ABOUT SOME PROPERTIES OF THE

POLY(N-VINYLCARBAZOLE) DOPED WITH A HALOGEN

(I, Cl, Br)

G. SAFOULA,1 J. C. BERNEDE,1* S. TOUIHRI2 and K. ALIMI3

1G.P.S.E, Equipe Couches Minces et Mate riaux Nouveaux, Faculte des Sciences et des Techniques, 2rue de la HoussinieÁ re, BP 92008, 44322 Nantes Ce dex 3, France, 2Departement de Physique, FaculteÂ

des Sciences de Sfax, Sfax, Tunisia and 3Faculte des Sciences de Monastir, Monastir, Tunisia

(Received 18 April 1997; accepted in ®nal form 28 October 1997)

AbstractÐAfter doping with a halogen (X = I, Br or Cl) of the PVK, there is charge transfer (CT)complex formation as shown by electron spin resonance (ESR), optical absorption and X-ray electronspectroscopy (XPS). The conductivity of this CT-complex depends on halogen. For halogen with higherelectronegativity (Br, Cl) there is not only CT-complex formation, but there is also some polymerdegradation by the halogen. This degradation is ampli®ed when the temperature during doping isincreased. It results in the formation of a new amorphous polymer with some crystallites of carbazoleand NH4X embedded in an amorphous polymeric matrix. The increase of the conductivity of these newpolymers is probably related to the formation of some p bonds. It appears that, after all, doping withiodine is more promising since there is some polymer degradation even at room temperature with otherhalogens. # 1998 Elsevier Science Ltd. All rights reserved

INTRODUCTION

Organic polymers o�er many advantages such aseasy processability, small density, large ¯exibility ofthe chemical structure, good mechanical propertiesand many doping ways which o�er a wide varietyof speci®c applications.Organic materials will probably be an important

part of the electronic and optoelectronic materialsof tomorrow.Since the discovery, in 1957, of the photoconduc-

tive properties of poly(N-vinylcarbazole) [1] manyworks have been performed on that polymer [2, 3].Many authors have investigated the electrical andoptical properties of PVK in relation with theirstructural properties. PVK has been used in xero-graphic systems and its nonlinear optical propertieshave also shown interest [14±19]. However, theabsorption of PVK which is entirely in the ultraviolet region (l< 350 nm) and its electrical proper-ties (carrier trapping e�ect) are limiting parametersfor new applications such as radioxerography [4, 9]or optoelectronics [10, 11]. One possibility to over-come these limitations is to form a charge transfercomplex (CT-complex) between PVK and a dopant.The PVK being an electron donor, the dopantshould be an electron acceptor. PVK/TNF (trinitro-¯uorenone) blends have been largely studied [3±9].Another simple way to form a CT-complex is theuse of halogen, since halogens have high electrona�nity w. Even if some works have been reportedon the complex PVK iodine [3, 12±16] and someothers on bromine doped PVK [3], not many works

have been reported in that ®eld of investigations.However, a synthesis paper related to the inter-action of halogens (Br, Cl, I) with carbazole substi-tuted polydiacetylenes has been published [1±7] andit will be very helpful for the discussion of the pre-sent experimental results.In this paper we review the properties of CT-

complex salts obtained by doping PVK with halo-gen: I, Br or Cl. The results are discussed with thehelp of the main property di�erences of halogenatoms: electronegativity, size, etc.

EXPERIMENT

Doping of the PVK powder

The poly(N-vinylcarbazole) was provided by Aldrich, itspurity was 99.99%. In order to dope the PVK, some pow-der was introduced in a Pyrex tube with a small amountof halogen. Then the Pyrex tube was sealed. The reactionduration was 24 h to one week in order to achieve the sat-uration state. The iodine was introduced in the tube in thesolid phase state. Br was ®rst sealed in a small capillarytube and this tube was broken by shaking the Pyrex tubeafter sealing, chlorine was introduced in the tube under a0.25 atm chlorine pressure.Sometimes the reaction was activated by heating the

Pyrex tube, the annealing temperature was varied from373 K to 573 K.

