The crystal and molecular structure of 1,2,3,5-tetramethylbenzobthiophenium tetrafluoroborate and some theoretical calculations on the hypothetical 1H-thiophenium cation

摘要

J.C.S. Perkin I1 The Crystal and Molecular Structure of 1.2,3,5-Tetramethylbenzob-thiophenium Tetrafluoroborate and Some Theoretical Calculations on the Hypothetical I H-Thiophenium Cation By R. Morrin Acheson," Richard J. Prince, Garry Procter, and John D. Wallis, Department of Biochemistry, South Parks Road, Oxford OX1 3QU David J. Watkin, Laboratory of Chemical Crystallography, 9 Parks Road, Oxford OX1 3PD The X-ray crystal structure of 1,2,3,5-tetramethylbenzobthiophenium tetrafluoroborate, refined to R0.043, shows the bonding about sulphur to be pyramidal and not planar. This structure and the 13C n.m.r. spectral data suggest that benzo blthiophenium salts should be considered as cyclic styrene derivatives. Theoretical calculations also showed an energy maximum for the planar configuration of the 1H-thiophenium cation. THIOPHENE,benzo btliiophen, and dibenzob,d thiophen have been converted into their S-alkyl derivatives by reaction with alkyl halides in the presence of silver tetrafluoroborate, though the yields of thiophenium salts are 1ow.l If the sulphur atom in the parent heterocycle is considered to have a formal sp2 hybridisation, then the salts might be expected to have a planar geometry about sulphur, so that maximum overlap in an aromatic type (4-12 + 2) x electron system could be retained.However the IH n.m.r. spectra of 2-substituted 1-ethylbenzob- tliiophenium salts e.g. (l) and the 3-bromo salt (2) show I Me (1) 2-Me (3) (2) 3-Br (4) 2,3-Me2 (5) 2,3!5-Me3 magnetic non-equivalence of the methylene protons of the ethyl group.This could be explained by the sulphur atom being pyramidal so the methylene hydrogen atoms would be diastereotopic. Diastereotopic hydrogen atoms could also be observed for a planar sulphur atom if there was restricted rotation about the S--CH, bond. This explanation would be unlikely for the 3-bromo compound (2) and furthermore 1-ethyl-2-metliylquinolinium iodide which is more sterically hindered than (1)does not show the phenomenon. To remove any doubt concerning the stereochemistry, the crystal structure of 1,2,3,5-tetra-methylbenzo:bthioplienium tetrafluoroborate (5) was therefore examined. RESULTS AND DISCUSSION Our X-ray crystal structure determination has shown that the bonds about the sulphur atom in the benzojbl- thiophenium salt (5) are arranged pyramidally (Figures 1and 2).The angles about sulphur are 104.0, 105.5, and 92.2', the last one being in the five-membered ring and is similar to those of the corresponding angles for the benzotliiophen (6) (91.7') and the thiophenium ylid (8) (92.2"). The other two angles are typical of a sul-phonium salt e.g. 100.8, 102.5, and 102.3" in (9).* The +S-CH, bond length (1.806 A) is also normal cf. 1.794 and 1.817 A and (11),61.800 A. The ring S-C C(51) I B FIGURE 1,2,3,6-Tetramethylbenzob1 thiophenium tetrafluoro- borate thermal ellipsoid plot. View perpendicular to the plane of the carbocyclic ring bond lengths (1.768 and 1.779 A) are shorter than in benzo'blthiophen 1,l-dioxide (12) (1.80 and 1.82 A) and the S-phenylsulphonium salt (10) (1.82 A).The five-membered ring is slightly puckered with the sulphur FIGURE View through plane of the carbocyclic ring of2 