Surface & Coatings Technology 231 (2013) 71–76

Contents lists available at ScienceDirect

Surface & Coatings Technology

journal homepage: www.elsevier.com/locate/surfcoat

Fabrication and characterization of water-based polyurethane/polyaniline conducting

blend films
Cheng-Ho Chen ⁎, Yi-Ting Kan, Ching-Fong Mao, Wei-Tung Liao, Ching-Dong Hsieh
Department of Chemical and Materials Engineering, Southern Taiwan University, Tainan City, Taiwan

a r t i c l e i n f o a b s t r a c t

Available online 27 March 2012 This study describes the fabrication and properties of water-based polyurethane/polyaniline (WPU/PANDB)
conducting blend films. Conducting polyaniline doped by dodecyl benzene sulfonic acid (DBSA) is prepared
Keywords: by using chemical oxidation polymerization. Conducting polyaniline doped by DBSA (PANDB) is then blended
Polyaniline with WPU dispersion to form WPU/PANDB conducting blend coating. The conducting blend coating is then
Dodecyl benzene sulfonic acid applied on a poly (ethylene terephthalate) (PET) film and dried to form WPU/PANDB conducting blend film.
Water-based polyurethane
Next, the thermal, chemical, morphological, and electrical properties of WPU/PANDB conducting blend films
Conducting blend film
are discussed. Electrical analysis results indicate that electrical conductivity of the WPU/PANDB blend film
increases when increasing the PANDB content. Additionally, according to attenuated total reflectance-
Fourier transform infrared (ATR-FTIR) analysis, hydrogen bonding forms between the \N\H of PANDB and
the C_O of WPU. Meanwhile, in view of the onset degradation temperature in TGA thermographs, the
thermal stabilities of all WPU/PANDB conducting blend films increase when increasing the PANDB content.
The resulting WPU/PANDB conducting films can be used as anticorrosive, antistatic, electrostatic discharge
(ESD), and electromagnetic interference (EMI) shielding materials.
© 2012 Elsevier B.V. All rights reserved.

1. Introduction the two components, i.e. the electrical conductivity of PANI with the
physical and mechanical properties of the polymeric matrix [7,8].
Intrinsically conductive polymers (ICPs) have become an efficient In this study, water-based polyurethane (WPU) dispersions are
alternative to inorganic conductors in many practical applications in adopted as the polymeric matrix for ES type PANI. This study fabri-
the recent decade [1]. Since the discovery of conducting properties cates WPU/PANDB conducting blend coatings. WPU dispersions are
of polyaniline (PANI) in the early 1980s, PANI has been an important chosen as the polymeric matrix of the conducting blend coatings due
member in the ICP family, owing to its ease of preparation, excellent to their versatility and wide range of excellent properties, including
environmental stability, various forms, interchangeable oxidation abrasion resistance, impact strength and low temperature flexibility.
states, versatility and economic cost, PANI has thus been extensively The coatings are versatile, safely handled and environmentally friendly
studied and applied in diverse areas such as secondary batteries, coating materials applied to many different substrates. WPU dispersions
biosensors, corrosion protections, anti-static packaging materials, have made them an effective replacement for solvent-based coatings.
and conducting coating [2–5]. In this study, n-dodecyl benzene sulfonic acid (DBSA) is adopted
MacDiarmid et al. elucidated the chemical structure of PANI, as the dopant for fabricating ES type PANI since DBSA is both a proton
as shown in Fig. 1 [6]. The conducting form of PANI can be easily acid and a surfactant. Therefore, PANI doped by DBSA (PANDB) can be
prepared by doping the emeraldine base (EB), which has a half effectively dispersed in the WPU matrix. Additionally, PANDB has
oxidized backbone (Fig. 1c), with a proton acid resulting in the many H-donators and H-acceptors in the backbones. PANDB can
emeraldine salt (ES) type PANI (Fig. 1b). However, the ES type PANI thus create various H-bonding types with WPU [9–11].
is intractable. Owing to the poor mechanical properties of ES type Despite the extensive attention paid to PU/PANI conducting
PANI, ES type PANI must be blended with a polymeric matrix for blends [8,11–21], WPU/PANDB conducting blend has seldom been
practical applications. One of the methods to process ES type PANI studied. This study attempts to integrate the desired properties of
without altering the polymer structure is by blending it with conven- two components, i.e. the electrical conductivity of PANDB with the
tional polymers. Such blends may combine the desired properties of physical and mechanical properties of the WPU to form WPU/PANDB
conducting blend coatings. Conducting blend coatings are then cast
onto polyethylene terephthalate (PET) substrates and dried to form
⁎ Corresponding author. WPU/PANDB conducting blend films. Additionally, the morphological,
E-mail address: [email protected] (C.-H. Chen). chemical, thermal and electrical properties of WPU/PANDB conducting

