physical chemistry - ACS Publications - American Chemical Society

---

THE JOURNAL OF

PHYSICAL CHEMISTRY Registered in

U.S. Patent Ofice

@ Copyright, 1968, by the A m e r i c a n Chemical Society

VOLUME 72, NUMBER 10 OCTOBER 15, 1968

Vaporization Characteristics of Zinc Chloride, Bromide, and Iodide by Donald W. Rice and N. W. Gregory Department of Chemistry, University of Washington, Seattle, Washington 08106

(Received May 17, 1968)

Comparison of the results of Knudsen and torsion-effusion studies of the equilibrium vapor pressures of ZnClz, ZnBrz, and ZnIt, respectively, does not indicate the presence of significant amounts of dimeric, or higher polymeric, species; it is concluded that such species constitute less t8han10% of the saturated vapors at pressures below 10-1 mm. Entropies derived from the vapor pressure data for the monomeric molecules are compared with those predicted from currently reported molecular constants. Condensation coefficients of all three solid phases and of liquid zinc chloride are near unity. Keneshea and Cubicciotti have recently reported equilibrium vapor pressures of liquid zinc chloride and zinc bromide, respectively. By comparison of quasistatic and transpiration data and from qualitative mass spectrometric evidence, they conclude that the saturated vapors are principally monomeric, but include small arnounts of dimeric molecules, ranging from, for example, 14% a t 412" to 4% at 575" for zinc chloride. Because of the relatively large uncertainty in the dimer partial pressures, the thermodynamic properties derived for the dimer molecules have large uncertainties. The total vapor pressures observed for zinc chloride are in good general agreement with those reported by other w~orkers~-~ but the conclusions about the molecular composition are not in good agreement. Cubicciotti and Eding note that thermodynamic properties of the monomeric vapor molecules, calculated from vapor pressure data, do not agree well with those predicted from molecular constant^.^ Because of these and other questions, we have made a Knudsen and torsion-effusion study (200-350") of the vaporization characteristics of the zinc halides. Comparison of such data provides information about the average niolecular weight of the vapor (the 412-575' results1 predict that Xzn2clt = 0.33 in the effusion range) I n addition, comparison of vapor pressures a t lower temperatures with those at higher temperatures allows one to assess the reliability of the heats and enI

tropies of vaporization deduced from van't Hoff treatments of the independent sets of data and hence to assess better the reliability of the molecular constants assigned to the zinc halides.6 The only previous report of effusion work on these solids is a Knudsen study by Niwa.' His results for ZnC12 and ZnBrz do not agree well with an extrapolation of the liquid-phase results. Zinc chloride is one of few substances with a vapor pressure at the melting point in the range convenient for effusion studies. This allows one to observe the effect on effusion steady-state pressures of a marked change in sample surface area (by melting a finely divided sample), which provides information of interest about the condensation coefficient.

Experimental Section The torsion-eff usion apparatus has been described p r e v i o u ~ l y . ~Tungsten *~ wires, 1 or 2 mils in diameter (1) F. J. Keneshea and D. Cubicciotti, J . Chem. Phys., 40, 191 (1964). (2) H.Bloom and B. J. Welch, J. Phys. Chem., 62, 1594 (1958). (3) H.Bloom, J. O'M. Bockris, N. E. Richards, and R. G . Taylor, J . Amer. Chem. SOC.,80, 2044 (1958). (4) H.I. Moss, Doctoral Dissertation, Indiana University, 1960. (6) D. Cubicciotti and H. Eding, J . Chem. Phys., 40,978 (1964). (6) L. Brewer, G. R. Somayajulu, and E. Brackett, Chem. Rev., 63, 111 (1963). (7) K. Niwa, J . Fac. Sci. Hobkaido Univ., Ser. ZZI,3, 17 (1940). (8) R.J. Sime and N. W. Gregory, J. Phys. Chem., 64, 86 (1960).

