Site-selective chemical-vapor-deposition of submicron-wide conducting
polypyrrole films: Morphological investigations with the scanning
electron and the atomic force microscope
F. Caciallia)and P. Bruschi
Dipartimento di Ingegneria dell’Informazione, Elettronica, Informatica, Telecomunicazioni,
Via Diotisalvi 2, 156126, Pisa (I), Italy
?Received 8 September 1995; accepted for publication 13 March 1996?
We report morphological investigations of polypyrrole thin films deposited by means of a
self-aligning vapor phase technique onto glass, silicon and silicon dioxide substrates, coated with an
oxidizing precursor. The variation of the deposition parameters allows the control of the film
microstructure which can be fibrillar and strongly anisotropic or globular and tendentially isotropic.
Patterning of the precursor by electron-beam lithography allows the production of submicron wide
lines as shown by both the scanning electron microscope and the atomic force microscope. © 1996
American Institute of Physics. ?S0021-8979?96?03612-2?
Conjugated polymers have received considerable atten-
tion over the last 20 years, for the novelty and the versatility
of the relative physical and electronic properties. As a result,
many applications have been proposed and realized, ranging
from conductive coatings for aeronautic purposes,1to thin
film transistors,2,3and, more recently, light-emitting diodes.4
Polypyrrole ?PPy? has been widely studied because of
the good long-term stability and the ease with which it is
possible to obtain high conductivities.5The polymerization is
thought to proceed through an oxidative reaction,6as for
other aromatic or heteroaromatic molecules such as benzene
Two main approaches are then possible: the chemical
oxidation of the monomer by a suitable precursor,7–11or the
electrochemical route.12,13Although the latter is the most
widely used, it suffers from intrinsic disadvantages and con-
straints, such as the need for a conducting substrate ?elec-
trode?, or the requirement of a minimum thickness, if the
film is to be removed and placed onto a non-conducting sub-
On the other hand, the chemical approach can be engi-
neered in a technique allowing a vapor phase deposition di-
The method relies on the oxidizing properties of an ap-
propriate agent produced, for example, by chlorination of a
metallic film deposited and patterned by the conventional
methods of the electronic technologies.10,14The subsequent
exposure of the metallic salts to the vapors of the monomer
leads to the polymerization on the sites of the precursor only
and gives a film which is also self-aligned with the precursor
pattern. The technique is sketched in Fig. 1. Other potential
advantages are the very high purity of the active material
?monomer?, which can be purified by distillation, without the
need for soluble electrolytes, and the compatibility with
other processes of integrated circuit manufacturing.
As an alternative to pure metal films, granular metal pre-
cursors based on polytetrafluoroethylene ?PTFE? and metals
such as copper ?Cu? or palladium ?Pd?, have been used in the
past14to improve the mechanical properties and prevent loss
of the pattern due to the hygroscopic nature of the salts.
Films prepared with this technique showed high conductivity
?up to 100 S/cm? and almost no metallic content remaining at
the end of the processing.
The sensing properties of such films, from pure and
granular metal precursors, have also been characterized in
terms of the relative variation of the resistivity, upon expo-
sure to water and other organic vapors ?methanol, ethanol,
propanol, buthanol, acetone?.15
More recently, investigations on the low-frequency re-
sistance fluctuations have been carried out both on unex-
posed films,16and after exposure to gas and vapors.17In
comparison with other disordered semiconductors18–20very
low 1/f noise levels have been measured on as-prepared
films, and we found reversible and selective changes of the
resistive fluctuations spectra induced by the exposure to or-
In order to characterize further the potential of the tech-
nique we have used electron-beam lithography to pattern
sputtered copper films which were then converted into the
oxidizing salts and used as precursors for the polymerization
In this paper we report the fabrication of polypyrrole
films with a minimum lateral dimension of about 800 nm.
