Electrical conductivity of halogen doped poly-(N-vinylcarbazole) thin films
G. Safoula a, K. Napo a, J.C. Bernede a,*, S. Touihri b, K. Alimi c
a Groupe de Physique des Solides pour lElectronique, Universite de Nantes, Equipe Couches Minces et Materiaux Nouveaux, FSTN,BP 92208, 2, Rue de la Houssiniere, 44072 Nantes Cedex 03, France
b Department de Physique, Fac. Sciences Sfax, 3038 Sfax, Tunisiac Fac. Sciences Monastir, 5000 Monastir, Tunisia
Received 3 September 1999; received in revised form 4 February 2000; accepted 23 August 2000
Thin films obtained by thermal evaporation of poly(N-vinylcarbazole) (PVK) powder have been doped with iodine
and chlorine either by using a doped powder or by doping deposited films. The doping level of the thin films has been
checked by X-ray photoelectron spectroscopy, then the samples have been electrically characterised by currentvoltage
and room temperature conductivity measurements. It is shown that chlorine reacts with PVK during doping which
induces new compounds formation.
Three dierent conductivity domains have been put in evidence, in the case of iodine doping. In the low field range
the current is dominated by space charge eect. In intermediary field the current is ohmic. In the high field range the
PooleFrenkel eect is dominating.
The evolution of properties of some samples is attributed to iodine ionisation of neutral iodine under high field.
These results, obtained on evaporated thin films, are compared to those obtained with spin-coated PVK
films. 2001 Elsevier Science Ltd. All rights reserved.
Keywords: Poly(N-vinylcarbazole); Thin films; PooleFrenkel; Space charge eect
Organic polymers are recently intensely studied for
their potential characteristics, such as the easy physical
manipulation and the great flexibility of chemical
structure (extraction). These two characteristics allow
processing cost of return to be reasonable. Conductor
and semiconductor polymers, particularly have many
doping possibilities for specific applications.
Poly(N-vinylcarbazole) (PVK) is mostly studied 
because of its remarkable photoconductivity properties.
One of the investigation aspects consists of optimising
the potential application in xerography field [7,8], which
needs an absorption in the visible wavelength by com-
plex salt formation .
Such absorption in the visible region could also be
very interesting in order to achieve solar cells. Solar cells
based on organic thin layers is now a current interest of
many researchers . This is mainly due to low cost,
simplicity of fabrication of large area, however many
progresses should be done to achieve acceptable e-
ciency and device stability. In order to progress in that
direction a good knowledge of the conduction processes
in the films is necessary.
The most important properties studied in this mate-
rial is charge generation from photo excitable centres,
charge transfer (CT) and the electro-optical eect.
Dierent methods to obtain PVK thin films using
chemical processes  and recently physical pro-
cesses, sputtering  or evaporation [19,20], are re-
ported in the literature.
European Polymer Journal 37 (2001) 843849
* Corresponding author. Tel.: +33-2-51-12-55-30; fax: +33-2-
E-mail address: email@example.com-
tes.fr (J.C. Berne`de).
0014-3057/01/$ - see front matter 2001 Elsevier Science Ltd. All rights reserved.PII: S0 01 4 -3 05 7 (00 )0 0 18 5 -3
With regard to what has been said before, we have
attempted to study PVK doped with iodine or chlorine
 to better understand CT process between PVK
and halogen as well as to compare the electrical prop-
erties of PVK thin films obtained by evaporation to that
of films obtained more classically by spin coating.
2. Materials and methods
In the case of iodine, we describe the electrical con-
ductivity of thin films obtained by using two dierent
techniques. In the first one  thin films are obtained
using vacuum thermal evaporation of PVK powder pre-
doped to 3 at.% iodine atoms. The second technique
consists of thin films deposited from pure PVK powder
and then doped in an iodine atmosphere. Thin films
obtained in this case are called iodine post-doped.
In order to dope the samples with chlorine, they were
exposed to chlorine at room temperature in a glass tube
under a 0.25 atm. chlorine pressure.
The result of a comparative study of the physico-
chemical properties of thin films obtained through these
two processes was published recently . It has been
shown that the post-doped films were more homogeneous
and, as discussed below, that more iodine was present in
these films. The doping level of the dierent samples has
been estimated through X-ray photoelectron spectros-
copy (XPS) analysis combined with argon ion etching.