336 1

DONALD W. RICEAND N. W. GREGORY

3362 for different sensitivities and approximately 70 cm long, served as suspension fibers. Four Pyrex effusion cells, ca. 1.4 cm i.d. and 4 cm long, with two nearly equivalent orifices on opposite sides and ends, were used. The total orifice areas were: cell 1, 7.18 X cm2; cell 2, 90.9 X cm2; cell 3, 0.978 X 10-8 cm2; and cell 4, 2.05 X lo-* cm2. Temperatures of the effusion cells were taken to be those indicated by calibrated thermocouples in an equivalent stationary cell, placed adjacent to the moving cell inside the radiation-shielded vacuum chamber. This assumption was based on an actual comparison, under virtual operating conditions, of the temperatures at various points on the two cells. The temperature of the reference cell was regulated (*0.5’) with a Leeds and Northrup Speedomax H controller. The torsion constants, k, relating pressure to angular displacement, 6 ( P = k6), for the various combinations of cells and fibers were determined by calibration with zinc. At effusion temperatures, the vapor pressure of zinc is comparable with that of the zinc halides and has been measured by several independent workers. We have made an additional Knudsen determination and obtained results in good accord with those of Barrow, et ~ 1 . ’ ~ Pyrex Knudsen cells of standard design were used.” To ensure temperature uniformity, thermocouples, placed at various locations on the cell walls, were monitored during the 5-18-hr runs. At the conclusion of a run, effusates, collected on a replaceable liquid nitrogen cooled finger, were removed after admission of dry argon and dissolved. The amounts of effused zinc were determined by EDTA compleximetric titration.12 The orifice areas (in square centimeters) and Clausing factors of the three cells were, respectively: cell 1, 14.07 X 0.99; cell 2, 2.24 X lob3,0.97; and cell 3, 4.20 X lo+, 0.98. Cells were ca. 3.8 cm long and had a 2.2 cm i.d. Orifice areas were measured with a microscope and a camera lucida. Samples of Baker’s Analyzed reagent grade zinc chloride, Fisher’s Certified reagent grade zinc bromide, and Fischer’s zinc iodide were vacuum dried and purified by vacuum sublimation prior to use; the composition was verified by analytical and mass spectrometric methods. In the torsion studies the samples (0.52.0 g) were degassed by raising the temperature slowly to the highest point of interest. Temperatures for subsequent measurements were selected at random. Equilibration occurred rapidly; several points could be obtained in 1 hr. A change in pressure (at constant temperature) as a function of time was never observed, even €or runs in which 40% of the sample had effused. I n the Knudsen apparatus, the samples were normally replaced after four measurements. Qualitative mass spectrometric studies were conducted with a special Knudsen cell and heating attachment, designed to fit into the entrance port of an MS-9 AEI mass spectrometer. Although the orifice The Journal of Physical Chemistry

(area ca. 5 X cm2) was placed 0.5 mm from the heated source block, a narrow 1.6-cm channel, 0.15 cm i.d., through the latter permitted only an estimated 5% of the effusing molecules to reach the ionization chamber without striking the walls of the channel. The device was not equipped with a shutter.

Results and Discussion Torsion-effusion steady-state pressures are shown graphically in Figure 1. For each combination of cell and fiber, data were summarized by an equation of the form shown in Table I. Constants a and b were determined by a least-square treatment and c was determined from the estimated A ~ P ’for the vaporization process. The heat capacities of solid ZnClz and solid ZnBrz have been measured by Cubicciotti and Eding, 17.6 and 18.4 cal deg-l mol-’, respectively, a t our mean temperature of 560°K.6 Values for the monomeric vapor molecules were estimated from molecular constants as 14.5 and 14.7. Values for ZnIz were estimated by comparison. The vapor pressures assumed to be equilibrium values are given by the equations so labeled in Table I and are shown as solid lines in Figure 1. These were obtained by plotting, at a given temperature, the reciprocal of the steady-state pressures for the various cells (based on the equations summarized in Table I) against the orifice areas and extrapolating to zero orifice area. Because vapor pressures over only a limited temperature range above the melting point of zinc chloride fell in the effusion range and because values from all cells where measurements could be made were virtually indistinguishable, equilibrium vapor pressures over liquid ZnClz were taken to be those measured with cell 4. This was the only cell for which the orifice area-fiber combination permitted an appreciable liquid range to be scanned. The steady-state torsion pressures for each of the compounds were not very sensitive to a change in cell orifice area. For example, only ca. a 10% decrease was observed for the chloride when the area was increased by a factor of 100. Hence the apparent condensation coefficients, a, as estimated from the ap, P e q u i l and proximationP,,,il = P,,[1 ( a / A a ) ] where P,, are the equilibrium and steady-state pressures, and a and A , the orifice area and cell cross-sectional areas, respectively, must be near unity. Values of CY = 0.6 0.3 were obtained for the chloride and bro-

