1115
GASEOUS EQUILIBRIUM BETWEEN CADMIUM AND CADMIUM(II) CHLORIDE
-
A Spectroplhotometric Study of the Gaseous Equilibrium between Cadmium and Cadmium(I1) Chlloride'
by Buddy L. Bruner and John DL Corbett Institute for Atomic Research and Department of Chemistry, Iowa State University, Ames, Iowa (Received November 92, 1963)
~~
A Beckman D U spectrophotometer has been modified to permit investigation of gaseous equilibria a t high temperatures, and the absorption spectra of the Cd-CdClz system and its components have been measured between 545 and '780'. The mixture displays new bands due to the moi~ochloridein addition to the transitions found with the pure components. CdCl,(g) = 2CdCl(g) was found to be 34.15 f 0.5 kcal. AH'looofor the reaction Cd(g) mole-' and, with the third-law entropy change, AF'looo is 23.8 f. 0.5 kcal. mole-'. Utilizing the solubility data for Cd in CdG12(1), 83.9 and 32.6 kcal. mole-' are obtained for AH oloooand AFolooo,respectively, of the presumed reaction Cd2C12(1) = 2CdCl(g). Values of 2.11 f 10.03 and 0.027 0.01 e.v. for D o o of CdCl and Cdz, respectively, are also obtained. Andysis of the CdCl vibrational bands yields 334.5 cm.-' for we" and 1.3 cm.-l for z,"~," of the 2.Z ground state, and 397 cm.-l for we' and 0.9 cm.-l for xe'we' of the zlI,,2excited state. The 3181-A. band is not observed and its previous assignment to CdCl is concluded to be erroneoun.
+
*
Introduction Although the solution of cadmium in its chloride has received considerable attention from investigators with regard to the nature of the solute species,2vavery little has been established about the gas phase equilibrium. Tarasenkov and Skulkova" reported that CdClz was significantly less volatile in a C12 stream than in one of Nz, and Walters and Barratt6 assigned to CdCl the two new band systems that they observed in the spectrum of the cadmium metal on introduction of CdClz a t 500 to 800'. These two observations suggest that a stable subhalide may exist under equilibrium conditions. The present paper reports a direct determination of the nature of the gas phase species above the solution of Cd in CdClZand of the thermodynamic parameters of the equilibrium. Since the objective of this investigation required both identification and measurement of the equilibrium species, preferably over a range of temperature and concentration, the electronic absorption spectra of the system were studied. The use of absorption spectra not only provides a means of satisfying the minimum requirements of the investi-
gation but in principle permits simultaneous and essentially independent measurements of all the species present. Although in practice only the relative concentration of the product CdCl could be measured reliably, even this would have been achieved only with difficulty by conventional pressure-measuring methods.
Experimental Materials. For accurate, unambiguous, and quantitative spectral measurements, it is necessary to use chemicals of high purity. Therefore, the materials used herein were transferred only under vacuum or in an argon-filled drybox after purification. Cadmium metal (A. D. MacKay, 99.999%) was fused under vacuum, cooled, and the resulting surface impurities (1) Contribution No. 1411; work was performed in the Ames Laboratory of the U. S. Atomic Energy Commission. (2) Summarized by J. D. Corbett, W. J. Burkhard, and L. F. Druding, J . A m . Chem. SOC.,8 3 , 76 (1961). (3) L. E. Topol, J . Phys. Chem., 67, 2222 (1963). (4) D. N. Tarasenkov and G. V. Skulkova, Zh. Obshch. Khim., 7 , 1721 (1937). (5) J. M. Walters and S. Barratt, Proc. Eooy. Soc. (London), A122, 201 (1929).
