Aug., 1942
CRYSTAL-CHEMICAL STUDIES OF THE ALUMS
3. Sodium chloridc, sulfnte and ferrocyanide gavr practically the theoretical potentials whereas low values were obtained with sodium citrate. This is ascribed to complex formation. 4. The behavior of mixtures of cations can be predicted and the motility ratios or the cations
1819
within the clay membranes have been determined experimentally in certain cases. Using these values, the experimental and theoretical potentials for mixtures of cations are found to be in good agreement. COLUMBIA,
[CONTRIBUTION FROM THE SCHOOL OF CHEMISTRY O F THE INSTITUTE OF
Crystal-Chemical Studies of the Alums. IV. Expansion'
M1sso~p.1
RECEIVED MAY4, 1042
TECHNOLOGY O F T H E UNIVERSITY O F MINNESOTA]
Coefficients of Linear Thermal
BY HAROLD P. KLUCAND LEROYALEXANDER
I n thc systenatic study of the crystal chexistry of the alums now in progress in this Laboratory, as e a n y as possible of the phlrsical and clierrical properties of their crystals are being investigated. The only t1iern.al expailsion data on the coenoner a l u m seem to be the sorr.ewhat - ~ detemined uncertain results of S ~ r i n g , ~who the change in density with tercperature for several aluas, and calculated the vulure changes therefrom. Spring coiicluded that his earlier values',3 were vitiated by partial dehydration of the alums and repeated the aeasure~llents.~He was unwilling, however, to claim that his final results solely expressed the volurre change brought abcut by thermal expansion because of the possibility of some dissociation of the hydrates with increasing t e n perature. I n view of the uncertainty of these results, and of the ease with which such mcasurercents can be made by rr,eans of X-ray diffraction, it seemed desirable to deternine the coefficients of expansion of a fcw alums by the XTray technique. Since the characteristic powder diffraction pattern of the alums is particularly sensitive to the effects of dehydration,' the X-ray method presents an important advantage over the pyknometric and dilatorretric methods. This communication presents the results of such reasurerents, for the approximate temperature range 20 to 50') on the following a l u r s : KAl(S04)2.12H20, NH4AI(S04)2*12HzO, T1Al(SO4)2*12H20and N H4Cr (S04)2e12H20. Experimental The potassium and ammonium alums used were from lots prepared for a previous study.5 The (1) (2) (3) (4) (6)
Paper 111. THIS J O U R N A L . 62, 2093 (1040). Spring, Bull. rlasse sei. acud. m y . belg., 131 3, 331 (1882). Spring, Ber., 16, 1254 (1882). Spring. i b i d . . 17, 408 (1684). KIug and Alexunder, Tam JOURNAL, 62, 1402 (1040).
thallium alum was a sample prepared by Dr. N. 0. SrrAith,G and presented by Professor J. E. Ricci for an earlier study.' The ammonium chrome alum was tlic reagent grade salt used without further purification. The experimental technique followed was that of Straumanis and c o - ~ o r k e r s ,which ~ involves precision determination of' the lattice constants a t two different temperatures by thermostating the camera. Details of the thermostat and X-ray technique have been described p r e ~ i o u s l y . ~ ~ ~ FeK radiation was used throughout the study except in the case of the chrome alum where CrK radiation was used. Results The mean coefficient of linear thermal expansion CY, the average increase per unit length per degree centigrade, can be obtained from the expression
where a, is the lattice constant at the corresponding temperature t. The results of the study are tabulated in Table I. For each alum the temperatures and corresponding lattice constants of the separate determinations are listed together with the mean value of a.106 obtained by using all possible combinations of lattice constants separated by a t least a 25' interval. The error is expressed as the probable error of the mean. Spring gives no data which can be compared directly with the values of a in Table J. When his best data4are recalculated, they lead to the follow( G ) Hill, Smith and Ricci, ibid.. 64, 888 (1940). ' (7) Straumanis, levins and Karlsons, 2.aizorg. allgem. Chcm., 488, 176 (1038).
Vol. 64
DAVIDH. VOLMAX TABLE
LATTICECONSTANTS AXD KAl(SOdv12H:O Temp., O C , a , A.
