pressure vessel. The amounts of about 0.1 mole were observations were cont'inuous in Lhe sense t,hat determined to 0.02%. The weighings were coreach run began where the previous one encled, thus rected for the buoyancy of the weights. The heats t'here were no unobserved regions. An almost triv- of vaporization were measured within about 0.2' ial correction to the measured heat capacity, of the boiling point and an appropriate correction was made. The results are given in Table IV. (3) W. F. Giauque and J. B. Ott, J . Am. Chem Soc., 82, 2089 (1980). (4) J. B. O t t and W.
F.Giauque, ibid., 82, 1308 (1960).
( 5 ) W. F. Giauque and H. L. Johnston, ibid., 61, 2300 (1929).
R.H. T
394
T
~
kG. ~ E. BRODALE, ~ ~ ~ . ~~ N D'ITi. ~
F.~ GIAUQUE ,
T'ol. 6G
TABLE \ 'I TABLE IV TRIFLUOROMETHANETHERMODYNAMIC P R O P E R T I E S O F TRIFLUOROMETHANE GAS IS GIBBS MOLE-^ Cal. mole-1, boiling point = 190.97'41. 1 gbE = 1 defined cal. deg.-l Moles Time of heat
1IE.4T O F TrAPORIZATION O F
evap.
input, min.
0.11399 .11215 .08612
30 34 34
AH
3993.5 3992.8 3995.5 Av. 3994 r t 4
T, OK.
-(PO CPO
S O
-T HOD)/
(HO - HoQ)/ T
7.949 15 36.501 28,552 7.949 7 ,949 20 38.788 30.839 7.949 7,949 25 40,562 32.613 7.949 7.949 42.011 7.949 30 34.062 7,940 35 35,287 43.236 7.949 Thermodynamic Properties of Solid and Liquid 40 7.949 36.349 44.298 7.949 Trif3uoromethane.-The thermodynamic properties 7.949 45 37.285 45.234 7.949 of solid and liquid trifluoromethane are given in 7.949 46.072 7.949 50 38.123 Table V. The extrapolation below 15°K. could be 7.949 7.950 55 38.880 46.829 done only by recognizing Einstein contributions to 7,949 47.521 7.952 60 39.572 the heat capacity in the region 15-3OoK. in addi7.951 7.962 48,748 40.797 70 tion to the expected Debye term. The properties 7,953 49,812 7.986 41.859 80 of the gas are given in Table VI. The moments 7,959 42.796 50,755 90 8.031 of inertia were taken from Bernstein and Herzberg.6 100 7 ,970 51,605 8.102 43.635 1-40 = IBO= 81.08 X ICO = 148.67 X 110 7.986 8.198 44.395 52.381 lO-4O g. cm.2. The fundamental frequencies were 117.97 8.003 8.293 44.955 52,958 taken from Plyler and B e n e d i ~ t . ~ The values used 120 8.009 45,091 53.100 8.320 XTere 507(2), 700(1), 1150(1), 1152(2), 1372(2), 130 8.038 8,465 45.733 53.771 and 3031(1). 8.074 54,404 8.629 46.330 140 'Values of the thermodynamic properties of the 150 8.117 55,006 8.809 46,889 gas at higher temperatures are given at 100" in- 160 8.166 47,414 9.002 55.580 tervals t o 1500°K. by Gelles and Pitzer.s 8.221 9.205 170 47.911 56.132 8,282 48 382 56.664 9.416 180 8.347 57.179 9,634 48.832 TABLE V 190 8.353 57,228 9.657 48.875 190.97 THERMODYNAMIC P R O P E R T I E S O F SOLID AND LIQCIDTRI8.417 49,262 57,679 9.858 200 FLUOROMETHANE IS GIBBSMOLE-' 8.491 49.674 58.165 10.085 210 1 gbs. = 1 defined cal. deg.