ANISOTROPIC OSCILLATIONS IN HEXAMETHYLENETETRAMINE CRYSTAL
July, 1947
tory and curves 1loo%' and 112'28' are poor, tfne
7
-tenth maximam, far example, having merged
with its neig&k0~ the right. (3u+ conclusion is that m the hexamethylenetetramine molecule in the gas phsse the C-N bond amgles are 109.5'
withalimitoferrorof *lo. The oQsei7rad go d u e s for maxima and minima are indicated by wwtkal arrows in Fig. 1 and are listed in Table I with the ratios q m . / q 0 for the finally accepted model with C-N = 1.48 A., and tetrahedral angle. Consideration of Table I, and of similar compsons for models with L CN-C equal to 107 58' and 110'58', leads to th result C-N = 1.48 A. with a limit of error o *0.01 A., if (seeabove) LC-N-C = 109'28' f aad Stosickl also found C-N = 1'. Ham1.48 A. by the correlation method, but reported C-N = 1.47 A. ( ~ 0 . 0 2A,) instead, perhaps on the basis of their imperfectly resolved radial distribution peak a t 1.47 A.
8
1557
68'0 63.0
1.W. 1.011.
68.6
1.008
8 9 10
74.7 1.009 78.8 0.990 11 81.6 .98' 11 86.0 .w1+ 12 89.6 1.003* 12 94.0 1.ooO 13 97.6 0.995 13 100.7 1.001 Averageb1.001 Average deviation 0.008 Values of Rampson and Stosick.1 For ow best model, qcdd./qd = O.QBf3, a. d. 0.029; if fust, second, fourth and fdth values are r&ed (inner dugs, rings adjacent t o A ) , qod&r,f 1.003, a. d. = 0.016. *If only &e six,-s sym&tric& and pres?mably most roiiablg &edwuable €&&wes rtpresaatad by ststrrtf values are considered,the s-ge is 1.001 and the a m deviation is 0.005. HEKAPI[ETHYLENEAularowlsdkpemL-We wish to express cwtr gratitude to Prdessix Linus Pauling for suggest-
10
!
TABLE I ELECTRON DIFFRACTION DATA FOR TETRAMNE
Max.
Min.
goto
1 1
2 2 3
9.4 13.6 17.8 23.2
A A 3
4
4 5
5
27.9 33.6 36.5 39.8 43.0
6
50.3
6
7
PO
5.7 9.3 12.7 17.9 22.3 26.5 27.8 29.7 33.6 36.2 39.1 43.5 47.3 51.4 53.9
QdCd./clO
(1 ,070) 0.989 0.986 1.005" 1.020 1.008 0.993 1.000 1.003 0.985 0.998 1.000* 1.013 1.012 1.018
--
ing this hWtStiiigratiaat, to Dr. R A. S p w and Mr. Ketlt Harmon for assistance in preparing the p h a t o p p h s a d cutvles, and .to the International B e - Machines Corporcpeion for the loan of the machines used in making the punched card calculstians.
s==w
A reinvestigation of hexamethylenete$xamine vapor by electron difFraction has led to re& (C-N = 1.48 * 0.01 A. a d LC-N-C = LNC-N = 109.5 A 1'; C-H 1.09 A. and LHC-H = 109'28' assumed) in agreement with but more precise than those of the earlier investigation, PASADENA, CALIFORNIA
RECEIVED J a y 29,194f3
[CONTRIBUTION FROM THE GATESAND CRELLIN LABORATORIBS OF CHIEMISTRY, C A L I F O ~INSTITUTE U OF TECHNOLOGY No. 10821
Anilsvttropic Oscillations in the Hexamethylenetetramine Crystal BY P. A. SHAFFER, JR.
Whereas the carbon-nitrogen distance found for gas molecules of hexamethylenetetramine by the electron Mraction method,' 1.48 A., agrees satisfactorily with the value predicted from the consideration of covalent radiiJ2previous investigations of the structure of hexamethylenetetramine crystals*have indicated a lower value, about (1) G. C. Hampaon and A. J. Stosick. TRISJOURNAL, 60, 1814 IlQa8); V. S c h o ~ k f vhd f P. A. sylo&r.Jr., Wd.,OS, 1555 (1947). (2) L. PasZhg, "The Natute of tbe C h r l i a I Bond," 2nd ed., Cornell Univcnitg Pres,IthDcp, N o r Yo&. 1840, p. 164. (8) R. G. D i c W n and A. L. Raymond, " m a JOURNAL, 46, 22 (1928); H. W. Don& and H. Mark, 2. p h y d k . Chcm., 107, 181 (lQ28t; R. W. G. Wnkeff and R.B.C o n y , E . Krist., re, 462 (1984); a.Brill, a.0.Grimm, C. Hermann, and A. Peters, Ann. Physik. 84,
898 (1989).
