COMMUNICATIONS TO THE EDITOR
Nov. 5, 1962
4167
(max.) 258-259 mH (min.)] and from the filtrate Solution of V in glacial acetic acid causes rapid depothere was obtained 193 mg. (67%) of 1,2,3,4-tetra- sition of colorless crystals, containing active halophenylcyclobutene (VI) (calcd. for C2sHzz: C, gen, tentatively identified as 3,4-dichlorotetra93.8; H, 6.2. Found: C, 93.6; H, 6.3), identified phenylcyclobutene (75%) (m.p. 193-195’; calcd. by its spectral properties (X”,t,”,“ 303 mp (19,500) for C2sH20C12:C, 78.7; H, 4.7; C1, 16.6. Found: n.m.r. [CC14]: 2 proton singlet a t 6 = 4.62, 10 Ph C, 79.0; H, 4.5; C1, 16.7), presumably by chloride protons a t 6 = 6.78, 10 Ph protons a t 6 = 7.0-7.5; attack on the unstable 3,4-diacetoxytetraphenylYRaman 1635 cm.-l) and by its conversion to 11 on cyclobutene. V is stable to approximately 150°, a t bromination. which temperature it decomposes to yield tin tetraThe solid dication V is stable to light and dry air chloride and 1,4-dichlorotetraphenylbutadiene, m.p. but reacts instantly when its solution is exposed to 181-153O, identical with the product obtained from water or other weak nucleophiles. Paralleling the the thermal decomposition of tetraphenylcycloknown chemistry of tropylium, aqueous solvolysis butadiene-palladium chloride complex.20 Further of V yields a mixture of products including tetra- studies on the fundamental nature of the dication phenylfuran, cis-dibenzoylstilbene and d i d e ~ y l . ‘ ~are in progress. (13) D. Bryce-Smith a n d N. A. Perkins, Chem. a n d Jnd., 1022 Acknowledgments.-The authors are grateful to (19.59). Dr. P. M. Maitlis for a stimulating discussion and (14) T h e unambiguous proof of structure of V I reopens t h e question for the sample of 1,4-dichlorotetraphenylbutadiene. of the identity of the hydrocarbon C ~ S H Dm.p. Z , 122-123O, obtained from t h e hydrogenation of t h e tetraphenylcyclobutadiene-nickel bromide complex.2 T h e similarity of the 6 values for the nonphenyl protons of V I and t h e 122’ hydrocarbon (two singlets a t 6 = 4.10 and 5 Y7)* suggests certain environmental similarities for these protons, b u t t h e nun-identity of t h e proton singlet of VI (8 = 4.62) t o either 01 these latter bands makes untenable our previous conclusion t h a t the 123’ hydrocarbon is a mixture of t h e cis and trans isomers of VI. (However, t h e importance of t h e proof of structure of t h e 1 2 2 O hydrocarbon a s evidence for t h e presence of a four-membered ring in t h e tetraphenylcyclobutadiene-nickelbromide complex, became an academic one in t h e light of t h e preparation of 11). (15) T h e f a c t t h a t tropyliumls and cyclopropenium1’ salts can be recovered f r o m their acidified, aqueous solutions, whereas t h e dication V cannot be so recovered, reflects t h e reversibility of t h e solvolysis of the former two species (eq. [ l ] )and t h e nonreversibility of t h e latter (eq. [2]). T h e formation of tetraphenylfuran is t h e expected intra-
R‘
t
2H20 C _
ROH
7
H30’
(1)
Ph
Ph
L Ph\
’
OH
vi“
1
,Ph
I
Ph
PI1
v 4
J
I”
vg molecular solvolysis product of t h e mono-hydroxy carbonium ion Vf, analogous t o t h e formation of ditropyl ether by t h e intermolecular a t t a c k of tropylium OD t h e initially formed tropyl alcohol.’a It is also conceivable t h a t t h e unstable glycol Vg is also formed which gives rise b y tautomeric rearrangement t o didesyl (VII). T h e latter may account for t h e presence of cis-dibenzoylstilbene, which b y analogy with t h e disproportionation of tropylium t o cycloheptatriene a n d tropone,’s probably arises b y hydride abstraction from V I 1 b y excess V. T h i s also accounts for t h e variation in production composition, which would be markedly dependent on t h e availability of excess dication V. A detailed s t u d v of this reaction is in nroeress. (16) W. von E. Doering a n d L. H. Knox, J. A m . Chem. S O ~ 79, ., 352 (1959).