+-

*

(9) R. J. Sime, Doctoral Dissertation, University of Washington, 1959; University Microfilms, LC59-3346, Ann Arbor, Mich. (10) R. F. Barrow. P. G. Dodsworth. A. R. Downie. E. A. Jeffries, A. C. Pugh, F. J. Smith, and J. M. Swinstead, Trans. Faraday Soc., 51, 1354 (1955). (1 1) J. L. Margrave, “Physicochemical Measurements at High Temperature,” J. O’M. Bockris, J. L. White, and J. D. MaoKenzie, Ed., Butterworth and Co. Ltd., London, 1959, p 225. (12) H. A. Flashka, “EDTA Titrations,” Pergamon Press Inc., New York, N. Y., 1959, p 76.

3363

VAPORIZATION CHARACTERISTICS OF ZINC CHLORIDE, BROMIDE, AND IODIDE

-I

n

-2

E E

c0

CT,

-0

-3

-4

I

1.60

I

1.7 0

I

1.80

I

1.90

I

2.00

1

2.10

I 2.20

I 000/ T(O K ) Figure 1 . Torsion-effusion results. Cell number and fiber diameter, in mils, respectively: A, 4 (1); X, 4 (2). The solid lines represent the pressures derived as equilibrium values.

mide; that for the iodide appears somewhat smaller, 0.3 f 0.5. No appreciable change in a with temperature could be detected. If the effective areas of the solid samples are much larger than the assumed cell cross-sectiona,l area, the actual condensation coefficients will be smaller than these values. This possibility was tested for zinc chloride by comparison of the steady pressures over a finely divided powder and those over the same sample after it had been melted in the effusion cell. No change was seen; neither did a sudden change in the steady-state pressure occur as the sample was taken through a temperature range including the melt-

0 , 1 ( 1 ) ; 8, 1 (2); 0, 2 (2);

+, 3 (1);

ing point. The large uncertainties in the apparent condensation coefficients reflect the insensitivity of P,, to variations of orifice area when a is near unity. The Knudsen-effusion results are also summarized in Table I and are shown graphically in comparison with the torsion-eff usion equilibrium lines in Figure 2. The Knudsen data agree well with torsion pressures if monomers are assumed to be the only species of importance. The equations in Table I are based on this assumption. The data shown in Figure 2 represent a composite of results from all the cells; no systematic dependence of Knudsen steady-state pressures on oriVolume 78, Number 10 October 1968

w.

DONALD RICEAND N.

3364

Table I : Torsion-Effusion and Knudsen Results (Constants a, b, and c for the Equation log P(mm) Cell

Fiber diameter,

no.

mils

=

Temp range, a

b

OK

K n u d sen

ZnClz(s) = ZnC12(g), c = 1.56, ACpo = -3.10 cal molb1 deg-1 1 15.770 4 0.354 7751 =!= 180 495-555 2 15.389zk0.313 7509 i 178 540-590 2 15.612i0.276 7666 i 148 500-565 1 15.449A0.383 7535 f 215 535-590 1 15.976 f 0.463 7850 f 257 520-580 Equil (extrapolated) 16.194 f 0.284 7967 f 219 495-590 1-3 16.268f0.174 8047 zk 110 535-590

Torsion

4

Torsion

1 1 2 4 4

Torsion

K nu dsen Torsion

Knudsen

1 1 2 3 4

l J

ZnCll(l) = ZnClz(g), c = 4.83, ACpo = -9.6 cal mol-ldeg-l 2 25.942 f 0.274 8399 i 167 591-640 ZnBrz(s) = ZnBr2(g),c = 1.88, ACpa = -3.74 cal mol-’ deg-1 1 17.498zk0.473 7848 i 235 485-520 2 17.510i0.375 7905 f 206 525-580 2 16.274 f 0.266 7250 & 137 485-540 1 17.193i0.397 7704 f 212 515-555 2 17.810f0.286 8040 f 170 550-610 Equil (extrapolated) 17.366 i 0.240 7793 i 153 485-580 17.583 i0.174 7963 i 110 525-610

