COMMUNICATIONS TO THE EDITOR
Jan. 20, 1954
pounds is further illustrated by their apparent dissociation constants.6 PK’a
66% D M F
11-Membered ring (Id) 9-Membered ring (Ib)
Water
7.8 10.6
9.1 9.2
The increment in pK’, observed for I b in changing solvent from 66% dimethylformamide t o water is reversed in sign from that usually observed for tertiary amine-H+ (compare Id, which is normal), but is typical for enols and acid types, wherein hydrogen is being removed from neutral oxygen, and is therefore evidence of the existence of the transan-
I
I
nular quaternary form, R-N+-C-OH,
631
N40e: C, 45.40; H, 4.90; N, 15.13. Found: C, 45.56; H, 4.86; N, 15.13); perchlorate, m.p. 260-261’ dec. (Calcd. for C ~ H & l N O ~ :C, 39.75; H, 6.67; N, 5.79. Found: C, 39.85; H, 6.51; N, 5.66). THE NOYESCHEMICAL LABORATORYNELSONJ. LEONARD UNIVERSITY OF ILLINOIS RICHARD C . Fpx URBANA, ILLINOIS MICHINORI OK1 STEFANO CHIAVARELLI RECEIVED SOVEMBER 24, 1953
SELECTIVE REDUCTIONS OF AROMATIC SYSTEMS TO MONOOLEFINS
Sir:
of the base
In an earlier publication from this Laboratory‘ we announced the surprisingly large uptake of lithconjugate acid in dimethylformamide. The 8-membered ring6 aminoketone, l-methyl- ium metal by aromatic ring systems in a solvent of 1-azacyclooctan-5-one (If), exhibits carbonyl ab- ethylamine. Thus, in some instances, six equivalents of metal were absorbed per benzene nucleus. sorption a t 1683 cm.-l and no absorption above This clearly indicated that far more extensive re3000 cm.-’; the perchlorate is transparent in the duction of the aromatic rings was occurring than 6 region; pKta9.75 in 66% DMF, 8.75 in water. had been observed previously in the sodium-amFormulations involving interaction of 3 amine system. and ketone functions are thus indicated for 8-, 9- monia We now wish to announce the application of this and 10-membered rings in which a full transannular discovery to practical organic syntheses. Thus, bond between N and Cco can create a 5- or 6-mem- employing the lithium-ethylamine reducing mebered ring within the larger cycle. Diethyl y, 7 ’-methylimino-bis-butyrate was dium, we have succeeded in reducing naphthalene, in treated with sodium in xylene under acyloin ring- in a one-step operation, directly to Ag~lO-octalin 52y0 yield. The nitrosyl chloride of our product closure conditions to give (53%) l-methyl-l-azacy- melted a t the same point as an authentic specimeq2 clononan-5-01-6-one (Ib), m.p. 95-97’ (Calcd. for C9H17N02: C, 63.12; H, 10.01; N, 8.18. Found: and a mixed m.p. was not depressed. A 2,4-diniC, 63.74; H, 9.98; N, 7.90); perchlorate7, (Calcd. trobenzenesulfenyl chloride derivative melted at for C9H18ClNOE: C, 39.97; H, 6.63. Found: 142-142.5’. Anal. Calcd. for C16H1904N2SC1: C, C, 39.70; H, 6.93). Similarly, from diethyl y,y’- 51.82; H, 5.14; N, 7.56. Found: C, 51.57; ethylimino-bis-butyrate was obtained (60%) 1- H, 5.46; N, 7.65. Likewise we have shown that tetralin is reduced ethyl-l-azacyclononan-5-ol-6-one (IC), b.p. 64-65’ in 68% yield to A93l0-octalinby lithium in ethylam(0.15 mm.), n Z o1.4999 ~ (Calcd. for CloH19N02: C, 64.83; H, 10.33; N, 7.56. Found: C, 65.12; ine, thus indicating that