This value compares favorably with those reported in the
literature for other techniques, such as the deposition by
means of the scanning electrochemical microscope.21
Morphological studies conducted by means of a scan-
ning electron microscope ?SEM? and by an atomic force mi-
croscope ?AFM? revealed that a very anisotropic fibrillar
structure is obtained when the precursor is kept in a humid
environment during the halogenation and the polymerization,
and then allowed to crystallize. This can be prevented by
controlling the deposition parameters as described later.
a?Author to whom correspondence should be addressed at: Cavendish Labo-
ratory, Madingley Road, Cambridge CB3 0HE, UK; Electronic mail:
70 J. Appl. Phys. 80 (1), 1 July 1996 0021-8979/96/80(1)/70/6/$10.00 © 1996 American Institute of Physics
Although both the AFM and the scanning tunneling mi-
manufacturing technologies, the AFM has the advantage
over the STM of being virtually insensitive to the local
work-function of the material. It is therefore the preferred
instrument for morphological characterizations.
Silicon and glass substrates were thoroughly cleaned by
three successive ultrasonic baths in trichloroethylene, ac-
etone and ethanol respectively, before drying in flowing ni-
Electron-beam patterning of the copper films was carried
out according to the procedure described in Ref. 22, with
only minor variations.
We used commercial
?PMMA? with a molecular weight of about 950 000 a.u. as
the positive resist, in conjunction with lift-off lithography to
reduce the exposure time to a minimum. The PMMA was
deposited by spin-coating so as to give 500 nm thick films
which were then exposed in an ISI-DS-130 SEM with an
electron beam operated at 30 keV. After developing, the
films were baked at 70 °C for 1 h.
With the use of a lift-off technique, the film adhesion to
the substrate becomes crucial, as the patterning requires the
adhesion to the substrate to be stronger than the cohesion of
the film along the pattern perimeter. Enhancement of the
surface adhesion strength has been achieved by means of a
mild plasma etch of the substrate, aimed to increase the sur-
face porosity, and with the use of rf sputtering instead of
Joule evaporation as the metal deposition technique.
The conversion of the metal films into the oxidizing
agent and the polymerization of the monomer ?pyrrole? were
conducted according to the techniques described in Refs. 10,
14, and 15 ?technique A? and to the upgraded version de-
scribed in Refs. 16 and 17 ?technique B?, with significant
differences in the resulting morphologies. In both cases we
used an electrochemical cell in order to produce the
Cl2/HCl/H2O vapors for the chlorination of the metal pattern,
but in technique A this reaction and the next polymerization
step were carried out at room temperature, whereas in tech-
nique B the vapors were heated up to ?200 °C during the
chlorination. The substrates were also heated up to tempera-
tures higher than 50 °C before being introduced into the re-
action chamber. This prevented water absorption by the
metal salts and the growth of CuCl2crystals which, if
present, act as a polymerization template for the monomer.
At the end of the chlorination step the substrates were slowly
cooled down to a temperature in the range of 50–80 °C, in a
dry environment. The polymerization was then carried out in
a chamber saturated with the monomer vapors, holding the
substrate at ?80 °C for the first 5 min and then lowering the
temperature to 35 °C for the rest of the polymerization time
?3 to 15 h?. A final rinse ?water or ethanol? is necessary in
order to remove the excess of the oxidizing agent.
Cooling down to room temperature and storage of the
samples in a dry environment after the end of the chlorina-
tion step is also effective in maintaining the effectiveness of
the precursor. The increase of the substrate temperature to
80 °C during the initial part of the polymerization is de-
signed to accelerate the reaction rate, but the temperature
then has to be reduced to ?35 °C in order to prevent en-
hanced oxidation of the polymer by the residual oxygen in
This upgraded version of the technique has proved effec-
tive also with gold ?Au?, palladium ?Pd?, and iron ?Fe?,
whose salts are very hygroscopic, whereas the polymeriza-
tion of the monomer did not take place when we used alu-
minium ?Al?, tin ?Sn?, lead ?Pb?, nickel ?Ni? or indium ?In? as
the metallic precursor.
