2. ARTICLE IN PRESS
G Model
SYNMET-13632; No. of Pages 10
2 J.M. Kinyanjui et al. / Synthetic Metals xxx (2011) xxx–xxx
cost alternatives that can also enhance the electronic properties of the role of solvent composition in a co-solvent system of ace-
PI. tone/DMSO for the formation of conductive PI/Carbon composite
Traditionally, carbon materials have been dispersed into PI films is assessed. The electrodeposition of Pt at the PI/Carbon
using DMSO (dimethyl sulfoxide) and other common solvents to electrode interface using cyclic voltammetry is verified. The elec-
obtain electronically conductive PI/Carbon composites [16–18]. trochemical properties of Pt, deposited at the PI/Carbon interface,
The use of low-volatility solvents such as DMSO and NMP (n- are investigated using the ferricyanide redox couple and methanol
methylpyrrolidone) is problematic because the homogeneity of oxidation reaction. The electrochemical measurements confirm the
the PI/Carbon composite is influenced by dispersion of the sec- electrochemical activity of Pt metal deposited from the reduction
ondary component and trapped solvent in the resulting polymer of PtCl4 2− .
film [17]. Uniform dispersion of the carbon black is also critical
to ensure that the mechanical integrity of PI is maintained, with 2. Experimental
high aggregation of the secondary component producing brittle
materials [19]. Therefore, a uniform dispersion of carbon and the 2.1. Chemicals and solutions
PI precursor [polyamic acid (PAA)] in a suitable solvent prior to
imidization increases the homogeneity of the material [17,20–22]. Poly(pyromellitic dianhydride-co-4,4 -oxydianiline)amic acid
Moreover, enhanced electrical conductivity has been previously (PAA) solution (Aldrich, 15.0 wt.% ±5 wt.% in NMP/aromatic hydro-
attributed to the formation of small fibrous carbon black aggregates carbons, Cat.# 575828, trade name: Pyre-M.L.® RC-5057), dimethyl
with chain-like structures within the polymer matrix [19]. Such sulfoxide, DMSO (Aldrich, ≥99.9%, Cat.# 154938), conductex® SC
microstructures can be instrumental in the transport of electrons carbon black (Columbian Chemical company), potassium tetra-
through insulating polymers and strongly influenced by solvent chloroplatinate, K2 PtCl4 (Strem, 98%, Cat.# 78-1970), perchloric
composition. acid, HClO4 (J.T. Baker, 69–72%, Cat.# 9652-33), sulfuric acid, H2 SO4
The combination of PI with metals is not specifically tied to the (Mallinckrodt, 98%, Cat.# 2876), acetone (J.T. Baker, 99.5%, Cat.#
formation of conductive PI using carbon. In fact, PI/metal com- 9006-7), potassium ferricyanide, K3 Fe(CN)6 , Methanol, CH3 OH
posites have also been reported previously for microelectronic (VWR, 99.8%, EM-MX085-9). All materials were used as received.
photoresist and biosensor applications using surface deposition or
impregnation of secondary metals on/into the insulating polymer
2.2. Preparation of PI/Carbon composites
[23–25]. Surface modification of insulating PI has been achieved
using solution hydrolysis, photoirradiation, and chemical vapor
The carbon/acetone mixture was prepared by combining 5.45 g
deposition [21] for a variety of metal species, including Ag [26,27],
of conductex® SC carbon black with 100 ml acetone and stirring
Fe [28], Cu [27,29–31], Al [32,33], Au [27], Pt [34–36], and
for 24 h. The solution was then sonicated for approximately 3 h.
Pd [27,34,35]. The electrochemical deposition of metal on free-
54.5 g of polyamic acid (PAA) was dissolved in 20 ml of DMSO and
standing PI substrates has not been extensively studied due to the
combined with the carbon/acetone solution to achieve a 10 wt.%
low conductivity of the polymer substrate material. The electrode-
PAA/Carbon composite.
position of metal on thin PI films has thus been predominantly
The PI/Carbon substrates were obtained by applying a uniform
performed on conductive substrates, including Au and glassy car-
coating of the PAA/Carbon mixture on a glass substrate and allow-
bon [23,37,38]. The method utilizes the inherent conductivity of the
ing it to dry for 24 h. The glass substrate with the PAA/Carbon film
substrate rather than the polymer to facilitate the electrochemical
was inserted into a vacuum oven at a temperature of 80 ◦ C for a
deposition of metal. In addition, PI films are often grafted onto con-
period of one week to ensure the removal of residual solvent. Ther-
ductive substrates using methods such as photo-irradiation [39]
mal imidization of PAA to PI was utilized to achieve polymerization
or by dipping [37] the substrates into the polymer precursor that
[17]. A temperature range of 250–300 ◦ C is typically utilized for the
forms thin films after polymerization.
thermal imidization of PAA to form PI [22,42,43], and therefore a
The electrodeposition of PAA in a solution containing metal salt
final thermal imidization was conducted at 280 ◦ C for 30 min to
at an electrode surface, thereby creating PAA/metal hybrids, has
form the PI/Carbon composite. The typical film thickness for the
also been documented. For example, the incorporation of Au and
delaminated PI/Carbon substrates obtained was 150 m.
Ag nanoparticles into electrochemically deposited PAA has been
used to produce PAA/metal composites [40]. The method utilizes
the affinity of PAA to reduce the incorporated metal cations during 2.3. Electrochemical apparatus and conditions
the polymerization process at the electrode surface. Specifically,
when the carboxyl group of the PAA reacts with triethylamine, a Pt was electrochemically deposited on a free-standing PI/Carbon
polyamate salt is formed with a cation that readily exchanges for working electrode using cyclic voltammetry. Cyclic voltammetry
the metal cation. This allows the metal precursor to be dispersed was also used to probe the ferricyanide redox couple of the resulting
into the polymer matrix. The metal cation can be both thermally PI/Carbon/Pt composite films. All electrochemical measurements
and electrochemically reduced to metal during polymerization. were conducted using a CHI 760 potentiostat/galvanostat with
The utilization of a PI/Carbon/Pt composite for fuel cell applica- included software. The experiments were performed in a one-
tions has been recently explored for sulfonated PI as the polymer compartment, three-electrode cell. All potentials are referenced to
electrolyte film. A mixture of Pt in carbon black has been incorpo- a Ag/AgCl electrode (3 M KCl filling solution). As counter electrode,
rated into sulfonated PI, cast onto an electrode, and used to examine we used a 0.5 mm-thick platinum sheet with an area exceeding the
the oxygen reduction reaction (ORR) [38]. Additionally, the elec- immersed area of the working electrode by a factor of two.