Experimental technique of measurements

A SETARM/TG-DTA 92 was used for the gravimetricthermal analysis (GTA) at a heating rate of 1 K minÿ1.The pyrolysis head was used as the heat source for thethermal degradation. Approximately 0.5 g of ®ne powderwas placed in a Pyrex tube connected to a liquid nitrogentrap and a vacuum pump. The vacuum was maintained at10ÿ2 mmHg. The polymer was outgassed under vacuum

Eur. Polym. J. Vol. 34, No. 12, pp. 1871±1876, 1998# 1998 Elsevier Science Ltd. All rights reserved

Printed in Great Britain0014-3057/98 $ - see front matterPII: S0014-3057(98)00050-0

*To whom all correspondence should be addressed.

1871

for 5 h prior to pyrolysis. The volatile pyrolysis productswere distilled from the polymer immediately after theywere formed and condensed in the cold trap.

Infrared spectra were obtained by the KBr disk methodusing an FITR spectrometer. The optical measurementswere carried out at room temperature using a ``CARY2300'' spectrometer. The optical density (OD) wasmeasured at wavelengths from 2 to 0.2 mm. Opticalmeasurements were also performed by the KBr diskmethod (pressed pellet of KBr mixed with a small amountof polymer).

XPS measurements have been carried out with aLeybold LHS-12 apparatus. (XPS measurements havebeen performed at the University of Nantes, Nantes,CNRS.) XPS data were obtained with the magnesiumsource radiation (1253.6 eV) operating at 10 kV and10 mA and the pass energy was set to 50 eV. High resol-ution scans with a good signal-to-noise ratio wereobtained in the C 1s, N 1s, O 1s and I 3d or Cl 2p or Br3d regions of the spectrum. In order to decrease thecharge e�ect the powders were ®xed to the substrate bypressing polymer powders on a sheet of indium. The quan-titative analyses were based on the determination of the C1s, N 1s and I 3d5/2, Cl 2p and Br 3d peak areas with, re-spectively, 0.2, 0.36, 6.4, 0.58 and 0.67 as sensitivity fac-tors. The sensitivity factors of the spectrometer were givenby the manufacturer. The vacuum in the analysis chamberwas about 10ÿ6 Pa. All the spectra were recorded underidentical conditions.

The decomposition of the XPS peaks into di�erent com-ponents and the quantitative interpretation were madeafter substraction of the background using the Shirleymethod [18]. The developed curve ®tting program permitsthe variation of parameters such as the Gaussian/Lorentzian ratio, the full width at half maximum(FWHM), the position and the intensity of the contri-bution. These parameters were optimized by the curve ®t-ting program, in order to obtain the best ®t.

The structure of the powders was examined with ananalytical X-ray system, type DIFFRACT AT V 3.1Siemens which used a graphics program EVA. The wavelength l=1.5406 AÊ with the Cu Ka source.

1H NMR (100 MHz) spectra were recorded on a VarianXL 100 spectrometer. TMS was used as the internal stan-dard and CDCl3 as the solvent.

The ESR measurements were carried out in a BrukerER 200 spectrometer operating at X bands. The conduc-tivity of the samples was measured in the dark on powderpellets at 5�106 Pa. The two faces of the pellets (diameterf = 8 mm) have been metallized by evaporation of goldwire under vacuum. An electrometer (KEITHLEY 617)was used to measure the DC conductance between 290and 440 K.

The limited temperature domain was due to the quitehigh resistivity of some pellets for the low temperaturelimit and to the polymer degradation for the high tem-perature limit.

EXPERIMENTAL RESULTS

Systematic characterizations have been pursuedon the PVK powders doped with di�erent halogens.Some of them have been earlier published [14±16].Therefore, these results will be summarized indi�erent tables in order to facilitate their discussion.

Gravimetric thermal analysis

The results are summarized in Fig. 1 where theweight losses (in per cent) are reported versus thetemperature of measurement. The samples are PVKpowder doped with iodine and bromine at di�erent

temperatures, doped with chlorine at room tem-perature and pure PVK powder. It can be seenthat, if 10% of weight loss is ®rst achieved in thedoped powder, for weight losses higher than 60%the results are opposite, except in the case of PVKpowder doped with iodine at room temperature.The ®rst rapid weight loss (10%) in doped poly-

mers can be attributed to halogen release from thepowder and also, as usually, to the loss of absorbedwater and gaseous species, which is not necessarilyassociated with any change of the structure.The smaller weight loss of doped powder at

higher temperatures will be discussed below.