Volume 68, Number 6
M a y , 1964
BUDDY L. BRUNER AND JOHN D. CORBETT
1116
(primarily CdO) dissolved in dilute "0%. The cleaned buttons were then distilled twice a t 415' under dynamic vacuum torr) in a Pyrex tube that had been degassed a t 500'. The CdC12 was prepared by direct chlorination of the metal2 and vacuum sublimed twice a t 550' after being chlorinated for 1 hr. a t 500'. After purification, both salt and metal were sublimed directly into the sample cell or into small glass fingers which were subsequently sealed off for storage. Apparatus. A Beckman DU spectrophotometer equipped with a 1P28 photomultiplier detector and a Hz source was modified for measurements on gas phase samples a t high temperatures by rearranging the components of the instrument to provide the "reversed optics'' necessary to minimize the effects of furnace radiation above 500°.6*7This and the enlargement of the sample compartment to contain the required furnace made i t necessary to modify the cell compartment mounting block and the light source arrangement. The mounting block was altered to accommodate the detector in a manner essentially identical to that described by Gruen and MeBeth.* The source was modified by removing the condensing mirror and cutting a large hole in the housing wall so that the light beam entered the cell compartment directly from the source. Since this produced a divergent beam over the 58.5 ern. between the source and the monochromator slit, substantially reducing the effective intensity, two fused silica condensing lenses (Perkin-Elmer, 35-mm. diameter, 8-mm. thickness, 55-mm. focal length) were then mounted in adjustable holders a t each end of the sample compartment to focus the beam. The wave length calibration was checked with a Hg discharge tube. The furnace compartment for the high temperature work was essentially a large box constructed from boiler plate, angle iron, and sheet metal with exterior Cu coils for water cooling, and was mounted to the monochromator through the alignment pegs of the modified mounting block. The source was mounted on the other end of this compartment. Access to the furnace and sample was attained by unbolting and removing one side and the top of the compartment. The cells were of cylindrical, fused silica construction (Pyrocell, 100-mm. path length, 22-mm. diameter) with a finger on the sample cell perpendicular to the light pa,th a t the longitudinal midpoint. A 16411. (40 cm.) hinged muffle furnace (Heavy Duty Electric) was rewired so that the 4-in. (10 cm.) end sections could be controlled separately from the central sections to allow compensation for the heat loss a t the ends of the furnace. A set of appropriately drilled firebrick plugs was also used within the end sections for the same The Journal of Physical Chemistry
purpose. A hole in the lower, central heating section permitted the installation of a third heater that was used to regulate the temperature of the liquid sample, if any, inside the finger of the cell during spectral measurements. A stainless steel block within the central section of the furnace acted as a holder for the absorption cells and as a body of large heat capacitk to minimize temperature variations. The block was constructed in halves to permit easy access and was drilled out to accommodate both a reference cell and a sample cell with its finger. Grooves in the top of the block and adjacent to the sample cell held thermocouples for control of the central heating section and for the determination of the temperature along the length of the cell. Cells with SO-, loo-, or 150-mm. path lengths could be accommodated by the use of a set of steel inserts. The small tubular heater positioned around the sample cell finger was fitted with a Lavite insulator and, a t the bottom, with a small cylindrical stainless steel block. iiwell was drilled into the center of the block along its lateral axis to accommodate the finger, and a second hole a t the base of and perpendicular to this well contained a measuring and controlling thermocouple. The temperatures of the central heaters of the furnace and of the small finger furnace, 15' below that of the main block, were regulated by two Brown proportioning controllers (Minneapolis Honeywell) while the end heaters were controlled by a Celectray off-on controller. Sample temperatures were measured with the aid of a Rubicon precision potentiometer. The entire furnace was on wheels and, as required by the Beckman instrument, was alternately positioned with sample and reference holes in the light path with the aid of preset stops. The absorption cells were filled after a careful regimen of degassing and chlorination (except for those used for metal samples since chlorination or washing with HC1 introduced spurious monochloride lines into the spectra). The samples necessary for condensed phase measurements were sublimed into the cells while the weighed samples for gas phase measurements were added directly to the cell in the drybox. I n all cases the cells were evacuated and sealed off prior to measurements. Data Correction. The difficulties of maintaining a closely matched reference cell prompted the use of the (6) B. R. Sundheim and J. Greenberg, Rev. Sei. Instr., 27, 703 (1956). (7) B. R. Sundheim and J. Greenberg, J . Chem. Phys., 28, 439 (1958). (8) D. M. Gruen and R. L. McBeth, J . Znorg. Nucl. Chem., 9 , 290 (1959).
