NOTES
1964
tive rate a t each position is therefore 600 X 5 = 3000; and the rate for p-xylene relative to that for benzene is 4 X 3000/6 = 2000. Similarly,.inasmuch as the single available position in pentamethylbenzene is influenced by one para, two meta, and two ortho methyls, the rate of chlorination of pentamethylbenzene relative to that of benzene is 870 X 5 X 5 X 600 X 600/6 = 1.3 X lo9. Such calculated relative rates are compiled in Table I, together with available experimental values’ for chlorination or bromination. The agreement between the calculated and the experimental values, which is good in view of the 350,000-fold range, inspires considerable confidence in the five values for which experimental confirmation is a t present unavailable. A product of relative rates as calculated above is mathematically related to a sum of activation energy differences calculated from Arrheniustype equations for the individual rate constants k = Ae--E/RT For
Val, 70
-
+
defined by the neutralization equation NO+ C1NOC1. As in the analogous case of carbonyl chloride,2 the chloride ion should not be appreciably solvated, for the electronic structure of NOCl (resonance between :O=N+C1- and :O=N-6 .. : will not be affected by an electrondonor so weak as chloride. In agreement with this prediction, potassium chloride is insoluble in liquid nitrosyl chloride, contributing no conductance effect. The nitrosyl ion, on the other hand, should be solvated very strongly, on account of the resonance structures :O=X-Cl-N=O:] . .. . .. .. +,
..
)
.
[.
Hence it is expected that nitrosyl salts, many of which have been recognizedJ4should be soluble in liquid nitrosyl chloride, with high electrical conductance. Experimental Results.-These ideas have been tested in a preliminary investigation of the electrical conductance of solutions of typical nitrosyl ki/k = (Ai/A)e(E-Ei)/RT salts in liquid nitrosyl chloride. The monoand nitrosyl salts NOAlC14, NOFeC14 and NOSbCle X(E - Ei) = ZRT In ( k i l k ) ZRT In (ASIA) = are readily soluble strong electrolytes, as deR T In l I ( k i / k ) - RT In lI(Ai/A) manded by the theory (see Table I). The solvaIf the several A terms are equal,‘ the relationship tion is indicated by the reaction NOA1C14(s) simplifies to NOCl(g) S NOAICh.NOCl(s), recently discovered in our laboratories by Donald E. McKenzie. Z(E - Ei) = RTlnJI(ki/k) According to the theory of absolute reaction ratess The dissociation pressure of the product is approximately 240 mm. a t 0’. k = (KT/h)e-F/RT Less favorable results were obtained with where K = Boltzmann constant, h = Planck con- (N0)&3nClG, (N0)2TiCle, and (NOS04H), all of stant, and F = difference in free energy between which appeared quite insoluble and non-conductinitial and activated states, or “activation free i n g Dinitrosyl pyrosulfate, (N0)2S207,6showed a energy.’, Consequently, without any assumption slight conductance which very slowly passed of constancy for an A term, or “frequency factoPS through a maximum (see Table 11). Sulfuric acid was inert. k i / k =s e(F-Fi)/RT The inert character of potassium chloride was and not changed by an “acid” solution of nitrosyl Z(F Fi) = RT In lI(kI/k) chloroantimonate, which, by analogy with carso that a product of relative rates is simply related bonyl “acidsJJin carbonyl chloride, should have to a sum of activation free energy differences. dissolved it. The reason might be a protective (4) Cf.Bradfield and Jones, J . Chcm. Soc., 1006, 3073 (1928); coating of KSbCls. Trans. Faraday Soc., 87, 737 (1941). I n an attempt to estabish chloride ions in liq(5) Glasstone, Laidler and Eyring, “The Theory of Rate Procuid NOC1, the base-action of water was tried. esses,” McGraw-Hill Book Co., Inc., New York, N. Y . , 1941. However, the water only dissolved without conBARTLESVILLE, OKLAEOIU RECEIVED NOVEMBER 28,1947 ductance and was recovered by evaporating the nitrosyl chloride. The usual hydrolysis reaction 2NOC1+ NO NO2 2HC1 is slow a t H20 Liquid Nitrosyl Chloride as an Ionizing Solvent room temperature and the products are nonconducting in nitrosyl chloride. BY ANTONB. BURGAND GEORGE W. CAMPBELL, JR.‘ (a) Germann and Gagos, J . Phys. Chcm., 28, 965 (1924); Liquid nitrosyl chloride is expected to be a (b)(2) Germann and Timpanp, ibid., 29, 1423 (1925); ( c ) Germann and fairly effective ionizing solvent, for its dielectric Birosel, ibid., 29, 1469 (1925); (d) Germann, THIS JOURNAL,47, constant (18.2 a t 12’)la is comparable to that of 2465 (1925); (e) Germann and Timpany, i b i d . , 47, 2275 (19251 (3) Ketelaar and Palmer, ibid., 69,2629 (1937). liquid ammonia. The acid-base system would be (4) (a) Willke-Dorfurt and Balz, 2.onorg. ollgem. Chem., 169,210
-
+
-
+
(1) This note is based upon a thesis submitted by George W. Campbell, Jr., to the Graduate Faculty of the University of Southern California, in partial fulfillment of the requirements for the degree of Master of Science, June, 1947. (la) Ketelaar, Rcc traw. chim., 69,289 (1943).
