J O U R N A L O F C H E M I S T R Y Materials Room temperature dedoping of conducting poly-3-alkylthiophenes Wu Chun-Guey,* Chan Mei-Jui and Lin Yii-Chung Department of Chemistry, National Central University, Chung-Li, Taiwan 32054, Republic of China. E-mail: T610002@cc.ncu.edu.tw Received 14th July 1998, Accepted 14th September 1998 Thin films of FeCl3 doped polyalkylthiophene (side chain length from 6 to 18) have been automatically dedoped in ambient atmosphere at room temperature.The dedoping rate is dependent upon the side chain length, dopant, ambient lighting and film morphology. The maximum dedoping rate occurred in the first 30 min and dropped oV rapidly after one hour. Dedoping is the reversible process of doping, bipolarons being reduced to polarons and then to the neutral state.The production of Fe2+ and HCl during the dedoping process indicates that moisture acted as the reducing agent. In other words, bipolarons were reduced by H2O to polarons. Polarons subsequently disproportionate to bipolarons and neutral polymer or reduced Fe3+ to Fe2+. Protons, produced from the oxidition of H2O, react with FeCl4- or FeCl42- to produce HCl and iron complexes.polymer chains. Therefore, after doping, the resulting films Introduction may have diVerent dedoping rates. The potential applications of conducting polymers in micro- The applications of conducting P3ATs require that we have electronics have been investigated over the past few decades1–3 a thorough understanding of their properties. In order to and a significant eVort has been made to obtain soluble and understand better the nature of the dedoping phenomenon, processable materials.4 Enhanced solubilization can be we investigated the eVects of structure regularity, alkyl side achieved by grafting long alkyl side chains on a polyaromatic chain length, dopant, solvent, and light on the room temperabackbone as has been demonstrated in polyalkylthiophenes ture dedoping of doped P3ATs.(P3ATs). Polyalkylthiophenes represent a class of conducting polymers that are soluble and processable, and yet retain a Experimental degree of the electrical conductivity of the insoluble parent, polythiophene.5–8 The solubility of polyalkylthiophenes arises Reagents from the decrease of attraction between polymer chains and 3-Bromothiophene (stored over molecular sieves), the introduction of favorable interactions between the substitu- Ni(dppp)Cl2, LDA, Mg, MgBr2, HgO, I2 and bromoalkanes ents and the solvent.Unfortunately, side chain substitution (with chain length ranging from 6 to 18), were purchased from also leads to thermal instability of the doped polymer since in Aldrich Chemical Co. or other commercial sources and used P3AT, the doped state has a higher energy than the neutral as received unless otherwise specified. All solvents are HPLC state.Thermal dedoping appears to be a critical technological grade and distilled over Na or CaH2 prior to used (bromo- problem for the application of P3AT. A considerable number alkanes were distilled over MgSO4). of studies have been made to understand the instability of doped P3AT. It was reported that the dedoping (or conduc- Preparation of P3AT films tivity) decay rate depends on the dopant,9,10 degree of doping and temperature11 as well as the humidity level of the surround- The monomers, 3-alkylthiophenes with a side chain length ing atmosphere.12,13 On the other hand, some studies also ranging from 6 to 18, were prepared by a published procedure20 showed that the dedoping process also depends on the arrange- (Ni catalyzed coupling of the alkyl Grignard with 3-bromoment of alkyl side chains in the polymer backbone.14 A thiophene).The purity of monomers was checked by 1H NMR reduction of the number of alkyl side chains in a regular way and found to be >95%. Regio-random P3ATs were prepared increased the stability towards thermal dedoping. by