Infrared spectroscopy

The IR spectrum of PVK powder indicates thepresence of the following structures: aromatic cycle(3045 cmÿ1, 1920 cmÿ1, 1885 cmÿ1, 1625 cmÿ1,1450 cmÿ1), CH2 (2930 cmÿ1, 1600 cmÿ1, 1480 cmÿ1,1405 cmÿ1), C±H (1315 cmÿ1, 1150 cmÿ1, 740 cmÿ1,715 cmÿ1) C±N (1230 cmÿ1), and C±C (1120 cmÿ1,1020 cmÿ1, 1000 cmÿ1, 920 cmÿ1). There is no sensi-ble modi®cation of the spectrum after doping atroom temperature with iodine or bromine. Afterdoping with chlorine new absorption bands between900 and 1100 cmÿ1, which can be attributed toaryl±para/Cl bonds, appears. Di�erences betweenthe reference spectrum of pure PVK and those ofiodine and bromine doped PVK appear and thenincrease with the annealing temperature.The inferences from the spectral changes that

occur during the thermal treatment of iodine dopedPVK may be summarized as follows: a progressivedecrease of the C3±N and of the aromatic skeletonvibrations while some new bands appear in the900±2000 cmÿ1 domain.In the case of bromine, when the doping tem-

perature increases the FWHM of the peakincreases. Then the wave numbers of absorptionpeaks are modi®ed and the intensity of some peaksdecreases while new peaks appear.

Optical density measurements

Typical optical density measurements arereported in Fig. 2. It can be seen that whatever thedopant is, a large absorption band appears in thedomain 400±1000 nm. This large band induces a

Fig. 1. Weight loss of the PVK powders versus the tem-perature. (a) Pure PVK; (b) iodine doped PVK at roomtemperature; (c) iodine doped PVK at 373 K; (d) iodinedoped PVK at 573 K; (e) chlorine doped powder at room

temperature; (f) bromine doped PVK at 423 K.

G. Safoula et al.1872

smoothing of the absorption edge situated at350 nm in the pure PVK. When the powders aredoped with iodine or bromine at temperaturesabove the room temperature some absorptionbands of the PVK powder disappear.

XPS results

The XPS results are summarized in Tables 1 and2. The XPS quantitative analysis shows that thesurface contamination of the polymer is small as

usual with polymer samples. It can be seen also inTable 1 that the atomic percentage of the dopantincreases with the halogen electronegativity in thesequence: I (3 at.%), Br (5 at.% to 10 at.%) and Cl(10 at.% to 20 at.%).For the chemical analysis the main component of

the C 1s spectrum (285 eV) is used as reference forthe other binding energies, since this value corre-sponds to the well-known hydrocarbon bindingenergy situated at 285 eV [19]. It can be seen that C1s can be decomposed into two components (C±C,C±N) in the case of pure PVK. One or two otherpeaks can be obtained in the case of small surfacecontamination. These peaks can be attributed toC±O±H and C = O and/or COOH groups. Thesame results can be obtained with the doped pow-der, except in the case of iodine doped sample at573 K, where the C±N has disappeared, and in thecase of chlorine doped samples where the bondsattributed to C±O±H and C=O and/or COOHbonds are systematically present and very intensive.The N 1s peak corresponds to two peaks after

doping, the ®rst one with the smaller binding energycan be assigned to covalent nitrogen and the otherone to positively charged nitrogen. However, in thecase of chlorine doping the two peaks can corre-

Fig. 2. Optical density measurements. (a) Pure PVK pow-der; (b) iodine doped PVK at room temperature; (c) iodinedoped PVK at 373 K; (d) iodine doped PVK at 573 K; (e)

chlorine doped powder at room temperature.