1.44 A. The work reported in t h i s paper was carried out in order to determine whether or not the reported difference in this interatomic disfance for the crystal and the gas is real. If the carbonnitrogen distance is indeed t h i s short in the crystal, it may be of interest in connection with the anomalously short carbon-nitrogen bond distances found in the amino acids.' In the first X-ray study of the crystal structure of hexamethylenetetramine Diclrinson and Raymond found the space group of the crystal to be I43fi, with the edge of the cubic unit equal to (4) For a review of observed carbon-nitrogen bond dfstmrrs m E. W. Hughes and W. N. Lipsaomb, T g ~JOWP.NY, s 68,1970 (1046).
P. A. SEAFFHR, JR.
1558
7.02 A. There axe two molecdes in the unit, the eight nitrogen atoms being in the positions (0, 0,O; 3,3,*) u,u,u; tZ,fZ,u; a,@; u,fZ,Band the twelve mrbon a@ms in the positions (O,O,O; 3,3,3) v,O,O; O,v,O; O,O,v; 5,0,0;O,e,O; O,O,S. The values of the parameters u and v reported by the four previous sets of investigators are given in Table I, together with the corresponding values of the carbon-nitrogen distance. TABLE I U
C N
1)
0.120 a0.003
0.235 a0.004
.12
*
.01
.26
*
.I2
a .01
.23
h
.1204
*
.2363 * .2360 *
.01
1.44 * O 04 Dickinsonand 1 48
h
Raymonda 08 Gone11 and
Mark
.1220
*
.wo8 .002
*
.08 Wy&oE and Carey .0004 1.443 * .OlO Brill, 61 ol. .002 1.448 a .016This paper
.01
1.42
“The probable errors assigned here are one-third of
the maximum esrws determined by Dickinson and Raymond by a conservative method.
‘
Careful spectrometric measurements of the intensities of r&ection from some planes of ‘hexarmethyleneklmminewere made with CuRa radiation by W y e k d and Corey, and another set of careful measurements was made with MoRa radiation by Brill, G.rimm, H e m a m and Peters. Wyckoffsn$ k e y did not attempt a precise evaluation of the parameters u and w from their data; precise parameter values were, however, reported by Brill and his ccllahrators. In the present investigation a redetermination of the parameters u and Y has been made with ttse of the data reported by W y c k d and Corey and those re rted by Brill and collaborators, analyzed by t e method of least squares. In addition to the introduction of a temperature factor of the usual form, a special temperature factor corresponding to rotational thermal vibration of the hexamethylenetetramine molecules in the crystal has been considered. The existence of this rotational vibration is indicated by new X-ray data obtained a t low temperature; it is also indicated by the ellipticity of the peaks of the Fourier projection of the crystal as reported by Brill, et U E . ~ The parameter values found by our least squares calculations, also given in Table I, differ slightly from those reported by Brill and collaborators, and correspond to the value 1.45 * 0.01 A. for the carbon-nitrogen distance in the molecules of hexamethylenetetramine in the crystalline substance. A Low Temperature Study.-Several 30’oscillation pfiotbgraphs of hexamethylenetetramine Were taken with CuKa radiation a t room temperature and a t a temperature approaching that of l i p i d air, the axis of oscillation being parallel to the mbe edge. This work was done in collaboration ~ 4 t hDr. Lindsay Helmholz; the apparatus used to maintain the crystal a t the lower temperature was that described by Helmholz.6
r
(6) L. Helmholz, J . Chcm. Phys., 8,740 (1935).