chemical oxidative coupling of the corresponding monomers The dedoping mechanism in P3AT has been explained by with FeCl3–CHCl3 according to the procedure described by the increasing reactivity of the conducting oxidized state Sugimoto et al.21 Removal of oligomers and impurities from towards reducing species, such as water.9,15 However, the obtained polymers was achieved by Soxhlet extraction Horowitz et al.16 found that the oxidized polymer (in solution) with MeOH.The dark red neutral P3AT was dried under is reduced to the neutral state when acetone was added, but vacuum. Regio-regular polymers were prepared by a literature the exact nature of this reaction is still unknown. Most studies procedure.22 The percentage of HT–HT dyad was measured have centered on the thermal dedoping process of P3ATs by 1H NMR spectroscopy.23 25 mg of dried P3AT was discontaining FeCl4- counter ions.Pei et al. believe that thermally solved in 2.5 ml of solvent and the solutions were cast onto a activated side chain mobility is the cause for dopant removal silicon wafer or a glass slide.All film thicknesses were between from the polymer backbone and that larger dopants will be 4750 and 5250 A° . removed more easily.17 However, Ciprelli et al. discovered that the size of the dopant ions does not appear to be a determining Doping of P3AT films with solutions of FeCl3 or AuCl3 in dry factor in the dedoping process10 and that the electronic propernitromethane ties of the counter anion play a key role in this process.Nevertheless, it was reported by Abdou et al. that for thin Chemical doping was performed by dipping the film (cast on films (<1 mm), photochemical dedoping dominates over ther- substrate) in a nitromethane solution of FeCl3 or AuCl3 for mal dedoping under ambient lighting.18 In addition, solvato- 30 and 60 min, respectively.After oxidation, all films were chromic behavior was observed for polyalkylthiophene.19 The rinsed several times with nitromethane and dried with N2. UV–VIS–NIR absorption spectroscopy was used to monitor solvents used to cast the polymer films aVect the packing of J. Mater. Chem., 1998, 8, 2657–2661 2657Table 1 Molecular weights, Dlmax, HT–HT dyad ratios and average the doping process.Completeness of doping was judged by dedoping rate in the 30 min of P3ATs the total disappearance of the p–p* transition of the neutral polymer. The degree of oxidation was determined by elemen- Side Average dedoping tal analysis. chain Molecular HT–HT rate in the length weight Dlmax b/nm ratio (%) first 30 min Rate of dedoping 6 4 000 39 53 1.86 The dedoping process was monitored by UV–VIS–NIR 7 66 000 32 55 2.88 spectroscopy. 8 100 000 65 67 5.54 8a 4 000 90 96 1.50 10 95 000 28 49 7.77 Physical measurements 12 70 000 34 58 5.86 12a 4 000 85 98 1.60 Fourier transform IR spectra were recorded for films on a Si 14 110 000 35 58 7.08 substrate using a Bio-Rad 155 FTIR spectrometer. The thick- 16 4 000 34 57 6.74 nesses of the polymer films were measured with a Dektak 3 18 3 200 34 48 8.85 surface profile measuring system.The scan length is 5 mm and aRegio-regular polyalkylthiophene. blmax(film)-lmax(CHCl3 solu- the thickness was calculated from the average of the length tion). scanned. The thickness of films was further calibrated by UV–VIS absorption measurements. UV–VIS–NIR spectra were obtained using a Varian Cary 5E spectrometer in the octadecylthiophene). The degree of doping determined from laboratory atmosphere at room temperature. Scanning electron elemental analysis is 0.25 which is similar to that reported in microscopy (SEM) and energy dispersive spectroscopy (EDS) the literature.28–31 The doping time and degree of doping of studies were performed with a Hitachi S-800 apparatus at regio-regular P3ATs are close to those of regio-random poly- 15 kV.X-Ray photoelectron spectroscopy studies were carried alkylthiophenes. out