Table 2. XPS qualitative analysis of PVK powders doped with a halogen (I, Br, Cl)

C 1s N 1s O 1s X

PVK powder C±C C±N C±OH C=O N N+ C=O C±O±H Xÿ X

Pure 285 286.3 400.4 532.873 27 100 ± 100

Iodinedoped(X = I) At room temperature 285 286.3 288.5 400.2 401.5 532.6 619.8 621.2

73 16 9 83 17 100 35 65At 470 K during 24 h 285 286.3 288.7 400.5 402.7 532.6 619.5 621

73 23 4 86 14 100 63 37At 470 K during 24 h 285 40.2 401.2 532.7 619.4 620.6

100 86 14 100 63 31

Brominedoped

(X = Br) At room temperature 285 286 287.1 289.2 400.6 402.3 531.9 533.8 68.8 70.776 18 3.5 2.5 83 17 64 36 20 60

At 470 K during 24 h 285 286 287.5 288.8 400.7 402.5 531.8 533.1 68.9 7176 18 3.5 2.5 81 19 28 72 40 60

Chlorine doped at room temperature (X = Cl) 285 286.2 287.1 288.8 401 402.5 531.8 533.1 198.8 20157 27 11 5 83 17 28 28 35 65

First line: binding energy (eV).Second line: relative atomic percentage.

Table 1. XPS quantitative analysis of PVK powders doped with a halogen (I, Br, Cl)

Relative atomic percentage

PVK powder C N O X

Pure 92 4 4 ±

Iodine doped (X = I) At room temperature 94 3 1 2At 470 K during 24 h 93 3 1 3At 570 K during 24 h 93 3 1 3

Bromine doped At room temperature 78.5 3.5 8.5 9.5(X = Br) At 470 K during 24 h 88 4 5 3

Chlorine doped at room temperature (X = Cl) 72 3 8 17

Halogen doped poly(N-vinylcarbazole) 1873

spond to charged nitrogen. The dopant signal canbe systematically decomposed into two components.One corresponds to an ionic halogen while theother one can be attributed to covalent halogen. Itshould be noted that in this case there is also astrong evolution of the atomic percentage of theionic component following the sequence: I(80 at.%), Br (20 at.% to 40 at.%), Cl (3 at.% to5 at.%).The occurrence of N+ 1s signal and of two com-

ponents in the halogen peaks is in good agreementwith the results obtained on the polycarbazole iod-ine complexes [20].

XRD measurements

The results are reported in Table 3. The secondbroad peak (d= 4.18 AÊ ) detected in the PVK corre-sponds to the disordered phase [21, 22], and, there-fore, we can conclude that the disorder increases inthe PVK powder after doping by chlorine at roomtemperature and by bromine at temperatures higherthan room temperature (Table 3).The other peak (d = 10.78 AÊ ) corresponds to the

crystallized PVK phase. Therefore, it can be seenthat doping induces an increase of the disorder inthe PVK powder, which has been already shown in

the case of iodine doping of the poly(N-methyl-3,3-carbazole) [20]. Moreover, new compounds are putin evidence by the XRD spectrum. In the case ofiodine doping at temperatures above 373 K, NH4Ihas been identi®ed while similarly NH4Cl can beproposed after chlorine doping. In the case of bro-mine, NH4Br is not stable and carbazole appears tobe one of the products of decomposition. Afterthese dopings the signals related to the new phasesare very small which shows that the main part ofthe polymer has become amorphous.

1H NMR study

First of all, if the PVK is soluble in CDCl3 sol-vent, its solubility decreases after doping. Thisdecrease is ampli®ed when the annealing tempera-ture during doping increases. Also, the NMR spec-tra change. In the case of iodine doping [15], thePVK signal progressively disappears, the carbazolesignal increases and then decreases and at lastvanishes when the annealing temperature is higherthan 473 K. Simultaneously, a new signal appearsin the case of iodine which can be attributed toNH4I. In the case of bromine after doping at 373 KNH4Br is detected. However, at higher temperaturesthis signal vanishes while the carbazole signal isdetected at 473 K.The signal is very weak and its intensity is smaller

than the impurity traces such as CHCl3 andacetone.

ESR and conductivity measurements

Before doping there is no ESR signal induced byPVK. After doping an ESR signal is systematicallyobserved (Table 4). It can be seen that after dopingat room temperature, the ESR signal depends onthe dopant. It is either isotropic with one g (bro-mine doping), isotropic with two g (chlorine dop-ing) or anisotropic with one g (iodine doping). Ofcourse, much care should be taken because theseidenti®cations of the ESR spectrum correspond tomathematical models and they should by physicallyjusti®ed. Moreover, all these signals are modi®ed byannealing treatments (Table 4). The shape of theESR signal varies with the localization of thedopant as shown in the Discussion.