1117
GASEOUS EQUILIBRIUM BETWEEN CADMIUM AND CADMIUM(II) CHLORIDE
empty reference "hole" in place of the cell so that a correction procedure was required to obtain the actual sample absorbance. Since the low temperature measurements (approximately 100') of a system essentially represented the transmission of an empty cell (sample absorptions were undetectable), these values were adopted as l o e ~ l / l r e f e r e n o e . The actual corrections were obtained by drawing a calibration curve of %T (transmittance) us. wave length for the low temperature values of each run and interpolating the correction values from this graph. This procedure implicitly corrected for any mismatching of the absorption cells in cases where a reference was employed as well as for other incidental errors such as a faulty positioning of the furnace blocks. The slight change in cell transmittance observed a t higher temperatures was accommodated by shifting the transmission values of the correction curve so that these were identical with those of the sample in the blank 4000 to 5000-8. region. The corrected transmittances were then obtained by dividing the measured transmittances by the values from the correction curve.
Results and Dirscussion Spectral measurements of the gas phase above condensed Cd, CdCl,, saturation solutions of Cd in CdClz, and of all-gas systems of Cd and of mixtures 2f Cd and CdClz were made between 2000 and 5000 A. I n all cases the measurements were made frolm room temperature to at least 750'. The choice of the spectral region was made after preliminary measurements cisclosed no interesting (detail from 5000 to 20,000 A., while temperatures above 800' were not used since at this temperature the absorption above condensed samples in a 100-mm. cell was already too large to be measured with good accuracy. Characteristic data for absorption peaks are tabulated in Table I for both all-gas and condensed phase samples. The spectrum of the metal vapor, described in Table 11, is in excellent agreement with previously reported observation^.^. The observed spectrum of gaseous CdClz presented no interesting details. It first appears a t 675'K. a t a pressure of 3 X torr and quickly forms a rather broad, structureless continuum centered a t 2000 8. and shaded to the red. With increasing pressure the band broadens and shifts to the red until a t 60 torr (1025'K.) the transition is apparently centered at 2400 8. and extends to 3600 8. Identzjication of CdCZ. A typical vapor phase spectrum of the Cd-CdCl, mixture is presented in Fig. I with the source of the Cd and CdCl, lines marked for reference. The most notable feature of these spectra is the observation of two series of bands extending to
Table I : Spectral Data for Gaseous Cd and for Cd-CdCL Mixtures (100-mm. Path Length) ,------Cadmium % T % T
T , OK.
989.2 1008.0 1048.5 928.4 968.0 987.8 1003.5 1018.3
T, OK.
945. 6d 965. 6d 987. 4d 1006.4d 1026.4d 955.6" 985. 4e 1016.1" 1045.6e 951. gf 977.7' 1001.9' 1027.7' 1052.4j 817.6 869.1 902.3 936.3 968.7 993.2 1018.0
metal----
(3261 A.) (3172 .&.) (Cdd (Cd)
8.7 8.8 9.2 19.2 7.6 4.8 3.8 3.0
83.8 84.1 84.3 94.2 85.8 80.5 72.1 64.1
P/Tb (Cd)
0.3167 0.3133 0.3080 0 .2052c 0,3412' 0.4242" 0.5062' 0.6037'
Absorbance5 (Cdd
%T (3070 A.)
0.0768 0.0752 0.0742 0.0155 0.066 0,0835 0.142 0.194
99.2 98.9 98.7 98.6 98.5
Cadmium plus cadmium dichloride %T % %T AbsorbanceD (3261 .&.) (3070 A.) (3170 I . ) (CdCI, (Cd) (CdC1) (CdC1) 3070 A.)