trocatalytic properties of Pt-deposited polyimide/carbon nanotube
films for methanol and nitrite oxidation have also been studied [41]. 2.4. X-ray photoelectron spectroscopy (XPS)
These studies document the unique combination of metals, carbon,
and PI to produce synergistic properties for novel applications. XPS was employed to identify the oxidation state of platinum in
In our study, the synthesis of a “free-standing” PI/Carbon/Pt the PI/Carbon/Pt composites. XPS measurements were performed
composite electrode is demonstrated. The dispersion of carbon using a Specs PHOIBOS 150MCD electron analyzer and Mg K
into PAA is examined and the conductivity of the composites pro- excitation. The base pressure was in the 10−10 mbar range for all
duced using different carbon loading is evaluated. Specifically, measurements. The electron spectrometer was calibrated using
Please cite this article in press as: J.M. Kinyanjui, et al., Synthetic Met. (2011), doi:10.1016/j.synthmet.2011.08.046
3. ARTICLE IN PRESS
G Model
SYNMET-13632; No. of Pages 10
J.M. Kinyanjui et al. / Synthetic Metals xxx (2011) xxx–xxx 3
XPS and Auger line positions of different metals (Cu, Ag, and Au)
[44].
2.5. Thermal analysis of PI/Carbon/Pt films
Simultaneous thermal analysis (STA) was conducted using
a Netzsch STA449 C thermal analyzer. Simultaneous differential
scanning calorimetry (DSC) and thermal gravimetric analysis (TGA)
was performed on PAA, PI, PI/Carbon, and PI/Carbon/Pt under air
as both the purge (50 ml/min) and protective (20 ml/min) gas at a
heating rate of 10 ◦ C/min. Samples were placed in alumina pans and
covered with an alumina lid with a centered pin-hole. The furnace
was evacuated to about 10−3 torr prior to introducing purge and
protective gases.
Platinum loading in the PI/Carbon/Pt composites with a uniform
carbon loading of 10% was estimated using TGA. The composite was
heated to a temperature of about 900 ◦ C, at which all organic com-
ponents of the PI/Carbon composite were decomposed. All mass
loss was assigned to the decomposition of the organic compo-
nent. The remaining mass in the test crucible at the end of the
experiment was then used to estimate the metal content for the
composite material. For statistics, a minimum of three runs was
conducted.
2.6. Scanning electron microscopy (SEM)
SEM images of PI/Carbon/Pt composites were obtained using
a JEOL 5600 electron microscope equipped with a backscattered
electron (BSE) detector. The films were affixed to the sample
holder using carbon tape and measurements were performed at
an acceleration voltage of 15 kV. Deposition of a metallic Au layer
to enhance conductivity was not required for SEM measurements.
Fig. 1. Photoacoustic FTIR spectra of (a) PAA, (b) PAA/Carbon, (c) PI, and (d)
2.7. Electrical conductivity characterization PI/Carbon (600–1900 cm−1 ).
Electrical contacts to the PI/Carbon and PI/Carbon/Pt films were The FTIR spectra of PAA, PAA/Carbon, PI, and PI/Carbon in the spec-
made using a Cascade Microtech C4S-64/50 probe head with tung- tral range between 600 and 1900 cm−1 are presented in Fig. 1a–d,
sten carbide electrodes. The four-point probe sheet resistance of respectively. The asymmetric C O band at 1780 cm−1 is observed
each film was then measured at locations across the surface of for both PI and PI/Carbon and appears as a shoulder on the broad
the substrate using an Agilent 34401A Digital Multimeter con- C O symmetric stretch band centered at 1740 cm−1 . In addition,
nected through a Cascade Microtech CPS-05 probe station. Constant a strong sharp band located at around 729 cm−1 , assigned to the
pressure for each measurement was maintained for the probe imide C O bending mode, appears in both PI and PI/Carbon. The
head contacting the substrate using a mechanical stop. Aver- intensity decrease of the C–NH band for PAA at 1546 cm−1 and the
ages for a minimum of five measurements at each location are increase in band intensity for the C O imide functional group at
presented, together with representative standard deviations and 730 cm−1 and 1778 cm−1 confirm the formation of PI after thermal
relative standard deviations. Electrical resistivity values collected imidization. Similarly, the spectra can be monitored in the range
using these methods were converted to electrical conductivity for between 1900 cm−1 and 3700 cm−1 (shown in the supporting doc-
clarity. umentation) for changes in the carboxylic acid OH group of PAA. In
addition, characteristic secondary amide NH stretch bands occur as
2.8. Fourier transform infrared (FTIR) spectroscopy multiple bands in the same region between 3065 and 3300 cm−1
[45]. A broad decrease in band intensity for OH and NH is observed
All FTIR measurements were performed on pristine samples for PAA after thermal imidization and the formation of PI/Carbon
using a Digilab FTS-7000 spectrometer and a photoacoustic detec- composite.
tor (MTEC). Each sample was scanned 64 times with a resolution
setting of 4 cm−1 , and scans were averaged to produce each spec- 3.2. Carbon loading and conductivity of PI/Carbon
trum.
The factors that influence the conductivity of polymer/carbon
3. Results and discussion black composites include the properties of the carbon species
utilized, aggregation of the carbon within the polymer, and the
3.1. Preparation of PI/Carbon heterogeneity and dispersion of the carbon in the polymer matrix.
Variations in the density and particle size of the carbon play a crit-
The preparation of PI was accomplished using thermal imidiza- ical role because they influence both aggregation and dispersion
tion of PAA and carbon incorporated into the co-solvent system [17–19,46,47]. Increasing the density of the carbon filler allows
of acetone and DMSO. The polymerization of PI using this mixture more compact aggregation and packing, leading to enhanced elec-
is evaluated using photoacoustic FTIR spectroscopy in Fig. 1 [21]. trical conductivity at lower percolation thresholds (i.e., the measure
Please cite this article in press as: J.M. Kinyanjui, et al., Synthetic Met. (2011), doi:10.1016/j.synthmet.2011.08.046
4. ARTICLE IN PRESS
G Model
SYNMET-13632; No. of Pages 10
4 J.M. Kinyanjui et al. / Synthetic Metals xxx (2011) xxx–xxx
of the volume of a conducting phase required to change the mate- Table 1
Electronic properties of PI/Carbon.