Table 3. XRD di�raction of PVK powders doped with halogen

Compound PVK NH4X Carbazole

dhkl (AÊ ) 10.78 4.18

Pure PVK powder S, B W, BPure PVK annealed at570 K S, N W, B

Iodine doped PVK powderAt room temperature S, B W, BAt 470 K W, NAt 570 K W, N

Bromine doped PVKpowderAt room temperature S, B W, BAt 420 K during 24 h W, B S, B W, NAt 470 K during 24 h W, N

Chlorine doped powder atroom temperature W, B S, B W, N

S, B: strong and broad; W, B: weak and broad; S, N: strong andnarrow; W, N: weak and narrow.

Table 4. ESR and room temperature conductivity results obtained on PVK powder doped with a halogen (I, Br, Cl) as a function ofdoping temperature

ESR results Conductivity results

DH Density of spins s at room temperaturePVK powder doped with a halogen g factor (Gauss) (cmÿ3 molÿ1) (O cm)ÿ1

Iodine At room temperature g1=2.004 8.00 2.5� 1017 2�10ÿ13

gx=gy=1.998Annealed at T = 370 K during 24 h g= 1.998 8.50 2�1017 2�10ÿ13

Annealed at T = 470 K during 24 h g= 1.999 7.40 5�1018 10ÿ6

Annealed at T = 570 K during 24 h g= 1.998 6.00 2�1018 2.5�10ÿ8

Bromine At room temperature g= 1.998 6.20 3�1020 10ÿ9

Annealed at T = 370 K during 24 h underbromine pressure

g= 1.998 8.00 4.6� 1020 10ÿ8

Annealed at T = 470 K during 3 days (3months after doping)

g= 2.002 5.30 2�1021 10ÿ6

Chlorine At room temperature g1=2.005 9.50 2�1019 10ÿ9

g2=2.000 10.50Pure PVK 10ÿ14>s>10ÿ16

G. Safoula et al.1874

It appears that the ESR signal intensity is relatedto conductivity. The room temperature conduc-tivities of the doped powders are reported inTable 4. It can be seen that the CT-complex for-mation is corroborated by a systematic increase ofthe conductivity. Experimental studies have shownthat in cases of Br and Cl doping, even at roomtemperature, there is not only CT transfer for-mation, but also a polymer chain attack by thehalogen. This polymer degradation increases withdoping temperature and a new amorphous polymeris obtained. Its conductivity is related to new conju-gated chain formation and to their spin density(Table 4).

DISCUSSION

After doping at room temperature, whatever thehalogen is, a CT-complex salt is obtained as shownby optical, XPS and ESR measurements.The new large absorption band appearing in the

optical density can be attributed to an absorptionband of a CT-complex. XPS has shown systemati-cally that at least a part or the dopant atoms is inthe anion state which demonstrates that it hasexchanged a charge with the polymer. This fact iscorroborated by the ESR signal which appears afterdoping. This signal corresponds to radicals for-mation in the polymer. However, this interactionwill depend on the halogen and the doping tem-perature. When the polymer structure is roughlypreserved after doping, the interaction is betweenthe dopant and the pendant carbazole group sincethere are no p bonds in the polymer backbone.When there is strong polymer degradation after thedoping process, which often occurs as we will dis-cuss below, the interaction cannot be so easily loca-lized.Therefore, halogen doping at room temperature

induces complex salt formation. However, the e�-ciency of this doping depends on the halogen. Theroom temperature conductivity of the iodine dopedPVK is increased by about three orders of magni-tude without any strong modi®cation of the PVKchains. In the case of Cl or Br doping, there is,even at room temperature, not only an increase ofthe conductivity, which can be attributed to theCT-complex formation, but also PVK degradation.In fact, XRD spectra have shown that these dop-

ings not only induce an amorphization of the PVKbut also induce new compounds formation (NH4X,carbazole), i.e. degradation of the polymer. Thesemodi®cations of the polymer are also shown byNMR measurements, infrared and visible absorp-tion. Moreover, XPS spectrum of chlorine dopedsamples have shown that abnormally high contri-bution of oxidized carbon has been found.Therefore, they should be attributed to C±Cl andCl±C±Cl2 bonds related to addition and/or substi-tution reactions. This is also true in the case of bro-mine for a doping at T>373 K.The increase of the conductivity of doped powder