0257-8972/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
doi:10.1016/j.surfcoat.2012.03.056
72 C.-H. Chen et al. / Surface & Coatings Technology 231 (2013) 71–76

Fig. 1. Four forms of PANI: (a) leucoemeraldine base (LEB), (b) metallic emeraldine salt (ES), (c) emeraldine base (EB), and (d) pernigraniline base (PNB).

blend films are studied. The resulting WPU/PANDB conducting blend 2.4. Characterization
can be applied as corrosion protection for metals, anti-static packaging
materials and conducting coatings for electronic products. 2.4.1. Transmission electron microscopy (TEM) analysis
The synthesized PANDB for TEM (JEM-1230, JEOL Ltd., Japan) analysis
2. Experimental was prepared by casting diluted dispersion of PANDB on a carbon coated
copper grid. The TEM acceleration voltage was set at 80 kV.
2.1. Materials
2.4.2. UV–vis analysis
Aniline, dodecyl benzene sulfonic acid (DBSA), and ammonium The synthesized PANDB was dissolved in chloroform at room
persulfate (APS) were purchased from Merck, Germany. Nonionic temperature. Since the WPU cannot be dissolved in chloroform, the
WPU dispersion, tackiness and curing agents for WPU were supplied WPU and WPU/PANDB solutions were prepared in aqueous phase.
by Hailsun Chemical Co., Ltd. For WPU/PANDB solutions, the PANDB was well dispersed in the
WPU aqueous solution. Optical absorbances of the WPU, WPU/PANDB,
and PANDB solutions in the wavelength range of 300–900 nm were
2.2. Preparation of polyaniline doped by dodecyl benzene sulfonic acid
determined using an ultraviolet–visible spectrophotometer (UV–vis;
(PANDB)
Shimadzu, model UV-2401 PC).
PANDB was directly prepared by chemical oxidation with ammonium
persulfate based on the modified procedure described by Cao et al. [22] 2.4.3. Attenuated total reflectance-Fourier-transform infrared (ATR-FTIR)
and our previous study [23]. analysis
8 g of aniline was mixed with 23 g of DBSA and 400 ml of distillation The chemical structures of pure WPU film, WPU/PANDB conducting
water to form a homogeneous milky white dispersion of anilinium– blend films, and pure PANDB were identified using an attenuated
DBSA complex in a 1000 ml four-necked flat bottom reactor under total reflectance-Fourier transform infrared (ATR-FTIR) spectrometer
proper stirring at room temperature. The APS solution (20 g APS (Spectrum One; Perkin Elmer, USA). The wavenumber ranged from
dissolved in 200 ml distillation water) was then added slowly into the 400 to 4000 cm− 1, and the scanning rate was 32/s.
reactor. The mole ratio of aniline:DBSA:APS is 1:0.8:1. After 24 h, the
darkish green PANDB dispersion was precipitated with 600 ml acetone. 2.4.4. Optical microscope examination
The reaction mixture was then filtered and washed with deionized The morphologies of pure WPU film and WPU/PANDB conducting
water several times until the filtrate was colorless. Finally, the product blend films were examined using an optical microscope (model
was collected and dried in an oven at 60 °C for 24 h. Olympus BH2-UMA, Japan).