rial from insulating to conducting). For example, highly crystallized
graphite (HCG) exhibits a higher conductivity and a lower percola- % Carbon Resistivity Conductivity % Relative standard
tion threshold than carbon blacks because of its higher density. In content (Ohm cm) (S/cm) deviation
addition, smaller particle sizes enhance conductivity because they 2.5% 2.1 × 105 4.7 × 10−6 8.7
allow higher aggregation and smaller gaps between the carbon cen- 5.0% 27 0.037 0.95
ters [17,19]. Significant changes in the conductivity are observed for 7.5% 3.4 0.29 0.32
10.0% 1.4 0.70 4.5
PI once a critical carbon content (percolation threshold) is reached
12.5% 0.80 1.2 0.10
[19,48]. Networks form after carbon reaches a critical mass, lead-
ing to coagulation and increased contact between the particles
[16,17,47]. The increase in conductivity is attributed to the forma- but the mechanical stability of the films is compromised at car-
tion of fibrous chain-like carbon networks in the polymer matrix bon loadings ≥12.5%, with the emergence of surface fractures and
and has been previously reported for PI/Carbon black composites decreased flexibility. Therefore, a carbon loading of 10% was utilized
[16,17]. for all subsequent measurements in our studies. The conductiv-
Previous studies of the influence of carbon loading on the elec- ity values between 8 and 16 wt.% carbon in PAA are significantly
trical conductivity have focused on dispersion using low-volatility higher when compared to previously reported values for PI films
solvents such as DMSO and NMP. These solvents are problematic containing the same carbon filler (SC conductex) used in this study
because they are often trapped in the polymer, leading to materi- [17].
als that are less homogeneous. In our study we evaluate the use of
a binary solvent, combining highly volatile acetone with DMSO to 3.3. Pt deposition on PI/Carbon
reduce the incorporation of solvent into the PI/Carbon composite.
In addition, the solvent is used to homogeneously disperse car- Deposition of Pt metal on freestanding PI/Carbon working
bon into the PAA prior to imidization to produce materials with electrodes was achieved using cyclic voltammetry encompassing
percolation thresholds at carbon loadings consistent with known potentials for the reduction of PtCl4 2− ion to Pt metal. The equation
PI/Carbon materials. and potential for the reduction of PtCl4 2− is as follows [49]:
The influence of the carbon content on the electrical conductiv-
ity of PI is presented in Fig. 2 for the PI/Carbon materials produced PtCl4 2− + 2e− → Pt(0) + 4Cl− , E ◦ = 0.493 V vs. Ag/AgCl
using the acetone/DMSO solvent system. The carbon content for The voltammetric response for a planar Pt reference and a
the composite is reported as the % carbon by weight in the polymer PI/Carbon electrode in solution containing 1 M HClO4 are provided
precursor, PAA. An approximate ∼104 increase in the conductivity in Fig. 3a and b (bottom), respectively. The Pt electrode shows a
is observed as the %Carbon is increased from 2.5% to 5%, as also voltammetric response at negative potentials, which is consistent
listed in Table 1. The data suggests that the percolation threshold is with proton adsorption at the metal surface. The response of the
achieved at carbon loadings between 2.5 and 5%. The conductivity PI/Carbon electrode is strongly diminished due to the absence of Pt
of the materials continuously improves above 5% carbon content, at the electrode surface. The electrochemical deposition of Pt from
solutions containing 20 mM PtCl4 2− in 1 M HClO4 are provided for
Pt and PI/Carbon electrodes in the center of Fig. 3. The voltammetric
response for both electrodes is very similar, with the characteristic
Fig. 3. (a) Cyclic voltammetry for a planar Pt electrode in 1 M HClO4 prior to depo-
sition (bottom), during Pt deposition in a solution containing 2.0 × 10−2 M PtCl4 2−
in 1 M HClO4 (center), and the same electrode after Pt deposition in solution con-
taining 1 M HClO4 (top). (b) Cyclic voltammetry for a PI/Carbon electrode in 1 M
HClO4 prior to Pt deposition (bottom), during Pt deposition in a solution containing
2.0 × 10−2 M PtCl4 2− in 1 M HClO4 (center), and the same electrode (PI/Carbon/Pt)
Fig. 2. Conductivity of PI/Carbon as a function of the percent carbon by weight in in solution containing 1 M HClO4 (top). The scan rate for all measurements was
PAA. 0.01 V/s.
Please cite this article in press as: J.M. Kinyanjui, et al., Synthetic Met. (2011), doi:10.1016/j.synthmet.2011.08.046
5. ARTICLE IN PRESS
G Model
SYNMET-13632; No. of Pages 10
J.M. Kinyanjui et al. / Synthetic Metals xxx (2011) xxx–xxx 5
electrochemistry associated with the Pt surfaces clearly visible. In contrast, the voltammetry associated with H adsorp-
Specifically, the formation and reduction of PtO is observed (A/A ). tion/desorption at the Pt metal deposited on each electrode is
In contrast, Pt reduction is not evident for either electrode (small significantly different. The reaction for the process is:
shoulder at potentials more negative than 0.20 V). Furthermore,
proton adsorption/desorption at the Pt metal surface is clearly vis- H+ + e− → H → Hads.
ible (B/B ) for both electrodes during this metal deposition process.
Finally, the voltammetric response of the Pt planar electrode and The potentials for the proton adsorption/desorption (B/B )
the PI/C/Pt composite electrode is evaluated in solution contain- for the Pt planar and PI/Carbon/Pt electrodes are observed at
ing 1 M HClO4 only (i.e., without the Pt precursor) to demonstrate approximately −0.07/−0.12 V and −0.02/−0.16 V, respectively,
that Pt deposition has been achieved at both electrodes. Variations with modest shifts in potential as a function of the number of cycles
in the Pt surface morphology likely influence the small variations employed. The peak-splitting for the Pt electrode is on the order of
observed in the voltammetry for the PI/Carbon and bulk Pt surfaces. 50 mV, while a value of 140 mV is observed for the PI/Carbon/Pt
However, the same voltammetric processes for the Pt surface are composite. In addition, the voltammetric waves associated with Pt
resolved at each substrate. Significant changes in the voltammetry deposition are broader relative to the planar Pt electrode. Proton
can be observed during Pt deposition for both electrodes. Specif- adsorption at a polycrystalline platinum electrode is a two-step
ically, each voltammetric cycle results in an increase in current process, involving weak adsorption followed by strong adsorp-
associated with an increase in the surface area of deposited Pt. The tion of hydrogen atoms at the Pt surface [54]. This process is
voltammetric waves associated with PtO (A/A ) [50,51] and Pt-Hads highly dependent on the crystal surface structure of Pt, with the
(B/B ) [50,52–54] increase with each deposition cycle. The reduc- voltammetric response determined by the number and orientation
tion of Pt (small shoulder at potentials more negative than 0.20 V) is of crystal planes of Pt that are exposed [53,55]. The presence of
observed for both electrodes during the Pt deposition and is absent defects and impurities can lead to disorder in the crystal structure of
when both electrodes are immersed in solutions containing only the electrode, further reducing the distinct adsorption/desorption
1 M HClO4 . The voltammetry is consistent with the deposition of voltammetry characteristics of different crystal planes of Pt. The
Pt through the reduction of the Pt metal precursor (PtCl4 2− ) at the data suggests that differences in Pt crystallinity may translate into
PI/Carbon electrode surface. different electrochemical responses for the oxidation/reduction of
More in-depth analysis of the voltammetry associated with PtO solution species.