26.2 26.5 26.8 26.8 27.5 28.2 28.4 28.5 28.7 26.2, 26.4 26.7 26.7 27.3 74.7 52.5 34.3 18.6 6.1 3.3 1.8
81.0 78.1 74.7 71.6 68.2 74.7 68.3 62.4 55 5 75.8 71.4 66.1 60.3 55.0 98.0 94.3 86.2 69.1 43.9 25.0 11.0
82.0 77.6 74.6 70.5 70.2 75.4 69.9 61.9 58.2 76.0 71.2 65.9 59.8 53.8 98.1 94.6 85.0 68.0 43.8 25.0 12.4
l'/Tbtc
(Cd)
0.078 0.094 0.014 0.133 0.153 0.107 0.146 0.189 0.236 0.102 0.128 0.162 0.201 0.241 0.039 0.089 0.144 0.227 0.340 0.451 0.592
Corrected for background. Torr deg.-'. Condensed phase run; P I T for Cd calculated from known vapor pressure data of F. D. Rossini, et al., "Selected Values of Chemical Thermodynamic Properties," National Bureau of Standards Circular 500, U. S. Govt. Printing Office, Washington, D. C., 1952, Run A; 0.1884 and 0.0090 torr deg.-l of Cd and CdC12, respectively. e Run D ; 0.1623 and 0.171 torr deg.-I of Cd and CdClz Run E; 0.1841 and 0.140 torr deg.-I of Cd and respectively. CdC12, respectively.
'
lower wave lengths from 3070 and 3170 A. These values and their accompanying fine structure form a positive identification of the presence of CdCl in that they correspond well to the CdCl lines reported by Walters and Basratt6 although the wave lengths ob-. (9) 9. W. Cram, Phys. Rev., 46, 205 (1934)
Volume 68,Number 6
May, 1964
1118
BUDDY L. BRUNER AND JOHN D. CORBETT
0
I Cd
4
80
2
-
70-
0
v) v)
f
60-
v)
z 4 a I-
I
CADMIUM-CADMIUM ( 0 ) CHLORIDE 50-
k
z W
#a
40-
30>
I
I
I
I
-
WAVELENGTH ANGSTROMS Figure 1. Typical spectrum of gas phase mixtures of Cd and CdC1, (with P I T of Cd and CdCl? equal to 0.2205 and 0.0187 torr deg.-’, respectively; T = 1026°K.).
served in this study are consistently about 2 A.smaller. The assignment of this doublet to CdCl is based on the predicted appearance of a doublet from the 211-2L: transitions anticipated for such a species and the very fact that discrete bands are observable rather than more or less of a continuum to be expected for a polyatomic species such as Cd2C12. The inability to detect the previously reported transition at 3181 A. is most likely due to an erroneous assignment to CdCl of a transition of Cd by Walters and Barratt.5 Their failure to record the detection of the subsequently observed 3178-a. lir1e93losupports this contention. No other continua or discrete bands could be found on careful scanning to 20,000 A.11 It therefore must be concluded that all other possible species are absent or present in undetectable amounts unless masked by the Cd CdC1, continuum. It is unlikely that any other components, particularly Cd2C12,would exhibit no transitions in the normal ultraviolet-visible region. There is also no indirect stoichiometric evidence to suggest that any other species was present. Vibrational Bands of CdC1. The aforementioned
+
The Journal of Physical Chemistry
discrepancies between the reported and observed CdCl vibrational bands indicated a need to re-examine the vibrational analysis presented by Howelll2on the basis of the earlier values. To this end, Deslandres tables were prepared using weighted averages of the wave lengths observed in a number of measurements, one of which is given in Table 111. Assuming that the CdCl systems observed were due to the 211-2L: transition and that only the &-heads were likely to be resolved, the intense 3070 and 3170 A. transitions were assigned to the 0 4 transitions for the anticipated doublet. It was then possible to assign all of the other bands in a very consistent manner (using as a guide the value of 330.5 cm.-’ for we’’ that Cornell13obtained (10) S. Mrozowski, 2. Physik, 6 2 , 314 (1930). (11) I t is interesting to note that the spectra of empty cells that had previously been used for a run that contained the metal always exhibited the 2288-b. resonance transition of Cd. The intensity corresponded to a pressure of some 1 0 - 3 torr of Cd(g), and this transition could not be eliminated by heating the cell under vacuum to 500° or
by washing the cell with acids. (12) €1. G. Howell, Proc. Roy. Soe. (London), A182, 95 (1944). (13) S. D. Cornell, Phys. Rev.,54, 341 (1938).
