amino group migrates in an open-chain intermediate. Compounds (3), (5), and (probably) (6) can be excluded because of the red color. 2. I R spectra: V ~ of(i) = ~151ocm-1 (cH,c~,), 1520cm-' (SO,); v C r N of (CH,),N--CN: 2206 in both solvents. From the red solution new bands appear at 1630cm-' (CHzCI,) or 1652cm-' (SO,). The structures ( 3 ) , (4), (51,and (6) are thus excluded: aminoazirines comparable ] ; (5) with ( 4 ) have a C=N band a t ' 1 7 6 0 ~ m - ' ~ ' ~from one would expect a C r N band at 2190 cm-' (cf. [15, 16]), and for ( 6 ) , as a ketene imine derivative["], a band at 2000cm-'. Of the structures proposed the only one remaining is thus (2) (in 2-alkoxyazetine~['~] vC=N has been observed at 1620cm-').-The band at 1630cm-' is not to be found in the IR spectrum of a pyrolyzate solution that has been kept for 2 days at room temperature, is colorless, and no longer shows an ion of m/e= 182 in the mass spectrum. 3. U Vspectra: The UV spectrum of a methylene chloride solution of the pyrolyzate, measured at -40 C, shows an intense absorption maximum at 527 nm with ~-2OO00 [ C I ): h,,, = 352 nm]. The disappearance of this maximum at 3 0 C follows second order kinetics with a half-life of 56 min (initial absorption 1.100 OD). From theoretical consideration^^'^^ a maximum would be expected for (2) at 670nm; one would predict shorter-wave absorptions for the other structures considered. 4. N M R spectra: The 'H-NMR spectra of the pyrolyzate in mixtures of CH2CI2 and CD2Cll show signals in the range 6 = 2.5-3.5 ppm. Singlets at 6 = 2.75 and 2.9 ppm can beassigned to the triazine ( I ) , and a singlet at 2.85 ppm to dimethylcyanamide. A 60 MHz 'H-NMR spectrum, measured at -60'C, shows, besides these signals, two singlets at 6=3.22 and 2.94ppm. An intensity ratio of 1 :2 can be found from the 220MHz spectrum. If the sample in the spectrometer is allowed to warm up to room temperature these two singlets disappear simultaneously, being no longer present after ca. 12h, whilst at the same time new signals appear at 6=2.31 and 3.35ppm; the solution then has only an extremely faint red color. As in the 'H-NMR spectrum, so in the 13C-NMR spectrum two new signals arise, at 138.29 and 128.28ppm (relative to TMS); these can be assigned to olefinic C atoms. According to the conditions of the pyrolysis and working up, further signals of lower intensity can be observed in the 'H-NMR spectra of the pyrolyzate; some of them decrease in intensity with increasing temperature. The following properties can thus be assigned to the red compound in the pyrolyzate solution: molecular weight 182; UV:h,,,=527 nm ( ~ ~ 2 0 0 0 0I)R; : vC=N=1630 cm-'; 'H-NMR: singlets at 6=3.22 and 2.94 ppm (1 :2). Thence we conclude that tris(dimethy1amino)azacyclobutadiene (2) is formed on pyrolysis of ( I ). From the 'H-NMR spectrum the yield can be estimated as approximately 30 Yo. The relatively high thermal stability and the NMR spectra of the product indicate that (2) should be described as the resonance hybrid ( 7 a ) - ( 7 d ) , where especial irnpor- RzNGNRz 0 R 2 N g R 2 - R 2 N E N R 2 R?N 7 ci: K = 848 CH, R2n' 17b) R2N ( 7 ~ j 0 RzN$NR2 (81 R2N 5. Chemical behavior: Experiments inwhich (2) was treated with methyl iodide, tetracyanoethylene, acetylenedicarboxylic ester, phenyl isothiocyanate, tosyl isocyanate, or Fe(COj5 led to inseparable mixtures of products. The difficulties in working up were partly due to the fact that the triazine ( I ) , which is present in the pyrolyzate in excess of (2) (>40%), reacts with electrophiles to give deeply colored, as yet unidentified Protonation of (2) also seems to take a complicated course. Received: June 22, 1973 [Z 886a IE] German version: Angew. Chem. 85, 918 (1973) [ I ] M . J . S. Dewar, M . C. Kohn, and N . Trinujst~i,J. Amer. Chem. SOC.93, 3437 (1971). [ 2 ] M . J . S. Dewar and N . Trinujstii. Theor. Chim. Acta 17, 235 (1970).  