ZnIz(s) = ZnIz(g), c = 2.21, ACpa = -4.4 cal deg-1 mol-’ 1 18.121 f 0.347 7444 i 167 465-500 2 18.089i0.348 7431 4182 510-545 2 18.350zk 0.240 7623 =t118 480-520 2 19.057i0.324 7944 f 180 530-580 Equil (extrapolated) 18.397 f 0.306 7578 f 219 465-580 11 2 18.733 i0.175 7785 4 110 485-555

1 1 2 4

fice area (varied by a factor of 7) was seen. Hence all results were weighted equally in a least-squares treatment to derive the equations given in Table I. For each of the three substances, the difference (Figure 2) in the vapor pressures obtained by the two independent methods appears within experimental error. The small apparent difference may in part result from the extrapolation procedure used for the torsion data which was not practical for the Knudsen data (orifice areas could not be varied over such a wide range). If, on the other hand, the differences were assumed to be real, it would be necessary for the zinc halides to decompose to form free halogen or to disproportionate to form molecules in which zinc has an oxidation state greater than 2 (incompatible with the known chemistry of zinc). We observed no evidence of decomposition of the zinc halides in our experiments and it is not expected on thermodynamic grounds.13 Zn-ZnClz mixtures were found to have vapor pressures equal to the sum of those of pure Zn and ZnClz; i.e., no evidence for formation of significant amounts of ZnCl(g) was seen. Qualitative mass spectrometric analysis of the equilibrium vapor between 180 and 330” gave no evidence for the presence of significant quantities of dimer molecules for any of the three halides. The apparatus available, although not suitable for quantitative determinaThe Journal of Physical Chemistry

a

w. GREGORY

- bT-1 - c log T ) Torsion constant (k), mm radian-’

7.43 x 10-3 1.46 x 10-l 1.34 x 10-2 7.71 x 3.72 X

7.12

x

10-1

8.05 x 1.48 X 10-l 1 . 3 8 x 10-8 3.72 x 7.75 x 10-1

6.67 x 1.35 x 1.39 X 7.78 x

lo-* 10-l

lom2 10-l

tion of the vapor composition, should have easily detected ions derived from dimeric parent molecules if these were present in amounts exceeding a few per cent. The electron ionizing potential was varied between 8 and 70 eV with no discernible effect on the apparent vapor composition. Appearance potentials were determined for ZnX2+ (Cl, 12.1 f 0.5 eV and I, 9.8 f 0.5 eV) and ZnX+ (C1, 13.2 f 0.5 eV and I, 11.6 f 0.5 eV) ions; the relative intensities of ZnXz+, ZnXf, and Zn+ were ca. 100: 11:6 in all cases. We conclude that the amounts of dimer in the saturated vapors in each case do not exceed 10% in the effusion vapor pressure range. This amount is substantially less than predicted by the least-squares equation given by Keneshea and Cubicciottil for the dimer equilibrium vapor pressures above liquid ZnCh and ZnBrz. However, our observations are easily within the uncertainty of the data; if, for example, we use the largest enthalpy permitted by their prescribed error limits and their reported entropy, only 3’% dimer is predicted for zinc chloride a t our temperatures. Thermodynamic Properties. Thermodynamic properties derived from a van’t Hoff treatment of the equilibrium data summarized in Table I are given in Table I1 (13) L. Brewer, Paper No, 7, “Chemistry of Miscellaneous Materials,” L. L. Quill, Ed., McGraw-Hill Book Co., Inc., New York, N . Y . , 1960.