naphthalene may pass H, 10.49; N, 7.69); perchlorate’ (Calcd. for through this intermediate in its reduction to A9,loCIOH~OC~NOB: C, 42.08; H, 7.08. Found: C 41.78; octalin. Certainly the above reactions appear to H, 7.13). From diethyl 6,6’-methyl-imino-bis- be the method of choice a t the moment for the prepvalerate was obtained (47%) l-methyl-l-azacyclo- aration of this olefin. It is noteworthy that sohendecan-6-01-7-one (Id), b.p. 85-86’ (0.2 mm.), dium in liquid ammonia is reported3 to reduce n 2 0 1.4926 ~ (Calcd. for CllHzlNOz: C, 66.29; H, naphthalene to tetralin. We have found that lith10.62; N, 7.03. Found: C, 66.52, H, 10.70; ium in ammonia acts similarly. The greater basicN, 7.21) ; perchlorate’ (Calcd. for CllHzd21NO~: ity of ethylamine compared to liquid ammonia may C, 44.04; H, 7.39. Found: C, 44.40; H, 7.67). be one of the factors causing more extensive reducDiethyl 6,6’-ethylimino-bis-valerateyielded (64%) tion in these cases, although we have not as yet l-ethyl-l-azacyclohendecan-6-ol-7-one (Ie), b.p. 65’ proved this point. In another case we reduced biphenyl to l-cyclo(0.02 mm.), n Z o 1.4907 ~ (Calcd. for C I ~ H ~ ~ NC,O ~ : 67.56; H, 10.87; N, 6.57. Found: C, 67.39; H, hexylcyclohexene in 66% yield, the nitrosyl chlo10.67; N, 6.45). Dieckmann ring closure of di- ride derivative giving an undepressed mixed m.p. ethyl y,$-methylimino-bis-butyrate a t high dilu- with an authentic ample.^ Sodium in liquid amtion in xylene using potassium t-butoxide,s followed monia is reported5 to reduce biphenyl to l-phenyl-lby hydrolysis and decarboxylation, yielded (20%) cyclohexene, again indicating a marked difference l-methyl-l-azacyclooctan-5-one,purified through in the two systems. Similarly benzene is reduced by lithium-ethylthe picrate, m.p. cu. 300’ dec. (Calcd. for C14H18amine to cyclohexene, showing again a tendency ( 5 ) We are indebted to Mr. Donald 0. Woolf, Jr., and Dr. Harold E. for the reagent to leave one unreduced double bond. Boaz of Eli Lilly and Company, Indianapolis, Ind., for the electromet-
I
I
tic titrations and for aid in their interpretation. (6) € C. I. Brown and E. A. Fletcher, THISJ O U R N A L , 78,2808 (195l),
/have demonstrated the coardination of l N . with B-0/ ’
‘0-
8-membered ring in triethanolamine borate. (7) Hygroscopic glaas. (8) N.J. Leonard and R. C. Sent:, Ibld,, 14, 1704 (1969).
across an
( 1 ) R. A. Benkeser, R . E. Robinson and H . Landesman, THIS JOUR7 4 , 5699 (1952). (2) P . D . Bartlett, F. E. Condon and A. Schneider, ibid., 66, 1558
NAL,
(1944). ( 3 ) P. Lebeau and M. Picon, Con#f. rend., 168, 1514 (1914). See also C. B. Wooster and F. E.Smith, THISJ O U R N A L , 63, 179 (1931). (4) R. Cdcgce, E. Vogel and H. Hoger, Be*., 811, 144 (1952). (6) W.Ruokol and H.Brrtmohnmidor, Ann., 640, 167 (1939).
GR2
COMMUNICATIONS TO THE EDITOR
Vol. 76
In this regard i t seems t o bear some resemblance to the calcium-ammonia reducing system, but differs from i t in being far more convenient to handle on a laboratory scale. Our reagent is also effective in reducing compounds which contain functional groups in the aromatic ring, but a fuller treatment of this subject will be announced later.