(A/A ) provides a measure of the Pt surface chemistry and the
involved electrochemical processes. Specifically, the formation and 3.4. Scanning electron microscopy (SEM)
reduction of PtO is observed at ∼0.73 V and ∼0.60 V for both elec-
trodes. The voltammetry is consistent with the following reaction SEM is used to investigate the morphology of the PI/Carbon/Pt
where shifts in potential are common based on the electrode com- composite and the distribution of Pt particles on the surface. The
position [53]: SEM images of PI/Carbon/Pt electrodes with Pt loading of 6 and 30%
are shown in Fig. 4. The images show the immersed interface with
PtO(s) + 2e− + 2H+ Pt(s) + H2 O, E = 0.78 V vs. Ag/AgCl Pt deposits (left and/or top) and the electrode material that was not
Fig. 4. SEM images of (a) PI/Carbon/Pt composite with low Pt loading (6%, ×2000), (b) PI/Carbon/Pt composite with high Pt loading (31%, ×2000), (c) the latter at higher
magnification (×5000), and (d) PI/Carbon/Pt with high Pt loading at low magnification (×35).
Please cite this article in press as: J.M. Kinyanjui, et al., Synthetic Met. (2011), doi:10.1016/j.synthmet.2011.08.046
6. ARTICLE IN PRESS
G Model
SYNMET-13632; No. of Pages 10
6 J.M. Kinyanjui et al. / Synthetic Metals xxx (2011) xxx–xxx
immersed (i.e., no Pt deposited) for clarity. The Pt metal deposits In contrast, the observed asymmetric lineshape further indicates a
at 6% (Fig. 4a, ×2000) loading appear as evenly spaced, discrete metallic environment of the Pt atoms—it is well known that metallic
clusters with diameters of ∼400 nm on the PI/Carbon substrate. In systems allow a continuum of low-energy excitations in the final
contrast, the Pt deposits at 30% Pt loading in Fig. 4b (×2000) show state of the photoemission process, giving rise to a characteris-
a coalesced deposit of metal with no discernible clusters on the tic asymmetric shape (commonly described by a Doniach-Sunjic ˇ ´
PI/Carbon substrate. The Pt deposits (6% loading) suggest that car- profile). It is not necessarily correct to assume that several initial-
bon dispersion in PI is uniform for the acetone/DMSO co-solvent state species, described by symmetric Gaussian or Voigt functions
system. In fact, carbon aggregation at the PI surface is not discern- (including a metallic species), are needed to describe the spectrum
able on the substrates at higher magnification (×8000) in Fig. 4c. properly, as has been done in earlier publications of related sys-
Previous studies in literature that have used SEM to evaluate the tems [57]. Furthermore, local differences in final-state screening
morphology of PI/Carbon films have shown large (approximately of the created photo-hole might lead to small shifts of core level
60 m diameter) surface defects when prepared with carbon black. lines as well, which might be responsible for the small difference
The defects were attributed to the formation of water vapor pock- between the here-observed value and the literature reference for
ets produced during imidization [17]. Alternatively, these defects bulk metallic Pt.
could also be created by the fast expulsion of residual low-volatility
solvents as the temperature is increased to initiate imidization. The
3.6. Thermal gravimetric analysis (TGA) of PI/Carbon/Pt films
lack of large-scale defects observed in the PI/Carbon substrates pro-
duced in these studies at much lower magnification (×35) in Fig. 4d
The thermal properties of PI (a), PI/Carbon (b), and two
suggests that the temperature program was effective in expelling
PI/Carbon/Pt composites (c and d) were evaluated using ther-
water vapor and/or the binary solvent prior to imidization. Fur-
mal gravimetric analysis (TGA), as shown in Fig. 6. The thermal
thermore, the absence of defects in the PI/Carbon substrates may
properties of PI, PI/Carbon, and PI/Carbon/Pt can be evaluated at
be attributed to the minimization of low-volatility solvents in the
temperatures that result in the thermal degradation and loss of
preparation. The results demonstrate that Pt deposition at PI/C
mass associated with the polymer. The residual mass at 900 ◦ C pro-
composite electrodes can be controlled electrochemically to pro-
vides an estimate of residual carbon in the PI/Carbon sample and
duce deposits with variable thickness and morphology.
metal loading in the PI/Carbon/Pt composites. For example, the final
mass of PI with no carbon reaches a value of zero, consistent with
3.5. X-ray photoelectron spectroscopy of PI/Carbon/Pt the complete loss of polymer during the thermal processes shown
in Fig. 6a (note that the origin of the ordinate was offset for better
X-ray photoelectron spectroscopy (XPS) is used to examine the viewing). In contrast, the TGA of PI/Carbon shows a residual mass
oxidation state of the Pt metal deposited on PI. The XPS survey of 2% relative to PI (Fig. 6b). Similarly, Pt loadings of 4% and 26% are
spectrum of a representative PI/Carbon/Pt film is shown in Fig. 5. obtained for the PI/Carbon/Pt composites when the residual carbon
It confirms the presence of Pt, C, O, and N at the surface, in addi- content of 2% is subtracted from the residual mass shown in Figs. 6c
tion to some minor Cl, S, and Zn impurities. While the Cl and S and d, respectively.
impurities are likely due to the precursor chemicals, solvent, and/or
electrolyte, the Zn impurities might have been adsorbed during
transfer in the XPS ultra-high vacuum system. A detail scan of the
Pt 4f core levels is shown in the inset of Fig. 5. The 4f7/2 peak maxi-
mum is found at 71.35 eV (±0.05), which is in good agreement with
reference values for bulk Pt metal (71.1 or 71.2 eV) [56] and sug-
gests that the Pt atoms are indeed in a metallic environment. Most
relevant Pt-containing compounds (with non-zero oxidation state)
exhibit binding energies of 72.5 eV and above. In particular, PtO and
PtO2 are expected to be found at 74.2 eV and 75.0 eV, respectively
[56], and would also be expected to exhibit a symmetric lineshape.
