Enhanced efficiency in dye sensitized solar cells with nanostructured Pt decorated multiwalled carbon nanotube based counter electrode

Article abstract of DOI:10.1016/j.electacta.2012.04.013

Nanostructured Pt decorated functionalized multiwalled carbon nanotubes (Pt/f-MWNTs) cast on fluorine doped tin oxide coated glass (FTO) is demonstrated as a novel counter electrode (CE) for dye sensitized solar cells (DSSCs). Pt particles (20 wt.%, size <5 nm) are uniformly decorated on f-MWNTs using a facile microwave assisted polyol reduction method. The DSSC fabricated with Pt/f-MWNTs CE (effective Pt loading ～10 μg cm^-2) exhibits an enhanced efficiency of ～5.4% (under an illumination of AM 1.5 G, 100 mW cm^-2), compared to ～2.3% and 4.5% for f-MWNTs and Pt CE based reference DSSCs. The CEs are fabricated without using conventional binders using a simple doctor blade method. A new equivalent circuit is proposed to model the electrochemical impedance behavior of porous f-MWNTs based CE. On the other hand, the behavior Pt/f-MWNTs based electrode is similar to that of Pt nanoparticles but with higher catalytic activity. Cyclic voltammetry studies reveal that Pt/f-MWNTs counter electrode offers greater catalytic surface area and facilitates faster reaction at its surface.

Full text of DOI:10.1016/j.electacta.2012.04.013

Electrochimica Acta 72 (2012) 199–206
Contents lists available at SciVerse ScienceDirect
Electrochimica Acta
journal homepage: www.elsevier.com/locate/electacta
Enhanced efficiency in dye sensitized solar cells with nanostructured Pt
decorated multiwalled carbon nanotube based counter electrode
Adarsh Kaniyoor, Sundara Ramaprabhu∗
Alternative Energy and Nanotechnology Laboratory (AENL), Nano Functional Materials Technology Center (NFMTC), Department of Physics, Indian Institute of Technology Madras
(IITM), Chennai 600036, India
a r t i c l e i n f o
a b s t r a c t
Article history:
Nanostructured Pt decorated functionalized multiwalled carbon nanotubes (Pt/f-MWNTs) cast on fluorine
doped tin oxide coated glass (FTO) is demonstrated as a novel counter electrode (CE) for dye sensi-
tized solar cells (DSSCs). Pt particles (20 wt.%, size <5 nm) are uniformly decorated on f-MWNTs using a
facile microwave assisted polyol reduction method. The DSSC fabricated with Pt/f-MWNTs CE (effective
Pt loading ∼10 ␮g cm−2) exhibits an enhanced efficiency of ∼5.4% (under an illumination of AM 1.5 G,
100 mW cm−2), compared to ∼2.3% and 4.5% for f-MWNTs and Pt CE based reference DSSCs. The CEs are
fabricated without using conventional binders using a simple doctor blade method. A new equivalent
circuit is proposed to model the electrochemical impedance behavior of porous f-MWNTs based CE. On
the other hand, the behavior Pt/f-MWNTs based electrode is similar to that of Pt nanoparticles but with
higher catalytic activity. Cyclic voltammetry studies reveal that Pt/f-MWNTs counter electrode offers
greater catalytic surface area and facilitates faster reaction at its surface.
Received 20 September 2011
Received in revised form 1 April 2012
Accepted 3 April 2012
Available online 10 April 2012
Keywords:
Pt nanoparticles
Carbon nanotubes
Efficiency
Impedance
Charge transfer resistance
© 2012 Elsevier Ltd. All rights reserved.
1. Introduction
surface area and therefore increased catalytic activity. However,
unless the Pt particles are strongly adhered to the substrate, they
The counter electrodes in dye sensitized solar cells (DSSCs) [1]
usually consist of a thin layer of Pt deposited on fluorine doped tin
oxide (FTO) coated glass substrates. Pt is chosen due to its excel-
lent catalytic activity toward tri-iodide reduction, a mechanism
essential for the continuous operation of a DSSC. The Pt counter
electrode is also chemically stable in the highly corrosive iodide
electrolyte used in DSSCs [2–4]. Generally, these films have a Pt
loading of ∼10–100 ␮g cm−2, a thickness of ∼10 nm and a charge
transfer resistance (Rct) of ∼1–2 ꢀ cm2 [2–6]. The low Rct observed
in Pt films results in a fast rate of tri-iodide reduction at the counter
electrode, thereby limiting the recombination reaction occurring
at the TiO2-electrolyte interface to a considerable extent [7]. The
resulting decrease in the dark current leads to an overall increase
in power conversion efficiency. Therefore, at present, the highly
efficient DSSCs employ Pt as counter electrodes [8–10].