SEM observations were made by means of the same mi-
croscope as that used for the electron-beam exposure, while
for the AFM experiments we used a Park SFM-BD2 micro-
scope. The response of the latter was calibrated along the
three axis, within ?5% accuracy, in the range of 144 nm to
1 ?m by using latex microspheres ?Interfacial Dynamics
Corporation? as a reference standard. No metallic coating
was applied to the polymers in order to prevent charging
effects and to improve the image quality in SEM observa-
Film thicknesses were measured with a stylus profilome-
In Figs. 2?a? and 2?b? we show, at different magnifica-
tions, the SEM lateral view of a PPy film on a silicon sub-
strate, produced with technique A. Fibrils with lateral dimen-
sions in the range of 0.5 to 2 ?m and lengths up to many tens
of microns are visible but maximum lengths up to 200 ?m
have also been observed. Both in Figs. 2?a? and 2?b? we note
bending of the rod-shaped structures, and fibril fusion is evi-
dent in some areas. At greater magnifications ?Fig. 2?b?? sev-
eral white spots appear dispersed on the fibril surface, and
we believe that they are residues of a non-conducting oxidiz-
ing agent, which have not been removed by the ethanol rinse.
FIG. 1. Schematical representation of the deposition technique: ?a? the sub-
strate; ?b? after the deposition of the metallic film; ?c? after the definition of
the geometries of the film ?lithography?; ?d? after the chlorination step; ?e?
after the exposure to the pyrrole vapors.
71J. Appl. Phys., Vol. 80, No. 1, 1 July 1996F. Cacialli and P. Bruschi
The use of technique B yields films with no appreciable
structure down to very small dimensions—of the order of
100 nm. In Fig. 3 we show SEM picture of one of these
films, which clearly shows granular morphology, but no evi-
dence for fibril formation as in the previous figures. No
structure was detected at lower magnifications. Although the
picture quality is affected by the relatively high magnifica-
tion and, possibly, by the lower film conductivity and capaci-
tance, the globular structures with characteristic diameters in
the range of 150–300 nm present similarities to those re-
ported in the literature for some of the electrochemically syn-
Purpose-built 4-probe test-patterns were employed to
measure the conductivity of the polypyrrole films, which was
found in the range of 1–10 ??1cm?1at room temperature.
The temperature dependence of the resistance of one of these
structures is reported in Fig. 4, as measured in the range of
?26.5–100 °C and shows a negative temperature coefficient,
suggesting a phonon-assisted conduction mechanism. This is
consistent with the results of previous studies on similar
films, made over larger temperature ranges.14,25
In Figs. 5 and 6 we show SEM pictures of the films
obtained by electron-beam patterning of a copper precursor
film ?40 nm thick. Each of the patterns in the lower corners
of Fig. 5 is composed of 6 lines with widths varying in range
from 4 ?m to 0.8 ?m. Figure 6 shows details of the thinnest
line in the lower left-hand corner of Fig. 5. Although the line
edges exhibit some irregularities, the metallic pattern is well
FIG. 2. SEM views of a film produced with technique A onto an oxidized
silicon substrate, at different enlargements. The white spots on the surface of
the fibers are thought to be residues of the oxidizing salt not removed by the
FIG. 3. SEM picture of a film produced by using the upgraded deposition
method ?technique B?. No structure is observable at higher magnifications.
No fiber formation is evident on the reported scale-length. The anisotropy is
FIG. 4. Temperature dependence of a polypyrrole film deposited with tech-
nique B, as measured by means of two- and four-contact technique. The
four-contact result has been multiplied by two to allow presentation on the
FIG. 5. SEM overview of the geometries in the case of electron-beam pat-
terning of the precursor.
72J. Appl. Phys., Vol. 80, No. 1, 1 July 1996 F. Cacialli and P. Bruschi
reproduced and the linewidth is less than 850 nm over the
whole segment. The lateral width of the Cu-precursor was
The AFM data shown in Fig. 7 have also been taken on
one of the thinnest lines of the patterns in the corners, and
show a lateral dimension which is consistent with that mea-
sured by the SEM, within the calibration error of the AFM.
The groove evident in the lower left-hand corner of Fig. 7 is
due to a dent in the substrate caused by an accident in han-
The height profile shows a thickness of the PPy films of
about 220 nm. This value is a factor of 2 or so lower than
that expected from the relation between the thickness of the
Cu-precursor and that of the PPy film as determined for films
with lateral dimensions of ?4 nm ?Fig. 8?. The roughness
values measured along the longitudinal dimension, at the
center of the PPy line are about 35 nm ?rms? and 30 nm
In Fig. 9?a? we also report the rendered three-
dimensional ?3D? image of the lower long horizontal line of
Fig. 5. The line is slightly larger than that in Fig. 7 and with
a tendency to have increased height with a larger width ?the
maximum measured height is ?250 nm; Figs. 9?b?, 9?c??.