on a Perkin-Elmer PHI-590AM ESCA/XPS spectrometer system with a cylindrical mirror electron (CMA) energy ana- Room temperature dedoping of P3AT lyzer. The X-ray sources were Al-Ka at 600 W and Mg-Ka at The FeCl3 doped P3ATs were stored at room temperature and 400 W.Depth profile secondary ion mass spectra (SIMS) were in ambient lighting. The dedoping of P3AT was monitored by measured with a SIMS Cameca, ims-4f instrument where a UV–VIS–NIR spectroscopy as shown in Fig. 1(a). It was Cs+ ion gun for sputtering was adjusted to 10 kV acceleration found that as the dedoping proceeds, the peak at ca. 1640 nm voltage. Gel permeation chromatography analyses were carried (due to charge-carrying bipolaronic states) decreased and a out on an Eldex model 9600 HPLC with a UV detector and peak at 480 nm (which corresponds to the p–p* transition of 30 cm length columns of Waters HT0&HT4 (molecular weight the neutral polymer) increased.However, the intensity of the range: 100–600 000). Polystyrene with diVerent molecular peaks at 264, 310 and 364 (the absorption peaks of FeCl4-) weights were used as calibration standards and THF was used showed no observable change in the first hour.Interestingly, as an eluent. the peak at ca. 800 nm increased initially [with a slight shift of the absorption maximum, as shown in Fig. 1(b)] then Charge transport measurements decreased as dedoping proceeds.The absorptions of polarons Direct current electrical conductivity measurements of the calculated from a theoretical model were 0.18, 1.23 and films on substrates (1.2 cm×1.2 cm square plate) were per- 1.51 eV.32 The peak at ca. 800 nm (1.7 eV) may be due to the formed in the usual four point geometry.24 The four points absorption of both polarons and bipolarons. At the beginning on the sample surface were in line at an equal spacing of of dedoping, bipolarons were reduced to polarons and 2 mm.Each point was adhered to a gold wire electrode. An although the polarons are also reduced to the neutral state, appropriate current (ranging from 1 nA to 1 mA) was main- the rate of reduction of bipolarons to polarons is higher than tained on the two outer electrodes. The floating potential that of reduction of polarons.Thus, the total concentration across the two inner electrodes was measured to determine the of polarons initially increases. After the concentration of conductivity.25 bipolarons decreased, the production of polarons from bipolarons was slower than the reduction of polarons to the Results and discussion neutral state, and hence the total concentration of polarons decreased.Preparation of P3ATs EVects of the structure and the side chain length on the dedoping Regio-random polyalkylthiophenes (P3ATs) with side chain rate lengths ranging from 6 to 18 were prepared by oxidative coupling of the corresponding monomer using iron(III ) chlor- In order to eliminate the influence of thickness on the dedoping ide as an oxidant. The ratios of HT–HT (head-to-tail to headrate, polymer films with similar thickness (5000±250 A° ) were to-tail ) dyad configuration23 and molecular weights are listed prepared.Since some of the UV–VIS absorption peaks may in Table 1. As shown in Table 1, there is no direct relationship be diYcult to identify during dedoping, the dedoping rate was between side-chain length and dyad ratio or molecular weight.calculated using the absorption peak at 1640 nm [eqn. (1) and Dark red films (thickness: 5000±250 A° ) of neutral polymers (2)]: were cast on glass slides or quartz disks from CHCl3 and THF solutions, then dipped in 0.1 M FeCl3–CH3NO2 for doping. R=(I1640-Ii1640)/I1640×100% (1) After doping, the original broad peak (due to the p–p* n=R/ti (2) transition of the neutral polymer) disappeared, while two broad absorption bands (at lmax ca. 800 and ca. 1640 nm) where R=relative intensity change, n=average dedoping rate within