Although several low cost counter electrode materials such as
CoS, Cu2S, Au, carbon in its various forms and conducting polymers
[11–15] have been developed, a suitable replacement for Pt counter
electrode has not yet been found. Pt being inevitable, high perfor-
mance at low costs can only be obtained by better utilization of the
surface area of Pt, thereby, facilitating lesser material usage. Use
of Pt nanoparticles instead of a film can, in principle, offer higher
can leach out and eventually deposit on the TiO2 surface, causing
short-circuit by catalyzing the recombination reaction [7]. In addi-
tion, unsupported Pt nanoparticles tend to oxidize easily. To avoid
these, Pt nanoparticles can be dispersed on suitable catalyst support
materials. Reports indicate the use of NiO [16], SnO2 [17], acetylene
black [18], carbon black [19], etc. as support materials for Pt. Even
with the above counter electrode materials, the performances of
DSSCs rarely exceed those of Pt based cells.
In this context, multiwalled carbon nanotubes (MWNTs) can
be an interesting choice as catalyst support material for DSSCs.
MWNTs are coaxial cylinders of graphitic carbon, discovered by
Sumio Iijima in 1991 [20]. They possess excellent properties such
as large surface area, high electrical conductivity, good thermal
and chemical stability [21]. Most importantly, they can be tailored
through functionalization and metal decoration to suit specific
application. As a result, they have been used directly or as catalyst
support material in different fields such as gas sensors, biosensors,
gas adsorption and storage and nanofluids [22–26]. Also, defect
rich species of MWNTs have shown excellent catalytic activity
toward tri-iodide reduction in DSSC [27]. Furthermore, in compari-
son to other one-dimensional carbon nanostructures such as single
walled carbon nanotubes, carbon fibres, coils or helices, high purity
MWNTs can be prepared in large quantities. Hence, in this paper,
we demonstrate the use of Pt decorated functionalized MWNTs
(f-MWNTs) as a novel counter electrode material for DSSCs. Fine
particles of Pt could be uniformly deposited on f-MWNTs by a
∗
Corresponding author.
E-mail address: ramp@iitm.ac.in (S. Ramaprabhu).
0013-4686/$ – see front matter © 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.electacta.2012.04.013

200
A. Kaniyoor, S. Ramaprabhu / Electrochimica Acta 72 (2012) 199–206
simple and rapid method. The DSSC with Pt/f-MWNTs based
counter electrode shows much higher power conversion effi-
ciency than the Pt or f-MWNTs based reference DSSCs. Detailed
electrochemical impedance spectroscopy and cyclic voltammetry
measurements are carried out to investigate this high performance.
based on that developed by Nazeeruddin et al. [31]. The resultant
paste was spread on FTO plates (TEC 22-15) using a glass rod. Again,
the dimensions and thickness were controlled using a framework
of Scotch 3M tape. The coatings were finally sintered at 450 ◦C in a
furnace for 1 h. The active areas of the photo anode were ∼0.25 cm2.
No other treatments (such as TiCl4 pre- and post-treatments) are
carried out. The TiO2 films were immersed in N719 dye solution for
∼12 h for complete dye uptake. The TiO2 photoelectrode and the
carbon based counter electrodes were clamped with a SX 1170-
25 film as spacer, cut into appropriate size and shape. The space
between the electrodes was filled with the iodide electrolyte.