2.3. Preparation of WPU/PANDB conducting blend films 2.4.5. Electrical conductivity analysis
The surface resistance (Ω/sq) of pure WPU film and WPU/PANDB
WPU dispersion was placed into a beaker equipped with a me- conducting blend films was examined at room temperature by using
chanical stirrer. Various amounts of the synthesized PANDB were a high-impedance meter (Agilent 4339B).
added to the WPU dispersion. The WPU/PANDB mixture was then
thoroughly mixed. Next, a tackiness agent was added to the mixture 2.4.6. TGA analysis
under mechanical stirring until the mixture viscosity reached around The thermal stabilities of pure WPU film and WPU/PANDB
7500 cps. A curing agent of 1.5 g was added to the mixture sequen- conducting blend films were examined using a thermogravimetric
tially and well completely mixed. The resulting mixture was degassed analyzer (TGA; Perkin Elmer model TGA 7). To eliminate the moisture
with a vacuum pump to eliminate air bubbles. Finally, the air bubble interference, the sample was heated from 50 °C to 120 °C, held for
free mixture was cast on a polyethylene terephthalate (PET) substrate 2 min, and then reheated from 50 to 700 °C at a heating rate of
and heated to 120 °C for 2 min, then 150 °C for 2 min to complete the 10 °C/min under a nitrogen atmosphere. Finally, all weight loss stages
curing reaction of WPU. were analyzed.
 C.-H. Chen et al. / Surface & Coatings Technology 231 (2013) 71–76 73

3. Results and discussion 355 nm

430 nm
1.2
850 nm
Fig. 2 shows the TEM image of the synthesized PANDB by chemical (e)
oxidation polymerization of aniline at room temperature. Notably, 1.0
the morphology of synthesized PANDB exhibits an interconnected

Absorbance (a.u.)
and branched network-like fibrous structure with fiber diameters 0.8
of 150–250 nm. The average diameter of PANDB is approximately
200 nm. 0.6
Fig. 3 displays UV–vis absorption spectra of WPU, WPU/PANDB, (d)
and PANDB solutions. There is no absorbance peak in UV–vis
0.4 (c)
spectrum of the WPU aqueous solution (Fig. 3a). Three characteristic
absorption peaks significantly appear in UV–vis spectrum of the
PANDB chloroform solution, at around 355, 430, and 850 nm 0.2 (b)
(Fig. 3e). The absorption peak appearing at around 355 nm can be (a)
ascribed to π–π* transition of the benzenoid rings, while the peaks 0.0
appearing at around 430 and 850 nm can be attributed to polaron–π* 300 400 500 600 700 800 900
transition and π–polaron transition, respectively [24]. Analysis results Wavelength (nm)
indicate that the synthesized PANDB is emeraldine salt (ES) type.
Three characteristic absorption peaks also appear in WPU/PANDB Fig. 3. UV–vis absorption spectra of (a) WPU, (b) WPU/9 wt.%PANDB, (c) WPU/17 wt.%
PANDB, (d) WPU/33 wt.%PANDB, and (e) PANDB solutions.
aqueous solutions due to the existence of PANDB (Fig. 3b–d). Moreover,
the intensities of characteristic absorption peaks of WPU/PANDB
aqueous solutions are increased with the content of PANDB. In Fig. 3, of \CH2\ group, and _C\O\ bending vibration in the PU main
it can be found that the intensities of characteristic absorption peaks chain, respectively [25].
of the PANDB chloroform solution are higher than those of WPU/ For pure PANDB (Fig. 4e), the characteristic peaks appearing
PANDB aqueous solutions. This is because PANDB can be dissolved at 1596 and 1495 cm − 1 are due to the stretching vibrations of
in chloroform. However, PANDB can only be dispersed in the WPU N_Q_N ring and N\B\N ring, respectively. The characteristic
aqueous solution. peak appearing at 1300 cm− 1 is attributed to C\N stretching vibration
Fig. 4 shows the ATR-FTIR spectra of a pure WPU film, WPU/PANDB of the secondary amine in the PANDB main chain. The peaks appearing
conducting blend films, and pure PANDB. For a pure WPU film, the at 998–1040 cm− 1 are due to the asymmetric and symmetric O_S_O
characteristic peak appearing at 3350 cm − 1 is due to the stretching stretching vibration of DBSA. The characteristic peaks appearing at
vibration of \N\H bond in the urethane segments (Fig. 4a). The char- 1100–1200 cm− 1 are due to the B\NH\Q bond or B\NH\B bond,
acteristic peaks appearing at 2840–2950 cm − 1 are attributed to sym- and in-plane bending vibration of benzenoid or quinonoid C\H
metric and non symmetric stretching of the C–H bond with carbonyl. (where B represents benzenic-type rings and Q represents quinonic-
The characteristic peaks appearing at 1700–1735 and 1687 cm − 1 are type rings). Since these absorption peaks overlap in the range of
due to the stretching vibration of C_O bonds of the urethane and 1000–1200 cm− 1, a broad peak appears. The peaks appearing at 800–
urea segments in the PU main chain, respectively. The characteristic 700 cm− 1 correlate with the characteristic feature of the B\NH\Q
peaks appearing at 1539, 1463 and 768 cm − 1 correlate with N\H bond or the B\NH\B bond, and out-of-plane bending vibration
bending vibration of the secondary amine, C\H bending vibration of benzenoid or quinonoid \C\H and \N\H bonds. Additionally,
for the peaks appearing at 2900–2950 and 3267 cm− 1 are due to
the stretching vibration mode of the \C\H (\CH3 or \CH2\) and
\N\H bonds, respectively. These characteristic peaks correlate well
with those in literatures [26,27].
For WPU/PANDB conducting blend films (Fig. 4b–d), with an
increasing PANDB loading, the peak appearing at 1730 cm − 1, which
is attributed to \C_O stretching vibration mode, is shifted to