The electrical measurements were carried out in dark
chamber. Several sets of measures were realised on each
sample to allow stability monitoring after electrical ex-
citation. Measurements of d.c. voltagecurrent (IV )
were accomplished in a wide voltage range (11000 V) at
room temperature (T 270 K). The d.c. voltage wasfurnished by a high voltage generator HP 6110 A d.c.
The current intensity was measured with a KEITHLEY
The conductivity measurements were done using the
same electrometer, coupled to a multimeter which is
connected to a cuivreconstantan thermocouple to en-
able reading the temperature.
The structure of samples is longitudinal. Golden elec-
trodes were deposited by vacuum evaporation. The dis-
tance between these two electrodes is d 1 mm. Thethickness of PVK iodine doped thin films varied from 1 to
5 lm.As mentioned before, the evolution of the polymer
after doping has also been checked by XPS measure-
ments. XPS measurements were performed with a
magnesium X-ray source (1253.6 eV) operating at 10 kV
and 10 mA. The energy resolution was 0.75 eV at a pass
energy of 50 eV. The quantitative XPS study was based
on the determination of the C1s, N1s, I3d5=2, Cl2p and
O1s peak areas, with 0.2, 0.36, 6.4, 0.58 and 0.61, re-
spectively, as sensitivity factors (the sensitivity factors
were given by the manufacturer (Leybold)). The de-
composition of the XPS peaks into dierent components
and the quantitative interpretation, were made after
subtraction of the background using the Shirley method
. The developed curve-fitting programs permit the
variation of parameters such as the Gaussian/Lorentzian
ratio, the full width at half maximum, the position and
the intensity of the contribution. These parameters were
optimised by the curve-fitting program, in order to ob-
tain the best fit. The binding energy positions were
corrected to annul the charging eects. The calibration
uses the binding energy of the C1s peak of the hydro-
carbon, the position of which is assumed to be constant
at 285 eV .
The iodine depth profile in the samples was studied
by recording successive XPS spectra after ion etching for
short periods. Using an ion gun, sputtering was ac-
complished at pressures of less than 5:1 104 Pa with a10 mA emission current and a 3 kV ion beam energy. In
such conditions all the surface of the films was sputtered.
3. Results and discussion
Room temperature conductivity of the dierent
samples is reported in Table 1. It can be seen that, while
there is systematically an increase of the conductivity
after iodine doping, after chlorine doping there is no or
uncertain conductivity increase. The dierences in con-
ductivity at room temperature between pre- and post-
doped PVK thin films are probably indicated by their
structure, stability and iodine concentration.
Before anymore electrical characterisation the eect
of the halogen doping on the films have been checked by
Since XPS spectra of PVK and iodine doped PVK
have been discussed elsewhere  only the most
significant results will be discussed here. In Fig. 1 the
C1s spectra of pure PVK (Fig. 1a), iodine doped PVK
(Fig. 1b) and chlorine doped PVK (Fig. 1c) are reported.
It can be seen that the signal obtained for pure PVK and
iodine doped PVK are more or less similar with, after
iodine doping a small relative increase of the two smaller
peaks by comparison with the highest one.