Vol. 69
The crystal, roughly 0.3 mm. in each dimension, was lacquared to a small aluminm wire which served as a c o n d u c w mppolit, a d it was necessary to mat the myskal itself with b e i p s to prevent its rapid sutpdimtiona t the fow presswe used. The evacuated vessel whi& $s.ltsed the crystal was provided wit$ windows (of film bsse) through which the incident and r&ect& X-ray beams could pass. The photographs do not pErmit intensity estimates to be made which approach the accuracy of the spectrometric determiaations of the earlier qapers. Direct estimation of inknities of ref3ection on the films yields the intensities uncorrected for e x t i n d m and perfection ( p r h b l y slight in the liquid air case) of the crystal and for non-uniform absorption of the reflected beam by the lacquer and windows. The intensities of the roomtemperature photographs are qualitatively in agreement with the intensities dekmined by use of the spectrometer. The liquid air photographs indicate the existence of a large average amplitude of thermal vibration a t morn temperature, the large angle reflections being much stronger a t the lower temperature than at the higher one. The edge of the unit cell a t the lower temperature is 6.95 A. A comparison of the ratios of intensities of some pairs of reflections for both temperatures was made. At liquid air temperature the reflection (640) is stronger than (730), but this relation is reversed at room temperature, despite the fact that the effect of the ordinary temperature factor would be to decrease (730) slightly more rapidly than (640) with increasing temperature. Calculated intensities based on u = 0,119 and v = 0.234 are in the order (640) > (730) a t low temperature as well as room temperature, and the expected slight increases in the values of the parameters due to contraction of the cell on cooling (assuming that the molecules do not change in size) are not capable of producing the observed inversion of order. The reflection (310) appears weaker a t the lower temperature than a t the higher one. This also is contrary to the usual temperature effect. The Calculation of the Structure Amplitudes. -A preliminary Fourier projection of hexamethylenetetramine was made in the direction parallel to a cube edge, with use of the data of Brill, et ul. The parameter values obtained from this projection are u = 0.119 and v = 0.234. Structure amplitudes were calculated for the (hkO) reflections with these parameters. The scattering factors used were those of Hartree which are tabulated in the “Internationale Tabellen zur Bestimmung von Kristallstrukturen,” Vol. 2, p. 571.8 For the nitrogen atoms the neutral N factors were used, and for the methylene groups the scattering factors were taken equal to those for
.
(6) Published by Gebriider Borntraeger. Berlin, 1935; revised edition, reprinted by Edwards Brothers, Inc , Ann Arbor, Michigan,
1944.
ANISOTROPIC OSCILLATIONS IN HEXAMFTTHYLENETETRAMINE CRYSTAL
July, 1947
1558
use of International Business Machines Corporation machines in about one how? M&btt@ the least square computatiow result in improved agreement of the calculated and experhhental F's as indicated by a several-fold decrease in the s u m of the squares of the residuals in all cases, somewhat Werent parameters q i s e from the two sets of data. The agreement between the ~ W Osets of results is improved by the omission of the first few reflections to avoid uncertainties due to exF = exp -B (4fcH, (cos 2 r her + cos 2 r ku + tinction. IEowever, there remain important discrepancies (1) the temperature inversion of the 1) + S ~ COS N 2 x h~ COS 2r Ku) for h + K = 2n (1) intensity relation between (730) and (M),(2) F=Oforh+k=2n+l the opposite of the expected effect of temperature where fcaSis the scattering factor for the methyl- on (310), and (3) the disagreement of the calcuene group, etc., and the factor exp( -B(sin%)/Xa)is lated and observed F's for (10.2.0) and (11.1.0). a temperature factor