J. Chem. SOC., Faraday Trans. I, 1987,83,43-49 The Use of Electron Spin Resonance and ENDOR and TRIPLE Resonance Methods for Structural Elucidation Isomeric 10,lO-Diphenylphenanthren-9( 1 OH)-ones Bernard0 J. Herold,* Maria J. Romiio and Jose' M. A. Empis Laboratdrio de Quimica Orgbnica, lnstituto Superior Tkcnico, Au. Rovisco Pais, 1096 Lisboa Codex, Portugal Jeffrey C. Evans and Christopher C. Rowlands Department of Chemistry, University College Cardig, P.O. Box 78, Cardig CFl 1 XL, Wales A series of substituted 10,lO-diphenylphenanthren-9( lOH)-ones has been reduced in ethereal solvents, and the resulting paramagnetic solutions have been examined by multiple resonance techniques. The values and relative signs of the various coupling constants proved useful and unambiguous in ascertaining the position of ring substitution in the starting material. These methods have therefore proved to be an alternative, if rather specialized means to determine which of several isomeric compounds is present.lO,lO-Diphenylphenanthren-9( 10H)-one [DPP, structure (3)] is the main product of the reaction of fluorenone (1) with diphenylchloromethyllithium (2) in tetrahydrofuran (THF) solution at low temperatures?! C H 6 5, /L' C6H5 (2) 0 A similar reaction starting with either 2- or hubstituted fluorenones could result in (a) either the 7-substituted DPP (5) and or the 2-substituted RPP (4), (b) the 3-substituted DPP (8) or the 6-substituted DPP (9), respectively (see below). 0 (4) (5) (6) 4, 5, 6: (a) X = C1 (b) X = OCH, (c) X = N(CH,), (d) X = F (e) X = CH, 4344 E.S.R.Resonance Methods for Structural Elucidation .x 0 (7) 7, 8, 9: (a) X = OCH, (b) X = F (c) X = CH, Predominance of isomers (6) and (9) over (5) and (8), respectively, means that substituent X will always maintain a through-resonance interaction with the carbonyl of the starting materials. The other possibility [i.e. predominance of (5) and (8)] means that such an interaction might arise between the substituent and the diphenylmethyl carbon. The full description of these reactions, as well as the yields of the various ketones (5, 6, 8 and 9)2 and their structures, previously determined by X-ray diffraction2, and by multinuclear n.m.r. spectroscopic techniques,2$ are given elsewhere. The structural similarity between the radical anion of DPP and that of an 'ortho- biphenylylketyl model ' (lo), enables the approximate electron spin distribution to be easily predictable.Hence, the spin distribution obtained from ENDOR and triple resonance experiments has been used to distinguish between isomerically substituted DPPS. Thus the conjugated system (10) can be considered as an odd-alternant radical bearing an anionic 0x0 substituent in position 9. In such a system, carbon atoms at even-numbered ring positions are expected to bear large positive spin densities, and those at odd-numbered ring positions, rather smaller and negative spin den~ities.~ Comparison of the substituted DPPs (5) and (6) with system (10) shows that in (5) the substituent occupies a low spin position, whereas in (6) it occupies a high spin position. The same reasoning shows that a low-spin-density carbon atom is substituted in the ketyl of (8), and a high-spin one in the ketyl of (9).Also, it has been established that substitution does not dramatically alter the electron spin distribution6$ in similar compounds. Structural elucidation therefore can in this case be obtained by determining whether the substituent is on a high positive or a low negative spin density carbon atom. Experimental 10,lO-Diphenylphenanthren-9( 1 OH)-one, (DPP) (3), was prepared as described earlier1 and purified to constant melting point. 7-Chloro-DPP (5a), 7-methoxy-DPP (5b) and 7-N,N-dimethylamino-DPP (5c) were obtained from the corresponding 2-substituted fluorenones (4a, 4b and 4c) and purified as describedB.J . Herold et al. 45 3-Fluoro-DPP (8b) and 6-fluoro-DPP (9b) resulted in the same way from (7b), 3-methoxy-DPP (8a) from (7a), 7-methyl-DPP (5e) and 2-methyl-DPP (6e) from (k), 3-methyl-DPP (&) from (7c), and 7-fluoro-DPP (6d) from (4); as above, the preparation, purification and identification of these compounds is described elsewhere.2 All other starting materials were used as commercially available or purified as previously Room-temperature electrolytic reductions of dimethoxyethane solutions of the various DPPs were carried out under vacuum. 0.1 mol dm-3 tetrabutylammonium perchlorate was used as a supporting electrolyte. The paramagnetic solutions obtained were stable for > 3 h at 200 K. E.s.r., ENDOR and triple resonance spectra were recorded at ca.200 K using a Varian E 109 spectrometer interfaced with a Bruker ENDOR/triple resonance unit. Computational work was carried out using a VAX/VMS system, and standard Huckel-McLachlan and linear regression analysis programs. Results 10,lO-Diphenylphanthren-9( 10H)-oneg (DPP) DPP (3) was reduced by a sodium film in dry oxygen-free tetrahydrofuran (THF) at 200 K. The brown solution had an e.s.r. spectrum comprising a featureless broad triplet (total width 1.5 1 mT). The ENDOR spectrum gave eight proton coupling constants and one sodium coupling constant [see fig. 1 (a)]. Their relative signs were determined by general triple resonance [see fig. 1 (b)]. The coupling constants are listed in table 1. These results are in agreement with others previously reported by Franco et aL9 Variable-temperature ENDOR and special triple resonance spectra [see fig.1 (c)] show that the smallest proton coupling is a multiple absorption, and that its optimum ENDOR temperature is higher than that observed for the other protons. An estimate of the multiplicity of this coupling was obtained by subtracting from the measured e.s.r. spectral width the width attributable to all the other coupling constants. In this way a multiplicity of 10 (k 1) was found for this hyperfine coupling. Electrolytic reduction of DME solutions of DPP gave a broad featureless e.s.r. spectrum. From this an ENDOR measurement gave seven coupling constants, one of which could be attributed to more than one proton. These values are similar to those obtained from the sodium reduced ketyl system.Substituted DPPs ENDOR, general and special triple resonance measurements on sodium reduced THF solutions of (5a), (5b), (5c) and (8a) gave the results listed in table 1. The measurements were made at 200 K immediately the solution became coloured. Complete sets of data were not obtained for all the other substituted DPPs, but the existing data were sufficient for st ruc t ure-elucidation purposes as described. Fluorine coupling constants were detected in the spectra obtained from various isomeric fluorinated DPPs. One of these gave an e.s.r. spectrum with a total width of 2.3 mT, and no apparent l9F ENDOR coupling, whereas the ENDOR spectrum of its isomer shows an unmistakable value of a(F) = 1.49 MHz. The compounds producing these results proved to be, respectively, (9b) and (8b); a(F) = 1.06 mT was measured from the e.s.r.spectrum of the ketyl of the former compounds. The ENDOR spectrum of the radical anion of (5d) shows a(F) = 8.01 MHz. For the various methylated DPPs, special triple and ENDOR spectra obtained at46 E.S.R. Resonance Methods for Structural Elucidation ( b ) 11 I MI 11 /I d d n a 5 MHz I ’ I 23 MHz I 0.1 MHz I 8 MHz Fig. 1. (a) ENDOR spectrum of a solution of sodium-reduced DPP in THF at 200 K. (b) General triple resonance spectrum obtained from DPP ketyl solution as above; the pump frequency was 13.013 MHz. (c) Special triple resonance spectrum obtained from DPP ketyl solution as above; the centre frequency was 13.875 MHz, starting at +O. 1 MHz. various temperatures easily pinpointed each isomer and confirmed the values of the methyl-proton coupling constants.These were determined to be 3.65 MHz for (5e), 2.49 MHz for (6e) and 0.328 MHz for (8c). Discussion Hiickel-McLachlan calculations of carbon spin densitiesg were carried out for the radical anions in table 1 using a set of literature values for the various resonance and Coulomb integral Calculations were performed on ‘substituted (10) model systems’, as described above, and no attempt was made to account for the electronic effects of the diphenylmethyl bridging between positions 9 and 10. As a consequence of thisTable 1. Hyperfine couplings obtained from ENDOR and general triple resonance spectra of some 10,lO-diphenylphenanthren-9( 10H)-one ketyl solutions, at 200 K ke tyl substituent/generation a,/MHz a,,/MHz 3'- none/electrolysis in DME - 15.02 -8.77 3.70 - 3.24 1 .86" 1.86" 0.86 O.3Ob - Sb'-Na+ 7-OCH3/THF -14.86 -10.56 5.36 -2.51 - - 1.47 0.72 0.29 1.26 3'-Na+ none/THF -15.54 -9.66 3.78 - 3.00 2.34 -1.80 0.90 O.3Ob 0.86 Sa'-Na+ 7Cl/THF -16.26 -10.08 3.37 - 2.63 - - 1.62 0.72 0.32 1.23 5c'-Na+ 7-N(CH3),/THF -14.23 -10.86 3.32 - 2.46 - - 1.44 0.72 0.29 1.14 8a'-Na+ 3-OCH3/THF -15.44 -9.67 3.77 - 2.77 2.22" 2.22" 0.85 0.28 0.95 a Double peak, as revealed by special triple resonance; the fact that no detectable general triple effect is observed is taken to mean the peak is due to two coupling constants of equal magnitude and opposite signs.Multiple peak (see text). Table 2. Calculated proton hyperfine couplings in (MHz), and linear least mean squares parameters for various lO,lO-diphenylphenanthren-9( 10H)-one ketyls position subs ti tuent ke t yl 6 8 5 2 7 4 QFH" rb SC ~~ - 3'-d - 14.52 - 8.07 4.86 -4.13 1.635 -3.10 - 7 1.32 0.990 75 > 99.9 - 3'-Na+e - 15.12 - 8.66 5.16 -4.12 1.966 -3.11 -74.01 0.988 55 > 99.9 7-C1 5a'-Na+e - 15.25 -9.57 5.04 - 4.23 - - 3.21 -75.24 0.980 63 > 99.0 7-OCH3 5b'-NaPe -13.40 -11.06 4.75 - 3.66 - - 2.68 -77.06 0.983 39 > 99.0 7-N(CH3), 5c'-Na+f - 14.57 - 11.37 5.07 - 3.92 - - 2.92 -76.27 0.978 57 > 99.0 3-OCH3 8a'-Na+e - 15.03 - 8.67 5.03 - 3.75 1.91 - 3.48 -72.91 0.990 21 > 99.9 a Q& is McConnell's constant, in MHz, determined by the least-mean-squares method for each case.r is the correlation coefficient. s is the significance of the correlation by Fischer's test, percentagewise.Huckel-McLachlan m.0. parameters used were, in units of pee:' 6, = 0.1; ~ 1 4 = 1.15; y,, = Y , , ~ , = 1.3; y11,12 = 0.8; y9,14 = 1.6. Same as above but 614 = 1.2, other parameters as in ref. (7). f Same as above but dC(C--N(CH3)2) = - 0.25, an arbitrary non-optimized value. P 2348 E.S.R. Resonance Methods for Structural Elucidation simplification, small equal calculated spin densities appear at positions 1 and 3. However, because of the approximate nature of the calculations, we prefer not to speculate as to the effect of this bridging group, even at the cost of disregarding some potentially useful data in the analysis that follows. In each case, the two smaller proton couplings were ignored. The others were compared with the calculated carbon spin densities using h e a r least-mean-squares analysis, which yielded best McConnell’s Q values and hence ‘ calculated ’ coupling constants.Table 2 lists these ‘calculated’ proton coupling constants and best Q values, in MHz, together with pertinent regression analysis data. These show excellent signifi- cance levels. One can point out that for the radical anions of compounds (5a, b and c ) listed in table 1 the triple resonance results would clearly have sufficed for structural elucidation, because a positive proton coupling, of ca. 2.3 MHz, is missing. This unmistakably indicates that substitution occurs at a low negative spin density carbon atom, i.e. at the 7 position and not 2. The remaining radical anion of (8a) (table 2) can similarly be proved to be a 3-substituted-DPP, because if it were the 6-substituted isomer the - 15.44 MHz coupling would not be present.Although incomplete, the data relative to the other compounds can easily be interpreted as compatible with what might be expected. Thus the fluorinated DPP ketyls, with a(F) values of ca. 1.060 mT, 1.49 MHz and 8.01 MHz, are, respectively, the 6-fluoro (9b), 3-fluoro (8b) and 7-fluoro (5b) ketyls. It is known that substitution of H for F in alternant systems can result in a ratio a(F)/a(H) = 2.1° This would confirm the hyperfine coupling constants for ring positions 7, 3 and 6. The same can be said for the methylated DPPs, where special triple resonance pinpoints the methyl-proton couplings of 3.65 MHz in the 2-methyl-DPP (6e) and 2.49 MHz in the 7-methyl-DPP (5c) ketyls.These results are in accordance with the fact that methyl-proton coupling constants closely match in magnitude the single proton coupling observed for the same position in the unsubstituted compound. The ca. 0.3 MHz absorption which is common to all these ENDOR spectra is assigned to the diphenylmethyl protons. This assignment not only yields a correct e.s.r. spectral simulation but is also comparable to results obtained by other authorsll in situations where 0-n: electron spin density transfer occurs. We believe that by confirming the earlier X-ray2V3 and n.m.r.2*4 data, these results show that ENDOR and triple resonance techniques can be of use in isomer structural elucidation. The scope of application is actually quite broad; in fact, the basic requirements for this rather simple approach to chemical structure elucidation are that aromatic or conjugated substrates be involved in rearrangement reactions whereby an ambiguity in migratory aptitudes might arise.Although admittedly specialized, the methods described are reasonably direct and, barring secondary chemical reactions, unambiguous. We thank M. L. Franco for helpful suggestions concerning the probable non-planarity of the DPP radical anion. Financial support for this work is gratefully acknowledged from NATO, for a cooperative research grant, and from ‘Instituto Nacional de Investigaqgo Cientifica’, through ‘ Centro de Processos Quimicos da Universidade Tecnica de Lisboa’. A grant from the S.E.R.C. towards the purchase of an ENDOR spectrometer by University College Cardiff is also gratefully acknowledged. References 1 J. M. A. Empis, M. L. T. M. B. Franco, B. J. Herold and J. J. R. P. Queiroga, Tetrahedron Lett., 1975, 2 B. J. Herold, M. J. Romgo, C. Krueger and R. Mynott, presented in part at 7 O Encontro Anual da 47, 4153. Suciedade Portuguesa de Quimica, Lisbon, July 1984, p. PCl1, to be published.B. J . Herold et al. 49 3 M. J. Romiio and C . Krueger, Acta Crystaflogr., Sect. C., in press. 4 B. J. Herold, R. Mynott and M. J. Romiio, to be published. 5 N. M. Atherton, in Electron Spin Resonance (Ellis Horwood, Chichester, 1973). 6 J. M. A. Empis and B. J. Herold, J . Chem. Soc., Perkin Trans. 2, 1986, 425. 7 B. J. Herold, J. M. A. Empis, J. C. Evans and C. C. Rowlands, J . Chem. Soc., Perkin Trans. 2, 1986, 8 W. Schlenk and E. Bergmann, Ann. Chem., 1928,463, 209. 9 M. L. Franco, B. J. Herold, C. C. Rowlands and J. C. Evans, presented in part at the 5 O Encontro Anuaf 431. da Sociedade Portuguesa de Quimica, Oporto, p. C29.58, (1982). 10 P. H. H. Fischer and J. P. Colpa, 2. Naiurforsch., Teil A, 1969, 24, 1980. 1 1 R. Biehl, K. P. Dinse, K. Moebius, M. Plato, H. Kurreck and U. Mennenga, Tetrahedron, 1973, 29, 363. Paper 61998; Received 22nd May, 1986