86
R. J. SIMEAND N. W. GREGORY
interfacial tension curve for the iodide is almost linear above 40". This seems to imply that the solubility of this compound in water, which must be small according to the large interfacial tension values, increases but little with the temperature. Accordingly its transition phase with water must, be small in thickness. From reference to Table 11, it is seen that the entropies, latent heats and enthalpies of the various interfaces all increase with the temperature. Plots of these intensive interfacial properties as functions of the temperature show that these relations are linear but very definitely not parallel.
Vol. 64
A comparison of the temperature rate of increase of the entropies of the interfaces of n-ethylbenzene and n-propylbenxene with the corresponding halogen substituted compounds of the present series, shows that, with the exception of the chloroethylbenzene, the entropies of the latter not only increase more rapidly but to a greater extent. This seems surprising in view of the expected restriction in the intra-molecular activity suggested by Adam8due to the addition of the relatively heavy halogen atoms on the alkyl chain. (8) N. K. Adam, in "The Physics and Chemistry of Surfaces," Oxford Univ. Press, 1941,p. 67.
VAPOR PRESSURES OF FeC12, FeBr2 AND FeL BY THE TORSION EFFUSIOK METHOD BY R. J. SIMEAND N. W. GREGORY Contribution from the Department of Chemistry at the University of Washington, Seattle, Washington Received July 9 , 1969
Total vapor pressures developed over solid FeC12, FeBr2and FeL, respectively, have been measured between 400 and 470' by the torsion effusion method. Results are discussed in relation to other published information concerning the vaporization of these substances.
A study of the vapor pressure of iron(I1) bromide (350 to goo"), using effusion, transpiration and diaphragm gage techniques, has been reported previously from this Lab0ratory.l Effusionpressures, measured near 400" in the Conventional manner, were found to be about 30% lower than transpiration pressures when monomeric FeBr, was assumed t o be the principal vapor phase molecular species; this assumption was suggested by comparison of transpiration and diaphragm gage data near the melting point (690"). In the present work additional information on the vapor pressure of FeBrz and new data for FeClz and Ferz have been obtained using a torsion effusion apparatus. In this method knowledge of the molecular weight of the effusing vapor is not required for evaluation of the total pressure within the effusion cell. Very recently, a mass spectrometric analysis of the vapor species in equilibrium with FeBr22and with FeClZ3has been published by Porter and Schoonmaker. Their findings confirm that the major vaporizing species for FeBrz (and also FeC12) in the effusion temperature range is the monomeric form but indicate that the dimer becomes increasingly important a t higher temperatures. Experimental Part The torsion effusion method4 has been actively used by a number of investigators in recent years.6 A brief descrip(1) R . 0.MacLaren and N. W. Gregory, THISJOURNAL, 69, 184 (1955). (2) R. F. Porter and R. C. Schoonmaker, ibid., 63, 626 (1959). (3) R. C. Schoonmaker and R. F. Porter, J . Chem. Phgs., 29, 116 (1958). (4) Introduced by M. Volmer, Z . physilc. Chem., Bodenstein Festband, 836 (1931). (5) See for example: (a) R. F. Barrow, D. G. Dosworth. A . R. Downie, E. A. N. 5. Jeffries, A . C. P. Pugh, F. J. Smith and J. M. Swinstead, T r a n s . Faraday Soc., 51, 1354 (1955): (b) R. D. Freeman and A. W. Searcy, J . Chem. P h y s . , 22, 762 (1964): (c) M. D.Scheer, THISJOURNAL. 61, 1184 (1967).
tion of our apparatus follows. Two holes of nearly equal area were symmetrically placed ca. 1.5 cm. from the axis of suspension on opposite sides of a quartz effusion cell (5 cm. long and 1.5 cm. 0.d.). Holes were blown in thinned sections of the quartz wall. Two independent cells, attached to tungsten wires ca. 55 em. long, were used; cell 1, hole areas 1.95 X 10-2 and 1.725 X cm.2, to a 2 mil. wire, torsion constant 1.820 dyne cm. radian-', and cell 2, hole areas 2.60 X 1 0 - 3 and 3.72 X cm.2, to a 1 mil wire, 0.1004 dyne cm. radian-'. At the bottom of each wire the cell assembly was attached by a clamp to which was bolted a small aluminum disk. To measure the torsion ronstant of the wires, rings of known moments of ineitia were placed on this disk and the entire assembly treated as a torsion pendulum system. A rigid tungsten wire, to which was attached a mirror and an aluminum vane (10 X 1 X 0.01 (cm.) which, through interaction with a magnet, aided in damping oscillations), connected the aluminum disk to a quartz rod attached to the cell. The cell was heated by directing radiation from three 500 watt projection bulbs onto a blackened copper cylinder, 6 cm. o.d., 7 cm. high, with walls 0.1 cm. thick. The top of the cylinder was a split disk which could be closed around the support rod of the cell after the latter was lowered inside. The projection bulbs were mounted so as to place their filaments a t the fori of eliptical reflectors, major axis 8 in., semi-major axis 6.5 in., placed at 120" intervals around the furnace.6 A dummy cell was placed in the jacket as close as possible to the suspended cell. The dummy was supported by three thermocouples (placed a t the center and each end); the temperature indicated by these couples was assumed to represent the temperature of the torsion cell. Thermocouple leads were brought out of the vacuum system through Stupakoff seals at the bottom of the apparatus. The entire assembly was surrounded by a 10 cm. 0.d. Pyrex tube and could be dismantled via an "0" ring seal. The pressure within the tube, monitored continuously with an ionization gage, was kept below 10-5 mm. The angle of twist was measured with a conventional telescope and scale assembly. Scale deflections were read within 0.05 cm., corresponding to an uncertainty of the torque angle of 5 X radian. Although the wire characteristics were determined by a oendulum treatment, the entire assemblv was finally calibrated to remove uncertainties associated"with slight irregu(6) L. R. Weisberg and G. R . Gunther-Rfohr, Rev. 5%. In&., 26, 896 (1966).