corresponding to isotropic oscillations of the atoms. In the calculations The Effect of Angular Oscillation of the Molecule which are given under F1 in Table 11, B is 4.5. In an effort to explain the temperature effects Wyckoff and Corey used no explicit temperature noted in the preceding paragraphs as well as to obfa6tor in their calcdations. tain a better over-all agreement of Calculated and TABLE I1 observed structure amplitudes, the roughly spherical molecule of hexamethylenetetramine was asFB F w Fw hkO (obs.) (obs.), (calcd.) Fi F2 F: sumed to undergo angularly symmetric oscilla110 52.8 52.6 52.8 48.7 47.3 44.1 tions about its center of gravity in addition to the 200 20.4 17.6 22.0 18.5 18.0 15.9 isotropic vibration of each atom about its as-12.1 -12.2 220 14.5 16.6 -12.9 -10.0 310 3.19 5.3 - 6.9 - 3.7 - 4.2 - 4.6 sumed rest point. For convenience, these two ef2.5 0 - 4.6 2.2 400 0 2.3 fects were treated separately, the isotropic part of 330 10.6 14.6 12.8 8.7 7.7 8.6 the temperature factor appearingas exp( --B(sin20/ 4.9 5.4 420 5.84 8.9 4.6 4.3 X2) as before; B, therefore, is in units of A.2. The - 2.5 0.7 0.5 0.2 510 1.20 0 440 21.1 23.1 26.0 20.2 22.2 21.0 introduction of a corresponding constant b which 10.0 530 9.16 13.0 11.5 8.6 9.2 is related to the degree of anisotropy due to angu600 0 0 2.0 2.4 0.4 0.1 lar oscillation is exemplified by the treatment for - 3.4 - 3.1 620 3.10 4.9 - 3.0 - 2.6 the methylene group assumed to be a t the point 550 7.04 10.4 10.9 7.9 8.6 7.7 710 2.76 3.7 4.7 2.8 2.0 2.3 (O,O,v) on the z axis. The time average distribu3.0 640 2.08 2.4 4.3 3.4 1.8 tion of this group is taken to be (neglecting the iso- 1.8 - 1.8 730 2.31 3.5 - 2.4 - 1.1 tropic factor)
difference t ~ & ~ e e nthe factors for the oxide iort aftd $or neuU oxygen. Table I1 contains the stmcture a obiserved by Brill, Grimm, Hermann and Peters, FB(obs.), those observed by Wydmff and Cm-ey, Fw(obs.), the calculated amplitudes of Wyckoff and Corey, Fw(ca1cd.), and those calculated using the above parameters, FI. The expression for the structure amplitudes of the (hkO) planes is carbon plus the
{
800 820 660 750 840 910 930 770 10.0.0 860 10.2.0 950 10.4.0 11.1.0 880
7.58 2.39 3.35 1.84
CT)I
-
0
3.57 0 0 0 0 1.34 0 0 1.06 2.28
-
7.1 8.7 2.4 1.2 0.7 - 2.6 0 - 1.3 0.7 0.8 3.0 4.4 0.2 0.5 0.2 - 0.8 1.3 0.7 0.2 0.4 0.2 - 1.1 0.3 0.6 0.2 - 0.5 0.1 - 1.2 1.3 1.9 4.5 2.25 0 2.5 0.119 .119 .234 .234
-
-
-
7.8 2.1 2.7 1.3 0.6 3.3 0.3 0.3 0 0.6 1.1 0.6 0.3 0.7 1.6 2.4 B 2.8 b .122u ,235r
where x and y are in fractions of the unit cube edge, a. The component of the structure am@tude which is due to this atom is therefore FCHr (o,o,V)
=
11
=P {2xi(hx -t Ky
ma
-m
+ ZZ))
Summed over the methylene groups for the (hM) planes, the final result is
The agreement of Fl with the observed F s
leavesmuch to be desired. In order to obtain the best possible fit, feas4-quares calculations were carried out with increments of the parameters SLS the variables in ttre manner described by Hughes'; in addition, the scsle factors which need to be applied to the o b e d amplitudes were adjusted as a part of the feast squares treatment. The least squares nurrnal eqzaations were obtained with the (7) E. W.Hughes, THIS Jomna, 68,1787 61941).
P(%Y) b d y
-m
cos 2r ko.exp
(- 'E2) + exp (FOE, =
0 for h
+k
'w) 1
for h + k 2s 1
5
+
0
2n (4)
In a similar way the total contribution of the nitrogen atoms is found to be (8) This procedure is described by P. A. Sttder, Jr.. VS c h o d u end Lintla Paulhg, J . Chnn. Phys., 14, 657 (1946).
P.A. -,
lsSa PH
Vol. 69
JR.
-qtie.xp { - ~b( h ~ + & ~ ~ ~ bkk e o s b ~ c o s clusion, 2 r h u it seem COS
2r ku
bhh - && Sin !h hU sin 2r &u BaO' 1 for h + & 2% (6)
F~=Ofork+k=28+1
-
The complete structure amplitude is now
F
+ FN)
Sins 9
-r;.)
-P ( - B
(6)
The W-widths of the distribution (assumed equal for CHt and N) are w/, (marr.) = 0.1325 d
d e @ that &e limits of ermt Szd.,mdkate the
assigtKd ti0 thirir F&Bulhrby lh&, exfentoftfieun
mtTQdved.