Fig. 5. Mg K XPS survey spectrum of PI/Carbon/Pt composite (inset: Pt 4f detail Fig. 6. TGA curves of (a) PI, (b) PI/Carbon, (c) PI/Carbon/Pt (6% Pt loading), and (d)
spectrum). PI/Carbon/Pt (31% Pt loading), using a heating rate of 10 K/min.
Please cite this article in press as: J.M. Kinyanjui, et al., Synthetic Met. (2011), doi:10.1016/j.synthmet.2011.08.046
7. ARTICLE IN PRESS
G Model
SYNMET-13632; No. of Pages 10
J.M. Kinyanjui et al. / Synthetic Metals xxx (2011) xxx–xxx 7
The thermal analysis of the materials also provides an important Table 2
Electronic properties of PI/Carbon/Pt.
measure of the thermal stability of PI, PI/Carbon, and PI/Carbon/Pt
materials. In this case, the thermal properties of all four samples %Pt content Resistivity Conductivity %Relative standard
are very similar, with onset temperatures of 576 and 570 ◦ C for PI (Ohm cm) (S/cm) deviation
and PI/Carbon, respectively, and ∼560 ◦ C for the PI/Carbon/Pt com- 0.0% 7.0 0.14 3.6
posites. A small mass change attributed to the loss of low-volatile 4.6% 4.2 0.24 0.029
NMP and DMSO solvents can be observed for all materials prior to 12% 0.93 1.1 5.0
22% 0.15 6.7 3.5
the onset temperature, corresponding to a mass decrease of ∼2–4%.
30% 0.091 11 0.58
For comparison, the operating temperature range for Kapton® HN 36% 0.051 20 0.69
film over long periods of use (as reported by the manufacturer) is
−269–400 ◦ C [1], which is well below the temperatures reported
in our study. This is due to the fact that PI has a glass transi- is recorded up to a value of 36%, as shown in Fig. 7 and Table 2. The
tion reported at ∼400 ◦ C, where Kapton® HN becomes rigid [58]. data suggests that Pt loading should be increased to maximize con-
The onset temperatures for the initial mass loss of the materials ductivity; however, costs associated with the metal deposition also
produced in this study suggest that they can be utilized at high need to be considered and a reasonable value of Pt loading must be
temperatures (up to 550 ◦ C) without severe thermal degradation. defined based on the conductivity required. Thus the conductiv-
However, it may be prudent to work at temperatures below the ity requirements of the specific application will dictate a suitable
glass transition to ensure that the materials remain mechanically “compromise” for the Pt loading.
stable.
3.8. Electrochemistry at PI/Carbon/Pt composite electrodes
3.7. Conductivity of PI/Carbon/Pt composites
The conductivity of PI/Carbon and, in particular, of the
The electrical conductivity of PI/Carbon/Pt composites as a func- PI/Carbon/Pt composites suggests that the materials can be uti-
tion of platinum loading is shown in Fig. 7. The values should be lized as freestanding electrodes to probe electrochemical reactions.
compared to the conductivity of the PI/Carbon material at 10% car- Specifically, the oxidation/reduction reaction involving the ferri-
bon loading (see Fig. 2, and first data point in Fig. 7). The percent cyanide couple can be examined [49]:
Pt loading was determined using TGA. As expected, the conduc-
tivity of the composite increases as the platinum loading on the Fe(CN)6 3− + e− Fe(CN)6 4− E ◦ = 0.161 V vs. Ag/AgCl
PI/Carbon surface is increased. The conductivity for PI/Carbon with The cyclic voltammetric response of a PI/Carbon electrode with-
10% carbon content is 0.140 ± 0.005 S/cm, which is lower than that out Pt, in solution containing 0.1 M K3 Fe(CN)6 in 1 M KNO3 , is
reported for PI/Carbon (0.700 ± 0.003 S/cm) for the same material in presented in Fig. 8a. The PI/Carbon film shows minimal elec-
Fig. 2, suggesting that batch-to-batch variations exist for the com- trochemical activity for the oxidation/reduction processes of
posite materials. However, both values are consistent with a ∼104
increase in conductivity when the carbon content in PI is increased
to 10%. As a function of Pt loading, a steady increase in conductivity
Fig. 8. Cyclic voltammetry of (a) PI/Carbon, (b) PI/Carbon/Pt (125 Pt deposition
cycles), and (c) a planar Pt electrode in solution containing 0.1 M K3 Fe(CN)6 in 1 M
KNO3 . The scan rate for all measurements was 0.01 V/s. Crosshair marks the zero
Fig. 7. Conductivity of PI/Carbon/Pt composite as a function of Pt loading. potential and zero current for each voltammogram.
Please cite this article in press as: J.M. Kinyanjui, et al., Synthetic Met. (2011), doi:10.1016/j.synthmet.2011.08.046
8. ARTICLE IN PRESS
G Model
SYNMET-13632; No. of Pages 10
8 J.M. Kinyanjui et al. / Synthetic Metals xxx (2011) xxx–xxx
Table 3
Oxidation/reduction potentials (Epc and Epa ), peak splitting ( Ep ), and number of
electrons transferred (n) for the oxidation and reduction [Fe(CN)6 4− /Fe(CN)6 3− ] of
ferricyanide at PI/Carbon, PI/Carbon/Pt, and a planar Pt electrode (∼1 cm2 ).
Electrode (1 cm2 ) Epc (V) Epa (V) Ep (mV) n
PI – – – –
PI/Carbon/Pt 0.157 0.414 257 0.230
Planar Pt electrode 0.183 0.308 125 0.472
ferricyanide. In contrast, the electrochemical response for the
PI/Carbon/Pt electrode (Fig. 8b) is clearly enhanced. For comparison
and reference, a planar platinum sheet electrode was also used to
evaluate the oxidation/reduction processes of ferricyanide, Fig. 8c,
which are clearly observed. The voltammetry is significantly more
resolved for ferricyanide at the planar Pt electrode when compared
to the same electrochemical reaction at the PI/Carbon/Pt electrode.