ti h, ti=dedoping time/h, I1640=peak intensity at grew in. These two peaks are believed to be the optical-induced electronic transitions involving charge-carrying bipolaronic 1640 nm immediately after doping and Ii1640=peak intensity at 1640 nm after dedoping for ti h.states.18,26 The peaks at 240, 316 and 368 nm correspond to the absorption of FeCl4- which is the counter anion of doped Typical average dedoping rates vs. dedoping times of polyalkylthiophene are shown in Fig. 2. The average dedoping rate polymers.27 The time required to completely dope the P3ATs is approximately the same (within 30 min), except for poly(3- was very fast in the first 2 h then decreased as dedoping 2658 J. Mater.Chem., 1998, 8, 2657–2661Fig. 1 The UV–VIS–NIR spectra of polyoctadecylthiophene during dedoping: (a) whole spectrum [(I ) just doped; (II ) after 1 h; (III ) after 21 h; (IV) after 250 h], (b) the polaron/bipolaron peak in the first hour. EVects of morphology on the dedoping rate of P3ATs The dedoping rate was aVected not only by side chain length but also by film morphology. It is found that for a given side chain length, the dedoping rate of a polymer film obtained from CHCl3 solution and doped with FeCl3 is higher than that from THF solution and doped with the same oxidant, (Fig. 2) and polyalkylthiophene films cast from diVerent solvents showed diVerent morphologies (Fig. 3). After doping, films obtained from CHCl3 solution had a rather smooth surface, indicating packing of straight polymer chains. By contrast the morphology of doped films obtained from THF solution appear as aggregates of many small polymer balls and fibers. The movement of polymer side-chains and migration of HCl gas (vide infra) are easier for polymers with Fig. 2 The average dedoping rate of FeCl3 doped polyoctadecylthiostraight chains, therefore films cast from CHCl3 solution have phene cast from ($) CHCl3, (+) THF solution. a higher dedoping rate. Furthermore, theoretical calculations have suggested that the stability of doped polythiophenes may proceeds. The maximum average dedoping rates of P3ATs all be maximized by a favored topology of the dopant molecule.36 occurred during the first 30 min and decreased with increasing In other words, the dedoping rate will decrease when the steric time.Table 1 lists the molecular weight, Dlmax (the diVerence arrangement of polymer backbone and dopant are favourable. in absorption maximum between polymer solution and poly- Therefore, the higher stability of doped polyalkthiophene films mer film), HT–HT ratios and average dedoping rate in the made from THF solution may result from the fact that these first 30 min of P3ATs with various side chain lengths.P3ATs polymer films have suitable steric arrangements to accommowith similar regio-regularity have similar absorption maxima date the FeCl4- dopant. In addition, the dissimilar morin solution [except for poly(octadecylthiophene)].The absorp- phology of polymer films obtained from THF and CHCl3 tion maximum for the p–p* transition of polymer solutions is solutions indicated that the interactions between polymer blue shifted with respect to the solid state value. This is due chains of those two types of polymer films are also diVerent. to the change in the planarity of conjugated backbone (solid This was displayed by the diVerent energy of polarons and state packing stabilizes the more planar conformation) which bipolarons in these two types of polymer films as shown in in turn depends on the structure regio-regularity of the poly- Fig. 4. Polymer films obtained from THF solution have lower mers.33–35 We found that the shift of the absorption maximum absorption energies for the polarons and bipolarons and from solution to the solid state film, Dlmax, paralleled the therefore have a lower dedoping rate.structure regio-regularity. However, the maximum dedoping rate is