(e)
(d)
Transmittance (a.u.)

3240
1725 (c)
(b)
(a)
3348

1730

4000 3500 3000 2500 2000 1500 1000

Wavenumber (cm-1)

Fig. 4. The ATR-FTIR spectra of (a) pure WPU film, (b) WPU/9 wt.% PANDB conducting
blend film, (c) WPU/17 wt.% PANDB conducting blend film, (d) WPU/33 wt.% PANDB
Fig. 2. TEM image of pure PANDB. conducting blend film, and (e) pure PANDB.
74 C.-H. Chen et al. / Surface & Coatings Technology 231 (2013) 71–76

WPU carbonyl are assigned to the effect of nearby PANDB molecules

100
to some extent [12].
Fig. 5 shows the TGA thermographs of pure WPU film, WPU/
80 PANDB conducting blend films, and pure PANDB in nitrogen. For
pure PANDB, there are three thermal degradation stages. The first
Weight (%)

60 weight loss at 100–250 °C was resulted from the evaporation of mois-

ture and free DBSA molecules. The second weight loss at 250–350 °C
was resulted from the evaporation and thermal degradation of the
40 e
DBSA molecules doped on the polyaniline backbone. The third weight
c loss starting at 450 °C represented the thermal degradation of the
20 polyaniline molecule. According to Fig. 5a–d, two steps of thermal
d
a
b degradation are observed for pure WPU film and WPU/PANDB
0 conducting blend films. For a pure WPU film, the first thermal degra-
dation, 200–400 °C, can be attributed to the degradation of hard
100 200 300 400 500 600 700 800 segment of WPU. The second thermal degradation, 400–600 °C, repre-
Temperature (oC) sents degradation of the soft segment of WPU [20]. For WPU/PANDB
blend films, the first thermal degradation step can be attributed
Fig. 5. TGA curves of (a) pure WPU film, (b) WPU/9 wt.%PANDB, (c) WPU/17 wt.% to degradation of the hard segment of WPU and DBSA of PANDB. The
PANDB, (d) WPU/33 wt.%PANDB conducting blend films, and (e) pure PANDB. second thermal degradation step is associated with complete degra-
dation of the soft segment of WPU and the main chain of PANDB.
1725 cm − 1, if PANDB content is increased to 33 wt.%. This slightly Owing to the onset degradation temperature observed in Fig. 5a–d,
shift is due to the hydrogen bonding formed between the \N\H the thermal stabilities of all WPU/PANDB blend films increase when
of PANDB and the \C_O of WPU. Moreover, the peak appearing increasing the PANDB content. This phenomenon is due to the interac-
at 3348 cm − 1, which is attributed to the stretching vibration of tion between WPU and PANDB molecules. Moreover, the weight loss
\N\H bond in the urethane segments, exhibits a more significant of the second thermal degradation also decreases when increasing
shift than that of the pure WPU film. The characteristic peak of the PANDB content. However, the WPU/33 wt.%PANDB conducting
\N\H is shifted from 3348 cm − 1 to lower frequencies: 3296, 3278 blend film has the largest weight loss in the first thermal degradation.
and 3240 cm − 1, for WPU/9 wt.%PANDB, WPU/17 wt.%PANDB, and This is because the amount of free DBSA molecules is increased with
WPU/33 wt.%PANDB blend films, respectively. This finding suggests the content of PANDB. The other possible reason is the reaction be-
that a certain degree of mixing is achieved and also that the carbonyl tween WPU and curing agent is interrupted as the content of PANDB
group of the WPU interacts more strongly with that of the \N\H is increased.
group of the PANDB than with the \N\H group of the parent mole- Fig. 6a–d shows the optical microscopic images of the pure WPU
cule. Obviously, not all of the \N\H groups are influenced by film and WPU/PANDB blend films. A pure WPU film is semitransparent
the other components, since the WPU is preponderant in the blend and has a milky white appearance (Fig. 6a). For WPU/PANDB blend
system. The broadness of the characteristic peak of \N\H implies films, the optical microscopic images reveal that PANDB can be
diverse environments. Moreover, changes in the absorption of the adequately dispersed and with an inter-connected formation in the