The peak separation Ep = Epa − Epc (Epa and Epc are the anodic Fig. 9. Cyclic voltammetry for the oxidation of methanol in solution containing 1 M
CH3 OH in 1 M H2 SO4 for (a) PI/Carbon, (b) PI/Carbon/Pt (2 Pt deposition cycles), (c)
and cathodic peak potentials, respectively) is an indicator for
PI/Carbon/Pt (12 Pt deposition cycles) and (d) PI/Carbon/Pt (2 Pt deposition cycles
electron transfer kinetics during the oxidation/reduction of fer- and 15 methanol oxidation cycles) The scan rate for all measurements was 0.01 V/s.
ricyanide at each electrode. For kinetically fast electrochemical
processes, Nernstian kinetics are expected, where Ep is equal
to (59/n) mV, n being the number of electrons transferred (n = 1 Methanol oxidation is not observed for PI/Carbon electrodes
for ferricyanide). The peak potentials (Epc and Epa ) and the peak without Pt in solutions containing 1 M H2 SO4 and 1 M methanol,
splitting ( Ep ) are obtained from the voltammetric response of as shown in Fig. 9a. However, methanol oxidation is observed in
ferricyanide at both PI/Carbon/Pt and Pt planar electrodes (and the same solution using a PI/Carbon free-standing electrode after
listed in Table 3). The voltammetry for the PI/Carbon/Pt electrode two electrochemical reduction cycles using PtCl4 2− , Fig. 9b. The
shows non-nerstian peak splitting of 257 mV (i.e., with a value sig- voltammetry confirms the formation of a PI/Carbon/Pt composite
nificantly greater than 59 mV). The data suggests that the electron electrode with an electrochemically active Pt surface. The surface
transfer is kinetically limited at the PI/Carbon/Pt electrode relative area and Pt content of the composite electrode can be varied by
to the planar Pt electrode. The values obtained for the Pt planar changing the number of Pt deposition cycles. For comparison, the
electrode (peak splitting of 125 mV) also suggest that the electron voltammetric response of a PI/Carbon electrode after twelve Pt
transfer is kinetically limited. However, the smaller peak splitting reduction cycles is presented in Fig. 9c. The oxidation of methanol
suggests that the electron transfer at the bulk Pt surface is enhanced is observed for both, the forward and reverse scans. In addition, the
relative to the PI/Carbon/Pt electrode. Although factors such as oxidation of methanol on the forward (positive) and the reverse
solution concentrations and non-linear diffusion are important, (negative) scan increases with increasing Pt content.
surface inhomogeneity, including defects on the electrode surface, For PI/Carbon/Pt after two Pt deposition cycles, the oxidation of
likely play a larger role in the more sluggish electron transfer kinet- methanol is characterized by two voltammetric waves centered at
ics for the PI/Carbon/Pt composite relative to the planar Pt reference 0.70 V for the forward scan and 0.57 V for the reverse scan (Fig. 9b).
[59]. After twelve Pt deposition cycles, the oxidation of methanol is
The catalytic oxidation of methanol on Pt was also probed observed at 0.95 V for the forward scan and 0.64 V for the reverse
using the freestanding PI/Carbon/Pt composites—the correspond- scan. The waves for both electrodes are more positive than the
ing voltammograms are shown in Fig. 9. The oxidation of methanol values reported for PI/CNT/Pt [41] and for a conductive polymer
is strongly influenced by the Pt surface morphology and can be used (polyaniline)/Pt composite [62]. The data suggests that oxidation
as a measure of the chemical reactivity of the PI/Carbon/Pt com- of methanol may be thermodynamically more favorable at the
posite. The oxidation of methanol occurs via two reaction paths, as PI/Carbon/Pt electrodes produced in this study.
presented below [60,61]: Further analysis of the peak currents, charge passed, and peak
splitting provides an overview of the oxidation of methanol using
Pt-CH3 OHads + H2 O → Pt + CO2 + 6H+ + 6e− PI/Carbon/Pt electrodes (given in Table 4). For example, both the
ratio of the peak currents (ipf /ipr = 1.01) and the charge-ratio asso-
Pt-CH3 OHads → Pt-COads + 4H+ + 4e− ciated with methanol oxidation (Qf /Qr = 1.09) obtained for the
PI/Carbon/Pt after two Pt reduction cycles are close to unity, indicat-
The first path involves the direct oxidation of methanol to CO2 , ing that the processes are chemically reversible. The peak splitting
while the second one involves the formation of a strongly adsorbed is on the order of 135 mV and can be used as an indicator for sur-
CO intermediate prior to oxidation. CO adsorption causes surface face poisoning at the metal surface. The peak splitting associated
poisoning by blocking Pt sites necessary for further oxidation of with the oxidation of methanol for conductive polymer polyaniline
CH3 OH. Additionally, methanol oxidation during the cathodic scan (PANI)/Pt composites was measured to be 45 mV [62]. Although the
is hindered by the formation of PtO at the electrochemical interface results confirm that PI/Carbon substrates are sufficiently conduc-
and does not occur prior to the reduction of the surface oxide. tive to allow the deposition of Pt metal to produce freestanding,
Table 4
Oxidation/reduction potentials (Epf and Epr ), peak splitting ( Ep ), current ratios (ipf /ipr ), and charge ratios (Qf /Qr ) for the forward and reverse scans (after 2 and 10 cycles,
respectively), associated with methanol oxidation at a PI/Carbon/Pt electrode (∼1 cm2 ).
Electrode (1 cm2 ) Epf (V) Epr (V) Ep (mV) ipf /ipr Qf /Qr
PI/Carbon/Pt (2 cycles) 0.705 0.570 135 1.01 8.18 mC/7.50 mC = 1.09
PI/Carbon/Pt (10 cycles) 0.955 0.640 315 0.89 141 mC/46 mC = 3.09
Please cite this article in press as: J.M. Kinyanjui, et al., Synthetic Met. (2011), doi:10.1016/j.synthmet.2011.08.046
9. ARTICLE IN PRESS
G Model
SYNMET-13632; No. of Pages 10
J.M. Kinyanjui et al. / Synthetic Metals xxx (2011) xxx–xxx 9
electrochemically active PI/Carbon/Pt surfaces, the polymer does [2] C.A. Kaufmann, A. Neisser, R. Klenk, R. Scheer, Thin Solid Films 480 (2005)
not provide enhanced protection against poisoning at the Pt surface 515–519.