The morphology of the films was always found to be too
corrugated to allow investigations on a molecular or even
atomic scale and all the attempts gave results impossible to
The adhesion of the films was found to be sufficiently
strong to prevent detachment and motion of the film under
the action of the probe, which was operated with a force in
the range of 1 to 10 nN.
The morphological studies we present here indicate a
strong dependence of the structure of the vapor-phase depos-
ited PPy films on the growth conditions.
The large anisotropy of the fibrillar formations, charac-
teristic of the deposition in humid environments at room
temperature ?technique A?, raises questions regarding both
the mechanism of formation and the possible control of the
phenomenon, aimed towards single fiber growth.
Observations of the samples before and after the poly-
merization step10show evidence that crystallization of the
chloride precursor is the most probable reason for the occur-
rence of the phenomenon. However it is not clear whether
the polymerization develops exclusively within the bound-
aries defined by the precursor crystals, or if the latter can
work as initiators or seeds of a ‘‘regular assembly’’ of pyr-
role molecules, which can then extend beyond the crystal
boundaries, at least for a certain range. Although we can
easily detect the presence of sharp fiber facets ?Fig. 2?b??,
pointing to the first interpretation, we observe that this is not
sufficient to rule out the PPy growth outside the crystal
boundary, and also that bending of the fibers could be an
indication pointing to the latter interpretation.
Other questions regarding the degree of intra-fiber order-
ing of the polypyrrole chains and the relation with the physi-
cal and electrical properties may be answered after develop-
ing a method to improve fiber alignment ?e.g., by stretching?,
or to produce and to test single fibers. The reduction of the
precursor dimensions up to levels comparable with those re-
ported here for the electron-beam patterned samples, could
FIG. 6. SEM picture of the thinnest line in the lower left hand corner of
FIG. 7. AFM pictures ?a,b? and height-profile ?c? of one of the thinnest lines
in a different corner of Fig. 5. The cross-section in ?c? is relative to the
position indicated by the black line in ?b?. The lateral dimension agrees with
the one measured by the SEM, within the allowed calibration error of the
73 J. Appl. Phys., Vol. 80, No. 1, 1 July 1996F. Cacialli and P. Bruschi
be a possible route to induce monocrystal growth and single
fiber formation. We note that our experiments were not
aimed at this, as we used technique B, which deliberately
prevents fiber formation ?Fig. 3?, but were instead concerned
with the demonstration of the realizability of submicron wide
continuous strips of polypyrrole.
Regarding the potential of the technique, we remark that
the dimensions reported here do not represent an intrinsic
resolution limit for the technique, but rather that of our litho-
graphic system, and the reduction of the precursor pattern
dimension should allow the reduction of the width of the
On the other hand, the morphology of the films appears
to be very disordered, from both AFM and SEM imaging,
and we suggest that this might arise from the uncontrolled
propagation of the polymerization reaction in the presence of
an excess of oxidizing salt. Blending of the oxidizing agent
with some inert material should then result in a more ordered
morphology and in a better reproduction of the precursor
We have described the use of a chemical vapor phase
deposition technique for the preparation of PPy thin films,
applicable to both insulating and conducting substrates. Thin
films of copper, deposited by rf sputtering were used to pro-
duce a salt precursor capable of inducing the site-selective
and self-aligned formation of the conductive PPy films on
the substrates. We have explored the possibility of reducing
the dimensions of the PPy films by using an electron-beam
patterning technique of the copper precursor and report a
minimum linewidth of about 800 nm. This resolution is not
intrinsic to the technique but is limited by the lithographic
system we used to pattern the initial copper films.
AFM observations of the same patterns confirm the SEM
results and provide an estimate for the height of the films in
the range of 220 to 250 nm ?5%, with a tendency to in-
crease with the lateral width.
We found that the morphology of the films is strongly
dependent on the growth conditions and we have shown that
it can be controlled by varying the temperature of the sub-
strates during the halogenation and the polymerization.