2. Experimental
2.1. Materials
P25 TiO2 was a generous gift from Degussa, Germany. FTO con-
ducting plates (TEC 22-7, TEC 22-15), SX1170-25 spacer, Pt catalyst
Plastisol, and Ruthenium 535 bis-TBA (N719) dye were purchased
from Solaronix, Switzerland. LiI was purchased from Merck. All
other reagents used were purchased from Sigma–Aldrich or Alfa
Aesar. High purity de-ionized (DI) water was used for all experi-
ments. The FTO plates were cut into 1.75 cm × 1.25 cm and rinsed
with soap water and DI water, followed by successive ultrasoni-
cation in DI water, acetone and ethanol. The electrolyte used in
DSSC consists of 0.1 M LiI, 0.05 M I2, 0.6 M DMPII (1,2-dimethyl-3-
propylimidazolium iodide) and 0.5 M TBP (tert-butyl pyridine) in
acetonitrile. The dye solution was prepared by dissolving N719 in
acetonitrile and tert-butanol (1:1) at a concentration of 0.5 mM.
2.4. Characterization
X-ray diffraction was carried out using a PANalytical X’PERT
Pro X-ray diffractometer with nickel-filtered Cu K␣ radiation as
the X-ray source. The pattern was recorded in the 2ꢁ range of
10–90◦ with a step size of 0.016◦. Identification and characteriza-
tion of functional groups were carried out using Perkin Elmer FT-IR
spectrometer in the range of 400–4000 cm−1. The morphology of
the sample was characterized by field emission scanning electron
microscopy (FESEM, FEI QUANTA 3D). Elemental analysis was car-
ried out by EDX analyzer in the FESEM system. High resolution
transmission electron microscopy (HRTEM) was carried out using
a TECNAI F 20 (S-Twin) microscope. For HRTEM measurements, the
sample was dispersed in absolute ethanol using mild ultrasonica-
tion and casted onto carbon coated Cu grids (SPI supplies). The I–V
characteristics of the fabricated solar cells were measured under
AM 1.5 G solar radiation using a Newport 150 W Class A Solar Sim-
ulator attached with Keithley 2420 source meter. The intensity was
calibrated to give an output of 1 Sun (100 mW cm−2) using a ther-
mopile detector and also using a NREL calibrated Si reference cell.
Cyclic voltammetry measurements were carried out in three elec-
trode system consisting of Pt wire as counter electrode, Ag/AgCl
as reference electrode and the catalyst coated FTO substrate as the
working electrode. The electrolyte consisted of 5 mM LiI, 0.5 mM
I2 and 0.1 M LiClO4 in acetonitrile. Electrochemical impedance
measurements were carried out using sandwiched symmetric
cell configuration in the frequency range of 100 kHz–0.1 Hz using
Solartron 1400 and 1470E Cell Test System. Under this condition,
two identical electrodes are prepared from the same catalyst. They
are clamped together with a spacer in between. The space is filled
by the same electrolyte that is used in the fabricated DSSCs. An ac
signal of 10 mV was applied over at 0 V dc bias to obtain the data.
The geometric area of the films was maintained at 0.5 cm2. The
obtained data were fit with appropriate equivalent circuits using
ZSimp 3.1 software.
2.2. Synthesis of MWNTs, f-MWNTs and Pt/f-MWNTs
MWNTs were prepared by catalytic chemical vapor deposi-
tion of acetylene over alloy hydride catalysts at ∼700 ◦C in the
presence of Ar. Details are given elsewhere [28]. As-synthesized
MWNTs were purified by air oxidation and acid treatment meth-
ods to remove the amorphous carbon and metallic impurities. The
acid treatment for purification was carried out by refluxing with
concentrated HNO3 for 24 h. For further functionalization, purified
MWNTs were ultrasonicated in concentrated H2SO4:HNO3 (3:1)
for 3 h. The sample was washed with DI water, filtered using a
nitrocellulose membrane and dried in vacuum to obtain fine pow-
der of f-MWNTs. To decorate Pt, 80 mg f-MWNTs were dispersed
in ethylene glycol:water (2:1) solution (30 mL) by ultrasonica-
tion. Calculated amount (5.3 mL) of 1 wt.% aqueous solution of
H2PtCl6·6H2O was added dropwise to the f-MWNTs dispersion and
the resulting solution was stirred 12 h. pH of solution was adjusted
to 12.5 by adding 1 M aqueous NaOH solution. The solution was
placed in a microwave oven and exposed to microwave (800 W,
pulses of 2 min) to reduce Pt precursor to Pt metal [29,30]. The
sample was washed with DI water, filtered and dried to obtain Pt/f-
MWNTs. Assuming complete reduction of Pt precursor, 20 mg Pt
can be obtained from 5.3 mL. Thus, the loading of Pt in Pt/f-MWNTs
is expected to be 20 wt.%.