i e of the Sitnwtwe The C-N distance in the hexamethylenetetramine molecule in the crystal is 1.448 A. The C-N-C angle is 107O19', and the closest approach of the carbon atoms is 2.33 A. The longer C-C distance, that between two carbon atoms in the same molecule and on the same cnbe edge, is 3.30 D
8. The N-N distance is 2.42 A., the N-C-N m ,"11, (min.) = 0.1325 a angle being 113'30'. The non-adjacent C-N distance is 2.78 A. Each molecule is surrounded by (7) 8 nearest neighbors a t 6.08 A. (alsng the body
It is to be noted that the atoms are assumed to vi- diagonals) and by 6 a t 7.02 A. (along the cube brate with random phases with respect to one edges). another. This is certainly not true, but the calcuA Fourier proiection of the.electron density dislations involvirtg the phases are more difficult and tribution in -thk crystal was made parallel- to a
the improvement in the representation is not expected to warrant the greater work. The structure amplitudes were calculated with the new equation with u = 0.119, u = 0.234, b = 2.5 and 4.5, and B = 0. Both sets of F s showed imprqvement over those calculated with b = 0. FOBcamparison, several sets of calculations are tabulated in Table I11 for pairs of reflections in which the temperature effect is most noticeable. The amplitudes calculated with B = 2.25 and b = 2.5 are listed in Table I1 under F2. It is seen that the inclusion of the anisotropic temperature effect improves the agreement considerably. TABLE I11 FB(0bS.)
hk0
310
Strongeratroom temperature 2.08 2.31 0 1.34 0
640
730
860 10.2.0 950
10.4.0 11.1.0
0
1.06 2.28
880
y)- F & y "&,4.7
4.7
6.8
11.1 4.0 2.1 2.5 3.3 3.1 2.1 23.3
5.5 3.5 1.4 3.6 2.0 1.7 4.8 8.3
3.0 3.1 0.2 3.3 1.6 2.2 4.9 4.2
To obtain the optimum values of u, u, B and b, a second set of least-squares calculations were made for both FB(obs.) and Fw(obs.). In this caee nw1y the same structure parameters result from the use of both the Brill data and the Wyckoff and Corey data. The averages of the parameters derived fmm these M o sets of data are u = 0.122, v 0.235, B = 2.4 and b = 2.8; the struccomputed f o r these values of the ture amp&* listed under F8in Table 11. Although no prctkble errors based on the residuals have been computed, the sensitivity of the results to changwin the k t i m of the data lead to estimated limits of erran of about k0.002 for the atam po&imea.nd roughly *0.2 for the temperature factor psram&ers. On the basis of this coni ;
cube edge, using the data of Brill, Grimm, Hermann and Peters and the signs given by column Fa in Table 11. The parameters obtained from the Fourier projection, u = 0.120 and v = 0.234, are each less than the values which resulted from the least-squares treatment. This may be attributed in part to the effect of the foot of the large peak at the origin in raising the inside of adjacent peaks and thus drawing the peaks in toward the larger one, and a similar effect of the nitrogen peaks in drawing in the outer methylene peaks. Brill, Grimm, Hermann and Peters have drawn attention to the distorted shape of the peaks, in their Fourier projection, which they attributed to a rotational oscillation of the molecule. The amplitude of this oscillation is given by the constants B and b. The root-mean-square amplitude of oscillation of each of the atoms along the radius vector connecting it with the molecular center of gravity is about 0.18 A. The root-mean-square amplitude in a direction normal to the radius vector is about 0.26 A. Each molecule is held by twelve weak hydrogen bonds from its carbon atoms to the nearest nitrogen atoms of four adjacent molecules and another twelve from its nitrogen atoms to adjacent carbons. Each nitrogen atom is thus connected by hydrogen bonds to three carbon atoms and each carbon atom is hydrogen-bonded to two nitrogen atoms. The bonding takes place (on the average) parallel to the body diagonal of the cubic unit, and the bonds are 3.88 A. in length, and form an angle of 20' with the body diagonal. The formation of weak hydrogen bonds between the methylene carbon atoms and the nearest nitrogen atom in an adjacent molecule must be assumed to accomt for the stability of the crystal, which has too small an amplitude of isotropic thermal oscillation for van der Waals binding. The decrease of the C-N-C bond angle from 109'28' to 107'19' may be due to the three hydrogen bonds, which pull the nitrogen atom away
July, 1947 CRYSTAL STRUCTURES OF TRXMRTBYL~LATINUM C h m m E AND TETRAMETHYLPLATINUM 1681