1,2,3,3-tetramethylbenzoI,thiophenium tetrafluoroborate deviating from the best plane tlirough the benzene ring in the opposite sense to the C(2) and C(3) atoms and the I-, 2-, and 3-methyl carbon atoms (Table 1). When compared with the molecular structure of the C(31) 1.L95t.4J C (31) FIGURE Bond lengths and angles for (5). E.s.d.s. in 3 parentheses benzobthiophen (7) (Figure 4) the salt (5)has longer (by 0.03 A) ring C-S bonds, a shorter (by 0.02 A) C(2)-C(3) bond and a longer C(3)-C(3a) bond, supporting the in- volvement of a sulphur lone pair in the x bonding of benzobthiophens.However the large e.s.d.s for the bond lengths of (7) (0.012 A) weaken the value of this (8) (10) + /$Ps-MeI' 00 (11) (12) comparison. A similar relationship holds between the thiophenium ylid (8) and thiophen with the ylid having longer C-S bonds (by 0.02 A), shorter formal double bonds (by 0.05 A), and a longer formal single bond (0.06 A). 1-Methylbenzob thiophenium tetrafluoro-borate (3) adds bromine trans to the 2,3-double bond1 while benzob thiophen is substituted 9 mainly at position 3. This is strong evidence for the relative lack of aromaticity in the fused thiophenium ring. Even if benzobthiophen is substituted by an initial addition of FIGURE Bond lengths and angles for (7).E.s.d.s. in4 parentheses bromine to the 2,3-bond, the corresponding bromine addition product (13) of the benzobthiophenium salt would be expected to lose hydrogen bromide more readily due to the greater acidity of a hydrogen OL to a positively charged sulphur. Although the loss of hydrogen bromide from (13) would involve increasing the steric interaction between the 1-and 2-substituents this is unlikely to be critical. Doubts are therefore cast on calculations which suggest that cations (14) and (15) should be stable.lO Further-more ab initio calculations, performed with a minimal basis set using STO-3G,11 on the hypothetical thio- phenium cation (16) for varying values of 8 gave an energy maximum for the planar conformation (0 180').The 13C n.m.r. spectrum (Table 2) of the benzob- thiophenium salt (4) shows the +S-CH3 resonance at 6 30.0 p.p.m. in good agreement with 6 27.5 p.p.m. for a trimethylsulphonium salt.12 The doublet resonances of the benzene ring are deshielded compared with the cor- TABLE1 Deviations from tlie best plane through the benzene ring of (5) Atom Deviation (A) Atom Deviation (A) S(1) 0.044 C(11) -1.557 C(2) -0.151 C(21) -0.279 C(3) -0.071 C(31) -0.041 Equation of plane: 7.40~-7.05~-3.102 = -5.36 responding benzobthioplien (6) and are similar to the doublet resonances of methyl phenyl sulphoxide l4 (Table 2). However C(2) and C(3) show smaller shifts when tlie sulphur is alkylated, the former being, sur- prisingly, upfield.This does not seem too unreasonable by comparison with styrene (Table 2) l5 in which the ter- minal olefinic carbon resonates at 6 112.3 p.p.ni. The 2-and 3-methyl groups would be expected to deshield C(2) by ca. 3 p.p.m. and the sulphonium group should effect a further deshielding of C(2) by ca. 18 p.p.m., TABLE2 13C: n.m.r. spectra measured at 22.63 MHz in deuteriocliloro- form S (p.p.m.) from internal NIe,Si. T-lH Attach- ments were confirmed by off-resonance clecoupling Compound (4) (6) l3Assignment Resonance positioiis 1-CH, 30.0 CH, 11.3 11.2 3-CH3 12.4 13.7 2-C 130.5 * 133.0 3-c 129.2 * 127.9 3a-C 143.0 : 140.9 4-, 