May, 1934
COMMUNICATIONS TO THE EDITOR
data with those obtained by Goodeve and Powney [J. Chem. Soc., 2078 (1932)l. The agreement
1251
abnormally low temperature by mixing our compound with another that decomposes a t this was excellent. In addition we have the qualita- lower temperature to give methyl radicals. tive observations that the melting point is below We have tried this experiment with a mixture of 190°K., and that the liquid dissolves slowly in acetaldehyde and azomethane. Decomposition water to give an acid solution which does not re- is actually induced in the aldehyde; a few per cent. act with sulfurous acid (indicating perchlorate of azomethane is sufficient to cause complete rather than lower valences of chlorine). The decomposition of the aldehyde at 300°, a t which amount of this oxide produced in a given reaction temperature the aldehyde alone is quite inert. mixture seems to be proportional to the initial Chain lengths, calculated from rate measurements, ozone pressure and to increase somewhat with the are found of the order of 30. The rate of the reacchlorine pressure; it is only slightly dependent on tion is found, within the accuracy of our few pretemperature between 20 and 30”. It is obvious liminary experiments, to be proportional to the that the formation of these oxides, CLO, and first power of the aldehyde pressure and to the C103, invalidates any interpretation of the be- square root of the pressure of the azomethane, havior of chlorine and ozone mixtures which as- just as the chain theory predicts. sumes that the pressure change measures the Our result strongly confirms the above explanaoxygen formed. tion of the decomposition of pure acetaldehyde as DEPARTMENT OF CHEMISTRY A. C. BYRNS a chain reaction, but this is not yet definitely UNIVERSITY OF CALIFORNIA G. K. ROLLEFSON proved since there may be some other faster reBERKELEY, CALIFORNIA action superimposed upon the chain reaction a t RECEIVED APRIL12, 1934 higher ’ temperatures. We expect, however, by further experiments with azomethane-aldehyde CHAINS IN THE ACETALDEHYDE DECOMPOSITION mixtures to establish this point by getting the Sir : activation energies of the chain steps. These A recent paper by F. 0. Rice and Herzfeld experiments may also throw light on the ques[THISJOURNAL, 56, 284 (1934)] gives an explana- tion of whether the azomethane decomposition is tion of the kinetics of the homogeneous thermal a chain reaction. decomposition of certain organic compounds on MALLINCKRODT CHEMICAL DARRELL V. SICKMAN AUGUSTINE 0. ALI;EN the basis of chain mechanisms, involving in some LABORATORY HARVARD UNIVERSITY cases very long chains. In many cases this theory CAMBRIDGE, MASS. RECEIVED APRIL16. 1934 comes into direct conflict with the generally accepted explanation (L. S. Kassel, “The Kinetics of Homogeneous Gas Reactions,” Chapter V ] of these reactions as unimolecular decompositions THE LOW PRESSURE EXPLOSION LIMITS OF DEUTERIUM AND OXYGEN “falling off” a t low pressures. In fact, most supposed examples of the ‘latter theory, such as the Sir: We have compared the explosion limits of ethers and azo compounds, must now be regarded as under suspicion of being due to chain mecha- deuterium-oxygen and hydrogen-oxygen mixtures a t low pressures at 500’ and 520” in potasnisms. There is a simple test, applicable in many cases, sium chloride-coated Pyrex according to our for the existence of such chains. One step in original technique [THIS JOURNAL, 55, 3227 these chains will nearly always be the reaction of (1933)]. Deuterium was generated by eleca methyl radical with the molecule of the sub- trolysis of pure heavy water; in some cases stance, and the activation energy for this must be additional care was taken to purify it by diffusion lower than that for the primary decomposition through palladium. Tank hydrogen was used into free radicals. Hence if we introduce free in alternate experiments. From Fig. 1 and Fig. 2 methyl radicals from some extraneous source into it is seen that the deuterium curve is much an organic vapor a t a temperature somewhat be- broader. At high pressures where there is no low that a t which it ordinarily decomposes, a appreciable deactivation a t the wall of the vessel, chain decomposition should be set up. Thus, the negative reciprocal slope of the explosion limit these decompositions should be induced a t an curve represents the relative deactivation efficien-
1701. 56
COMMIJNICATIONS TO THE EDITOR
1252
cies ZHz/ZOz and ZDz/ZOz in the chain-breaking O2 M, where X is the chainprocesses X carrier and M is Hz, D, or OL. Experimentally these values are 3.1 and 2.1, respectively; whence ZHz/ZD2 = 1.5. This value compares favorably with the value ZHz/ZDz = d p D z PHZ= &/2 = 1.4, roughly calculated from the simple theory of
+
+
I
1
I
I
i
..
relative rates of chain branching and chain breaking process controlling the explosion limit, must cancel one another. It is unlikely that this would be the case if D and H or D2 and Hz were the chain carriers. FRICKCHEMICAL LABORATORY ~
~
~
ARTHURA. FROST HUBERTN. ALYEA ~
~
~
RECEIVED APRIL17, 1934
THE PREPARATION OF HYDROGEN COMPOUNDS
c
3 20
OF SILICON
2?
60 80 100 mm. Fig. 1.-Explosion limits at 500": X, hydrogen and oxygen; 0, deuterium and oxygen; hydrogen diffused through palladium; @, deuterium diffused through palladium. 0
20
40
PO2'
1x1,
three-body collision frequency [Grant and Hinshelwood, Proc. Roy. Soc. (London), A141, 29 (1933) 1, assuming equal collision diameters €or Hz and D2. For according to this, the ratio ZH2/ZDZ is a function only of the reduced masses, p, of the respective complexes (X 0 2 HS) and (X O2 Dz); and the reduced masses are
+ +
+ +
0
20
40
60
80
100
PO*> mm.
Fig. 2.--Explosion limits at 520': X, hydrogen and oxygen; 0, deuterium and oxygen.
approximately the masses of HZ and Dz when X 0 2 is large compared with Hz and Dz. The whole of the observed change due to deuterium, therefore, can be accounted for on the basis of relative deactivating efficiencies. This leads to the observation that any additional changes, due to the substitution of deuterium for hydrogen, in the
+
Sir : In our studies on the kinetics of the catalytic decomposition of the hydrogen compounds of some of the fourth and fifth group elements it became necessary to prepare rather large quantities of silane, SiH,. Being acquainted with the very favorable results obtained in the preparation of the germanes in liquid ammonia [Kraus and Carney, THIS JOURNAL, 56, 765 (1934)], i t was thought advisable to follow the same procedure in an attempt to prepare some of the silicon compounds, since the former were obtained in unusually high yields, three to four times as great as those found in the aqueous method [Dennis, Corey and Moore, ibid., 46,657 (1924)l. Magnesium silicide, prepared by the direct combination of the elements, was dropped into a solution of ammonium bromide in liquid ammonia. A reaction was noticed to proceed immediately with the liberation of considerable quantities of gases, which upon examination were found to consist chiefly of hydrogen, silane and disilane, and of small amounts of trisilane. To date ap. proximately 30 liters of these silanes have been prepared by this method, with yields ranging from 65 to SO% based on the silicon used in the preparation of the magnesium silicide. The yield obtained here is about three times as great as that found by Stock and co-workers [Stock, "Hydrides of Boron and Silicon," Cornel1 University Press, Ithaca, New York, 1933, p. 211, who employed the analogous aqueous method. THEGEORGE HERBERT JONES WARREN C. JOHNSON LABORATORY T. R. HOGNESS UNIVERSITY OF CHICAGO CHICAGO, ILLINOIS RECEIVED APRIL18, 1934
~