Room temperature conductivity
Thin film Pure PVK Iodine doped Iodine post-doped Chlorine doped Chlorine post-doped
r (X cm)1 6 1010 108 106 6 1010 10106r6 108
844 G. Safoula et al. / European Polymer Journal 37 (2001) 843849
In the case of chlorine doped sample the C1s peak
shape is strongly modified, which induces a very dierent
peak decomposition. In Fig. 1a and b the peak situated
at 285 eV can be attributed to CC bonds , that with
a binding energy of about 1 eV more can be assigned to
CN  while the third one can be attributed to some
contamination COH. When a CT complex salt is
obtained positive radicals situated on some carbons and/
or nitrogens of the polymer chains appear. Therefore the
increase, after doping, of the peak situated at 286 eV can
be attributed to C formation. The increase of the thirdpeak corresponds to some oxygen contamination of the
sample during the doping process. In the case of chlorine
the fourth peak are of similar area, which means that
there is a strong oxidation of the carbon chain. The
chlorine reacts with the polymer to give dierent chlo-
rinated compounds. This high reactivity of chlorine with
PVK explains the instability of the measured conduc-
tivity and the small eect of chlorine on the polymer
conductivity, since no stable CT complex salt forms, but
chlorine compounds. Therefore no more electrical study
could be done on chlorine doped sample and only
electrical properties of iodine doped PVK will be dis-
About iodine doped PVK thin films, the iodine depth
profiles are presented Fig. 2. It can be seen that, after a
small decrease of iodine concentration at the beginning
of the etching, there is a stabilisation of the concentra-
tion in the bulk whatever the sample. However it can
also be seen that, if there is about 2 at.% of iodine in the
post-doped films, there is only 0.3 at.% in the films ob-
tained from the pre-doped powder (iodine pre-doped
Fig. 3 shows the results of measured density of cur-
rent (J ) as a function of voltage (V ) for dierent thin
films. The logarithm transform of these two variables
( lnJ , lnV ) portrays a linear relationship.In the case of pure PVK thin films (Fig. 3a), the slope
of this relationship equals 1, meaning that conduction is
of ohmic type. This pattern does not change when
considering other sets of measurements.
Contrarily, in the case of PVK thin films iodine
doped (Fig. 3b and c), if the relationship remains linear
it presents three dierent rates of change, allowing to
subdivide the plots in three sections according to the
electrical field E.
The rate of change observed in lower electrical field
(E < 7 104 Vm1) is of the same magnitude (slope 0:4) independently of the technique used. This value is
Fig. 1. C1s peak of (a) pure PVK (- - -) experimental curve, (b)
iodine post-doped PVK (- - -) theoretical curve and (c) chlorine
post-doped PVK ( ) dierent components.
Fig. 2. XPS iodine profile in PVK thin films: (a) pre-doped film;
(b) post-doped film.
G. Safoula et al. / European Polymer Journal 37 (2001) 843849 845
close to the value (0.5) reported in the literature [28,29],
which indicates space charge limited conduction.
In the high electrical field domain (E > 2:5105 Vm1), disparate results are obtained. PVK thinfilms iodine post-doped, in particular, have a non-ohmic
behaviour as portrayed in Fig. 3c (slope 1:4). Pre-doped films presented also a non-ohmic conduction at
the second measure (slope 1:6, Fig. 3b).Lastly, in section with intermediate electrical field,
7 104 < E < 105 V m1, the behaviour of PVK thinfilms iodine doped using these two techniques is as fol-
lows: curve in (Fig. 3a and b) shows a rate of change
equal to 1, while curve in Fig. 3c is characterised by
slope greater than 1 (1,3), meaning, that the pattern is
not of ohmic type.
Measurement replicates for these two treatments did
not show any variation of the evolution of lnJ f lnV , for post-doped thin films. However, for pre-doped thin films, the shift that occurs under high elec-
trical field is not reversible.
It should be reminded that dierent conduction
mechanisms of polymeric materials have been suggested
, and these are:
SchottkyRichardson emission consisting of an injec-tion of carriers from electrodes,
PooleFrenkel eect which is related to the trap ioni-sation applying an electrical field,
Space charge eect.
In this study specifically, the electrical conductivity in
relation to an applied electrical field, showed up three
sections of conduction:
At low electrical field (E < 7 104 Vm1) the elec-trical conductivity is governed by space charge as de-
scribed by J / E2 law.At intermediate electrical field (7 104 < E< 2:5
105 Vm1), after stabilisation, the conductivity is ohmic.At high electrical field (E > 2:5 105 Vm1) the
conductivity regime is dominated by electrical field ef-
In the following paragraph, we will discuss these
dierent regimes as a function of the mentioned con-
The eect of space charge is due to the dierence in
work function between the metal (electrode) and the
insulator (polymer). If the insulator does not contain
traps the work functions injected carriers remain free
and thus contributes to space charge current, in the same
way, as a non-connected diode. Whereas, the presence of
traps in the insulator reduces the current by trapping
In these two cases, the current density follows an
expression described by the Mott and Gurney law :
J a U2
where a 9=8ee0l in trapless insulator or a 9=8ee0lh with the presence of traps in the insulator, h:fraction of free carriers reported to trapped carriers, e0:vacuum permittivity, e: material permittivity (e 3 forPVK), l: mobility of carriers and d: films thickness (interelectrode distance in a longitudinal structure).