19.3 1 2 . 1333 19.9 12,1335 2c5.0 12.13313 50, ,5 12.1372 51.1 12. 1.378 52,2 12.1373 ol.lOd = 11.0 * O.:i
1
COEPFICIEh’TP O F E X P A S S I O N O F YI€E
‘e.
T e m p . . “C.
12.2141 12.2112 12,2148 12.2180 13.2181 12.21i9 r r . 1 0 ~= 9.5 * 0.2
a.
NH$Cr(SOd:.12FIrO Temp., ‘ C . a A.
A.
18.8 12.2047 19.0 12.z015 19.1 12. so40 25.0 12.2050 50.6 12.2095 51.1 12.2093 a,106 = 13.1 * 0.3
19.2 19.4 1!>, 6 5(.). 1 50. 8 51.4
ing values of a.106for the first three alums listed above: 3.3, (i.8 and 18.3, respectib-ely. These values are of the right order of magnitude but, otherwise, are in poor agreement with the X-ray values. No data are available for comparison with the result for ammonium chrome alum Potassium chrome alum was also studied a t this same time but the inability, after numerous attempts, to get photographs in the vicinity of 30’ without dehydration led to its abandonment. The authors wish to express their thanks and appreciation to Dr. K. 0. Smith and Professor J. E. Ricci for the sample of thallium alum. They
ALUMSBETWEEN 20-50’
TIAI (SOdz.1 L”rO
1;Hd AI (SOq)?.12 Hi20 Temp, a, A.
22.1 25.0 50.5 51.1
13.2501 12.2510 12.2539 12.2543
a.106 = 10.6
*
0.4
also wish to acknowledge with gratitude a grant from the Graduate School of the University of Minnesota under which this study was carried out. Summary The linear thermal expansion coefficients a for several alums have been measured for the approximate range, 20-50°, by means of X-ray diffraction. The values of (r.106observed are as follows: potassium alum, 11.0 * 0.3; ammonium alum, 9.3 * 0.2; thallium alum, 13.1 * 0.3; and ammonium chrome alum, 10.6 0.4. MINNIAPOLIS, MINNESOTA
RECXIVED APRIL 27, 1942
[CONTRIBUTION FROM THE D I V I S I O X O F CHEMISTRY, COLLEGE O F AGRICULTURE, UNIVERSITY O F CALIFORNIA]
The Vapor Phase Photo Decomposition of Methyl Formate BY DAVID H.
Recently, experiments on the photolysis of methyl formate have been reported by Royal and Rollefson.2 Experiments 011 this same problem had been in progress in this Lsboratory. The results for the photolysis products in this investigation are in substantial agreement iT-ith those giien by Royal and Rollefson. In addition, data on the determination of methanol in the reaction proclucts, and an approximate evaluation of the quantum yield for the reaction were obtained. In a subsequent paper on the photolysis of methyl acetate, Roth and Rollefson3 have given data on the determination of methanol as a photolysis product by oxidizing the alcohol to formaldehyde and treating with Schifi reagent. However, they were unable to analyze for the small amounts of methanol usually obtained iri the decomposition runs, and were forced to obtain (1) On leave of absence. Present address: The Technolngieal Institute, Northwestern University, Evanston, Illinois. (2) Roy31 and Rollefson,T H NJOURNAL, 63, 1521 ilDSI! (3) R o t h and Iiollefson. i b i d . , 64, B Y 0 (1042).
\;OLMAN~
comparatively large amounts of decomposition products by carrying out the reaction in a threeliter bulb. The method given below using a Grignard reagent suflices for the small amounts usually obtained and enables analyses to be made for all of the dccompssition experime!its condxted. Experimental Method The apparatus employed was essentially the same as already reported4 The source of radiation was an Hanovia Alpine mercury lamp. Quan tum yields were determined approximately by using nisnochloroacetic acid as an actinometer in a manner previously described4s5using the reliable value for the hydrolysis quantum yield gh-en by Smith, Leighton and Leighton.‘j Methyl formate was synthesized by allowing formic acid and methanol to react in the presence (4) V o l m s n , ibid., 63, 2000 (1941). (31 N‘ei~monti,Berginann and Hirshberg, ibid.. 68, 1675 (1036). (0) S m i t h . Lciyliton und ].eighton, iDid., 61, 2209 (1939).