Fig. 6. Optical micrographs (200× magnification) of (a) pure WPU film, (b) WPU/9 wt.% PANDB, (c) WPU/17wt.%PANDB, and (d) WPU/33 wt.%PANDB conducting blend films.
 C.-H. Chen et al. / Surface & Coatings Technology 231 (2013) 71–76 75

blend film depends on the amount of PANDB added to WPU matrix.

A higher PANDB content blended with WPU implies a darker
appearance of WPU/PANDB blend film. Although PANDB can be
adequately dispersed in WPU, it still can be observed in that dark
green area and light green area in Fig. 6b–d. It implies that a phase
separation occurs in the WPU/PANDB blending films.
Fig. 7 illustrates how the PANDB content influences the surface
resistance of WPU/PANDB conducting blend film. The surface resis-
tance of WPU/PANDB conducting blend film is decreased when
increasing the PANDB content. This phenomenon suggests that the
electrical conductivity of WPU/PANDB conducting blend film is in-
creased when increasing the PANDB content. The surface resistances
of pure WPU film, WPU/9 wt.%PANDB, WPU/17 wt.%PANDB, and
WPU/33 wt.%PANDB blend films are around 6.0 × 10 10, 1.1 × 10 10,
1.7 × 10 7, and 1.5 × 10 5 Ω/sq., respectively. This is due to the forma-
tion of more conducting paths through the WPU/PANDB blend film
containing higher PANDB content (Fig. 6b–d).
Fig. 7. Electrical conductivities of WPU/PANDB conducting blend films. Fig. 8 shows the antistatic comparison between the pure WPU
film (Fig. 8a) and the WPU/17 wt.%PANDB conducting blend film
WPU matrix (Fig. 6b–d). WPU/9 wt.%PANDB blend film is less semi- (Fig. 8b). The two films are abraded, and then placed among foamed
transparent than the pure WPU film. Additionally, its color is green polystyrene spheres. Clearly, the spheres are adsorbed on the pure
(Fig. 6b), due to PANDB with a rich green color. If the PANDB content WPU film, yet are not found on the WPU/17 wt.%PANDB conducting
is increased to 17 wt.%, the WPU/17 wt.%PANDB blend film has a blend film. This phenomenon indicates that the WPU/17 wt.%PANDB
darkish green color; in addition, it is not transparent (Fig. 6c). If conducting blend film can function as an electrostatic discharge.
the PANDB content is increased to 33 wt.%, the WPU/33 wt.%PANDB
blend film also has a darkish green color. Obviously, it is almost not 4. Conclusions
transparent (Fig. 6d). We conclude that the color of the WPU/PANDB
Conducting WPU/PANDB blend coatings and their films are
prepared in this study. PANDB can be well dispersed in the WPU
dispersion. Significant hydrogen bonding is formed between the
\N\H bond of PANDB and the \C_O bond of WPU. Both electrical
conductivities and thermal stabilities of WPU/PANDB conducting
blend films are increased when increasing the PANDB content. Since
WPU coatings have made them an effective replacement for the
solvent-based coatings due to environmental legislations and techno-
logical advances, conducting WPU/PANDB blend coatings fabricated
in this study are promising for anticorrosive, antistatic, electrostatic
discharge (ESD), and electromagnetic interference (EMI) shielding
applications.

Acknowledgments

The authors would like to thank the National Science Council of the
Republic of China, Taiwan, for financially supporting this research under
Contract No. NSC 95-2221-E-218-032-MY3. Ted Knoy is appreciated
for his editorial assistance.