Fig. 4 represents the variation of the current density J
versus E2 of PVK thin films iodine doped, in the regionof low electrical field. Estimated mobility of the car-
riers from the line slopes, using trapless insulator, is
1:1 106 cm2 V1 s1. This value was found in severalstudies . In these studies the PVK films have been
obtained by spin-coating which means that the PVK has
a high molecular weight. In the present work, the films
been obtained by evaporation, which means that, as
discussed in others papers [24,3942] the chain length of
the polymer is far smaller. Therefore it can be concluded
from the above similarity of the carrier mobility value in
Fig. 3. Voltagecurrent characteristics lnJ lnV of PVK thinfilms obtained from: (a) pure PVK powder, (b) PVK obtained
from iodine doped powder (3 at.%) first measure (), secondmeasure (s) and (c) PVK film obtained from post-doping
846 G. Safoula et al. / European Polymer Journal 37 (2001) 843849
the PVK films, whatever their origin, confirms that the
carriers are moving, not along the polymer chain, but by
hopping between carbazole groups. This explains that
similar mobilities are obtained from samples with very
dierent molecular weight.
In high electrical field regimes, the experimental re-
sults are often interpreted by either the Schottky eect or
the PooleFrenkel eect. These two theories although,
dierent in their principles, result in expressions of the
same form: ln J / E1=2. These two theories are essen-tially dierent from each other by the following points:
Schottky eect is limited to the surface of electrodes, PooleFrenkel eect is a volume eect. It corre-
sponds to the thermal excitation of most carriers
(holes) of trap level to band valence with dropping
of coulombian barrier under the action of the applied
electrical field. The dropping DE0 of the barrier is ofthe form:
1=2 bPFE1=2 2
where bPF is a PooleFrenkel constant and e an ele-mentary charge.
The corresponding current density J is described by
the following relationship:
J A exp bPFE1=2 EakT
where A: analogous constant of density current, E: ap-
plied electrical fieldV/d, V: applied voltage, k: Boltz-mann constant, Ea: energy of activation, d: distance
between electrodes, e0: vacuum permittivity and e: ma-terial permittivity.
Schottky eect corresponds to the modification of the
potential barrier of the metalinsulator interface.
Emission of carriers is analogous to thermoionic emis-
sion, the applied field contributes to decrease extraction
work of metalinsulator.
The corresponding current density is described in this
J AT 2 exp bSE1=2 UkT
where A: SchottkyRichardson constant (A 120Acm2), U: work function of metal and bS: Schottkyconstant. It is calculated from this expression:
PooleFrenkel constant is twice as Schottkys. Thisratio is explained by the fact that in the first case, the
carrier is trapped, whereas in the second case, the carrier
has its image in the other side, once it crosses the barrier.
bPF 2bS 6
In PVK material particularly, the dielectric constant
taken to be equal to 3 [33,34] gave a theoretical value of
bPF 4:38 105 eV m1=2 V1=2.In practice, it is often very dicult to distinguish
between these two mechanisms of conduction solely, on
the basis of these curves lnJ f lnE2. A solution tothis problem consists, in using PooleFrenkel conduc-
tivity as a function of electrical field.
In high electrical field regimes, our structures were of
longitudinal type. Therefore PooleFrenkel eect should
be more dominating as compared, to the surface eect
(Schottky eect). In this case, the electrical conductivity
under high electrical field must be pursuing Frenkels law.
rPF r0 exp bPFE1=2 EakT
where r0: electrical conductivity corresponding to lowelectrical field.
Fig. 4. Density of current versus electrical field J f E2 of PVK thin films iodine doped obtained by: (a) pre-doping technique (3at.%. iodine in pure PVK powder) and (b) post-doping technique.
G. Safoula et al. / European Polymer Journal 37 (2001) 843849 847
The electrical conductivity of our samples was de-
duced from the following relationship:
rPF JE 8
where J: through film current density (Am2) and E:applied electrical field (V m1).
Fig. 5 represents the result obtained lnr f E1=2 .From this figure, two regions of distinguished conduc-
tion can be seen (lower and higher electrical field).