~-
(17) R. Breslow, J. Lockhart and Hai Won Chang, ibid., 83, 2373 (1961), and refs. cited therein. (18) W. von E. Doering and L. H. Knox, ibid., 7 6 , 3203 (1954). (19) K. Ikemi, T. Nozoe and H . Sugiyama, Chem. a n d Ind., 932 (1960); K. M. Harmon, et ut., J. A m . Chcm. Soc., 84, 120 (1962). (20) A. T. Blomquist a n d P. 11.Maitlis, ibid., 84, 2329 (1962).
H. H. FREEDMAN THEDow CHEMICAL COMPANY EASTERN RESEARCH LABORATORY A. M. FRANTZ, JR. FRAMINGHAM, MASSACHUSETTS RECEIVED JULY 18, 1962
FORMATION OF IMINES BY ELECTROPHILIC SUBSTITUTION AT SATURATED CARBON IN THE BECKMANN AND SCHMIDT REACTIONS’
Sir: Although the insertion of neutral, electrondeficient nitrene intermediates into unactivated C-H bonds is well known,2 the replacement of paraffinic hydrogen by a catioiiic nitrogen species has not been noted.3 We now wish to report such an electrophilic substitution involving intramolecular attack by an iminium ion,4 generated under Beckmaiin or Schmidt reaction conditions, on a non-adjacent saturated carbon atom. The key model compound, 4-bromo-7-t-butyl-li n d a n ~ n ewas , ~ chosen because normal Beckmann rearrangement of its oxime (I) would be expected to proceed with great difficulty6 and also because (1) This research was supported by Grant No. 550-A of t h e Petroleum Research F u n d , American Chemical Society. Grateful acknowledgment is made t o t h e donors of this f u n d for their generous support. (2) (a) D. H R . Barton and L. R. Morgan, J . Chem. SOC, 6 2 2 (1962); (b) G. Smolinsky, J . A m . Chcm. Soc., 83, 4717 (1960); (c) J. W. ApSimon and 0. E. Edwards, Can. J. Chem., 40, 896 (1962). (3) In two cases, acid-catalyzed reactions of ketoxime sulfonates have yielded as minor products pyrazines, which are allegedly derivable from azacyclopropenes (P. A. s. Smith, J . A m . Chem. SOC., 70, 323 (1948)). Such possible intermediates could result from a n insertion of t h e type discussed here o r by cyclization of a vinyl nitrene (rf. C. Smolinsky, ibid., 83, 4483 (1961)). (4) T h e term “iminium ion” is used here essentially for brevity and is not meant t o infer t h a t polyphosphate ion is first lost from t h e esterified oxime, yielding a discrete cationic intermediate which inserts in a subsequent step. It has been pointed out, however (D. E. Pearsun and R. M. Stone, ibid., 83, 1715 (1961)), t h a t bond-breaking is well advanced in t h e rearrangement of oxime polyphusphates; moreover, aryl participation is not of importance in t h e indanone SYStem (ref. 6). Thus, usage of the term iminium ion appears justified. ( 5 ) L. F. Fieser and D. K . Snow, ibid., 60,176 (1938). (6) L. G. Donaruma a n d W. 2. Heldt, “Organic Reactions,” R. Adams, Ed., Volume 11, John Wiley and Suns, Inc., S e w York, N. Y., l N 0 , pp. 11-12.