VAPORIZATION CHARACTERISTICS OF ZINC CHLORIDE, BROMIDE, AND IODIDE Table I1 : Thermodynamic Properties for the Vaporization Processes ZnXds) ---------ZnCl-Knudsen

T,OK

560 3.1 35.1 4 0 . 5 38.6rt0.8 35.840.7 4 0 . 3 f 1.1

AZ,,', ea1 mol deg-l ART', kea1 mol-' ASTO, cal mol deg-l ARozsg, kcal mol-l ASOZOS,cal mol deg-l

Torsion

Knudsen

Torsion

560 3.1 34.7 f 1 . 0 3 8 . 2 4 1.3 35.441.2 39.9 4 1 . 7

560 -3.7 34.3 f 0 . 5 39.9 f 0 . 8 35.0 -i: 0 . 7 41.5 4 1 . 1

560 -3.7 33.6+0.7 39.0 4 1.1 34.3 f:0 . 9 40.6k1.4

560 -4.4 33.3rt0.5 40.6 rt 0 . 8 34.040.8 42.1 f 1 . 3

560 -4.4 32.4ztl.O 39.2 -f 1 . 4 33.1k 1.3 40.7 & 1 . 9

ti -2 2 0

Q

- 3.

I.'SO

7-

Knudsen

E E

1!eo

ZnIz--------

~--ZnBr~------

c

I

= ZnXdg)

Torsion

-I

1.70

3365

2!00

2110

IO OO/T(% 1 Figure 2. Knudsen effusion pressures (data points and associated. line) in comparison with the torsion-effusion equilibrium lines (---): 0, ZnClz; 8, ZnBrl; 0 , ZnIz.

together with estimated ACpo values. Only small changes are observed as the molecular weight of the halogen increases; the enthalpies of sublimation decrease slightly and the entropies increase. This behavior is intermediate between that of ionic alkali halides and covalent halides with only van der Waals interaction between molecules. The enthalpies of sublimation for ZnClz and ZnBrz predicted from relative enthalpy data5 and the previously reported vaporization data for the liquids, 35.5 and 34.2 kcal mol-', respectively, agree well with our values. If the total pressures reported in the liquid range' are attributed entirely to monomers and extrapolated to our temperakure interval, the values

predicted for the sublimation pressures of the solid are also in good agreement with our observations. On the other hand, our pressures are considerably higher than those reported by Niwa;' in his work, vapor pressures for solid ZnC12 are reported a t temperatures 10" above the known melting point, 318", which suggests that the measured temperatures may have been in error. The standard entropies of the ZnX2(g) monomers at 298"1i, calculated from entropies reported for the solid phases and the sublimation entropies we observed, are considerably greater than those calculated from molecular constants reported for these molecules : 66.6, 73.6, and 78.1 eu vs. 62.2, 67.7, and 71.7 eu, respectively. The reason for this discrepancy has been of concern to us. The standard entropies of the solid phases are based largely on emf measurements of Bates. l4 Ishikawa, et uZ.,l6 have performed similar emf experiments, but their results do not agree well with those of Bates. We have used Kelley's recommended values of 26.5 and 32.5 eu for ZnClz(s) and ZnBrz(s), respectively, which are in substantial agreement with Bates' results.'6 Kelley's value for ZnIz(s) is 35.0 eu, whereas Bates found a value of 38.5 eu; we have elected to use 36.7 eu, an intermediate value derived from Latimer's rule with CdI, as a reference substance. Bates,14 in a similar emf study of CdIz, obtained results in reasonable agreement with the entropy evaluated from low-temperature heat capacity data;" hence the entropy for CdIz(s) seems relatively well established. It is unlikely that uncertainty in the entropies of the solid phases at 298°K can account for the entire discrepancy. The molecular constants proposed for the monomeric vapor molecules have been tabulated by Brewer, et u Z . ~ The only serious question about them appears to be the value of the important doubly degenerate bending frequency, v2. Biichler, et QZ.,'~ reported an absorption (14) R. G. Bates, J . Amer. Chem. Soc., 61, 1040 (1939). (15) F. Ishikawa, G. Kimvra, and T. Murooka, Sci. Rep. Tohoku Univ., Ser. A., 21, 455 (1932); F. Ishikawa and T. Yoshida, ibid., 21, 474 (1932). (16) K. K. Kelley and E. G. King, Bulletin 592, U. S. Bureau of Mines, U. S. Government Printing Office, Washington, D. C., 1961. (17) A. S. Dworkin, D. J. Sasmor, and E. R. Van Artsdalen, J . Amer. Chem. Soc., 77, 1304 (1955). Volume 72,Number 10 October 1968

