efficient method of forming five- and six-meinbered cholanic acid and subsequent alkaline hydrolysis). rings. Sodium in liquid ammonia has not pre- The non-crystalline diester I11 was obtained by viously been considered to lead to practical yields nitric acid oxidation of 12-hydroxycholanic acid of acyloins from esters.' or by chromic acid oxidation of an 11,12-acyloin To a solution of 0.8 g. of sodium in 325 nil. of mixture to 11,12-seco-cholane-ll,12,24-triacid7~6 anhydrous ammonia and 229 ml. of ether (under (1n.p. 259-261'), followed by esterification with pre-purified nitrogen) was added in one hour 1.8 g. diazomethane and selective saponification. Anal. of dimethyl niarrianolate methyl ether (I) dis- Calcd. for C26H~zOa:C, 69.30;H, 9.40;neut. equiv., solved in 223 ml. of ether. After complete re- 4.50. Found: C, 69.28;H, 9.25;neut. equiv., 4-46 riioval of the ammonia by evaporation, the mixI n other experiments employing a ratio of four ture WLS acidified with hydrochloric acid. Purifi- equivalents of sodium to one of the diesters, the catioii afforded OS90 g. (60%) of 1G-keto-a-estra- consumption of sodium appeared to be instandiol-3-methyl ether (11))m.p. 1G2.0-lG2..4°, [ c Y ] ~ ~ Dtaneous. I n each series only one of the four pos-SS" (c = 1.0, 95%) ethanol). Anal. Calcd. for sible isomeric acyloins was obtained. C1yII2403: C, 75.97; 13, 8.03. Found: C, 75.79; (7) 11, Wieland a n d P. Weyland, 2. ghysiol. Chem., 110, 141 (1920). €1, S.09. Conventional heterogeneous acyloin re(8) 13.B. Alther a n d T. Reichstein, Hclr. Chim. Acta., 85, 816 (1942). action conditions (e.g., finely dispersed sodium in DEPARTMENT OF CHEMISTRY JOHN c. SHEEHAN refluxing xylene) did not yield detectable quantities MASSACHUSETTS INSTITUTE OF RICHARD C. CODERRE TECHNOLOGY LOUIS A. COHEN of acyloin product from I. R. C. O'NEILL 39, MASSACHUSETTS The acyloin I1 and I1 acetate showed a depres- CAMBRIDGE RECEIVED OCTOBER 2, 1952 sion of 111.p. on admixture with 16-keto-@-estradiol3-methyl ether4 and the corresponding acetate, respectively. Reduction of I1 acetate by the CRYSTAL STRUCTURE OF CYANURIC ACID ethanedithiol-Raney nickel n i e t h ~ d followed ,~ by Sir: saponification and chromic acid oxidation, afThe infrared spectra of cyanuric acid and forded estrone methyl ether, identified by cnm- deuterocyanuric acid, which were recently pubparison with an authentic s a m p l ~ . lished in THISJOURNAL, by Newman and Badger,' create doubts concerning the values of the interatomic distances found in an early X-ray structure determination by Moerman and Wiebenga.2 The crystal structure of cyanuric acid was reinvestigated in our laboratory by G. A. Croes, ,4. J. van Gent, R. P. van Oosten and D. W, Smits about two years ago. They started from new accurate measurements of the intensities of all X-ray reflections and refined the atomic coordinates by three dimensional Fourier syntheses. This work is not yet published, because we still want to study certain details of the electron density distribution. In connection with Newman and Badger's work, however, i t may be interesting to report here the new values for the bond lengths. These are subject t.0 a standard deviation of approximately 0.02 A., which means that the probability that a given bond length is in error by more than 0.04 A., or a bond angle by more than 3' is approsimately 5%. The molecules are all planar and situated almost esactly in parallel sheets (101). The old and new interatomic distances and bond angles are shown in I11 111 Fig. 1 and Fig. 2,respectively, which show one moleT3y the same technique, 1 1,12-seco-chol3ne-24- cule and its connections to others in the same layer. acid-11,12 dimethj-1 ester (111) was cyclized in 75-The deviation from a trigonal symmetry