independent of regio-regularity of regio-random P3ATs. EVects of light, dopant and moisture on the dedoping rate The dedoping phenomenon is the process of reduction of the partially oxidized polymer backbone.One can expect that a Photochemical dedoping of FeCl3 doped P3ATs thin films is known18 and it is believed that the photolability of the FeCl4- high degree of conjugation of the polymer backbone (or better stacking of the polymer chains) will lead to a more stable dopant is the cause. We studied the dedoping rate of FeCl3 doped P3AT thin films (ca. 5000 A° ) in the dark and under the doped state, and therefore give a low dedoping (reducing) rate. This is well demonstrated in the dedoping rate between ambient lighting at room temperature. It was found that the dedoping rates follow similar time dependent patterns in both regio-regular and regio-random P3ATs (Table 1). Therefore, the independence of the dedoping rate on structure regio- environments.However, we did not observe the formation of polymer bound alcohol or cross-linked polymer as seen by regularity of regio-random polyalkylthiophene indicated that some other factors may have a large influence on the dedoping Abdou and Holdcroft2 in the photolyses of P3AT in ambient air. Light as well as temperature and humidity14 are known rate of those polymers.As expected, the side chain length plays a major role in to be factors which cause the dedoping of P3ATs. By contrast with the literature report,2 we found that the dedoping rate of determining the dedoping rate of P3ATs. In general, the longer the side chain length the higher the dedoping rate. It is the sample stored in the dark is slightly faster than that in ambient light. The eVects of light on the dedoping rate of the believed18 that the dedoping phenomenon is due to thermally activated side chain mobility, which causes the products (of polymer films were also independent of the film morphology. Replacement of FeCl4- with a less photolabile dopant, the dedoping process) to be removed from the polymer backbone.At room temperature, longer alkyl side chains have a AuCl4-, decreased the dedoping rate.Nevertheless, the dedoping rates are only 4–15 times lower than for samples doped higher ability to remove the dedoping products, and therefore lead to a higher dedoping rate. with FeCl3. This implies that the photolysis of dopant may J. Mater. Chem., 1998, 8, 2657–2661 2659Fig. 3 SEM micrographs of doped (with FeCl3–CH3NO2) polyoctadecylthiophene cast from (a) CHCl3, (b) THF solution.those reported by Abdou et al.18 The released HCl was detected by pH paper or trapped by an active metal film, such as aluminium foil, and the resultant AlCl3 analyzed by EDS. Depth profile SIMS and ESCA spectra of dedoped P3AT showed that the distribution of the elements is fairly homogeneous through the whole film (Fig. 5). This indicates that during dedoping no significant ion migration or aggregation occurred. Moreover, depth profile Auger studies of P3AT during dedoping (after dedoping for 2 h) showed that although the concentration of Cl is higher on the surface, the Fe concentration was the same throughout the film. Reduction of the polymer backbone was also proved by the decrease in the binding energy of C 1s and S 1s.Detailed ESCA and Auger studies of P3ATs during dedoping will be reported elsewhere.37 Moreover, in situ UV–VIS–NIR studies showed that at the beginning of dedoping, the peak intensity at ca. 800 nm increased in the first 30 min and then decreased (vide supra). This indicates that polarons were generated in the dedoping Fig. 4 The UV–VIS–NIR spectra of doped (with FeCl3–CH3NO2) process.Such bipolaron to polaron transformations have been polydodecylthiophene cast from (a) CHCl3, (b) THF solution. observed in the thermal dedoping of thiophene oligomers.38 Both polaron and bipolaron peaks disappeared totally after