In region of high electrical field, the value of Poole
Frenkel coecients calculated from the line slopes is
bPF 4:74 105 eVm1=2 V1=2 (Fig. 5a) and bPF 4:53 105 eVm1=2 V1=2 (Fig. 5b) respectively for PVKthin film iodine pre and post-doped. These values con-
form to the theoretical value.
Therefore, here also, as in the case of spin-coated
PVK films [1,34,43], PooleFrenkel eect is the most
dominating conduction mechanism at high electrical
field for PVK films iodine doped with either technique,
however, in the case of thin films obtained from pre-
doped method, an application of intense electrical field is
necessary before the stabilisation mechanism, comes up.
Intense electrical field induced probably ionisation of
iodine, which provokes the behaviour observed. Subse-
quently, modification of the electrical state of iodine,
following the application of the electrical field has been
already reported .
The dierence noticed in structure and behaviour
between these two types of iodine thin films doped is
probably related to the dierence of iodine applied
quantity in these films. Post-doped films contained more
iodine and thus more ionised iodine in comparison to
the pre-doped films. We have already showed that a
ratio of I3 =I2 80%  exists in our doped films.
The dierence of the electrical behaviour can be ex-
plained by the variation in morphology and structure of
these films, as post-doped films have a better surface
state compared to the others . Post-doped thin films
(T 370 K) have a reticular structure . Similarlypre-doped thin films, which are electrically unstable,
evolve probably to reticular structure, following the
application of an intense electrical field.
Therefore, after stabilisation of the iodine doping,
whatever the sample is, the electrical behaviour of the
evaporated PVK films is similar to that of spin-coated
films. Since the molecular weight of the latter is far
smaller than that of the former, it can be concluded that
this parameter is not predominant in the case of PVK.
This result should be attributed to the fact that PVK is a
saturated polymer. In that case the conductivity process
does not take place along the polymer chains, but by
hopping from one chain to another one. More precisely
in the case of PVK from one carbazole group to another
one. The good agreement between the present electrical
study and that of spin-coated samples confirms that the
carbazole group are not destroyed during the evapora-
tion process, which corroborates earlier optical charac-
terisation of these films .
In the case of chlorine, its high electronegativity in-
duces a reaction between the dopant and the polymer
chain. Therefore no stable CT complex between chlorine
and PVK can be obtained, but chlorine compounds
form. The chlorine attacks systematically PVK films and
only iodine doped PVK samples could be extensively
studied by electrical measurements.
Fig. 5. Electrical conductivity versus electrical field ln r f E1=2 of PVK thin films iodine doped obtained by: (a) pre-dopingtechnique (3 at.% iodine in pure PVK powder) and (b) post-doping technique.
848 G. Safoula et al. / European Polymer Journal 37 (2001) 843849
In the case of iodine doping it is shown that the io-
dine concentration present in the films depends on the
doping process. The iodine concentration in the post-
doped films (2 at.%) is nearly one order of magnitude
higher than that present in the films obtained by evap-
oration of a pre-doped powder.
Therefore this study concerns mainly the electrical
behaviour of PVK thin films iodine pre and post-doped
under an electrical field. Post-doped PVK thin films
seemed to be more stable than those obtained from pre-
doped powder. The latter required an intense electrical
field treatment before stabilisation.
PVK thin films iodine doped present a conduction,
which is controlled either by space charge or by ohmic
conduction in both low and moderate field. Whereas the
conduction mechanism seems dominated by Poole
The formation of a charge transfer complex salt
between PVK and iodine contributes to increase of
conductivity by two and four order of magnitude re-
spectively for pre- and post-doped thin films. The in-
crease of conductivity is limited by saturated PVK
chains and probably by the localisation of carriers cre-
ated from atoms of carbazole group.
The present work shows that the electrical properties
of evaporated PVK films are similar to that of spin-
coated PVK films. The main dierence between these
two families of samples is the molecular weight of the
polymer. It is far higher in the case of spin-coated films
than in the case of evaporated films. This is explained by
the fact that conductivity processes in PVK are related
to the carbazole groups and not to the chain length of
the polymer. Therefore the present study confirms that
the carbazole group is preserved during the evaporation
 Pearson JM, Stolka M. Poly(N-vinylcarbazole) polymer
monographs, vol. 6. New York: Gordon and Breach, 1985.
 Chapoy LI, Munck DK. J Phys C 1983;3:644.