54, 127.2,t 133.5, 121.0, 123.7, ti-, 7-c 129.9,t 124.4 123.3, 121.0 7a-C 144.5 138.1 PhSOiCle l4 Styrene 15 I-C 146.3 138.2 2-, 6-C 123.G 126.7 3-, 5-c 129.4 128.9 4-c 131.0 128.2 1'-c 135.8 3J-C 112.3 S-CH, 44.0 *#+I $ Tridicate interchaiigeable assignments. J.C.S.Perkin I1 compared with styrene or a benzob thiophen, which could be accounted for by back-donation of x electrons from the benzene ring into sulphur 3d orbitals. On the other hand, the S(l)-C(7a) bond (1.768 A) is comparable with the S-C (phenyl) bond (1.763 A) in 2,3,5,6-tetra- chloro-4-cyanophenyl methyl sulphide.lG Shortening of the C-S bond in the latter compound by conjugation of the sulphur 3fi orbitals with the aromatic x system is precluded by the conformation of the molecule.The S-C (phenyl) bond length in the sulplionium salt (10) (1.82 A) is surprisingly long hut must be regarded cautiously due to the large e.s.d.s (0.02 A). The S-C bond lengths in the ylids (8) and (17),17 1.72and 1.71 A, respectively, are considerably shorter than the com-parable length in (5). The possibility that 3d orbitals are involved to a greater extent in the first two compounds than in the last cannot be established as coulomhic inter- actions would be expected to shorten this bond in any event. The conformations of the substituents about the S-C bonds of these ylids, and in the salt (lo),in the crystalline state are very similar with the lone pair of the sulphur atom almost at 90" to the $-orbital on the adjacent carbanion, or aromatic carbon atom for (lo). It may be concluded that the crystal structure for (5), the easy addition of bromine to (3),and the spectral data suggest that these benzothiophenium salts are best con- sidered as cyclic derivatives of styrene rather than as highly stabilised lox aromatic systems.EXPERIMENTAL Crystal Structure lleterunimtion for Compound (5).-Crystals of (5), C,,H,,HF,S, prepared as described,' were grown by diffusion of ether into a dichlorometliane solution. Oscillation and Weissenberg photographs suggested that the crystal was triclinic. It was transferred to an Enraf Nonius CA4D4four-circle diffractometer, and the crystal system, actually monoclinic (2= 8, p = 2.73 cm-l), and ;xcurate cell dimensions a = 16.37(1), b =-7.91(1), c = 22.01(2) A, p = 111.37(4)o~were determined.Diffraction data were collected using Mo-K, radiation by a m-20 scan and standard reflections were checked every hour. The space group (C2/c) was assumed from systematic absences and tlie number of formula units (8) per unit cell. The crystal density, 1.40 g CI~--~(theoretical value 1.40 g cm3), was determined using a Westphal balance. Lorentz and polarisation corrections were applied to tlie 2 599 reflections, equivalent reflections merged, and structure amplitudes judging from the effect of a sulphoxide substituent on a derived for 1 528 reflections with I:3a(I). The structure was solved by direct methods using RIIULTAN 77,18 andbenzene ring carbon in methyl phenyl sulphoxide.It is difficult to estimate the contribution sulphur refined hy full-matrix least squares to an I? value of 0.178 ming isotropic temperature factors. Further refinement 3d orbitals make to the bonding in benzob thiopheniurn using anisotropic temperature factors lowered the H value to salts. On the one hand the benzene ring is deshielded 0.075, after wliicli a differential Fourier synthesis