4168
Vol. 84
COMMUNICATIONS TO THE EDITOR
no features conducive to oxime fragmentation' were present. Moreover, since iminium ions have been shown to react intramolecularly with such nucleophiles as aromatic rings3 and alkenes, there was reason to expect C-C or C--H insertioti into the proximal t-butyl group of oxime I in the Beckmann rearrangement or in the related Schmidt reaction. Reaction of I with polyphosphoric acid resulted in i5-80% conversion to a 2: 1 mixture of imine I1
(2.5 cm.-l) in the
Br 1
Br I11
and lactam 111. In a typical experiment, 0.63 g. of oxime Il0and 10 g. of polyphosphoric acid were heated (130') for one minute until the oxime had melted and dissolved. After standing for 16 hours at room temperature, water was added and the aqueous solution extracted with ether. Workup of the extract gave 0.2 g. of crude I11 (see below). The aqueous solution then was made alkaline and again extracted with ether. Careful processing gave the imine I1 (vmax 1656 cm.-'; no vinyl proton signal in n.m.r. spectrum) as an extremely unstable material which darkened rapidly. Isolation was possible, however, by conversion to the perchlorate (0.40 g.), m.p. 204-205' (Ymax 1680 cm-') . Sodium borohydride reduction of IIperchlorate afforded the saturated amine IV as an oil (no >C=S-absorption a t 1655 cm.-I); IV yielded a crystalline p-nitrobenzamide, m.p. 186187'. Hydrogenolysis of IV over palladium-oncarbon in ethanol containing triethylamine gave the debrominated amine V, again isolated and analyzed as the p-nitrobenzamide, m.p. 174-175'. Structures 11, IV and V were unambiguously confirmed by an independent synthesis of the latter from 1-indanone and methallylamine. Crude mine VI (vmax 1655 cm.-l) was reduced with borohydride and the unsaturated amine (Vmax 3300 cm.-'; no absorption a t 1655 cm.-l) cyclized to 1' with polyphosphoric acid at room temperature. The 9-nitrobenzamide of V prepared in this manner was identical in all respects with the material obtained from 11. The imine, rather than enamine, structure for I1 was confirmed by the n.m.r. spectrum, and the similarity in magnitude of the shift (7) (a) R. K. Hill, J . O U R .Chem., 27, 29 (1962); ( h j R. K. Hill and 0 T. Chortyk, J . A m Chem. Soc., 8 4 , IO64 (1962) ( 8 ) Y . Arata and S. Sugasawa, Chem. Phavm. B d . , 9, 104 (1961). (9) (a) J. Meinwald, PYOC. Ciiem. Soc., 286 (19.58); (b) R , Griot a n d T. Wagner-Jauregg. H e l a . Chim. Acta, 42, 121, 605 (1959). (IO) ,411 new compounds gave satisfactory C,H analyses.
>
C=N-stretching absorption in
going from the base to the perchlorate salt when comparing I1 and l-methy1-3,4-dihydroisoquinoline.'l The evidence for the structure of the lactam 111, m.p. 134-135', (vmaX 3300 ern.-', 1660 cm.-l) rests mainly on its conversion by hydrogenolysis t o 8-t-butyl-3,4-dihydroisocarbostyril, m.p. 124-126', whose n.m.r. spectrum was clearly different in the -CHz- region from that of 3,4dihydrocarbostyril, the former showing a multiplet centered a t 3.1 ppm. downfield from TMS due to methylene protons adjacent to amide nitrogen, l 2 as in the spectrum of 3.-l.-dihydroisocarbostyril. The Schmidt reaction of the parent indanone was carried out a t ca. 55' in polyphosphoric acid. Under conditions selected to prevent decomposition, over 2Sc7, of ketone was recovered and a fiO-Xi70 yield of I1 and I11 was isolated as above, with I11 now the major product (ca. 2 : 1). Interestingly, 111 results from net methylene migration whereas 1-indanone gives only aryl migration when subjected to the Schmidt reaction.13 The mechanism of the electrophilic bond insertion leading to I1 may be similar to that postulated by Corey for a similar reaction involving positive oxygen, wherein a three-center process was invoked, e.g.
Rr
Such a transition state, in which little neophyl carbonium ion character is developed, would be compatible with the absence of alkyl or aryl migration.'j iVe are presently investigating the mechanism, scope and limitations of this interesting insertion reaction, which may be of appreciable (11) On t h e other hand, t h e pyrrolidine enamine of 1-indanone shows a 60 cm.-i shift t o higher frequency when converted t o t h e salt. (12) L. M. Jackmann, "Applications of Nuclear Magnetic Resonance Spectroscopy in Organic Chemistry," Pergamon Press, New Uork, AT. Y . , 1959, p. 5 6 . Also, for several exemplary spectra, see Yarian Spectra Catalog, Varian Associates, 1962, spectra 68 and 116. ( 1 3 ) L.H. Briggs and G. C. De Ath, J . Chefrt. Soc., 456 (1937). (14) E. J. Corey and R . W. White, J . Ant. Chem. S o c . , 80, 6686 (19.3). (1.5) I n alternative b u t less likely mechanism wuuld invrilve t h e formation o f a vinyl nitrene (see footnote 3) and then insertion.