[3] F. Kessler, D. Rudmann, Solar Energy 77 (2004) 685–695.
relative to PANI/Pt composites [62]. For example, the charge ratio [4] R.E. Southward, D.S. Thompson, D.W. Thompson, M.L. Caplan, A.K. Stclair,
for the oxidation of methanol between the forward and reverse Chemistry of Materials 7 (1995) 2171–2180.
scans (Qf /Qr = 3.09) indicates that poisoning is more prominent for [5] A.N. Tiwari, A. Romeo, D. Baetzner, H. Zogg, Progress in Photovoltaics 9 (2001)
211–215.
this electrode. The peak splitting associated with the forward and [6] A.C. Vasko, X. Liu, A.D. Compaan, 34th IEEE Photovoltaic Specialists Conference
reverse process is also significantly larger (315 mV) in comparison (PVSC) (2009) 001552–001555.
to the PI/Carbon/Pt electrode obtained from two Pt reduction cycles [7] Z. Li, H. Qing, J. Wei-Long, L. Chang-Jian, S. Yun, Chinese Physics Letters 25
(2008) 734.
and PANI/Pt [58]. The data indicates that the surface contaminants
[8] G. Gebel, G. Meyer, L. Gonon, P. Capron, D. Marscaq, C. Marestin, R. Mercier,
must be fully reduced and desorbed from the Pt to observe the Journal of Power Sources 157 (2006) 293–301.
oxidation of methanol on the reverse scan. Furthermore, the peak [9] K. Okamoto, Journal of Photopolymer Science and Technology 16 (2003)
247–254.
currents decrease with each subsequent cycle in methanol for the
[10] D. Hill, Y. Lin, L.W. Qu, A. Kitaygorodskiy, J.W. Connell, L.F. Allard, Y.P. Sun,
PI/Carbon/Pt electrode obtained from two Pt reduction cycles as Macromolecules 38 (2005) 7670–7675.
shown in Fig. 9d. This signifies that surface contaminants the cat- [11] X.W. Jiang, Y.Z. Bin, M. Matsuo, Polymer 46 (2005) 7418–7424.
alyst efficiency is reduced with successive cycling due to surface [12] M. Shigeta, M. Komatsu, N. Nakashima, Chemical Physics Letters 418 (2006)
115–118.
poisoning. The data suggests that the deposition of Pt at conduc- [13] K.E. Wise, C. Park, E.J. Siochi, J.S. Harrison, Chemical Physics Letters 391 (2004)
tive PI/Carbon electrodes should be explored more thoroughly to 207–211.
fully understand how the number of Pt deposition cycles influences [14] B.K. Zhu, S.H. Xie, Z.K. Xu, Y.Y. Xu, Composites Science and Technology 66 (2006)
548–554.
the Pt morphology and efficiency of methanol oxidation. In addi- [15] M. Lebron-Colon, M.A. Meador, J.R. Gaier, F. Sola, D.A. Scheiman, L.S. McCorkle,
tion, the co-deposition of Pt-Ru in a 1:1 composition ratio should ACS Applied Material Interfaces 2 (2010) 669–676.
be investigated. This ratio has been found optimal in the decreas- [16] Y. Imai, T. Fueki, T. Inoue, M.A. Kakimoto, Journal of Polymer Science. Part A:
Polymer Chemistry 36 (1998) 1031–1034.
ing surface poisoning by removal of CO species [63]. However, [17] J.S. Lin, H.T. Chiu, Journal of Polymer Research (Taiwan) 9 (2002) 189–194.
the results confirm that Pt can be electrochemically deposited at [18] D. Pantea, H. Darmstadt, S. Kaliaguine, C. Roy, Applied Surface Science 217
conductive PI/Carbon substrates to metallize the surfaces and to (2003) 181–193.
[19] W. Zhang, A.A. Dehghani-Sanij, R.S. Blackburn, Journal of Materials Science 42
provide chemically reactive and thermally stable PI/C/Pt compos-
(2007) 3408–3418.
ites. [20] T. Lee, S.S. Park, Y. Jung, S. Han, D. Han, I. Kim, C.-S. Ha, European Polymer Journal
45 (2009) 19–29.
[21] M.K. Ghosh, K.L. Mittal, Polyimides, Marcel Dekker, New York, 1996.
4. Conclusions [22] Y. Zhai, Q. Yang, R.Q. Zhu, Y. Gu, Journal of Materials Science 43 (2008) 338–
344.
The studies demonstrate that conductive PI/Carbon films can be [23] S. Metz, M.O. Heuschkel, B. Valencia Avila, R. Holzer, D. Bertrand, P. Renaud,
Engineering in Medicine and Biology Society, 2001. Proceedings of the 23rd
prepared through the thermal imidization of PAA, premixed with Annual International Conference of the IEEE, vol. 761, 2001, pp. 765–768.
carbon black from a binary (5:1 acetone:DMSO) solvent system. [24] H. Zhou, G. Li, X. Sun, Z. Zhu, B. Xu, Q. Jin, J. Zhao, Q.-S. Ren, Sensors and Actuators
The solvent system minimizes the use of low volatility solvents that A: Physical 150 (2009) 296–301.
[25] Y.G. Seol, J.G. Lee, N.E. Lee, Organic Electronics 8 (2007) 513–521.
are difficult to expel during the curing and imidization process and [26] S.L. Qi, W.C. Wang, D.Z. Wu, Z.P. Wu, R.G. Jin, European Polymer Journal 42
provides high carbon dispersion. In addition, the electrochemical (2006) 2023–2030.
deposition of Pt was demonstrated at free-standing PI/Carbon elec- [27] T. Sawada, S. Ando, Chemistry of Materials 10 (1998) 3368–3378.
[28] J.J. Bergmeister, J.D. Rancourt, L.T. Taylor, Chemistry of Materials 2 (1990)
trodes to form conductive PI/Carbon/Pt composite films. Thermal
640–641.
gravimetric analysis of PI/Carbon and PI/Carbon/Pt demonstrates [29] J. Cardoso, O. GomezDaza, L. Ixtlilco, M.T.S. Nair, P.K. Nair, Semiconductor Sci-
that the thermal stability of PI is not compromised with either car- ence and Technology 16 (2001) 123–127.
bon incorporation or Pt deposition. Electrochemical measurements [30] K. Horn, A.M. Bradshaw, K. Doblhofer, S. Krause, G. Weinberg, H.M. Seiden-
spinner, R. Schulz, Fresenius Zeitschrift Fur Analytische Chemie 333 (1989)
using ferricyanide show that the PI/Carbon/Pt film is electroac- 590–595.
tive. However, the electron transfer for ferricyanide at PI/Carbon/Pt [31] J.H.G. Ng, M.P.Y. Desmulliez, K.A. Prior, D.P. Hand, Micro & Nano Letters 3 (2008)
was kinetically limited relative to a bulk Pt electrode. The sur- 82–89.