located all 1981 269 C ka6 FIGURE Stereoscopic packing diagram for (5)5 the hydrogen atoms. The structure was refined to con- vergence by bloclred-matrix least squares, including an iso- tropic extinction parameter, and with constraints applied to the hydrogen atoms C-H dipnce constrained to 1.000 k with an e.s.d.of 0.01 A, C--C(sPZ)-H constrained to their common mean of 120.1,s A with an e.s.cl. of 2.00deg;. For each methyl group the H-C-H angles wp-e constrained to their common means (Table 3) and C(S)-C-Hconstrained to their TABLE3 Mean values of H-6-H and C(S)-t-H for methyl groups in (5) Methylcarbon Rlean*value for I-I-c-I3 (") Mean v,alue for C(S)-C-H (") c-11 106.73 7 111.345 c-12 109.774 108.724 c-13 106.302 112.038 C-15 11 1.821 106.745 coninion mean, both with an e.s.d. of 2.00". The com-ponents of the temperature factors of hydrogen atoms and of the carbon atorus to which they were bonded were con- strained l9 to have the same niagnitude in the direction of the bond, with an e.s.d.of 0.004 A.2 Weights were com- puted from the Chebyshev series w = 49.13t0 (X)+ 67.68S1 (X)+ 19.69t, (X)-l where (X) = Fo/F,,1,,~20Data re-duction and structnre refinement were done using CRY-STALS 21 and all calculations were made on the Oxford University ICI, lR0GA computer. Figure 5 shows a structure packing- diagram. Final atomic co-ordinates, anisotropic temperature factors, and structure factors are available in Supplementary Publication No. SUP 22999 (13 pp.).* * For details of Snpplenicntary Publications see Notice to Authors No. 7 in J.C.S. PcvF;iit 11, 1979, Index Issue. fie thank the M.1Z.C. (J. 1).W.) for a studentship. 0/410 Received, 13th March, 19801 REFERENCES R. M. Rcheson and D. R. Harrison, J. Chem. Soc. (C),1970, 1764.J. H. C. Hogg and H. euro;1.Sutherland, Acta Cryst., 1974, 80B, 2058. R. J. Gillespie, A. E. A. Porter, arid W. E. Willmott, J.C.S. Chem. Comm., 1978, 85. J. W. Cornforth, S. A. Reicharcl, P. Talaty, euro;I. L. Carrell, and J. P.Glusker, J. Amer. Chem. Soc., 1977, 99, 7292. A. Lopez-Castro and M. R. Truter, Acta Cryst., 1964, 17, 465. K. Gerdil, Helv. Chim. Acta, 1974, 57,489.' R. L. R. Towns and S. H. Simonson, Cryst. Struct. Comm., 1974, 3, 373. 8 B. Bak, D. Christensen, L. Hansen-Nyagaard, and J. Ras-trup-Andersen, J. Mol. Spectroscopy, 1961, 7, 58. B. Tddon and R. M. Scrowston, Adu. Heterocyclic Chem., 1970, 11, 263 and references therein. lo R. Zahradnik, Adv. Heterocyclic Chem., 1965, 5,1. l1 0. Eisenstein, G. Procter, and J. D. Dunitz, Heh. Claim. Acta, 1978, 61, 2538. 1* G. Barbarella, P. Dembech, A. Garbesi, and A. Fava, Org. Magnetic Resonance, 1976, 8, 108. l3 P. I). Clark, D. F.Ewing, and 12. W. Scrowston, Opg. Mag-netic Resonance, 1976, 8, 263. l4 G. W. Ruchanan, C. Reyes-Zamora, and D. E. Clark, Caizad. J. Chem., 1974, 52, 3895. l5 0. Kajimoto, amp;I Kobayashi, and T. Fueno, Bicll. Chem. Soc. Japan, 1973, 46, 1422. l6 D. R. Carter and I;. P. Boer, J.C.S. Pevkin 11, 1972, 2104. l7 A4.T. Christensen, T. ilrild, and W. G. Whitmore, Acta Cryst., 1969, B25,78. 18 G. Germain, P. RIain, and M. JI. Woolfson, Acta Cryst., 1971,A27,368. l9 J. S. Rollett, ' Crystallographic Computing,' ed. I;. R. Ahmed, Munksgaard, Copenhagen, 1969, p. 167. 2o J. I. Carruthers and D. J. Watkin, Acta Cryst., 1979, A35, 698. J. R. Carruthers, ' CRYSTALS User Manual,' Oxford Uni- versity Computing Centre, 1975.