 Pai DM. J Chem Phys 1970;52:2285.
 Mort J. Phys Rev B 1972;5:3329.
 Nakasawa Y, Ito N, Hoshino K, Hanna JI, Kokado H.
Jpn Appl Phys 1991;30:1002.
 Nakasawa Y, Ito N, Hoshino K, Hanna JI, Kokado H.
Jpn Appl Phys 1992;31:305.
 Rencroft PJ, Takahashi K. J Non-Cryst Solids 1975;17:71.
 Chen C, Mort J. J Appl Phys 1972;43:1164.
 Hanna JI, Inoue E. Photo Sci Engng 1983;27:51.
 Al-Mohamad A, Soukieh M. Thin Solid Films 1995;
 Sicot L, Fiorini C, Nunzi JM, Raimond P, Sentein C.
Synth Metals 1999;102:991.
 Brabec CJ, Padinger F, Hummelen JC, Janssen RAJ,
Sariciftci NS. Synth Metals 1999;102:861.
 Brabec CJ, Padinger F, Sariciftci NS, Hummelen JC.
J Appl Phys 1999;85:6866.
 Gill WD. J Appl Phys 1992;43:50.
 Pfister G, Gritits CH. Phys Rev Lett 1978;40:659.
 Ikida M. J Phys Soc Jpn 1991;60:2021.
 Papez V. J Electro Anal Chem 1990;282:389.
 Stulik P, Biederman H, Slavinski D, Fejfar A, Chudacek I,
Mackus P. J Electron 1991;70(3):509.
 Napo K, Chand S, Bernede JC, Safoula G, Alimi K.
J Mater Sci 1992;27:6222.
 Bernede JC, Taoudi H, DAlmeida K, Delvalle MA, DiazF. Polymer thin films obtained from evaporated monomers
recent research developments in polymer science. Trans-
world Research Network, Trivandrum, India, 1998.
 Bernede JC, Alimi K, Safoula G. Polym Degrad Stab
 Safoula G, Alimi K, Touihri S, Bernede JC, Napo K.
J Chem Phys 1995;92:146.
 Alimi K, Safoula G, Bernede JC, Rabiller C. J Polym Sci
Part B: Polym Phys 1996;92:845.
 Touihri S, Safoula G, Bernede JC, Le Ny R, Alimi K. Thin
Solid Films 1997;304:16.
 Safoula G, Touihri S, Postic M, Bernede JC, Molinie Ph.
J Chem Phys 1997;94:1602.
 Shirley DA. Phys Rev B 1972;5:6219.
 Beanson G, Briggs M. High resolution XPS of organic
polymers the Scienta ESCA 300 Database. Chichester,
UK: Wiley, 1992.
 Mott NF, Gurney RW. Electronic processes in ionic
crystal. New York: Dover Publications, 1940.
 Mott NF, Davis EA. Electronic processes in non crystal-
line materials. Oxford: Clarendon Press, 1971.
 Chand S, Radhakrishanan S, Mechdru C. J Phys D: Appl
 Reuter R, Franke H. J Appl Phys B 1989;45:219.
 Ahmad A, Collins RA. Thin Solid Films 1992;21:75.
 Ikida M. J Phys Soc Jpn 1991;60:2031.
 Enachescu M, Dima I, Motorga V, Botila T, Ciurea ML,
Korony CH. Rev Roum Phys 1991;36:65.
 Regensburger JP. Photobiology 1968;8:429.
 Pearson JM. Pure Appl Chem 1977;49:463.
 Reimer B, Bassler H. Phys Stat Sol (a) 1979;51:445.
 Safoula G, Bernede JC, Alimi K, Molinie P, Touihri S.
J Appl Polym Sci 1996;64:1733.
 Blanchet GB, Fincher Jr CR. Adv Mater 1994;6:881.
 Ivanov VF, Nekrasov AA, Gribkova OL, Vannikov AV.
Synth Metals 1996;83:249.
 Tamada M, Asano M, Yoshida M, Kumakura M. Polymer
 Tamada M, Omichi H, Okui N. Thin Solid Films 1995;
 Drajeva II, Klein-Szymanska B. Comptes Rendus de
lAcademie Bulgare des Sciences 1984;37:35.
G. Safoula et al. / European Polymer Journal 37 (2001) 843849 849