for from the center of the molecule along the body with the calculations, to Mr. Kent E prspating the Fourier projBction, and to D r. &gcnd. The same effect would not be. teptized with the carbon atoms because the two hydrogen Lindsay Helmholz for help with the low-kmbpnds conaecting with each carbon atom make an perature photographs. angle of 150’8’ with one another and the two tenS-w sions cancel for the most part. A comparison of X-ray data obtained a t low The existence of the hydrogen bonds between adjacent molecules, however, while expected to temperature with those obtained a t roam temreduce the pssibility of isotropic t h d vibra- perature indicates the existence of rotational thertion, does not greatly interfere with rotational mal vibrations in the hexamethylenetetramine vibrations of the molecule. Each nitrogen is a t crystal. Structure amplitudes calculated on the the top of a triangular pyramid whose base has a basis of assumed rotational vibrations are iu carbon atom of each vertex. The hydrogen bonds closer agreement with the observed amplitudes for form the three near-vertical edges of this figure. room temperature than are those calculated withDisplacement of the nitrogen 0.26 A. ig any di- out the use of a rotational temperature factor. rection parallel to the base may vary a hydrogen A slightly altered structure which provides a Mtbond length by not more than 0.11 A., and leaves ter fit than any other structure considered is derived by the method of least squares. The bond the average virtually unchanged. angles and bond distances are C-N = 1.45 f Acknowledgment.-I wish to express my grati- 0.01 A.; C-N-C = 107’; N-C-N = 113’30’. tude to Professor Linus Pauling for suggesting The carbon-nitrogen bond distance is about 0.03 this work to me and for the many helpful sug- A. less in the crystal than in the gas. gestions which he made during its course, and to CALIFORNU RBCELVBD(MARCH1,1847 Professor Verner Schomakpr for valuable discus- PASADENA, (9) Original manuscript received July 29, 1946. sions. I am indebted to my wife for assistance [CONTRIBUTION FXOY
THE
GAITS AND CRBLLINL A B O R A T OOF ~ ~CHBXISI’XY, ~S CALIFORNIA INSTITUTE OF l ” m y ,
No. 10871
The Crystal Structures of Trimethylplatinum Chloride and Tetramethylpiatixmm BY R. E. RUNDLE’AND J. H.STURDNANT All reported structures of quadrivalent platinum compounds have involved octahedral coordination about platinum. Presumably, however, quadrivalent phtinum is czapable of forming d’sp tebonds, and the platinum alkyls, tetramethylplatinum, hexamethyldiplatinum, trimethylplatinum chloride, etc., were expected to have this configuration.* Cox and Webstera determined the correct unit cell with a = 10.52 A. for trimethylplatinum c h l e d e from powder diagrams, but no real attempt was made to examine the structure. Since no structures involving d2@ tetrahedral bondhg have been verified, it seemed important to investigate the structures of some of the above-named compounds. Experimental Prepurtion of the Compounds.-Trimethylplatinum chloride was prepared by the action of platinum tetrachloride on methyl Grignard reagent as described by Pope and Peachey,‘ except that the platinum tetrachloride was made according to the directions of Kharasch and Ash(1) Resent addnar: Department of Chemistry. Iowa State College, Am-, Iowa. (2) Linw Pading, “The Naturc of the C M c a l Bond,” 2nd ed., Cornell University Pres, Ith.a, New York, 1944, p. 102. (8) E. 0. Cox and K.C. Webata, 2 . Krisl., W, 661 (leSa). (4) W.J. Pope and S. J. Pachey, 1. C h . Soc., $6, 571 (1909).
ford,l and was added to the Grignard reagent as the dry powder as recommended to us by Professor Gilman. The material had the properties described by Pope and Peachey‘ for trimethylplatinum chloride, and X-ray powder linesproduced by it agreed with the measurements of Cox and WebSter.’
Tetramethylplatinum6was furnished us by Professor Gilman. Unlike other compounds of the type M a it is a crystalline material. It is likewise less soluble in organic solvents than are other compounds of this formula. Crystal System and Unit Ceh.-Both trimethylplatinum chloride and tetramethylplatinum crystallize from benzene in anisotropic plates which quickly lose solvent of crystallization to form isotropic powders. From aliphatic hydrocarbons and the ethers they crystdim 8s rhombic dodecahedra of the cubic system. This habit was confirmed by the optical goniometry of several crystals of each compound. When heated the crystals decompose without melting. By the Laue method two, three, and four-fold axes were found for both crystals in positions corresponding to the goniometry, and Oh-m3m was (5) M.5. Kharasch and T.A. Ashford, Tma ]om&, I , 1 m (1938). (6) Henry Gilman and M. LichtenrrPlter, THIa Jouunt, W, (1938).