-' 8.569 50,071 10.317 58.640 220 - ( FO - HoQ)/ ('HO - HoQ)/ T T S O T, OK. CPO 8.650 50,454 10.552 59.104 230 0 668 0 174 0.494 2.601 15 8.734 10.789 50.824 59.558 240 1 400 ,384 1.016 3,370 20 8.821 60,003 11.029 51.182 250 2 315 .676 1.639 4.870 25 8.910 51.530 60,440 11.270 260 1,033 2,287 3 320 6.146 30 9,002 51.868 11.513 60.870 270 1,433 2.921 4 354 7.268 35 9.032 51.972 61,004 11.589 273.15 1.860 3.529 5 389 8.229 9,096 40 52.197 61.293 11.756 280 2,309 4.096 6 405 9,012 9.192 45 52.518 11.999 61.710 290 4,822 2.768 7 390 9.697 9.271 50 52.774 62,045 12,198 298.15 3.232 5.112 8 344 9,290 10,309 55 52.831 62,121 12.243 300 5.568 3.697 9 265 10.854 9.389 60 53.137 62.526 12.485 310 6.390 4.618 11 008 11.750 9.490 53,437 70 62,927 12,727 320 7.110 5.519 12 629 9.592 12,541 80 53.730 63.322 12.967 330 7,759 6,395 14 154 13,390 9.695 54,018 90 63.713 13.205 340 8.369 7.244 15 613 14.331 9.798 100 64.099 54.301 13.441 350 8,964 8.070 17 034 9.903 15,583 54.578 110 64.481 13.674 360 9.453 8.714 18 167 16,853 10,008 54,861 117.97 64.859 13.905 370 10.113 55,119 65.232 14,132 380 liquid 10,219 65.602 55,383 14.356 390 8.714 17.675 20,410 26 389 117.97 10.326 55.643 65,969 14.576 400 9,012 17.725 26 737 20,340 120 10,856 56.890 67.746 Xi. 617 450 10,442 17.912 28 354 20.094 130 11,380 58.061 69.441 500 16.554 18.064 29 839 11.775 20.016 140 The Vapor Pressure and Gas Imperfection of Trifluoro18.193 31 219 13.026 20,005 150 methane.-The vapor pressure %-as measured from 145'41. t o 18.308 14,204 32 512 20,049 160 the boiling point. h cathetometer was used t o compare a 18.413 15 317 33 730 20,176: 170 16-mm. diameter mercury manometer Kith a standard 18.517 16.372 34 889 20.369 meter bar. 180 Correctionti were applied for the meniecus depressions 18.621 17.376 35 997 20.632 190 and also for the weight of the column of fluoroform in the 18.631 17,471 36 102 20,662 190.97 tube loading t o the calorimeter. g was taken as 979.973lO (6) H. J. Bernstein and G. Herzberg, J . Chem. Phys., 16,30 (1948). (7) E. K. Plyler and W. S. Benedict, J . Research ~ V a t lBur. . Standa r d s , 47, 202 (1951). ( 8 ) E. Gelles and K. S. Pitzer, J. A m . Chem. Snc., 75, 5259 (1953).
(9) W. Cawood and H. S. Patterson, T r a n s . Faraday Snc., 29, 514 (1933). (10) Landolt-Bornstein-Roth, " Physikalische-Chemische Tabellen," Verlag Julius Springer, Berlin, 1923.
THERMODYNAMIC PROPERTIES OF TRIFLUOROMETHANE
March, 1962
for {,hi8location and go = 980.665 cm. sec.-2. of mercury was takenfrom the 1.C.T.''
The density
We have analyzed the vapor pressure data by a method which p r e ~ i o u s l y ~has ~J~ been shown to be very sensitive in showing errors in such measurements. For pressures near or below one atmosphere Berthelot's equation may be written in the approximate form
+
AFO/T =
PV = RT bP/TZ .-R In flu(]) = - R In P - b'P/T3 -
(1)
V m (1 - P ) / T (2) where b' (cal. deg.2 atm.-l mole-l) corresponds to b (cm.8 deg.2 mole'), f represents the fugacity of the gas, and a ( l ) the activity of the liquid referred to the standard state a t one atmosphere.
7
=
FO - Ho
-
FO-HO
395
AHoo
+T(3)
Let
where AHoo = 6,013.5, b' = -6.0 X lo5, V(1) = 48 ems3 estimated. The experimental observations are compared with equation 6 in Table VII. TABLEVI1 OBSERVED VAPORPRESSURES OF TRIFLUOROMETHANE T,OX.