of the SOYo yield to an 11,12-acyloin (m.p. 142--143'), central ring (Fig. 1) has disappeared (Fig. 2 ) ; identical with the lower melting keto1 prepared by the C-N bond lengths in the ring have practically Iieichstein's procedure6 (brotnination of 1'-ketonot changed and are still all essentially identical. On the other hand, a striking difference is ob(1) E. Chablay, A n n . chim., (9) 8 , 205 (1917), was unable t o isolate served between the new C=O distances and those a n y acyloin material trom such a reaction, and M. S. Kharasch, E. Sternfield and F. R. M a y o , J. Org. Chcm., 5, 362 (IWO), found tlie previously reported. In the previous structure yield of linear acyloin to be much lower t h a n in t h e case of t h e heterodetermination, which was based on estimated geneous reaction. intensities and the use of the trial and error method (2) J. Hew, J . R. Billeter and IC. Miescher, N t l u . Chim. Acta., 88, qnlp, we founod two different C==O distances (1.24 991 (1825). ( 2 ) !V 3 . Sr,hnsi,n and R . C, Christinnseii, TIIISJ i ) u a N A r . . 73, 5511 A. and 1.31 A.), each with an estimated probable (1951) error of 0.06 A., so that the difference was not (4) hf. Iliiffmnn and M. II. L o t t , J. H i d . Chein.. 169, 107 (1917). ?m3
DL---'
(51 17, IIaiiptrnnnn, T r z r ( I > ) I . Iluriirtt iiiiil '1'. I(
( 1 ) R iT;eucrn.rn :mil R ?Didger, I TFIISIOURNAL,74, 3545 (1952) ( 2 ) t I f I V w t 1 r u j I atnil S F \Ii,errnun 7 Krist , 99, 217 (1938).
COMMUNICATIONS TO THE EDITOR
Dec. 5 , 1952
& I
o c
H
o c
I
[ 2.83
7':
I2.81
bT
Q
[I OT]
K
I
I
b
6157
-+
Fig. 1.-Part of crystal structure of cymuric acid. Hydrogen bonds are indicated by dashed lines; atomic coordinates 1938.'
considered to be significant. The recent investigation proves that the $wo bonds haye essentially the same lengths (1.21 A. and 1.215 A., respectively), which is in agreement with Newman and Badger's prediction on the basis of the infrared spectra. The new bond length, however, is much smaller than the mean value (1.28 8.)of those previously observed. It is significantly smaller than the value (1.29 8.)which is expected from mesomerism between the usually accepted resonating structures, with standard lengths 1.47, 1.266, 1.43 and 1.215 8. for the pure C-N, C=N, C-0 and C=O bond, respectively. It may be remarked, that small values for the length of a resonating C=O bond are also observed in othe; carefully investigated and structures, namely, 1.19 A. in N-a~etylglycine~ 1.21 A. in dl-alanine4. (3) G. B Carpenter and J. Donuliue, T r i m J O U R N A L , 72, 231.5 (19.50). (4) H A Levy and K Y Corey, rbrd , 63,POL)> (1041); J Uonohue, i b i d , 72, 940 (1980).
i 0
[I oil
--+
Fig. 2.-As Fig. 1, but with the atoiiiic coorditirttcs takcii from a rrinvestigcttion (1950) of the srructurc.
....
Two kinds of hydrogen bonds NH 0 connect a molecule with others in the same plane. The first type, designated as class A b y Newman and Badger, is directed parallel to the b-axis of the crystal; the second, class B, connects the molecules in the [loll direction. In class A the NH bond is directed exactly toward the hydrogen-bonded oxygen, in class B this is almost exactly the case. In our opinion, however, i t is not possible to predict with certainty which of the two bonds is the shorter, without taking into account the complicated interaction of a molecule with all neighboring atoms. As is seen from Fig. 2 we find that hydrogen bond A is smaller than 2.85 A. and possibly significantly shorter than bond B. This would be in disagreement with Newman and Badger's interpretation of the infrared spectra. LABORATORr O F INORGANIC AND I'HYSICAL UNIVERSITY OF GKONINGEN
CIILSMIS'I'RY
E.11. \ \ ~ I E U U . Y t i A
THEXETIIERLANDS RECEIVED NOV~~JIBLCR 4, 1%2