R. Hoffann, Chem. Commun. 1969, 240.  Th. Curtius and K . Thun, J. Prakt. Chem. 44, 175 (1891). [ 5 ] M . Or.starreich, Ber. Dtsch. Chem. Ges. 30, 2255 (1897).  E. Ahderhalden and M . Puyuin, Ber. Dtsch. Chem. Ges. 53, 1129 ( 1920).  L. A. Paquertr. T Kakihuna, and J . F. Kelly, J. Org. Chem. 36, 435 (1971)  E. A. Chandross and G. Smolinsky, Tetrahedron Lett. 1960, 19.  G. L. Closs and A. M. Harrison, J. Org. Chem. 37, 1051 (1972). [lo] B M . Adyer. M . Krating, C. 19: Rees, and R. C. Storr, J. C. S . Chem. Commun. 1973, 19. [I I ] R. Gompprr and G. Srpbold, Angew. Chem. KO, 804 (1968): Angew. Chem. internat. Edit. 7 , 824 (1968). [I21 Cf. [8, 91 and H. Neunhoeffer, H.-D. Vdtter. and H . Oh/, Chem. Ber. 105, 3695 (1972). 1131 R. Gomppcr. U . Jersak, and K . Schiinafnger, unpublished results. [I41 M . Rens and L. Ghost-, Tetrahedron Lett. 1970, 3765.  R . Gompprr, E . Kutter, and W Tiipfl, Chem. Ber. 95, 2871 (1962).  R . Kunz, Dissertation, Technische Hochschule Stuttgart 1964, p. 28 1171 C. L. St<W!7,Sand J . C. French, J. Amer. Chem. Soc. 75, 658 (1953). [I81 G. Pfleri, P. Cousonni, G. Prlrzzn, and E. Tesfa, J . Heterocycl. Chem. 4, 619 (1967). [ 191 H.-U. Wagner, Angew. Chem. 85,920 ( I 973), Angew. Chem. internat. Edit. 12, 848 (1973). 1201 H.Q U U S ~ ,E. Schmirt, and R. Frank, Angew. Chem. 83, 728 (1971); Angew. Chem. internat. Edit. 10, 651 (1971). The Structure of Triaminoazacyclobutadiene[**] By Hans-Ulrich Wagner"] Cyclobutadiene ( 1 ) is antiaromatic; it can be stabilized by push-pull substitution''? Substitution as in compounds of type (2) has been found especially advantageous in this respect12].Analogously, azacyclobutadiene ( 3 ) , which 0 -R2N@NR2 c) tance attaches to the structure ( 7 d ) [ " J . Provisionally the possibility cannot be excluded that the product has one of the angular structures derived from (7d), e.y. (a), although the IR spectrum contraindicates this (cyclopropanimine has C=N bands at 1 7 7 0 ~ m - ' [ ~ ~ ~ ) . R2N ( 7~1) [*] Dr. Institut H -U. furWagner Orgdnische Chemie der Universitdt 8 Munchen 2. Karlstrasse 23 (Germany) [**I This work was supported by the Deiitcche For~chunesgemeinschaft. We are particularly grateful to the Leibniz-Rechenzentrum of the Bayrische Akademie der Wissenschaften. Anyew. Chem. internat. Edit. 1 Vol. 12 ( 1 9 7 3 ) / No. 10 is energetically favored in comparison with ( I ) by introduction of the ring-nitrogen that acts as acceptor, can be further stabilized by two donor groups at positions 2 and 4 [see (4)]. From this point of view a third donor group, as in ( 5 ) , has an unfavorable effect. Introducing a donor group at position 3 relative to the acceptor position, as in ( 6 ) , however. causes distortion of the ring to a rhombus[3' by increasing the bonding interaction and bringing positions 2 and 4 closer together. COOR 2 6 i 2) 1) P13 = O P24 = 0 N R,NQ,,, . O- C OOR -0.40 +0.29 +0.46 NRZ R2N& N NIZz (4) (31 -0.16 Lt2NN dkNR2 ( 5) x N (61 p13 = -0.35 -0.34 -0.39 pZ4 = +0.25 +0.35 +0.51 The transannular interaction of atoms at opposite corners of the four-membered ring can certainly no longer be neglected since thedistance between them of approximately 2 A comes in the region of normal bond lengths. The energy change including the effect of the diagonal interaction can be estimated to a first approximation from the bond orders in the H M O scheme (see the values under the Even in the push-pull-stabilized cyclobutadiene (2) closer approach of centers 2 and 4 und further separation of centers 1 and 3 should bring an increase in energy. As shown by X-ray structure analysis[51of ( 2 ) the measured angles (1-2-3 =93, 263-4= 87") and bond lengths (2-4=2.00,1-3=2.11 A)indicate that the four-membered ring is in fact distorted to a rhombus[61. Although such a distortion does not lead to an energy gain in unsubstituted cyclobutadiene ( I )['I, considerable stabilization