THEVAPORPRESSURE OF IRON HALIDES
Jan., 1960
larities in the holes and their placement in the cells. Zinc was found a convenient calibrating substance6a; its vapor pressure is known with good accuracy? and the effusion temperature range is reasonably close to that of interest for the iron halides. The calibration provides an apparatus constant which relates the pressure to the torque angle, P = ke; kl = 0.0323 f 0.0013, kt = 0.0129 f 0.0006. FeBr2 and FeIz were prepared (by others in this Laboratory)s by direct reaction of reagent grade iron with reagent grade (ACS) samples of the respective halogens at elevated temperatures. The iron content indicated a purity of better than 96% for FeTz and 99% for FeBr,. FeClz was prepared by dehydration of FeCl2.4Hz0 ( J . T. Bakers Analyzed, 99.9) in vacuo, initially a t ca. 120' and finally by resubliming the anhydrouig FeC12 a t ca. 500". All samples were purified by vacuum resublimation; material was introduced into the torsion cell through a small filling tube attached to the bottom; this tube was sealed off quickly and subsquent evacuation of the apparatus accomplished within a very few minutes to minimize hydrolysis of the samples, exposed to the atmosphere only through the pinholes. Hydration of FeClz and FeBrP is easily reversed i n vacuo (presumably also the iodide, although the system FeIrH20 has not been studied in detail) without appreciable oxidation. In the previous work with FeBr.2 no evidence was observed to indicate a dependence of the vapor pressure on the presence of small amounts of oxide; a similar behavior has been assumed for FeClz and Fer2. As the temperature of the samples was raised, degassing was followed by observing the angular deflection of the cell. Vapor pressures were calculated from the minimum torsion deflection reached a t a given temperature on heating, or from the deflection observed as the sample was cooled in steps after being held a t the highest experimental temperature for 1 to 2 hr. The vapor pressures obtained were the same with both procedures; the latter method was faster and hence more convenient.
Results and Discussion Results are shown graphically in Figs. 1, 2 and 3.'O Lines based on points obtained with cell 2 alone, may be represented by equations of the form = -AT-' B. Values of A and B, and log Pmm. associated thermodynamic properties, if the vaporization process is assumed to be of the simple form FeXs(s) FeX2(g), are summarized below (temperature range 1370-740"K.). A comparative plot of sigma functions ( - R In P - 8 In T ) ,where AC, (sublimation) has been estimated from the behavior
87
o Cell I Cell 2
I
I
1.47
1.44
Fig. 1.-Vapor
I
I
I
1.41 1.38 1.35 loOo/T, OK. pressure of iron(I1) chloride.
2
+
-
h
i E
TABLE I
:3 bn 3
I
9890 10220 960
FeCl2 FeBrz FeIz
11.10 11.95 11.82
45.312 4 6 . 7 i2 (44.7f4)
37.6 41.5 (40.9)
of similar halides11t12as -8 cal. mole-' deg.-', is shown in Fig. 4. The relative volatilities are as expected for solids in which the bonding is largely ionic. Bond energies of the vapor molecules have been estimated from the extrapolated heat of sublimation and other relevant data a t 298°K. 2(Fe-X)
=
-
-AHO(sub. FeX2) ANo(form. FeX2(s))1* AHO(sub. Fe)14 AHo(diss. X2)lS
+
+
(7) IACTIONOF IRON(II1) FROM ACID PERCHLORATE SOLUTIONS BY DI-(8-ETHYLHEXYL)-PHOSPHORIC ACID I N n-OCTANE BY C. F. BAES,JR., AND H. T. BAKER' Contribution from the Oak Ridge National Laboratory, Oak Ridge, Tennessee, Operated by Union Carbide Corporation for the U.S. Atomic Energy Commission and Texas Woman's University, Denton, Texas Received July IO, 1969
The distribution of iron( 111)between acidic aqueous perchlorate solutions and solutions of di-( 2-ethylhexyl)-phosphoric acid [(R0)2P02H, DPA] in n-octane has been examined as a function of various extraction variables a t constant aqueous ionic strength. The organic solutions involved have been investigated by means of isopiestic molecular weight determinations and viscosity measurements. The distribution behavior a t low iron levels (iron:DPA