not be the major factor for the dedoping of polyalkylthiophene dedoping for 3 months. However, the dedoped P3AT has a in ambient light.room temperature conductivity of 10-7 S cm-1, two orders of It was suggested by Ciprelli et al.10 that H2O was the magnitude higher than the neutral polymer. This increased reducing agent during the dedoping of P3ATs. In order to test conductivity may due to a trace amount of polarons and salt the eVect of water molecules on the dedoping rate of polyalkyl- impurity in the polymer film.Scheme 1 shows a simple dedopthiophene, the doped films were stored in a glove box (water ing mechanism derived from the above observations: concentration <1.0 ppm). The dedoping rate of such polymer bipolarons are reduced by H2O to polarons which then dispro- films was much lower than that of samples in the ambient portionate39 to bipolarons and neutral polymer.FeCl4- reacts atmosphere. Unfortunately, since the UV–VIS–NIR spec- with H3O+, produced from the oxidation of H2O, to form trometer was exposed to air, the dedoping of P3AT occurred Fe3+ complexes and HCl. Polarons or neutral polymer may when the UV–VIS–NIR spectra were measured. Reversibility of the doping/dedoping process: compared to the neutral polymer film, the intensity of the p–p* transition absorption decreased slightly (and was slightly blue shifted) after total dedoping.This implied that P3AT film degraded slightly during the dedoping process. Dedoping mechanism Polyalkylthiophenes in the neutral state are thermodynamically more stable than that in the doped state. In the presence of a large amount of good oxidant, however, the polymer backbone is oxidized.After the removal of oxidant, the polymer has the tendency to reduce back to the neutral state and if a reducing agent such as H2O is present, the polymer backbone will be reduced automatically. Nevertheless, some contradictions were observed in various studies.28 In order to understand more about the dedoping mechanism, we studied the products of the dedoping process very carefully.We found that Fe2+ and HCl were produced during the dedoping process (the ESCA spectrum showed no observable Fe2+ peak in the fully doped P3AT film37). The coexistence of Fe2+ and Fe3+ inside the polymer Fig. 5 Depth profile secondary ion mass spectra (SIMS) analysis of polydodecylthiophene after dedoping for 3 months. films was verified by ESCA studies. The results are similar to 2660 J.Mater. Chem., 1998, 8, 2657–26616 M. Leclerc, F. M. Diaz and G. Wegner, Makromol. Chem., 1989, 190, 3105. 7 G. Zerbi, B. Chierichetti and O. Inganas, J. Phys. Chem., 1991, 94, 4646. 8 C. Roux, J. Y. Bergeron and M. Leclerc, Makromol. Chem., 1993, 194, 869. 9 Y. Wang and M. F. Rubner, Synth. Met., 1990, 39, 153. 10 J.L. Ciprelli, C. Clarisse and D. Delabouglise, Synth. Met., 1995, 74, 217. 11 M. T. Loponen, T. Taka, J. Laakso, K. Vakiparta, K. Sunronen, P. Valkeinen and J. E. Osterholm, Synth. Met., 1991, 41, 479. 12 M. C. Magnoni, M. C. Gallazzi and G. Zerbi, Acta Polym., 1996, 47, 228–233. 13 M. R. Andersson, Q. Pei, T. Hjertberg, O. Inganas, O. Wennerstrom and J. E. Osterholm, Synth. Met., 1993, 55, 1227. 14 T. Taka, Synth. Met., 1993, 57, 4985. 15 K. Yoshino, S. Morita, M. Uchida, K. Muro, T. Kawai and Y. Ohmori, Synth. Met., 1993, 55–57, 28. 16 G. Horowitz, A. Yassar and H. J. Von Bardeleben, Synth. Met., Scheme 1 The room temperature dedoping process of poly- 1994, 63, 245. alkylthiophene. 