References
[1] E. Hrehorova, V.N. Bliznyuk, A.A. Pud, V.V. Shevchenko, K.Y. Fatyeyeva, Polymer
48 (2007) 4429.
[2] C.H. Chen, C.F. Mao, S.F. Su, Y.Y. Fahn, J. Appl. Polym. Sci. 103 (2007) 3415.
[3] S. Sathiyanarayanan, V. Karpakam, K. Kamaraj, S. Muthukrishnan, G. Venkatachari,
Surf. Coat. Technol. 204 (2010) 1426.
[4] E. Armelin, Á. Meneguzzi, C.A. Ferreira, C. Alemán, Surf. Coat. Technol. 203 (2009)
3763.
[5] E.C. Gomes, M.A.S. Oliveira, Surf. Coat. Technol. 205 (2011) 2857.
[6] J.P. Pouget, M. Laridjani, M.E. Jozefowicz, A.J. Epstein, E.M. Scherr, A.G. Macdiarmid,
Synth. Met. 51 (1992) 95.
[7] M. Ilberman, G. Titelman, A. Siepann, Y. Haba, M. Narkis, D. Alperstein, J. Appl.
Polym. Sci. 66 (1997) 243.
[8] K. Lakshmi, H. John, K.T. Mathew, R. Joseph, K.E. George, Acta Mater. 57 (2009)
371.
[9] T. Vikki, L.O. Pietilac, H. Osterholm, L. Ahjopalo, A. Takala, A. Toivo, K. Levon,
P. Passiniemi, O. Ikkala, Macromolecules 29 (1996) 2945.
[10] O.T. Ikkala, L.O. Pietilac, P. Passiniemi, T. Vikki, H. Osterholm, L. Ahjopaloc, J.E.
Osterholm, Synth. Met. 84 (1997) 55.
[11] K.S. Ho, K.H. Hsieh, S.K. Huang, T.H. Hsieh, Synth. Met. 107 (1999) 65.
[12] P.C. Rodriguesa, L. Akcelrud, Polymer 44 (2003) 6891.
Fig. 8. Antistatic tests: (a) pure WPU film, and (b) WPU/17 wt.%PANDB conducting [13] I. Sapurina, J. Stejskal, M. Špírková, J. Kotek, J. Prokeš, Synth. Met. 151 (2005) 93.
blend film. [14] H. Yoshikawa, T. Hino, N. Kuramoto, Synth. Met. 156 (2006) 1187.
76 C.-H. Chen et al. / Surface & Coatings Technology 231 (2013) 71–76

[15] J. Luo, X. Wang, J. Li, X. Zhao, F. Wang, Polymer 48 (2007) 4368. [21] P.K. Prabhakar, S. Raj, P.R. Anuradha, S.N. Sawant, M. Doble, Colloids Surf. B:
[16] C.V. Bouanga, K. Fatyeyeva, P.Y. Baillif, C. Khaokong, J.F. Pilard, M. Tabellout, Biointerfaces 86 (2011) 146.
Macromol. Symp. 290 (2010) 175. [22] Y. Cao, A. Andreatta, A.J. Heeger, P. Smith, Polymer 30 (1989) 2305.
[17] V. Aboutanos, L.A.P. Kane-Maguire, G.G. Wallace, Synth. Met. 114 (2000) 313. [23] C.H. Chen, J. Polym. Res. 9 (2002) 195.
[18] S.M. Yang, S.D. Chow, D.S. Lin, Synth. Met. 121 (2001) 1305. [24] T. Abdiryim, X.G. Zhang, R. Jamal, Mater. Chem. Phys. 90 (2005) 367.
[19] H.L. Frisch, J. Ma, H. Song, K.C. Frisch, P.C. Mengnjoh, A.H. Molla, J. Appl. Polym. Sci. [25] W.C. Chan, S.A. Chen, Polymer 11 (1988) 1995.
80 (2001) 893. [26] Z. Zhang, M. Wan, Synth. Met. 128 (2002) 83.
[20] X.M. Rao, G.W. Xu, J.J. Bao, H.F. Zhou, L. Li, Paint Coat. Ind. 37 (2007) 23. [27] S. Zhou, T. Wu, J.Q. Kan, Eur. Polym. J. 43 (2007) 395.