at room temperature can be explained by theincrease of the carrier density since it correspondsto the apparition of the ESR signal. The origin of

spins is CT-complex formation between PVK andhalogen. The carriers so formed are localized on thecarbazole units, since there are only s bonds in thealiphatic chains. The di�erent shapes of the ESRsignal could be attributed to the radical environ-ment modi®cation in the carbazole group.In the case of iodine doping, the electronegativity

of the dopant is small and the N+ radicals area�ected by its environment. When the electronega-tivity of the dopant increases the shape of the signalchanges. With bromine the radicals become sym-metrical because of the strong interaction betweenN+ and Brÿ. As shown earlier, chlorine dopinginduces, even at room temperature, some decompo-sition of the polymer. Therefore the signal corre-sponds not only to the N+ radicals but also tosome C+ radicals. These C+ radicals appear in thepolymers coming from the partial decomposition ofthe poly(N-vinyl-carbazole).The polymer degradation proposed after chlorine

or bromine doping is con®rmed by the study of thein¯uence of the annealing doping temperature forall the halogens. All the experimental resultsobtained after annealing converge towards the sameconclusion: there is polymer degradation with for-mation of a crystallized phase embedded in anamorphous polymer. As shown by GTA, the newpolymer is more stable than PVK; it is also less sol-uble, it is amorphous (Table 3) while the C±Nbonds are hardly detected as shown by IR absorp-tion and XPS measurements. In fact, N is presentin the small crystallites embedded in the amorphousmatrix. Therefore, PVK doping with a halogeninduces easily polymer degradation with reticulationreactions and nitrogen compound formation.The amorphous polymer is more conductive and

its ESR signal corresponds to the carbon radical,since g = 2.000 without any anisotropy. The appa-rition of carbon radical is probably related to theformation of C=C bonds with p orbitals.Therefore, the carriers are free along these new con-jugated chains and the hopping process is neededonly to jump from one conjugated chain to another,which explains the improvement of the conduc-tivity.It has been shown by quantitative analysis that

the atomic percentage of the dopant which can beintroduced in the powder, increases when theatomic weight of the halogen decreases. This factcan be related to the smaller atomic radius ofhigher elements. However, another property of thehalogens should be taken into account: the electro-negativity (wCl=3>wBr=2.8rwI=2.5). This evol-ution of the electronegativity not only induceshigher doping concentration but also induces higherchemical reactivity. Therefore, while iodine dopingneeds high temperatures during the doping processto degrade the polymer, the degradation is e�ectiveeven at room temperature in the case of Cl doping.These conclusions are in close agreement with ear-lier studies on other polymers.In the case of iodine, Sandman et al. [17] have

shown that there is no ESR signal and no reactionat all between iodine and poly(1,6-di-N-carbazolyl-2,4-hexadyne).

Halogen doped poly(N-vinylcarbazole) 1875

In the present work the ESR signal demonstratesthe occurrence of an interaction between iodine andPVK, but all the other results demonstrate that, atroom temperature, there is not any other reactionthan this CF-complex formation by electronexchange. However, Lewis and Taylor [23] haveshown that not only iodine includes donor±acceptoraction with the polymer but also the linking in thechains of polyethylene. When the PVK doping tem-perature is increased there is probably a similare�ect.More precisely, thermal polymerization of sol-

utions of N alkyl carbazole monomers in liquid iod-ine was put in evidence by Jenekhe et al. [24]. Theyshowed that liquid iodine acts as solvent and that itis clearly the initiator of a polymerization. Theseresults have been obtained with liquid iodine whichis in accordance with our experimental results. Atroom temperature there is only CT-complex for-mation, when the iodine is melted by annealing,there is modi®cation of the polymerization.In the case of Br and Cl with the donor±acceptor

action there is also degradation of the PVK.Sandman et al. [17] showed that the ®rst reaction ofbromine with carbazole units is a bromination ofthe carbazole rings. This bromination will allowreticulation of the polymers as proposed by Blocket al. [25]. The same process will work with chlor-ine. Moreover, the experimental evidence of NH4Xcompounds in the polymers demonstrates that thereis no reticulation of the polymer, but also, at least,partial destruction of it.