1119
CASEOCM EQUII,IXXIVM RETWEEN CADMIUM AND CAD;MITJM(~ I) CHLORIDE
Table I1 : Spectral Systems of Gaseous Cadmium Metal
1
TraneiDescription
Appaaranco“
tion
2288 A.
4.5 X 475°K.
torr
The major remnarice peak, hroadens slowly and syrnmetrirally a t the hme (100% ?’) from 0.5 torr. By 10 torr the peak begins to broaden asymmetrically to the red and by 13,5 torr it has formed a sharp edge a t 2200 and a wing extends to 3000 A. in the red The secondary resonanre line broadens symmetrically from the hase a t &5 torr and asymmetrically to the red above 500 torr A sharp peak which broadens very slowly and synirrirtrirally from its first appearance A sharp peak which 1s absorbed into the 2288-A. hand a t 135 torr A sharp line which tend! to be absorbed by the 3261-A. line at pressuros above 1 atrn.
A-Run A 0-Run D 0-Run E
4.
A.
0.8 torr 58h’K.
2123 A.
2 . 5 torr 700°K.
32fil
2214
d.
3178
A.
2 . 5 torr
700°K.
.io torr
840°K.
0 Prestww and the= cwrreqmnding tempcratures for the minimum detection of the vapor spectrum through 2t 1 0 0 h m . path ler1yth.
I
as
from emission measurernents in a different region). The results differ from Howell’s assignments and yield apparcnt valucs of 334.5 and 1 . 3 cm.-’ for we‘‘ and xe”we”, respectively, of the ground state as well as 397 and 0.9 em. for we’ and z,’w,’ of the 2111,2 state. A valuc of 420 f 10 cm. - l is suggested for me’ of the 2r~a,,statc on thc hasis of the 0-1 transition observed. Table 111: Vibriitional Rand Heads of the CdC1 Spectrum
A. 3170 31fil ( 4 ) 3150 (1) 3141 ( 0 )
3128+ (00) 31021 (1)
A.
A
3095 (1) 308‘3 (0) 3082 (00) 3070 ( I O ) 3OM (7) 3058 (3)
3052 ( 2 ) 3040 (0) 3033 (1 ) 3027+ ( I ) 3022 (00) 3016 (00)
+
Relative intensities in parentheses.
Calculation of Equilibrium Thermodynamics. A series of preliminary calculations indicated that the intended method of quantitatively analyzing the all-gas system on the basis of calibration curves for the metal and dichloride could not be applied, principally because
I
1.00
I/T’K lo3
I
IC6
Figure 2. Plot of log K , us. 1/7‘ for the reaction Cd(g) CdCl,(g) = SCdCl(g).
1 +
of extensive line broadening of the CdCl, spcctrurn by
Cd. Therefore the following procedure was adopted. The concentration of CdCl produced \vas assumed to be too small to decrease the initial, known concentrations of the reactants significantly. Acting on this assumption, a set of pseudo-cquilihrium constants were calculated over a range of temperatures and concentrations from the rclationship K , = (Ac,,cI) */ (CC,I)(CC~CIJ, where A c d c l is the absorbance of CdCl corrected for background absorption of Cd and CdClz and the C yuantilies represent the concentration of the corresponding reactants calculated from thc amount originally added to the cell.14 A plot of log K , us. 1/T should yield a straight line parallel to such a plot of the true equilibrium constant but diff.:ring in the intercept by a value of 2 log ab since absorbance is the product of path length ( b ) , absorptivity ( a ) , and concentration. This plot of data for three separate (14) The assumption here of substantially only monomeric CdCh s is confirmed by the recent vapor pressure study by in the g ~ phase F. J. Keneshea and D. Cuhicciotti, J . Chem. Phya., 43, 1778 (1964).