2.3. Fabrication of DSSCs
3. Results and discussion
The photoanode and the counter electrode were fabricated on
pre-cleaned and cut FTO substrates. The counter electrodes were
fabricated as follows. f-MWNTs and Pt/f-MWNTs were dispersed
in ethanol:Nafion solution (10:1 by volume) by ultrasonication for
1 h to give a concentration of 5 mg mL−1. The slurry was spread on
masked FTO plates (TEC 22-7) using a glass rod. Two parallel lay-
ers of Scotch 3 M tape served as the mask. The layer was allowed
to air dry for a few minutes. The deposition was repeated to get
∼10 ␮g cm−2 of Pt (i.e. 50 ␮g cm−2 of Pt/f-MWNTs catalyst). The
films were sintered finally at 200 ◦C on a hot plate. The final area of
counter electrode was reduced used to 0.5 cm2. The Pt counter elec-
trode was fabricated using commercially available Platisol solution.
The solution was brush coated on masked FTO plates and dried on a
hot plate. The final sintering was done in a muffle furnace at 400 ◦C
for 30 min. For the photoanode, a TiO2 paste formulation was made
3.1. Decoration of Pt nanoparticles
Pt decoration on f-MWNTs was carried out by microwave
assisted polyol reduction method instead of the conventional
refluxing method. The schematic of the same is shown in Fig. 1.
The purified MWNTs were treated with concentrated acids to pro-
duce surface functional groups such as carboxyl, hydroxyl and
carbonyl. Generally these functional groups have a site ratio of 4
(
OH):2 ( COOH):1 ( C O) [32]. The OH groups make f-MWNTs
highly hydrophilic [33], and therefore dispersible in aqueous sol-
vents such as ethylene glycol:water (2:1) used here. The COOH
and CO groups, on the other hand, affect the decoration of metal
nanoparticles [34] by acting as anchoring sites. The metal precur-
sor, H2PtCl6·6H2O when added to the solution dissociates to give Pt
ions. These ions get attached to the negatively charged functional

A. Kaniyoor, S. Ramaprabhu / Electrochimica Acta 72 (2012) 199–206
201
Fig. 1. Metal decoration. Schematic depicting the attachment of functional groups
on the surface of MWNTs (f-MWNTs) and decoration of Pt nanoparticles on func-
tionalized MWNTs (Pt/f-MWNTs).
Fig. 3. Formation of composite. Powder X-ray diffraction patterns of (a) f-MWNTs
and (b) Pt/f-MWNTs confirming the formation of the composite.
groups such as carboxyl and carbonyl via ion exchange or coordina-
tion mechanism [35]. This is the nucleation step. Upon microwave
heating, the polyol ethylene glycol decomposes into formaldehyde
which acts as a reducing agent. Consequently, the surface Pt ions
get reduced to well-dispersed Pt nanoparticles [36]. This bond is
generally, strong enough to withstand even ultrasonication pro-
cess. Therefore, the Pt particles are held strongly on the nanotube
surface. pH of the solution plays an important role in determining
the particle size and distribution. At pH higher than 10, Pt nanopar-
ticles with a narrow size distribution can be obtained. The detailed
mechanism on the effect of pH can be found in the work of Lee et al.
[36] Li et al. [37] and Yang et al. [38] Hence, in the present work, a
pH >12 was maintained. In addition, the microwave procedure used
here assists is uniform and rapid decoration of metal nanoparticles.
involves treatment with acids. On further functionalization (dashed
curve), various functional groups get attached to the surface, mak-
ing them hydrophilic. The vibrations that are present in f-MWNTs
are OH stretching at ∼3430 cm−1
,
CHx antisymmetric and sym-
OH bending at
metric stretching at 2920 cm−1 and 2850 cm−1
,
1620 cm−1, C O at 1720 cm−1, C O at 1250 cm−1 and 1050 cm−1
[22,23]. From this we can deduce that the functional groups that
are mainly present are hydroxyl, carboxyl, and carbonyl. On Pt dec-
oration (solid curve), the intensities of the functional groups such
as COOH and CO reduce significantly, suggesting that metal dec-
oration occurs preferentially at these sites. OH groups are present
to a lesser extent than in f-MWNTs, as indicated by peak intensity.