is to be expected for (3) in view of the bond orders shown. In the language of resonance theory this distortion favors the polar resonance structure ( 7). In (7) it can be clearly seen how introducing a donor group at position 3 buttresses the transannular interaction between centers 2 and 4 by stabilizing the positive charge [resonance structure (8)]. This effect can also be deduced from the changes in H M O bond orders on going from (3) to ( 6 ) and from (4) to ( 5 ) . The unstable red product of molecular formula C9HL8N4 obtained on flash pyrolysis of 4,5,6-tris(dimethylamino)1,2,3-triazine (9u)['] can have one of the structures ( I O U ) to (15 a). To aid clarification of the structure the electronic spectra to be expected for ( 1 0 ) to (15) were calculated Aiigew. Chem. intrmaf. Edif. 1 Vol. 12 (1973) / N o . 10 by a CNDO-Cl procedure (Table I); but since it can be assumed that the methyl groups have no appreciable influence on the chrornophore of the red compound, the CH3 groups were replaced by H for the calculations["']. Table I . Calculated wavelengths of the longest-wave electronic transitions for various structures C3H6N4(CoH,,NI). __ Structure L [nm] (9hi 386 flOhl 670 f1lhl 737 f l 2 h ) f13hi 194 393 f14h) 282 - fljh) 440 In hexane solution the longest-wave absorption maximum for the triazine ( 9 u ) that was used as reference molecule appears at 352 nm'"]. The calculated value is shifted to longer wavelength from the experimental value['21.With this trend in mind it is clear from Table 1 that the red compound in question (k$y:c'2= 527 nm) can have only structure (10) or ( I 1). The structures ( 1 0 ) and ( 1I ) are topologically closely related. On minimization of the energy of ( 1 0 ) and ( I I ) with respect to ull the coordinates by the MINDO/2 proced ~ r e [ ' ~the ' , most favorable structure with topology close to that of (10) is found to be a symmetrical planar fourmembered ring with strongly distorted diagonal distances as required by the considerations above (1-3 =2.07, 2 to 4 = 1.84A) but approximately equal lengths for the sides: Minimization starting from the geometry of the less stable structure ( I 1 ) discloses a tendency towards planarity, giving a planar rhombus (10). A planar structure ( 1 0 ) with localized single and double bonds is also energetically less favorable than a symmetrical delocalized structure. Bond fixation of the second order appears to play no part in (1o)[l41. According to these results the unstable red compound should have structure (10),as assigned also from the physical data[''. However, the MIND0/2 calculation gives Q an astonishingly low bond index of 0.962 for the C=NH, 849 Specific Heat and Formation of Kinks in Bimolecular Films of Long-Chain n-Alkyl Compoundsr**] By Go.htrrr/ L L I < / LSfr,p/itrri I/~.. Fir:. and 4rn7ir7 W,r.ss[*] bond of (11 h ) ;comparing this value with the bond indices of 1.507 for ( 1 6 ) and 1.537 for ( 1 7 ) shows that the IR wave number of the C=N stretching vibration of ( 1 1 ) is not necessarily so high as for (16) and ( 1 7 ) . Received: June 28. 1973 [ Z 886b IE] German verslon: Angew. Chem. X5, 920 (1973) [ I J J . D. Rohwrrs, cited in J. Chem. Soc. Spec. Publ. No. 12, I I I (1958).  R. Gonipper and G. Seyhold, Angew. Chem. 80, 804 (1968): Angew. Chem. internat. Edit. 7. 824 ( 1968); M. Nt,iirn.s~h~,urider and A. NirdcrIiuirirr. Helv. Chim. Acta 53, 519 (1970); R. Hofiiiuiin, Chem. Commun. 1969. 