Volume 68,h’umber 6 M a y , 1964
1120
BUDDYL,BRUNER AND JOHND. CORBETT
experiments, Fig. 2 (Table I), yields an apparent heat of reaction of 34.15 kcal. mole-1 a t 1000°K. for the reaction CdClz(g) Cd(g) = 2CdCl(g). The complete agreement between different experiments a t varying pressures as well as the shape of the peaks support the presumption that a(CdC1) is not affected by the other components. The reliability of this measure of AHolooocan be established by its int,ernal variation or self-consistency and by the agreement between the original assumptions and the equilibrium concentrations calculated with the use of this heat and a third-law entropy for the reaction. A least-squares analysis of the data shown in Fig. 2 gives a 95% confidence limit of ~ 0 . kcal. 5 and a 99.5y0 level of 1 0 . 7 5 kcal. for random errors in AHO1OOO. Utilizing the experimental heat and the statistical ASo~ooo of 10.4 e.u.’5-19 for the reaction, AFO1OOO is 23.8 kcal. mole-’, from which K P equals 6.3 X lop6. The latter value corresponds to a change of less than 1% in the concentrations of the reactants in the three runs and thus is consistent with the assumption that relatively little CdCl is produced. In principle, it would be possible to accommodate these small changes and to modify slightly the resulting AH”, but such a correction (0.2 kcal.) is probably not too meaningful considering the magnitude of the over-all experimental error. The results of Taresenkov and Skulkova4 that suggested a more substantial amount of a reduced species are thus inconsistent with the present findings as well as with recent studies of the vapor pressure of CdC12.i4,20 It is possible to analyze similarly the absorbance of the 3178-A. band of CdzDto obtain a value of Doo for this molecule. The experimental data yield a AHoluo0 value of 1.67 f 0.2 kcal. mole-’ for Cdz 2Cd, which compares with 2.0 =t0.5 kcal. mole- for AHO,l~~s obtained by Kuhn and ArrheniusZ1by essentially the same method. With A(Holfln0- Hod = (5 - 4.47)R kcal. (a 40 cm.-‘ for Cd?), the enthalpy change found here gives about 0.6 kcal. mole-1 or 0.02’7 e.v. for Dooof Cdz. Additional Thermodynamic Calculations. Having experimentally determined AHolooo for the gaseous reaction of Cd with CdC12, it is worthwhile to consider some other reactions for which thermodynamic values may now be derived or improved. The enthalpy obtained in this study may be combined with that for other processes to give a much better value for the dissociation energy of CdCl than has heretofore been available. For the present reaction CdCl,(g) Cd(g) = 2CdCl(g), the foregoing AFolooomay be combined with -11.68 kcal. mole-’ deg.-l for A ( F o T - H’xIs)/ to give a AHozesof 35.5 kcal. T a t 1000°K.15~17-19
Table IV : Heat of Atomization of CdCl(g) at 298’K.
+
-
+
The Journal of Physical Chemistry
kod.
Ref
-35 5 93 00 -43.6 f 0 . 2 57 88 26 75
This work 8
A.Y0essl
2CdCl(g) = CdClz(8) = CdCl,(g) = = ClS(P) Cd(s) =
+ +
Cd(g) CdClz(g) Cd(8) Cli(g) CdCIS(8) ZCl(g) Cd(g)
2CdCl(g) = 2Cd(g)
+ 2Cl{g)
__ 98 53
14 15
15
I
mole-l. The reactions in Table IV give th.e heat of atomimtion of CdCl(g) a t 298OK., and this together with A ( H 0 m - H00)15yields 122 48.6 f 0.7 kcal. mole-’ (2.11 f 0.03 e.v.) for DO0 of CdC1. The principal source of the estimated uncertainty is from the reaction studied here. This value of Doo compares and indiwith Gaydon’s estimate of 2.0 f 0.5 cates that the 2.8 e.v. value obtained by a linear BirgeSponer extrapolationz4is too large. A comparison of the Cd-CdCla reaction in gaseous and liquid systems is also informative, Using the solubility da,ta of Topol and Landis26 a heat of reaction of 5.67 kcal. mole-’ a t 1.000”K.can be estimated for the presumed reaction of Cd(1) with CdC12(1) to form an ideal solution of CdzClz(l), This combined with the heats of vaporization