Professor A. Nannini, Dr. A. Yoffe, Dr. C. Cecconi, and
Dr. M. G. Harrison are gratefully acknowledged for useful
discussions and proof-reading the manuscript. F.C. acknowl-
edges economical and technical support from Technobiochip
1M. Kaiser, S. Roth, J. Ghosh, D. Bormann, and P. Bernier, State of the Art
of Conductivity of Conducting Polymers ?NEOME-Network of Excellence
on Organic Materials for Electronics, Brussels, 1995?, pp. 1–15.
2J. H. Burroughes, C. A. Jones, and R. H. Friend, Nature 335, 137 ?1988?.
3G. Horovitz, X. Z. Peng, D. Fichou, and F. Garnier, J. Appl. Phys. 67, 528
4J. H. Burroughes, D. D. C. Bradley, A. R. Brown, R. N. Marks, R. H.
Friend, P. L. Burn, and A. B. Holmes, Nature 347, 539 ?1990?.
5T. Hauber, D. Schmeisser, H. Naarmann, and W. Go ¨pel, Synth. Met. 67,
6A. F. Diaz and J. Bargon, in Handbook of Conducting Polymers, edited by
FIG. 8. The relation between the thickness of the PPy films and that of their
precursors as measured for pists with lateral dimensions of ?4 mm.
FIG. 9. AFM pictures ?a,b? and height-profile ?c? relative to the long lower
horizontal line in Fig. 5. The cross-section in ?c? is relative to the position
indicated by the black line in ?b?. The width is slightly larger than that in
Fig. 7 and the thickness is also greater.
74J. Appl. Phys., Vol. 80, No. 1, 1 July 1996F. Cacialli and P. Bruschi
T. A. Skotheim ?Marcel Dekker, New York, 1986?, p. 86. Download full-text
7A. Angeli and L. Lutri, Gazz. Chim. Ital. 50, 128 ?1920?.
8K. Yoshino, S. Hayashy, and R. Sugitomi, Jpn. J. Appl. Phys. 23, L899
9A. Mohammadi, M. A. Hasan, B. Liedberg, and I. Lundstro ¨m, Synth. Met.
14, 189 ?1986?.
10A. Nannini and G. Serra, J. Mol. Electron. 6, 81 ?1990?.
11A. Mohammadi, I. Lundstro ¨m, O. Ingaa ¨s, and W. R. Salaneck, Polymer
31, 395 ?1990?.
12A. Dall’Olio, Y. Dascola, V. Varacca, and V. Bocchi, C. R. Acad. Sci. C
267, 433 ?1968?.
13A. F. Diaz and J. Bargon, in Ref. 6, pp. 81–119.
14A. Nannini, F. Cacialli, and P. Bruschi, J. Mol. Electron. 7, 21 ?1991?.
15P. Bruschi, F. Cacialli, and A. Nannini, Sensors Actuators A 32, 313
16P. Bruschi, F. Cacialli, A. Nannini, and B. Neri, J. Appl. Phys. 76, 3640
17P. Bruschi, F. Cacialli, A. Nannini, and B. Neri, Sensors Actuators B 19,
18M. Celasco, A. Masoero, P. Mazzetti, and A. Stepanescu, Phys. Rev. B
17, 2564 ?1978?.
19B. Pellegrini, R. Saletti, P. Terreni, and M. Prudenziati, Phys. Rev. B 27,
20O. A. Skaldin and O. A. Selezneva, JETP Lett. 56, 31 ?1992?.
21C. Kranz, M. Ludwig, H. E. Gaub, and W. Schuhman, Adv. Mater. 7, 568
22G. H. Bernstein, D. A. Hill, and W. P. Liu, J. Appl. Phys. 71, 4066 ?1992?.
23P. N. Bartlett and S. K. Ling-Chung, Sensors Actuators 19, 144 ?1989?.
24B Tieke and W. Gabriel, Polymer 31, 20 ?1990?.
25F. Cacialli, Ph.D. thesis, University of Pisa, 1994, pp. 70–76.
75 J. Appl. Phys., Vol. 80, No. 1, 1 July 1996F. Cacialli and P. Bruschi