+
+
(1927); (b) Gall and Mengdehl, B e . , 60B,86 (1927); ( c ) Rheinboldt and Wasserfuhr, ibid., 60B,732 (1927); (d) Angus and Leckie, Proc. Roy. Soc. (London), A160, 615 (1935); (e) Klingenberg, Rec. fro%chim., 66, 749 (1937). (5) Jones, Price and Webb, J . C k m . SOC.,186, 812 (1929).
Methods and Numerical Results.-All re- with the maximum viscosity of pure sulfur, 93,200 agents were prepared and handled with complete centipoises, a t 186-188'. exclusion of moisture. The nitrosyl chloride Fluorine was not tried in the earlier experiments was prepared by the usual reaction NOS04H because of the di5culty in obtaining this element. HC1+ H2S04 NOCl, and purified by means of a It was assumed, however, that because of its posihelix-packed column. With electrodes of bright tion in the halogen group it would be much more platinum wire 1 mm. thick, having 12 mm. ex- effective than chlorine. Recently, sulfur hexaposed length, and 6 mm. apart, polarization ef- fluoride became available commercially and it was fects caused a 15% uncertainty in the evaluation used in the same manner as sulfur chloride. Beof the cell constant (0.78) by reference to aqueous cause of its high thermal stability, sulfur hexafluoKCI, but the various NOCl solutions gave results ride, if effective, would be an attractive reagent consistent to 1%. These results accordingly are for reducing the viscosity of sulfur. presented in Tables I and I1 for purposes of comExperimental parison. Quantitative data were not obtained as the naTABLEI ture of the work was exploratory. The effect of CONDL'CTANCESI N LIQUIDNITROSYL CHLORIDE the sulfur fluorides was determined by visual comT$mp., Spec. cond., Molar Solute Molarity C. mhos. cond. parison of the fluidity of the treated sulfur with NOAlCt 0.098 -20 1.17 X lo-* 119 that of pure sulfur2which a t 180-195' shows very NOFeCL .0099 -20 1.34 X 10-3 136 little or no flow in the space of a few minutes. NOFeC14 .0094 -21 1.26 X 10-8 134 Sulfur hexafluoride was bubbled for five and NOFeClr ,0094 -44 1.00 X 10-3 106 one-half hours through pure sulfur heated to 310' NOSbCle .140 -20 2.35 X lo-* 168 during which period little, if any, decrease in visNOSbCls .140 -44 2.20 X 157 cosity was noted. On cooling from 310 to 194' NOCl Pure -20 2.88 X lo-' with the gas still bubbling through, the sulfur gelled indicating that the sulfur hexafluoride was TABLEI1 comparatively ineffective in reducing the viscosity CONDUCTANCES OF (-10)2Sz0, (SATD.IN NOCI) AT -20" of sulfur. Apparently the stability and inertness Time, hr. 0 0.75 2 15 46 of the hexafluoride is such that its reaction with Spec. cond. X lo* 4 . 9 5 25.0 48.8 23.8 14.7 the long sulfur chains is negligible under the conditions of the experiment. DEPARTMENT OF CHEMISTRY The direct fluorination of sulfur gives in addiTHEUNIVERSITY OF SOUTHERN CALIFORNIA Los ANGELES 7, CALIFORNIA tion to the hexafluoride much smaller and variable RECEIVED DECEMBER 11,1947 concentrations of SF4, SzF2, and (about 1%) S2Flo. As revealed by S c h ~ m bthese , ~ lower fluorides are removed before the purified hexafluoride The Effect of Sulfur Fluorides on the Viscosity of is compressed into the cylinders. These lower Sulfur fluorides are less stable and more reactive than the hexafluoride and might react with the long sulfur BY R.FANELLI chains to appreciably reduce the viscosity. Recently, it was shown,lV3that the halogens, The effect of the gases SF4 and SaF2 was deterchlorine, bromine, iodine and also hydrogen sulfide mined by heating in an electric furnace mixtures of and hydrogen persulfides, even when present in pure sulfur with metallic fluorides in rotating, minute amounts, reduce the viscosity of sulfur sealed, heavy-walled Pyrex tubes (free volume above 160'. The degree of effectiveness de- about 15 cc.). For the preparation of SF4 the creased, with respect to the halogens, in the order following reaction, reported to yield this comshown. It was also indicated that the increasing was employed viscosity of pure sulfur with rise in temperature 'iC0F.q S +4COFz SFq might be due to polymerization of sulfur into long straight chains and that reduction of the viscosity Under the conditions of this experiment, probably was due to the rupture of these chains by the hal- some of the cobaltous fluoride was converted to ogens which take up terminal positions of the seg- sulfide. For the formation of SSFStwo reactions ments. For example, in the case of chlorine ad- reported to yield this halide6-'were tried ded in the form of sulfur chloride we have 2HgF 4 3s HgzS %Fs
+
+
'
+
$(chain)
+ &Clz --t Cl-S-S-S Cl-s-s.s
--
*
S-s-s.
*
S-S-SC1 and/or
Sulfur mixtures containing 2% chlorine give viscosities under 15 centipoises throughout the entire liquid range, 115-445'. This is in sharp contrast (1) R. F. Bacon and R. FPnelli, Tnrs JOURNAL, 65,639 (1943). (2) R. F. Bacon 8nd R. Fanelli, Ind. Eng. Ckcm., 81,1043 (1942). (3) R. Fanelli, ibid., S8, 99 (1946).
2AgF 4 3s +AgzS
+
+ + %Fz
(4) W. C. Schumb,