84.0, 88.5; mol. wt. cryoscopic (benzene) 78.0, 78.0,) As hydrolysis even under very mild conditions yields HCOOH and NH3 quantitatively,s any formula with C-C or N-N groups is definitely excluded. Most obvious is a symmetrical ring structure I suggesting that the “dimeric hydrocyanic acid” is in fact the yet unknown 1,3,5-triazine, the parent compound of so many technical imporCONTRIBUTION FROM THE CHEhfICAL ROBERT A . BEXKESER tant substances which for a long period many chemLABORATORIES OF PURDCE USIVERSITY ists have tried in vain to prepare. I , - \ l ~ h Y E ‘ T T E , ISDIANA ROBERT E. R O R ~ N S O N The ultraviolet and the infrared absorption DALEM. SAUVE OWEN13. THOMASspectra of C3H3N3are closely related to other simple 1,3,5-triazine derivatives of established structure, RECEIVED NOVEMBER 27, 1053 e.g., 2,4,6-trimethyl-lj3,5-triazine and 2,4,6-trichloro-173,5-triazine. The highly symmetrical 1 ,J,S-TRIAZINE’.’ formula I representing a symmetrically shaped .Sir : molecule is in agreement with the extreme volatility Our recently published experiments on the mixed and the comparatively high melting point 186’) polymerization of nitriles to unsymmetrically sub- of C3H3N3. stituted triazines 11 \T with hydrogen halides as catalyst3 + Br-f-Br .HRr or failed with the sims /x N /N plest nitrile known, A vzz., hydrocyanic H Br acid. This led us I IIa IIb to a more detailed study of the reaction €I of hydrocyanic acid CtjHz-N--N-CeHs --+with hydrogen chloN W N ride the first product I 1 of which is the soc1 H +Me called “sesqui-hydroIV V I11 chloride of hydrocyanic acid,” 2HCX.3HC1, recently formulated as Bromination of C3H3N3yields a very sensitive CHC12-NH-CHCl-NH2. -1s demonstrated by bromo compound which contains three atoms of Hinkel, et ~ l .this , ~ compound easily loses two moles bromine per one C3N3unit, but is not identical with of HC1 to a compound 2HCN.HC1, regarded as the known cyanuric bromide. (Calcd. for C3HzN3CHCl=N-CH=NH, which splits off its last mo- Br3: Br, 74.97. Found: Br, 72.25, 73.18.) This lecule of HCI by treatment with dehydrohalogenat- compound reacts with aniline to 2,4-dianilido-1,3,5ing agents, e.g., quinoline, to the so-called “dimeric triazine(II1) (m.p. 316’. Calcd. for Cl~H13N6: hydrocyanic acid”, CzHzN2, first obtained by a dif- C, 68.40; H, 4.98; N, 26.60. Found: C, 68.36, ferent route by J. U. Nef.6 C ~ H ~ is Nconsequently Z 68.45; H, 4.95, 5.17; N, 26.60, 26.62) which we then formulated as C=N-CH=NH, imino- have also prepared from 2,4-dianilido-6-chloro1,3,5-triazineg (IV) through the intermediate V f ormyl-carbyl-amine. However, some of the physical as well as chemical (1n.p. 170-171’. Calcd. for C16HI6N6S: C, 62.11; properties of this substance raised serious doubts as H, 4.89; N, 22.64. Found: C, 61.83, 61.80; H, to its constitution. Therefore we reinvestigated 5.04, 5.05; N, 22.78, 22.89) by a method recently the molecular weight of the “dimeric hydrocyanic developed for the replacement of halogen by hydroacid” which led to the surprising result that this gen in the triazine series.’O Both compounds were apparently well known substance is really a trimer found to be identical which adds a further chemical of hydrocyanic acid, C3H3N3.7 (Cakd. for C3H3K3: support to the assumed triazine structure of C3H3N3 mol. wt., 81. Found: mol. wt. (Rast), 89. 9, 90.1, and also makes it Probable that the bromo cornpound is either 2,4-dibromo-173,5-triazine hydro(1) This article is based on work performed under Project 116-B of bromide (Ira) Or 2,4,6-tribromo-1)2-dihydro-1,3,5the Ohio State University Research Foundation sponsored by the triazine (IIb). Mathiesen Chemical Corporation, Baltimore, Md. (2) Triazines VI: communication by Ch. Grundmann, L. SchwenThere still remains the task of explaining the nicke and E. Beyer, B e y . in press. formation of triazine from the adducts of hydrocy(3) Ch. Grundmann, G . Weisse and S. Seide, Ann., 677,77 (1952). acid which occurs so anic and ( 4 ) L. E. Hinkel and G . H. R. Summers, J. Chcm. SOL.,2813 (1952). ( 5 ) L. E. Hinkel, et OL., 1834 (1930); 2793 (1932); 674 (1935); and under certain conditions with such good 184 (1936); 407 (1940): 1953 (1949): 2813 (1952). yields. Work along these lines is now in progress. (6) J. U . N e f , A n n . , 287, 377 (1895). A reinvestigation of the known reactions of C3H3N3 17) The discoverer of this compound, J. U. Nef,e based the formula
0
hT
&
\
CsHiNp on a single cryoscopic determination in benzene, which gave a value of 64 (instead of 54 for CnHnNs). Feeling unsatisfied with this result which he attributed to the poor quality of the benzene available in his laboratory, Nef promised a redetermination as aoon as he could obtain pure benzene, but apparently this has never been published.
(8) L. E. Hinkel, E. E. Ayling and J. H. Beynon, J . Chem. S O L . , 676 (1935). (9) J . T . Thurston, e1 id.,T H I SJ O U R N A L , 73,2981 (1951). (10) Ch. Grundmann, H. Clriuh and A. Kreutzberger, Bcr., 86,
181 (1953).