145.348 158.080 164.511 169.182 172.824 176.426 179.349 182.000 183.882 185 917 187.889 189.551 191.177 ~
Inter. om. Poba.
Poalod.
2.04 7.10 12.34 17.88 23.45 30.25 36.88 43.93 49.51 56.12 63.34 69.92 76.89
2.03 7.13 12.36 17.85 23.43 30.26 36.88 43.94 49,53 56.20 63.33 69.91 76.93
where AH,' is an approximate value of the heat of TABLEVI11 sublimation of trifluoromethane at the absolute zero, obta,ined by ignoring gas imperfection. VAPOR PRESSURE OF SOLID AND LIQUID TRIFLUOROMETHANE The values of AHo' are plotted against P / T 2 , and the intercept should give the true AHo' for P sublimation at the absolute zero. The above plot Liquid is found t o be a straight line, which gives a reason6.49 x 10-2 3 62 x 10-94 120 40 able basis for the use of the form b / T 2 in estimating 3.04 x lo-' 130 2 09 x 1 0 4 7 50 temperature coefficients for use in correcting en1.11 140 6 51 X 60 tropy and heat content €or gas imperfection. 3.31 1 02 x 10-9 150 70 The va,lue of AHoO was found to be 6013.5 cal. 8.45 2.47 x 10-7 160 80 mole-1 for the heat of sublimation at OOK. 6' = 19.01 170 1 70 x 90 -6.0 X 105 cal. atm.-l degUz mole-'. 180 38.54 4 90 x 10-4 100 There is another method of evaluating the second 190 71.85 7 42 x 10-3 110 virinl coefficient a t the boiling point, where the Boiling point Triple point heat of vaporization was measured 190.97 4,56 x lo-* 76.00 117.97 V(d
AH
dT
= T x -dP
+
'(1)
=
7+ T2
RT
b
(5)
IS the heat of vaporization corresponds to the boiling point V(1)= 48 cm.3 estimated for CHFJ at 190.97"K. (b.p.) b' is found to be -6.67 X 106 cal. atm.-l deg.-' mole-'
AH
Although the above value b' = -6.67 X lo5 has less uncertainty in its evaluation, it depends on the assumed 1 / T 2 form. We prefer to use the value -6.0 X 106 obtained from the vapor pressure data, because jt is more intimately connected with temperature coefficients, and the value of b which is obtained by that method will to some extent correct for an error in the assumed temperature dependence. Values of the vapor pressure at even temperatures for solid and liquid are given in Table VIII. They were calculated from the equation R In P(&trn) =
---
(I 1 ) "International Critical Tables," Vol. 2, MoGraw-Hill Book Co., New York, N. Y., 1920, p. 457. (12) R. H. Busey and W. F. Giauque, J. An. Chem. Soc., 76, 800 (1953). (13) R. H. Sherman and W. F. Giauque, zbzd., 7'7, 2154 (1966).
The Entropy of Trifluoromethane from Calorimetric Data.-A summary of the entropy calculation is given in Table IX. TABLEIX THEEXTROPY OF TRIFLUOROVETHANE IN GIBBSMOLE-' 1 gbs. = 1 defined cal. deg. -1 0-15°K. (extrapolation) 0 67 15-117 97" (graphical integration) 17 50 Fusion, 970/117.97 8 22 117.97-190.97 (graphical integration) 9 71 Vaporization, 3994/190.97 20 91 Entropy of CHF3gas at b.p., 190.97'K. 57 01 Corr. to ideal gas, AS == -2b'P/Ts 0 17 Entropy of ideal CHFa gas a t b p. 57 18 gbs. mole-' So from molecular data at 190 97°K. 57 23 At 298.15'K., Sornol data. 62.045 Soexp= 62 00 gbs. mole-l ~
~
In computing the value of S o given as experimental at 298.lS0K., the increment above the boiling point has been taken from the spectroscopic data. We thank John P.Chan for assistance with the measurements and calculations.