[32] S.H. Xie, B.K. Zhu, J.B. Li, X.Z. Wei, Z.K. Xu, Polymer Testing 23 (2004) 797–
face reactivity of Pt for the composite materials was evaluated as 801.
a function of the number of Pt reduction cycles and increasing Pt [33] G.M. Sessler, B. Hahn, D.Y. Yoon, Journal of Applied Physics 60 (1986) 318–
content. The oxidation of methanol provides a measure of com- 326.
[34] G. Maggioni, S. Carturan, D. Boscarino, G. DellaMea, U. Pieri, Materials Letters
plex chemical reactions at Pt surfaces electrochemically deposited 32 (1997) 147–150.
on the conductive PI/Carbon composite. The results confirm that [35] S. Yoda, A. Hasegawa, H. Suda, Y. Uchimaru, K. Haraya, T. Tsuji, K. Otake, Chem-
the electrochemical formation of PI/Carbon/Pt composites with istry of Materials 16 (2004) 2363–2368.
[36] N.T. Xuyen, H.K. Jeong, G. Kim, K.P. So, K.H. An, Y.H. Lee, Journal of Materials
electrochemically active Pt metal surfaces has been achieved. The Chemistry 19 (2009) 1283–1288.
electrochemical measurements suggest that the Pt surface reac- [37] M.R. Anderson, R.M. Davis, C.D. Taylor, M. Parker, S. Clark, D. Marciu, M. Miller,
tivity may be influenced by factors including deposit morphology, Langmuir 17 (2001) 8380–8385.
[38] K. Miyatake, T. Omata, D.A. Tryk, H. Uchida, M. Watanabe, The Journal of Phys-
thickness, and overall Pt loading. These parameters can be con-
ical Chemistry C 113 (2009) 7772–7778.
trolled electrochemically, and thus it is possible to optimize the [39] S.Y. Oh, J.K. Park, J.W. Choi, C.M. Chung, Molecular Crystals and Liquid Crystals
properties of conductive PI/Carbon/metal composites for a variety 377 (2002) 241–244.
of applications. [40] D. Andreescu, A.K. Wanekaya, O.A. Sadik, J. Wang, Langmuir 21 (2005)
6891–6899.
[41] X. Zhang, X. Shi, C. Wang, Catalysis Communications 10 (2009) 610–613.
Acknowledgements [42] H.R. Kricheldorf, O. Nuyken, G. Swift, Handbook of Polymer Synthesis, Marcel
Dekker, New York, 2005.
[43] L. Li, L. Qinghua, Y. Jie, Q. Xuefeng, W. Wenkai, Z. Zikang, W. Zongguang, Mate-
This work was support by the U.S. Department of Energy through rials Chemistry and Physics 74 (2002) 210–213.
the UNLV FCAST program under Grant No. DE-FG36-05GO85028. [44] D. Briggs, M.P. Seah, Practical Surface Analysis, Wiley, New York, 1990.
[45] R.M. Silverstein, F.X. Webster, Spectrometric Identification of Organic Com-
pounds, John Wiley & Sons, New York, 1998.
References [46] P. Murugaraj, N. Mora-Huertas, D.E. Mainwaring, Y. Ding, S. Agrawal, Compos-
ites Part A: Applied Science and Manufacturing 39 (2008) 308–313.
[1] http://www.2.dupont.com/Kapton/en US/assets/downloads/pdf/HN [47] M. Narkis, A. Vaxman, Journal of Applied Polymer Science 29 (1984) 1639–1652.
datasheet.pdf (2007). [48] M. Ghofraniha, R. Salovey, Polymer Engineering and Science 28 (1988) 58–63.
Please cite this article in press as: J.M. Kinyanjui, et al., Synthetic Met. (2011), doi:10.1016/j.synthmet.2011.08.046
10. ARTICLE IN PRESS
G Model
SYNMET-13632; No. of Pages 10
10 J.M. Kinyanjui et al. / Synthetic Metals xxx (2011) xxx–xxx
´
[49] P. Vanysek, in: W.M. Haynes (Ed.), CRC Handbook of Chemistry and Physics, [56] C.D. Wagner, in: D. Briggs, M.P. Seah (Eds.), Practical Surface Analysis: Auger
CRC Press, 2010. and X-ray Photoelectron Spectroscopy, John Wiley & Sons, 1990.
[50] E. Aschauer, R. Fasching, G. Urban, G. Nicolussi, W. Husinsky, Journal of Elec- [57] C. Roth, M. Goetz, H. Fuess, Journal of Applied Electrochemistry 31 (2001)
troanalytical Chemistry 381 (1995) 143–150. 793–798.
[51] A.J. Bard, R. Parsons, J. Jordan, Standard Potentials in Aqueous Solution, Marcel [58] M. Ree, T.L. Nunes, G. Czornyj, W. Volksen, Polymer 33 (1992) 1228–1236.
Dekker, New York, 1985. [59] X. Ji, C.E. Banks, A. Crossley, R.G. Compton, ChemPhysChem 7 (2006)
[52] J.M. Kinyanjui, N.R. Wijeratne, J. Hanks, D.W. Hatchett, Electrochimica Acta 51 1337–1344.
(2006) 2825–2835. [60] L. Li, Y. Xing, The Journal of Physical Chemistry C 111 (2007) 2803–2808.
[53] A.J. Bard, L.R. Faulkner, Electrochemical Methods, John Wiley & Sons Inc., Hobo- [61] E. Herrero, W. Chrzanowski, A. Wieckowski, The Journal of Physical Chemistry
ken, NJ, 1980. 99 (1995) 10423–10424.
[54] J.M.D. Rodriguez, J.A.H. Melian, J.P. Pena, Journal of Chemical Education 77 [62] D.W. Hatchett, R. Wijeratne, J.M. Kinyanjui, Journal of Electroanalytical Chem-
(2000) 1195–1197. istry 593 (2006) 203–210.
[55] J.O.M. Bockris, B.E. Conway, R.E. White, Modern Aspects of Electrochemistry, [63] A.S. Arico, V. Baglio, A. Di Blasi, E. Modica, P.L. Antonucci, V. Antonucci, Journal
Springer, New York, 1992. of Electroanalytical Chemistry 557 (2003) 167–176.
Please cite this article in press as: J.M. Kinyanjui, et al., Synthetic Met. (2011), doi:10.1016/j.synthmet.2011.08.046