of CdI5 and CdC1d4 and the heat determined for the gas phase reaction gives 83.9 kcal. mole-1 for CdzClz(l) = 2CdCl(g). A similar procedure yields AXoluou = 51,3 e.u. The former compares with 70.5 f 3 kcal. mole-’ for Hg2ClZ(1)= 2HgCl(g).26$27I n principle, a comparacould be detive value of the corresponding AFolooo (15) D. R. Stull and G. C. Sinke, “Thermodynamic Properties of the Elements,” American Chemical Society, Washington, U. C., 1956. (16) Solno = 85.55 e.u. for CdCh was calculated stat,istically from molecular constants given in ref. 17. (17) L. Brewer, G. R. Somayajulu, and E. Braokett, Chem. Rev., 63, 111 (1963). (18) K, K. Kelley. U , S. Bureau of Mines Bulletin 584, U . 5. Govt. Printing Office, Washington, D. C., 1960. (19) K. K. Kelley and E. G. King, U. 9. Bureau of Mines Bulletin 592, U. S. Govt. Printing Office, Washington, D. C., 1961. (20) J. L. Barton and H. Bloom, J . Phys. Chem., 6 0 , 1413 (1956); 62, 1594 (1958). (21) H. Kuhn and S. Arrhenius, 2 . Phusik, 82, 716 (1933). (22) A value of 2316 cal. mole-’ for CdCl(g) was calculated statistically from molecular constants given in ref. 19. (23) A. G. Gaydon, “Dissociation Energies and Spectra of Diatomic Molecules,’’ 2nd Ed., Chapman and Hall, Ltd., London, 1953, p. 223. (24) G. Hersberg. “Spectra of Diatomic Molecules,” 2nd Ed., D. Van Nostrand Co., Inc., Princeton, N. J., 1950, p. 516. (25) L. E. Topol and A. L. Landis, J . A m , Chem. Soc., 82, 6291 (1960). (26) K. Newmann, 2. Physik. Chem., 191A, 284 (1942). (27) S. J. Yosim and S. W. Mayer, J . Phys. Chem., 64, 909 (1960).
1121
GASEOUS EQUILIBRIUM BETWEEN CADMIUM AND CADMIUM(I1) CHLORIDE
rived from the spectral measurements on liquid CdC12 saturated with Cd (Table I) with the aid of the absorptivity of CdCl calculated from the data for the gaseous equilibrium and the reported saturation limits for the melt. In practice, the aforementioned broadening of particularly the CdC12 band by Ccl is so large because of the higher pressures of the components that the background correction for the CdCl bands is uncertain, or, in other words, the latter are not as well resolved as in the all-gas systems studied. Although such calculated results are in semiquantitative agreement with the gaseous systems, the spectral studies above the melts are generally not as useful for this study because of the dependence of solution composition on temperature as well as the higher pressures of the condensed components. Finally, it would be most interesting to be able to derive thermodynamic data for the dissociation CdzC12(1) = 2CdClcl) as well as its gas phase counterpart from the foregoing information on Cd2C12(1) = 2CdCl(g). With condensation data for InC128as a reasonable stand-in for those for CdC1, AHoloooand AXolooo
for the liquid phase dissociation of CdzClzare estimated to be about 40 kcal. mole-1 and 3 e.u., respectively. This corresponds to KD = lo-*.', in agreement with the lack of magnetic evidence for the monomer in these melts.29 On the other hand, vaporization data for Cd2C12necessary to consider the gaseous dissociation can only be guessed. However, if AXolooo for this vaporization is taken to be a plausible 24 e.u., a 1% limit of detection of Cd2C12(g)in the spectral studies gives a maximum value of 35 kcal. mole-l for AHolooo of the gas phase dissociation and a lower limit of 49 kcal. for vaporization of 1 mole of Cd2C12. The apparently complete dissociation in the gas phase vs. the opposite behavior in the melt thus appears to be mainly a consequence of a reasonable difference in the entropy change in the two cases, 27 us. 3 e.u., respectively, rather than the result of any marked difference in the enthalpy.
(28) Refer to work of F. D. Rossini, et al., quoted in Table I, ref. c. (29) N. H. Nachtneb, J. Phya. Chem., 66, 1163 (1962).
Volume 68, Number 6
May, 1964