240. Alkyl chains may change their conformation via rotation about C-C bonds. Among these isomerizations, the simultaneously occurring transformation of two trans bonds (t) into +gauche (8) and -gauche bonds (g), for example ...ttt ... + ...gtg ... or ...ttttt ... + ...gtttg ... etc., appears to be important in crystalline, paracrystalline, or partially crystalline compounds. In these cases the two parts of chain beyond the ...gt _..S... conformation remain parallel to each other. Conformations arising in this way were termed “kinks” by Pechhold’! They have been discussed as structural elements in polymersE2];their formation and rearrangement were assigned to definite relaxation regions (cf. e. g. Ref. [31). n l!j v H 6, ttttt tttt t it a l l - trans tttgtttttttt tttqtqtttttt isolated gauche -bond igi- kinks M 2s 1 tttgtgtgtttt 3 9 2 - kmks 26, tttgtttgtttt 292-kinks 3 6: tttgtttttgtt 293 -kinks [28871/ Fig. I . Conformations of alkyl chains with isolated g u d 7 e bonds and kinks (6,= 1.78, 6,= 1.27, 6;= 1.27 A).  If the lengths of the C-N form corrcsponds t o a kite. and C-C sides are different, the distorted  F o r summands of the first order 6t=2pp,,Ppv: cf E. Nrilhroriricr and H. Boc h . Das HMO-Model1 und seine Anwendung. Verlag Chemie, Weinheim 1968, p. 146. [ S ] H. J . L i ~ l t i ~ ,and r 5 . roii Grus\, Angew. Chem. X3, 489 (1971); Angew. Chem. internat. Edit. 111, 490 (1971).  T h e antibonding interaction in the cyclobutadiene ( 2 ) is further diminished by the facts that the acceptor substituents d o not lie in the plane ofthe ring and that 2- -4 overlap is thus reduced by transhybridization.  In alternant hydrocarbons all bond orders to second nearest neighbors are zero dd. U . J m a h . and R. &mpp<,r, Angew. Chem. X i , 918 (1973); m. internat. Edit. I?. 847 11973). [Y] R. L. Ellis, G. K u c h i l m z . and H . H. Juffc;, Theor. Chim. Acta 26, 131 (1972); J . Kroiirr and D. Proch. Tetrahedron Lett. IY72. 2537, and literature cited therein. [lo] Standard coordinates were used for all the molecules. It was found that for ( 1 0 ) the position of the longest-wave band was not appreciably altered by use of minimum coordinates. As an analytical method the calculations give only qirulirutice information about the spectra t o be ex pectcd. [ I I ] R. Gonipprr, C . Jersuk. and K . Sdiijiiufi~iyer,personal communication. [ 121 An exact analysis of this electronic spectrum is in progress. [I31 MINDO:2: N . Bodor, M . J . S. DLwar, A. Harqer, and E . Hu\elbuc/i, J . Amer. Chem. Soc. 92,3854 (1970). Minimization by the QCPE program of A . Komornicki and J . W M c I w r , Jr.. Chem Phys. Lett. 10, 303 ( I 97 1 ). 1141 The largest eigenvaluecalculated by us for the bond-bond polarirability matrix is relatively small. For the theory see: G . Biiisch, I . Tumir, and R. D. Hill, J. Amer. Chem. Soc. 9 / , 2445 (1969). 850 We succeeded in proving the existence of kinks experimentally by measurements of the shortening of the overall chain length upon kink formation141.As models we have used bimolecular films of n-alkylammonium ions and n-alkanols spread out between silicate layers. Kink formation was shown by a stepwise decrease of the film thickness with increasing temperature, measured by X-ray methods. Kink formation is a cooperative process in which the number of kinks per alkyl chain increases by 1 within a relatively small temperature range. For steric reasons, the kinks must be assigned to neighboring chains, so that kink blocks are formed. The cause of kink formation is the increasing thermal energy. We have now been able to elucidate kink formation by measuring the specific heat of the same models (e.g. ntetradecylammonium-beidelliteln-tetradecanol) in an adiabatic calorimeter. A small flask containing the sample was immersed in paraffin as bath fluid, and the specific heat was determined from the change in temperature of the paraffin (measured by an NTC resistance), while a definite amount of heat was supplied by electrical heating. The sensitivity of the calorimeter was better than 20mcal/h. The experimental results are shown in Fig. 2: curve a) [*] Priv-Doz. Dr. G. Lagaly, DipLChem. S . Fitz, and Prof. Dr. A. Weiss Institut f ~ Anorganische r Chemie der Universitiit 8 Munchen 2. Meiserstrasse I (Germany) [**I This work was supported by the Deutsche Forschungsgemeinschaft.