CONCLUSION

The e�ect of halogen doping on PVK powder hasbeen studied. It has been shown that after dopingthere is systematic CT-complex formation. The onlyhalogen which does not attack the PVK is iodine.Halogen doping induces an increase of the conduc-tivity and the occurrence of a large absorption bandin the visible absorption domain. In the case of bro-mine and chlorine doping, the CT-complex for-mation is escorted by polymer degradation. Thisfact is related to the higher electronegativity ofthese two halogens. The degradation e�ect isstrongly increased, whatever the dopant is, by heat-ing during the doping. The polymer can be comple-tely destroyed and small crystallites of ``nitrogenhalogen dopant'' compounds precipitated in theamorphous matrix of the new reticulated polymer.

AcknowledgementsÐThe authors wish to thank MrRabiller, Mr Molinie and Mr Godoy for the RMN, ESRand ATG measurements.

REFERENCES

1. Hoegl, H., J. Phys. Chem., 1965, 69, 755.2. Gill, W. D., in Photoconductivity and Related

Phenomena, ed. J. Mort and D. M. Pai. Elsevier, NewYork, 1976.

3. Pearson, J. M. and Stolka, M., (ed.) in Poly(N-vinyl-carbazole). Gordon and Breach, New York, 1981.

4. Chen, C. and Mort, J., J. Appl. Phys., 1972, 43, 1164.5. Mort, J., Phys. Rev., 1972, 5, 3329.6. Borsenberg, P. M. and Mey, N., J. Non-Cryst. Solids,

1978, 28, 305.7. Reimer, B. and Bossler, H., Phys. Stat. Solids, 1979,

51, 445.8. Reuter, Ru. and Franke, H., Appl. Phys. B, 1989, 48,

219.9. Wang, Y., Nature, 1992, 356, 584.10. Wada, T., Zhang, Y., Sochoi, Y. and Sasabe, H., J.

Phys. D., Appl. Phys., 1993, 26, B221.11. Wang, G., Yuan, C., Wu, H. and Wei, Y., J. Appl.

Phys., 1995, 78, 2679.12. Nakazowa, Y., Ito, N., Hoshina, K., Hanna, J. I. and

Kokado, H., Jpn J. Appl. Phys., 1992, 31, 305.13. Hermann, A. M. and Rembraum, A., J. Polym. Sci.

C, 1967, 17, 107.14. Bernede, J. C., Alimi, K. and Safoula, G., Polym.

Degrad. Stab., 1994, 46, 269.15. Alimi, K., Safoula, G., Bernede, J. C. and Rabiller,

C., J. Polym. Sci. B: Polym. Phys., 1996, 34, 845.16. Safoula, G., Bernede, J. C., Alimi, K., Molinie, P. and

Touihri, S., J. Appl. Polym. Sci., 1996, 60, 1733.17. Sandman, D. J., Elman, B. S., Homll, G. P., Hefter,

S. and Velazquez, C. S., ACS Symp. Series, 1987, 337,118.

18. Shirley, D. A., Phys. Rev. B, 1972, 5, 4709.19. Beamson, G. and Briggs, D., in The Scienta ESCA

300 Database. John Wiley and Sons, Chichester, 1992.20. Wellingho�, S. T., Deng, Z., Kedrowoski, T. J., Dick,

S. A., Jenekhe, S. A. and Ishido, A., Mol. Cryst. Liq.Cryst., 1984, 106, 289.

21. Kimura, A., Yashimito, S., Kana, A., Hirato, H.,Kusahayaski, S., Nikawa, H. and Kasai, N., J. Polym.Sci. A, 1970, 2(8), 643.

22. Gri�ts, C. H., J. Polym. Sci., 1975, 13, 1167.23. Lewis, T. J. and Taylor, D. M., J. Appl. Phys., 1972,

5, 1664.24. Jenekhe, S. A., Wellingho�, S. T. and Reed, J. F.,

Mol. Cryst. Liq. Cryst., 1984, 105, 175.25. Block, H., Cowd, M. A. and Walker, S. M., Polymer,

1977, 18, 781.

G. Safoula et al.1876