Nov. 5 , 1962
CORIMIJNICATIONS TO THE EDITOR
4160
propyl ring is not substituted.5 The coinplex proton magnetic spectra of the two compounds are very similar, differing only in slight shifts of peak positions and in relative intensities. These n.ii1.r. spectra have not yielded to analysis, perhaps because several stereoisomers are present. However, protons in a t least eight different environments are found, consistent with the nortricyclene structure (16) Kational Aniline Research Fellow, 1961-1962 and inconsistent with alternative pi-complex forDEPARTMESTOF CHEMISTRY UNIVERSITY OF BUFFALO PETERT LANSBURY mulations. BUFFALO14, SEW YORK JAMES G COLSON~” Upon treatment with water or bases the compounds undergo a remarkable fragmentation reRECEIVED AUGUST27, 1962 action,€with the reformation of bicycloheptadiene and a tin-containing fragment. This reaction TIN TETRAHALIDE ADDUCTS OF may be explained as a delta-elimination brought BICYCLOHEPTADIENE AND THEIR about by attack of the base at tin FRAGMENTATION’
value in the synthesis of heterocyclic rings. I t should be pointed out that insertion products such as I1 may have been formed in other Beckmann or Schmidt reactions and were inadvertently missed since they would remain in the oft-discarded diluted acidic solution from which the usual neutral products are removed by filtration or extraction.
Sir: r iQ Slthough olefins generally are known to be attacked by electrophilic reagents, few studies have been published on the interaction of olefins with strong Lewis acids other than boron hydrides.2 Recently it has shown that BC13 and C6H5BC12will J I\ add t o several reactive olefins, including bicycloBSnX3 heptadiene. We now wish to report the addition of SnC1, and xo SnBr, to bicycloheptadiene (I) to yield crystalline 1: 1 adducts, which like the boron halide adductsY The fragmentation is quantitative and was used are formulated as substituted nortricyclenes (11). for approximate analysis of the compounds, sufficient to show that they are 1:1 adducts. In a typical hydrolysis, the SnC&adduct gave bicycloheptadiene, 25.9; Sn, 35.8; C1, 38.0 (calcd. for C7HI0SnCl4; bicycloheptadiene, 26.5; Sn, 33.4; C1, 40.0). The SnBr, adduct gave bicycloheptaI I1 diene, 16.3; Sn, 19.6; Br, 60 (calcd. for CwHloThese products presumably arise through homo- SnBr4, bicycloheptadiene, 17.6; Sn, 22.3; Br, conjugative addition of the type described by 60.1). Conventional combustion analyses were Winstein and Shatavsky. The compounds are impractical because of the instability of the comprepared simply by mixing the reactants in ap- pounds. proximately equimolar proportion in dry isopenThe adducts are stable pure or under hydrotane a t 0’ and allowing the solution to stand. The carbons a t 0’ or below, but a t room temperature colorless crystals which form may be recrystal- or in contact with polar solvents they decompose lized from fresh isopentane; m.p. for C7HloSnC11, with the formation of tarry masses apparently 60.5-61.5’; for C7HloSnBr,,58-59’. containing polymerized material It seems not The infrared spectra of the two compounds pro- unlikely that similar adducts may be formed vide strong support for the nortricyclene structure. during the polymerization of diolefins by SnCl?. Both compounds have strong bands a t 3070 and 1020 cm.-l, which are diagnostic for the cyclo- Experiments are under way with other Lewis acids and olefins to determine the generality of propane ring in nortricyclene~.~-4 band also the addition reaction. Preliminary results indiappears in the 800-818 ern.-' region in both comform adducts with pounds, consistent with the fact that the cyclo- cate that GeBr4 and C6H5SnC& bicycloheptadiene, but that SnI,, Sic14 and ((26( 1 ) This research was supported by t h e directorate of Chemical H5)*SnC12 do not. Sciences, Air Force Office of Scientific Research, under grant KO.AF-
&7
AFOSR-62.244. (2) H. C . Brown, “Hydroboration,” W . A. Benjamin, Inc., 9 e w York, pi. Y . , 1902 ( 3 ) F. Joy and F . M. Lappert, Proc. Chepn Soc. ( L o n d o n ) , 353 (1960) (4) S. Winstein and 42. Shatavsky, Chem. arid I n d . , 56 (1986). ( 3 ) G. E. Pollard, Spectuockim. Acla, 18, 887 (1962). Cf. J. D. Roberts, E. R. Trumhull, W. Bennett and R.Armstrong, J . A m . C k e m . Soc., 74, 3116 (1950); E . R. Lippincott, ibid., 73, 2001 (1951).
+
DEPARTMENT OF CHEMISTRY OF WISCONSIN FREDRIC M. RABEL UNIVERSITY MADISON6, WISCOSSIN ROBERTWEST RECEIVEDSEPTEMBER i , 1962 (6) C. A . Grob and W. Baumann, H e l u . C h i m . A c t a , 38, 594 (1955); C . A. Groh, “Fragmentation in Solvolysis Reactions,“ in “Theoretical Organic Chemistry” (papers presented a t t h e Kekule Symposium), Butterworths, London, 1959, p p , 114-126.