17 Q. Pei, O. Inganas, G. GustaVson, M. Granstrom, M. Anderson, T. Hjerberg, O. Wennerstorm, J.E. Osterholm, J. Laakso and H. Jarvinen, Synth. Met., 1993, 55–57, 1221. also reduce some FeCl4- to FeCl42- which react with H3O+ 18 M. S. A. Abdou and S. Holdcroft, Chem. Mater., 1994, 6, 962. to form HCl and stable Fe2+ complexes. The redox reaction 19 K. Yosghino, P. Love, M. Omoda and R. Sugimoto, Jpn. J. Appl. between water and the polymer backbone and the release of Phys., 1988, 27, 2388.HCl gas could be the rate determination steps. P3ATs with 20 K. Tamao, S. Kodama, I. Nakajima, M. Kumada, A. Minato and shorter alkyl side chains form denser films and such films or K. Suzuki, Tetrahedron, 1982, 38, 3347. films with spherical morphology retard the release of HCl and 21 R. Sugimoto, S. Takeda, H. B. Gu and K. Yoshino, Chem. Express, 1986, 1, 635.the diVusion of water out of/in the polymer backbone, resulting 22 (a) R. D. McCullough and R. D. Lowe, J. Chem. Soc., Chem. in a lower dedoping rate. This mechanism also explains why Commun., 1992, 70; (b) R. D. McCullough, R. D. Lowe, at the start of dedoping, the concentration of FeCl4- remains M. Jayaraman and D. L. Anderson, J. Org. Chem., 1993, 58, 904. constant and the concentration of polarons increases. 23 (a) T.-A. Chen, X. Wu and R. D. Rieke, J. Am. Chem. Soc., 1995, 117, 233; (b) R. M. Souto, K. Hinkelmann, H. Eckert and F. Wudl, Macromolecules, 1990, 23, 1269. Conclusions 24 (a) S.-A. Chen and H. T. Lee, Macromolecules, 1993, 26, 3254; (b) G. B. Street, in Handbook of Conducting Polymers, ed. Dedoping of conducting polyalkylthiophene films at room T.A. Skotheim,Marcel Dekker, New York, 1986, vol. 1. p. 224. temperature occurred when the oxidized polymer was reduced 25 F. M. Smiths, Bell System Technical J., 1958, 710. by water. The dedoping rate depends on the polymer side 26 (a) M. S. A. Abdou and S. Holdcroft, Synth. Met., 1993, 60, 93; chain length, film morphology, structure regio-regularity, (b) H. Stubb, E.Punkka and J. Paloheimo, Mater. Sci. Process., dopant and light. During dedoping, some Fe3+ is reduced to 1993, 10, 119; (c) N. Colaneri, D. Nowak, D. Spiegel, S. Hotta and A. J. Heeger, Phys. Rev. B, 1987, 36, 7964. Fe2+ and at the same time HCl is produced. The dedoping 27 (a) H. L. Friedman, J. Am. Chem. Soc., 1952, 74, 5; process involved the conversion of bipolarons to polarons and (b) T.B. Swanson and V. W. Laurie, J. Phys. Chem., 1965, 69, 244. then to the neutral state. However, since polarons have a 28 Y. Cao, P.Wang and R. Qian, Macromol. Chem., 1985, 186, 1903. relatively low oxidation/reduction potential, (or may be 29 G. W. HeVner and D. S. Pearson, Synth. Met., 1991, 44, 341. trapped in defect sites), complete conversion to the original 30 S.Hotta, T. Hosaka, M. Soga and W. Shimotsuma, Synth. Met., neutral state may not occur at room temperature. 1984, 9, 87. 31 H. Neugebauer, G. Nauer, A. Neckel, G. Tourillon, F. Garnier and P. Lang, J. Phys. Chem., 1984, 88, 652. Acknowledgments 32 M. Boman and S. Stafstrom, Synth. Met., 1993, 55–57, 4614. 33 G. Horowitz, B. Bachet, A. Yassar, P. Lang, F. Demanaz, J. Fave This work was supported by the National Science Council of and F. Gariner, Chem.Mater., 1995, 7, 1337. the Republic of China via grant NSC-86-2113-M-008-005. 34 C. Wang, M. Benz, E. LeGoV, J. L. Schindler, T. J. Allbritton, C. R. Kannewurf and M. G. Kanatzidis, Chem. Mater., 1994, 6, 401. References 35 M. D. Curtis and M. D. McClain, Chem. Mater., 1996, 8, 936. 36 M. C. Magnoni, M. C. Gallazzi and G. Zerbi, Acta Polym., 1996, 1 S. X. Cai, J. F. W. Keana, J. C. Nabity and M. N. Wybourne, 47, 228. J. Mol. Electron., 1991, 7, 63. 37 M.-J. Chan and C.-G. Wu, manuscript in preparation. 2 M. S. A. Abdou and S. Holdcroft, Synth. Met., 1992, 52, 159. 38 M. G. Ramsey, D. Steinmuller and F. P. Netzer, Synth. Met., 3 W.-S. Huang, Polymer, 1993, 35, 4057. 1993, 54, 209. 4 (a) R. J. Jensen and J. H. Lai in Polymers for Electronic 39 (a) M. Deussen and H. Bassler, Synth. Met., 1993, 54, 49; Applications, ed. J. H. Lai, CRC Press, Boca Raton, FL, 1989, ch. (b) L. M. Tolbert, Acc. Chem. Res., 1992, 25, 561. 2; (b) J. R. Reynolds, J. Mol. Electron., 1986, 2, 1. 5 K. Yoshino, K. Nakao and R. Sugimoto, J. Appl. Phys., 1989, Paper 8/05462B 28, L490. J. Mater. Chem., 1998, 8, 2657–2661 2661