3366 band for ZnCl2 with a frequency of 295 cm-' which they assigned to v2. Matrix isolation and diffusion experiments by McNamee suggest that this frequency may be due instead to a polymeric species.lg If, as pointed out by Cubicciotti and Eding, v2 for ZnC12(g) is of the order of 100 cm-I rather than 295 cm-' (similarly, from our work, 58 cm-l instead of 225 cm-' and 38 cm-' instead of 180 cm-l for ZnBr2(g) and Zn12(g), respectively), the entropies of the monomeric gas molecules predicted by the two methods would be in good agreement. These lower values are close to but slightly larger than those reported for the mercuric halides.20* 21 It is possible that the discrepancy may be due to errors in the sublimation vapor pressure data. However, for this to be the only source, a temperature error of from 5 to 6.5" (ZnC12-ZnIJ, developed over a ca. 60" interval, would be necessary. Considerable care was taken in the experimental design to prevent such temperature differentials and to verify that such temperature differentials did not develop. The agreement between the Knudsen and torsion methods is good; it is unlikely that errors of the same magnitude and direction would develop in the independent experimental setups. Furthermore, the completely independent transpiration work of Keneshea and Cubicciotti is also in agreement with our results; the effect of the difference in the estimate of the amounts of dimer does not affect this correlation significantly. I n addition, the zinc calibration vapor pressures were obtained in the same Knudsen apparatus over virtually the same temperature interval and with the same temperature-measuring and -controlling devices; these results agree well with Barrow and others and show no sign of a temperature-dependent error.10'22 Our results for the vapor pressure above liquid

The Journal of Physical Chemistry

DONALD W. RICEAND N. W. GREGORY ZnC12 give an enthalpy of vaporization (mean temperature, 615°K) of 32.5 kcal mol-' and a standard entropy of vaporization of 34.2 eu, which agree well with the results of Keneshea and Cubicciotti' and of Bloom, et a1.,2,ameasured by transpiration and boiling point methods at higher temperatures. The calorimetrically measured heat of fusion for zinc chloride a t its melting point (591°K) is 2.45 kcal a value appreciably lower than those reported for other similar compounds.6 The heat of fusion obtained from the van't Hoff treatments, our vaporization, and sublimation data is 2.1 & 1 kcal mol-', in good agreement with the calorimetric result.

Acknowledgment. We acknowledge with thanks the financial support for this work from the National Science Foundation, Grant GP-6608~. (18) A. Btichler, W. J. Klemperer, and A. G. Emslie, J. Chem. Phys., 36,2499 (1962). (19) R. W. McNamee, Lawrence Radiation Laboratory Report UCRL-10451, University of California a t Berkeley, Berkeley, Calif., Sept 1962. (20) M. Wehrli, H e h . Phys. Acta, 11,339 (1938); 13, 153 (1940). (21) H. Sponer and E. Teller, Rev. Mod. Phys., 13, 75 (1941). (22) After this article was submitted for publication, it was learned, by private communication, that Professor 0. Schnepp (University of

Southern California), A. Loewenschuss, and A. Ron have recently completed some matrix isolation studies of the zinc halides. (0. Schnepp, A. Loewenschuss, and A. Ron, J . Chem. Phys., in press.) They found bending frequencies for ZnClz and for ZnBrz close to those estimated by Cubiccotti and Eding and those suggested by our results: ZnCla, 103 and 101 cm-1; ZnBrz, 76 and 73 om-'; and ZnIt, 62 cm-1. These values for w, together with the other molecular constants cited above, lead to standard entropies (298OK) for the vapor molecules of 66.1, 71.9, and 75.9 eu, respectively, which, when taken with our entropies of sublimation based on the torsion-eff usion studies (Table II), yield values for the solids of 26.2, 31.3, and 35.2 eu, respectively. These results now agree well within experimental uncertainty with the values based on emf data (26.5, 32.5, and (estimated) 36.7 eu). We thank Professor Schnepp for providing us with this information in advance of publication.