3.2. Spectroscopic analysis
3.3. Structural analysis by X-ray diffraction
The FT-IR spectra presented in Fig. 2 further supports this argu-
ment. MWNTs show minimal amount of functional groups as can
be seen from the figure (dotted curve). The lack of functional groups
makes them hydrophobic in nature. The small amount of func-
tional groups present is due to the purification procedure which
X-ray diffraction (XRD) patterns of f-MWNTs and Pt/f-MWNTs
are presented in Fig. 3. f-MWNTs exhibit the crystalline peaks of
graphite and have been indexed according to JCPDS file no. 89-8487.
The most intense peak occurs at a 2ꢁ of ∼26.5◦. This corresponds to
(0 0 2) plane of hexagonal graphite and has a d-spacing of 0.34 nm.
In the case of f-MWNTs, this d-spacing corresponds to the distance
between the walls. Peaks corresponding to (1 0 0) and (1 0 1) planes
can be seen at 42.9◦ and 44.5◦, respectively. The other graphitic
peaks are negligible in intensity. After metal decoration, peaks cor-
responding to f.c.c. Pt can be seen along with the peaks of MWNTs.
The diffraction profile of Pt was indexed according to the JCPDS file
no. 87-0646. The peaks at 2ꢁ ∼40.1◦, 46.4◦, 67.9◦ and 81.5◦ corre-
spond to (1 1 1), (2 0 0), (2 2 0) and (3 1 1) planes, respectively. The
crystallite size of Pt was determined from the (2 2 0) peak using
the Scherrer equation to be ∼3.7 nm. This peak was chosen over
the highly intense (1 1 1) peak as this peak suffered no overlap
with neighboring peaks of either MWNTs or Pt. It is to be noted
here that no peaks corresponding to either platinum oxide or plat-
inum carbide were observed. As mentioned earlier, unsupported
Pt nanoparticles get oxidized immediately. Therefore, absence of
any oxide peaks suggests that Pt particles are bonded to MWNTs.
Absence of carbide peaks suggests that Pt does not form regular
bonds with MWNTs, suggesting that the interaction is electrostatic.
3.4. Electron microscopy
Fig. 2. Functional groups. FT-IR spectra of (a) MWNTs (red dotted line), (b) f-MWNTs
(green dashed line), and (c) Pt/f-MWNTs (blue solid line) showing the presence of
various functional groups such as hydroxyl, carboxyl, carbonyl, etc. on the surface
of the nanotubes. (For interpretation of the references to color in this figure legend,
the reader is referred to the web version of this article.)
High resolution electron microscopy studies were carried out to
analyze the morphology of the samples. Fig. 4 shows the FESEM
and EDX images of f-MWNTs and Pt/f-MWNTs. f-MWNTs and

202
A. Kaniyoor, S. Ramaprabhu / Electrochimica Acta 72 (2012) 199–206
Fig. 4. Microscopy and elemental analysis. FESEM images of (a) f-MWNTS and (b) Pt/f-MWNTs. (c) EDX spectrum of Pt/f-MWNTs showing the presence of ∼20 wt.% Pt on
f-MWNTs.
Pt/f-MWNTs (Fig. 4a and b) show a densely packed morphology
consisting of randomly oriented MWNTs of nearly uniform diam-
eters. Extremely small particles of Pt decorated on the nanotubes
can be seen in the FESEM image of Pt/f-MWNTs. The loading of
Pt on f-MWNTs is expected to be ∼20 wt.%, based on the amount
used during synthesis. To confirm the metal loading, EDX analysis
was carried out (Fig. 4c). The peaks corresponding to C and Pt can
be seen. The weight percentage of Pt was found to be ∼19.2 wt.%,
close to the expected value. This value was determined by taking
the average of the values obtained at several spots.
The TEM and HRTEM images of the samples are shown in Fig. 5.
Fig. 5a shows clearly the hollow nature and thick walls of MWNTs.
MWNTs were found to have outer diameters in the range 20–30 nm
and inner diameters of ∼10 nm. This was also confirmed from
the HRTEM image of f-MWNTs in Fig. 5b and ∼26 walls could be
counted. The interlayer separation was calculated to be 0.34 nm.
Fig. 5. Particle size and decoration: (a) TEM and (b) HRTEM images of f-MWNTs showing their hollow and multiwalled nature. (c) