J. CHEM. SOC. FARADAY TRANS., 1994, 90(2), 327-332 Structural Studies on Paracyanogen and Paraisocyanogen Leonardus W. Jenneskens," Jan W. G. Mahyt and Edward J. Vlietstra Debye Institute, Department of Physical Organic Chemistry, Utrecht University, Padualaan 8, 3584 CH Utrecht, The Netherlands Simon J. Goede and Friedrich Bickelhaupt Scheikundig Laboratorium, Vrije Universiteit, De Boelelaan 1083, 1081 HV Amsterdam, The Netherlands Polymers derived from the isomeric C,N, monomers cyanogen (NCCN) and isocyanogen (CNCN) are investi- gated. Although both paracyanogen [poly(NCCN)] and paraisocyanogen [poly(CNCN)] consist of carbon and nitrogen in a close to 1 : 1 ratio, thermogravimetric (TG) and spectroscopic analyses (DRIFT, EPR, UV-VIS-NIR and XPS) reveal that their molecular structures are markedly different.Despite the occurrence of n-conjugation, no regular ladder structures are found. In line with the spectroscopic data, conductivity measurements show that pristine poly(CNCN) and poly(NCCN) are an insulator and a semiconductor, respectively. The theoretical prediction (extended Huckel bandstructure calculations) by Whangbo et ul.,' which was corroborated by valence effective Hamiltonian (VEH) bandstructure calcu- latioIw2 that paracyanogen [poly(NCCN)] with the regular ladder structures shown in Scheme 1 possesses metallic behaviour, has initiated a quest for this potentially intrinsic organic metal. However, despite the fact that poly(NCCN) was already prepared in 1815 by Gay Lussac by heat treat- ment of mercury(r1) cyanide, its intractable nature hampered its structural el~cidation.~ Only in the last decades has the synthesis of paracyanogen from different starting materials and its characterization received attenti~n.~ Nevertheless, there is still a paucity of experimental data and little is known about its structure and properties, especially with respect to its anticipated intrinsic metallic behaviour.In 1988 the synthesis of another C,N, isomer prepared by flash vacuum thermolysis (FVT, 773 K at lo-' Torr) of nor- bornadienone azine in 60% yield was rep~rted.~ Although initially the C2N2 isomer was identified as diisocyanogen (CNNC) since only one signal typical for is~cyanides'*~ was observed in its 13C and 14N NMR spectra at 173 K, the use of other precursors' and 15N labelling* in combination with spectroscopic analysis (high-resolution IR, microwave and '5N NMR spectroscopy), unequivocally revealed that another C,N, isomer, i.e. isocyanogen (CNCN), is the primary low-molecular-weight product.Apparently, at some stage during the thermolysis of the norbornadienone azine, an efficient isonitrile-nitrile rearrangement takes place (Scheme 2). Analogous rearrangements have been invoked previously to rationalize the behaviour of related compounds under thermolysis conditions.' In this respect, it is note-worthy that careful spectroscopic analysis (I3C and "N NMR, 173 K) of the norbornadienone azine pyrolysate showed that, besides the major product isocyanogen, addi- tional minor side products, such as hydrogen cyanide (2%), cyanogen (8%) and ethyne (2%), are present as impurities even after low-temperature distillation (temperature range 143-173 K).' The identification of these side products pro- vides evidence that other fragmentation pathways, such as retro-Diels-Alder cleavage,' are also operative in the ther- molysis of the norbornadienone azine.In line with theoretical predictions,' O isocyanogen (CNCN) was found to be considerably less stable than its isomer cyanogen (NCCN): It already polymerizes in solution at 193 K! At this temperature typical isonitrile reactions do not have a competitive rate." Hitherto, only two distinct low- molecular-weight reactions of isocyanogen have been report- ed, i.e.formation of N-cyanodibromoformaldimine by reaction with bromine' and its chromium pentacarbonyl complex.l2 Combustion analysis of pristine paraisocyanogen [poly(CNCN)] confirmed qualitatively the theoretically expected C :N ratio of 1 : 1 [experimental (YO)C: 50.15, H: 2.12, N: 41.37, (CN), calculated (%) C: 46.14, N: 53.86; experimental C : N ratio = 1.4 : 11.' The discrepancy between the theoretical and experimental data can be attributed to the incorporation of low-molecular-weight side-products in the polymer (vide infra). For paracyanogen prepared from several cyanogen precursors combustion analysis gave C and N values in the range of 32-34% and 32-44%, re~pectively.~" Although also for paracyanogen the theoretical and experi- mental results are at variance, it should be stipulated that it is well documented that its combustion analysis presents con- siderable experimental dific~lties.~Nevertheless, the data show that both paracyanogen and paraisocyanogen are pri- marily composed of carbon and nitrogen and may be looked upon as representative of heteroatom-substituted carbon-aceous materials. Hence, we were prompted to study in more detail the molecular structure and properties of paracyano- gen and paraisocyanogen.NC: cis trans FVT -46H6 -&Scheme 1 Proposed ladder structure for para~yanogen'*~ t Present address: Akzo Research Laboratories Arnhem, P.O. Box Scheme 2 Formation of isocyanogen by FVT of norbornadienone 9300,6800 SB Amhem, The Netherlands. azine5v8 Here we report the results of an investigation of both paracyanogen and paraisocyanogen with polarization microscopy, wide-angle X-ray powder diffraction (WAXD), thermogravimetry (TG), and diffuse reflectance Fourier-transform infrared (DRIFT), solid-state optical (UV-VIS- NIR, diffuse reflection technique), electron paramagnetic resonance (EPR) and X-ray photoelectron (XP) spectro-scopies.In addition, the electrical conductivity of pristine paraisocyanogen is determined and compared with that of paracyanogen. Experimental Synthesis and Polymerization of Isocyanogen Isocyanogen was obtained by FVT of norbornadienone azine at 773 K and Torr as described in detail elsewhere.8 Paraisocyanogen was prepared via two routes. (1) In a high- vacuum system isocyanogen (52 mg, 1 mmol) was dissolved in diethyl ether (5 ml) at 173 K and slowly warmed to ambient temperature during which paraisocyanogen precipi- tated (yield 60%).(2) In a high-vacuum system isocyanogen (52 mg, 1 mmol) was sublimed into the gas phase in a reac- tion vessel (volume 50 ml) at 253 K and warmed to ambient temperature during which pariasocyanogen precipitated on the glass surface of the vessel (yield 10-50%). Spectroscopic analysis revealed that both routes lead to identical paraisocy- anogen. Solubility experiments showed that paraisocyanogen, like paracyan~gen,~-~?*an intractable material insoluble in is common organic solvents and concentrated sulfuric acid.Synthesis and Polymerization of Cyanogen Cyanogen was prepared from mercury(I1) cyanide and poly- merized folowing reported procedures ;4a mercury(I1) cyanide (4 g, 14.8 mmol) was heated at 713 K in a sealed ampoule (Pyrex, volume 20 ml) for 24 h. After cooling to ambient tem- perature the ampoule was opened in a nitrogen atmosphere and paracyanogen was separated from mercury(0). Pristine paracyanogen was additionally heated at 473 K for 8 h in uucuo to remove traces of mercury(0) (yield 95%). The absence of mercury(0) in paracyanogen was established with XPS (vide infra). Characterization of Paraisocyanogen and Paracyanogen For the TG experiments a Perkin-Elmer TGS-2 equipped with an autobalance AR-2 was used (temperature program 323-1073 K, heating rate 20 K min-I).DRIFT spectra were recorded on either a Bio Rad FTS-7 or a Mattson Galaxy Series FT-IR 5000 spectrophotometer using a diffuse reflec- tance accessory; the samples were diluted with optically pure potassium bromide. Powder EPR spectra were recorded on a Bruker ESP 300 X-band spectrometer operating at 9.6 GHz. The g values of the peaks and the EPR spectrometer fre- quency were calibrated against solid diphenylpicrylhydrazyl radical (DPPH) standard assuming g = 2.0036.13 The EPR samples were sealed in high-purity quartz capillaries. Solid- state UV-VIS-NIR spectra were measured on a Varian Cary 5 dispersive spectrometer using a diffuse reflectance acces- sory (Praying Mantis); the samples were diluted with opti- cally pure potassium bromide.XP spectra were measured on a VG Escalab MkII spectrometer with non-monochromated Mg-Ka X-rays. The powder samples were mounted on a standard holder using two-sided tape. The absence of inter- fering signals from the tape was monitored and ascertained J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 by the observation of the spectral region in which Si from silicone present in the tape is expected. Data handling was performed using the VGS 5250 EDP software package. Results and Discussion Polarization Microscopy and X-Ray Powder DifFracton Polarization microscopy (crossed polarizers, magnification 400x ) indicates that both materials are amorphous solids. This is supported by wide-angle X-ray powder diffraction measurements ; no reflections are discernible ! This shows unequivocally that neither polymer possesses a three-dimensional ordering with regular repeat distances in the range of 2-10 A.14 The absence of reflections typical for graphite-or poly(acene)-like stacking, anticipated if both paracyanogen and paraisocyanogen possess a regular ladder- type structure as previously proposed for the former by Whangbo et a/.' and Bredas et al.,' suggests that the poly- mers are composed of small ordered arrays at most.Note that the latter may well generate diffuse scattered intensity at wide Bragg angles. Thermal Stability Although paracyanogen has been prepared before and has been the subject of several investigations, few experimental details concerning its thermal stability are a~ailable.~,~ To our knowledge only paracyanogen prepared via electro-polymerization of cyanogen has been subjected to controlled heat treatements with the objective to assess its conversion into carbon fibres.I5 To gain insight into the thermal stability of paracyanogen and paraisocyanogen both polymers were studied with TG under inert (N2) and thermo-oxidative (air) conditions.After loss of water of hydration (2 wt.%) at 373 K, the TG (N2) curve of paracyanogen shows no weight loss up to 673 K. Above 673 K, gradual thermal degradation and volatilization of the polymer takes place and is complete at 1123 K; no residue is found. In contrast, for pristine parai- socyanogen, which contained ca.6 wt.% water of hydration, loss of weight sets in already at 423 K followed by thermal degradation and volatilization yielding a residue of 18.5 wt.% at 1123 K. Under thermo-oxidative conditons the onset tem- perature for weight loss decreases for both polymers (paracyanogen: 573 K and paraisocyanogen: 373 K), fol-lowed by gradual thermal degradation and volatilization of both polymers leaving no residue at 1073 and 873 K, respec-tively. The differences in thermal behaviour indicate that the structures of pristine paracyanogen and paraisocyanogen are markedly different. However, it should be remembered that paracyanogen, in contrast to paraisocyanogen, is prepared by heat treatment of mercury(r1) cyanide in a sealed ampoule.Therefore, we have also studied the effect of a similar heat treatment of pristine paraisocyanogen at 713 K both in a sealed ampoule and by isothermal TG (N2). The isothermal TG (N2) curve showed a weight loss of ca. 50% within the first 30-40 min of the experiment after which the weight of the sample remained constant. TG (N2) analysis of heat- treated paraisocyanogen from the isothermal TG and the ampoule experiment gave identical TG curves. The heat- treated polymer is stable up to 773 K. Above this tem-perature thermal degradation occurs yielding a residue of 16.8 wt.% at 1123 K. Note that the total weight loss found with TG (N2) after heat treatment in either the isothermal TG or in the ampoule experiment agrees with the total loss of weight determined with TG (N2)for pristine paraisocyano- J.CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 gen. These results suggest strongly that the molecular struc- ture of pristine paraisocyanogen is altered by the heat treatment (cf.next section). Molecular Structure of Paracyanogen and Paraisocyanogen DRIFTS To gain insight into the molecular structure of both polymers DRIFT spectra were measured. The DRIFT spectrum of paracyanogen contains only broad absorption bands C2170- 2230 cm-' (w), 1750-1000 cm-' (s) and 800-700 cm-' (w), Fig. l(a)]. The absorption band centred at 2200 cm-' pro-vides evidence for the presence of pendent isonitrile (-NC) and nitrile (-CN) groups and the strong broad absorption band centred at 1500 cm- ' suggests that paracyanogen also contains imine (-C=N-) and alkene (-C=C-) type structural units.These data are in excellent agreement with those previously reported for paracyanogen prepared from a variety of cyanogen precursor^.^ For pristine paraisocyano- gen, a markedly different IR spectrum is obtained. Besides broad absorption bands at 2170-2240 cm-' (w) and 1750- lo00 cm-' (s) indicating the presence of pendent isonitrile (-NC) and nitrile (-CN) groups, and imine (-C=N-)- and alkene (-C-C-) groups, respectively, an additional strong, broad, featureless absorption band is found in the region 2500-3500 cm- ' (s) [Fig. l(b)]. Its shape suggests the I I I I I I I 3500 3000 2500 2000 1500 1000 500 ' wavenum ber/cm -Q) c-2 0,-n 0 3560 3000 2500 2000 ti00 1000 500 wavenumber/cm-' Fig.1 DRIFT spectra of (a)paracyanogen, (b)(-) pristine para-isocyanogen and (---) paraisocyanogen after heat treatment at 713 K presence of a wide variety of amine (-NHR) and carbon- hydrogen (-CH, and =CH,) groups in considerably differ- ent molecular environments. Apparently during polymerization of isocyanogen, the hydrogen-containing minor side-products, hydrogen cyanide and ethyne, which still contaminate isocyanogen isolated from the crude pyroly- sate by low-temperature distillation are efficiently incorpor- ated in the This is supported by the combustion data of pristine paraisocyanogen (experimental C : N ratio = 1.4 : 1) which deviate from the theoretically expected ratio = 1 : 1 (uide supra).Heat treatment of pristine parai- socyanogen at 713 K either in a sealed ampoule or with iso- thermal TG (NJ, leads to a reduction of the absorption bands at 2500-3500 cm-' and 2170-2240 cm-' and an increase in the 1750-1000 cm-' band [cf. Fig. l(b): the inten- sity ratio of 3250 and 1750 cm-' for pristine paraisocyano- gen is 0.52 and for paraisocyanogen after heat treatment is 0.221. Our data are in agreement with reported results on the effect of a heat treatment of poly(acrylonitrile).' Upon heat treatment, a decrease in intensity of the nitrile (-CN) absorption band concomitant with an increase of the broad absorption band at the position for imine (-C-N-) and alkene (-C=C-) type units was observed, which was attributed to nitrile cyclization.In the case of paraisocyano- gen, heat treatment undoubtedly will also lead to isonitrile- nitrile rearrangement^.^*'* ' Presumably, the nitrile substituents will subsequently cyclize under the high-temperature conditions.' Note, however, that even after heat treatment DRIFTS indicates that the structure of heat-treated paraisocyanogen differs from that of paracyanogen (Fig. 1). Nevertheless, DRIFTS suggests that both paracyano- gen and paraisocyanogen are amorphous networks derived from n-conjugated chains of different length consisting of coupled imine (-C=N-) and alkene (-C=C-) building blocks with, especially in the case of paraisocyanogen, a varying amount of pendent nitrile, isonitrile and amine-type substituents. Moreover, our DRIFT analysis provides evi- dence that heat treatment of pristine paraisocyanogen leads to an increase in unsaturation in combination with the for- mation of heteroatom-substituted cyclic n-conjugated struc- tures due to nitrile cyclization.EPR Powder EPR spectroscopy of pristine paraisocyanogen and paracyanogen showed the presence of a single broad reson- ance with g values of 2.0019 and 2.0017, respectively (reference DPPH, g value 2.0036) with a moderate peak to peak linewidth of 6 G at ambient temperature. The g values indicate the radicals to be carbon ~entred.'~ The absence of hyperfine splitting can be attributed to line broadening in the solid-state samples.In addition, the EPR signal for paracya- nogen is unsymmetrical, which suggests that the EPR spec- trum of paracyanogen is derived from different, but structurally related radical centres. In contrast to the powder EPR spectrum of paracyanogen, the EPR spectrum of pris- tine paraisocyanogen changed upon exposure of the sample to air. Besides a considerable decrease in intensity, a change in g value to 2.0035 (peak to peak linewidth 8 G at ambient temperature) was observed. This g value suggests the forma- tion of iminoxyl or nitrogen radical centres upon exposure to air.13 Unfortunately, owing to the small amounts of polymer available, we were hitherto unable to obtain reliable esti- mates of the spin density of paracyanogen and paraisocyano- gen before and after exposure to air.Nevertheless, the peak to peak linewidths found in the EPR spectra for both pristine polymers are indicative of the occurrence of delocalization of the radical centres and suggest, in agreement with the DRIFT analysis, that they are composed of n-conjugated structural units.'6a In passing, we would like to remark that as a conse- quence of the presence of unpaired electrons in both materials solid-state 3C (CP) MAS NMR measurements were thwarted. Electronic Absorption Spectra Based on the structure of both C2N2monomers and sup- ported by the DRIFT and EPR analysis, it is expected that paracyanogen and paraisocyanogen consist of n-conjugated chains. An estimate of the amount of n-electron conjugation in paracyanogen and pristine paraisocyanogen was obtained from solid-state diffuse reflectance UV-VIS-NIR measure-ments.The spectra show that the level of z-electron delocal- ization is markedly different for both polymers. Optical band gaps (Eg,cut-off energies) of ca. 0.85 and 1.45 eV are derived from the optical spectra for paracyanogen and pristine para- isocyanogen, respectively [Fig. 2(a) and (b), solid line].' In accord with the TG and DRIFT results, heat treatment of pristine paraisocyanogen leads to considerable changes of the solid-state UV-VIS-NIR spectrum. For the heat-treated paraisocyanogen samples prepared either with isothermal TG (N2, 713 K) or in a sealed ampoule (713 K) an increase in n-electron conjugation is found; similar UV-VIS-NIR spectra are obtained with an optical band gap of ca.0.96 eV [Fig. 2(b), broken line]. Hence in agreement with our DRIFT data, solid-state UV-VIS-NIR spectroscopy supports the conclusion that heat treatment of pristine paraisocyanogen leads to an increase of n-electron delocalization as a conse- quence of an increase in unsaturation (extrusion of hydrogen) in conjunction with nitrile cyclization reactions. In addition, the difference in optical band gap between heat treated parai- socyanogen (Eg 0.96 eV) and paracyanogen (Eg 0.85 eV) sup- ports our contention that even after heat treatment the structure of paraisocyanogen differs from that of paracyano- gen. XPS and Electrical Conductivity Measurements The conclusions derived from DRIFT and solid-state UV- VIS-NIR spectroscopy are corroborated by XPS measure-Fig.2 ments. In Table 1, the surface element concentrations (atom Yo)derived from XPS wide-scan spectra and the results of the decomposition of the C 1s peak shapes (XPS narrow-scan spectra) are presented. Although 0 is detected besides C and N, the C : N ratio for both paracyanogen and pristine para- isocyanogen are in satisfactory agreement with the bulk data obtained by combustion analysis (XPS: paraisocyanogen, C : 50.7, N: 37.4 and 0: 11.9, C :N ratio = 1.4 : 1; paracyano-gen, C: 46.8, N: 46.3 and 0:6.9, C :N ratio = 1 : 1).Based on the observed 0 1s binding energy (0Is, 534 ev) in com-bination with the 0 1s peak shape, the oxygen present can be attributed to physisorbed water (cf.also Thermal Stability section).I8 A superposition of the C 1s peak shape as well as the N 1s peak shape for paracyanogen and paraisocyanogen, respectively, shows directly that, in line with the DRIFT results, the bonding situation in both polymers is consider- ably different (Fig. 3). Analysis of the C 1s and N 1s peak J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 500 1500 2500 A/nm Afnm Solid-state UV-VIS-NIR spectra of (a) paracyanogen, (-) pristine paraisocyanogen and (---) heat treatment at 713 K shapes reveals that pristine paraisocyanogen contains carbon atoms mutiply bonded to nitrogen (C Is, 289 eV, 68%)and carbon atoms bonded directly to carbon (C Is, 285 eV, 32%).For paracyanogen the related values are 87% and 13Y0, respectively. Moreover, it should be stipulated that for pris- tine paraisocyanogen, both the carbon-to-carbon and carbon-to-nitrogen contribution to the C 1s peak shape is symmetrical. In contrast, paracyanogen possesses a distinct asymmetrical C 1s peak shape; a broad shoulder on the higher-binding-energy side, i.e. an energy-loss feature, is observed in the spectrum. The energy-loss feature can be attributed to interband transitions involving n states which are excited by some of the photoemitted electron^.'^ Similar observations also apply to the N 1s peak shape of para- Table 1 Surface elemental concentration (XPS wide scan) and relative functional group concentrations derived from XPS (narrow scan) C 1s peak-shape decomposition sample paracy anogen' paraisocyanogen (CNL C 1s binding energy 285 eV.18 surface concentration (atom YO) C N 46.8 46.3 50.7 37.4 50.0 50.0 C 1s binding energy 289 eV.'* relative concentration (Yo) 0 c=c/c-C" -CN/-NC/-C-N-' 6.9 13 87 11.9 32 68 -No mercury(0) could be detected by XPS.Theoretical composition. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 30 mz 25 X Y 20 38 15 10 I I I I I I I 280 285 290 295 300 305 EbP c XI I I 1I I 1 I 1 390 395 400 405 410 41 5 420 EbP Fig. 3 (a) C 1s and (b) N 1s peak shapes obtained with XPS of (---) paracyanogen and (-) pristine paraisocyanogen cyanogen with respect to that of pristine paraisocyanogen (Fig.3). The results strongly support that n-electron conjuga- tion is much more pronounced in the case of paracyanogen. The presence/absence of an energy-loss feature in the XPS spectra of paracyanogen and pristine paraisocyanogen, respectively, is indicative of the former being a semiconductor and the latter an ins~lator.'~ In agreement with this interpre- tation, electrical conductivity measurements (four-point-probe method)2o show that pristine paraisocyanogen and paracyanogen are an insulator and semiconductor, respec- tively (paraisocyanogen, ts = 1.0 x S cm-' and paracy- anogen, Q = 2.9 x S cm-' 'I). To assess if the experimental Q values for paraisocyanogen and paracyanogen are in line with anticipated intrinsic Q values the following crude calculations were made.Using the standard equa- tion for an intrinsic semiconductor ts = nep with n = no exp(-E$2k, T) and E, = 1.45 eV for paraisocyanogen and 0.85 eV for paracyanogen, under the assumptions that no, the number of electrons in the conduction band, is cm3 and p, the sum mobility of electrons and holes, is lod5 cm V-' s-1,22.23intrinsic Q values of 8.8 x lo-'' S cm-I and 1.0 x lop9S cm-', respectively, are calculated. Clearly, for both polymers the estimates are orders of magnitude lower than the experimental ts value and suggest that extrin- sic doping either by unintended impurities incorporated during polymerization or by oxidation in air has occurred. In the case of pristine paraisocyanogen doping with iodine vapour gave a substantial increase in electrical conductivity (298 K, cr = 1.0 x S cm-') which showithat the addi- tional extrinsic conductivity is p-t~pe.'~ Conciusions The results presented show that polymers derived from iso- cyanogen and cyanogen, respectively, i.e.paraisocyanogen and paracyanogen, possess different molecular structures and properties. Spectroscopic analysis shows that n-electron delo- calization is considerably less in the case of pristine para- isocyanogen. However, a heat treatment of pristine 33 1 paraisocyanogen leads to an increase in n-electron delocal- ization presumably due to hydrogen extrusion (increase in unsaturation) in conjunction with nitrile cyclization reactions Despite the presence of n-conjugated structures, regular ladder structures as initially proposed for paracyanogen by Whangbo et al.are not found.',' Hence, both polymers can be represented as amorphous heteroatom-substituted carbon- aceous materials. Experimental contributions of Dr. E. E. Havinga (Philips Research Laboratories, Eindhoven, The Netherlands), A. Schouten (Utrecht University, The Netherlands) and E. Neven (Vrije Universiteit, Amsterdam, The Netherlands) are gratefully acknowledged. References 1 M. H. Whangbo, R. Hoffman and R. B. Woodward, Proc. R. SOC.,London A, 1979,366,23. 2 J. L. Bredas, B. Tremans and J. M. Andre, J. Chem. Phys., 1983, 78, 6137. 3 L. J. Gay Lussac, Ann. Chim. (Paris), 1815,95, 175.4 (a)L. L. Bircumshaw, F. M. Tayler and D. H. Whiffen, J. Chem. SOC.,1954, 931; (b) J. Peska, M. J. Benes and 0.Wichterle, Coll. Czech. Chem. Commun., 1966, 31, 243; (c) M. M. Labes, Mol. Cryst. Liq. Cryst., 1989, 171, 243 and references therein; (d) J. J. Kampa and R. J. Lagow, Chem. Muter., 1993,5,427. 5 T. van der Does and F. Bickelhaupt, Angew. Chem., 1988, 100, 998. 6 0.Grabandt, C. A. De Lange, R. A. Mooyman, T. van der Does and F. Bickelhaupt, Chem. Phys. Lett., 1989,155,221. 7 F. Stroh and M. Winnewisser, Chem. Phys. Lett., 1989, 155, 21; F. Stroh, B. P. Winnewisser, M. Winnewisser, H. P. Reisenauer, G. Maier, S. J. Goede and F. Bickelhaupt, Chem. Phys. Lett., 1989, 160, 105; M. K. Scheller, H.G. Weikert and L. S. Ceder-baum, J.Electron Spectrosc., 1990, 51, 75; K. M. T. Yamada, M. W. Markus, G. Winnewisser, W. Joentgen, R. Kock, E. Vogel and H-J. Altenbach, Chem. Phys. Lett., 1989, 160, 113; G. Maier, H. P. Reisenauer, J. Eckwert, C. Sierakowski and T. Stumpf, Angew. Chem., 1992,104,1287. 8 S. J. Goede, F. J. J. De Kanter and F. Bickelhaupt, J. Am. Chem. SOC.,1991,113,6104. 9 R. W. Hoffmann, A. Riemann and B. Mayer, Chem. Ber., 1985, 118, 2433; A. Riemann and R. W. Hoffmann, Chem. Ber., 1985, 118,2544. 10 M. Sana and 0. Leroy, J. Mol. Struct., 1981, 76, 259; M. T. Nguyen, Chem. Phys. Lett., 1989, 157, 430; K. K. Sunil, J. H. Yates and K. D. Jordan, Chem. Phys. Lett., 1990, 171, 185; P. Botschwina and J. Flugge, Chem. Phys. Lett., 1991, 180, 589; W. B. De Almeida and A.Hinchliffe, J. Mol. Struct. (THEOCHEM), 1990, 65, 77; P. Botschwina and P. Sebald, Chem. Phys., 1990, 141, 311; F. M. Bickelhaupt, N. M. M. Nib-bering, E. M. van Wezenbeek and E. J. Baerends, J. Phys. Chem., 1992,%, 4864. 11 T. Saegusa and Y. Ito, Zsonitrile Chemistry, ed. Y. Ugi, Academic Press, New York and London, 1971, ch. 3 and 4. 12 G. Christian, H. Stolzenberg and W. P. Fehlhammer, J. Chem. SOC., Chem. Commun., 1982, 184; G. C. Schoemaker, D. J. Stuf- kens, S. J. Goede, T. van der Does and F. Bickelhaupt, J. Organomet. Chem., 1990,390, C1. 13 C. Wentrup, Reactive Molecules, Wiley, New York, 1984, ch. 2. 14 JCPDS Powder Diffraction File, International Centre for Dif- fraction Data, Swarthmore, PA, 1989; M. G. Northolt, L.H. Veldhuizen and H. Jansen, Carbon, 1991,29,1267. 15 J. H. Chen and M. M. Labes, Macromolecules, 1985, 18, 827; J. H. Chen and M. M. Labes, Macromolecules, 1986, 19, 1528 and references therein. 16 (a) A. Bhuiyan and S. V. Bhoraskar, J. Mater. Sci., 1989, 24, 3091; (b) T. Usami, T. Itoh, H. Ohtani and S. Tsuge, Macro-molecules, 1990, 23, 2460 and references therein ;(c) Encyclopedia of Polymer Science and Engineering, ed. J. I. Kroschwitz, Wiley, New York, 2nd edn., 1990, index vol., pp. 170-175. 17 C-J. Yang and S. A. Jenekhe, Chem. Mater., 1991,3,878. 18 K. Siegbahn, C. Nordling, G. Johanson, J. Hedmann, P. F. Heden, K. Hamrin, U. Gelius, T. Bergmark, L. 0. Werme, R. Manne and Y. Baer, ESCA Applied to Free Molecules, North-Holland, Amsterdam, New York, 1969, ch. 5. 332 19 F. R. McFeely, S. P. Kowalczyk, L. Ley, R. G. Cavell, R. A. Pollak and D. A. Shirley, Phys. Rev. B, 1974, 9, 5268 and refer- ences therein; W. M. Tong, D. A. A. Ohlberg, H. K. You, R. Stanley Williams, S. J. Anz, M. M. Alvarez, R. L. Whetten, Y. Rubin and F. N. Diederich, J. Phys. Chem., 1991,95,4709. 20 L. J. Van der Pauw, Philips Res. Rept., 1958, 13, 1. 21 Ya. M. Panshkin, T. P. Vishnyakova, A. F. Lunin and S. A. Nizova, Organic Polymeric Semiconductors, Wiley, New York, 1974,ch. 1. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 22 N. W. Ashcroft and N. D. Mermin, Solid State Physics, Interna- tional Edition, W. B. Saunders, Philadelphia, 1976. 23 E. E. Havinga, W. ten Hoeve and H. Wynberg, Synth. Met., 1993,55-57,299. 24 J. M. G. Cowie, Polymers: Chemistry & Physics of Modern Materials, Chapman and Hall, New York, 1991, ch. 17. Paper 3/05809C; Received 27th September, 1993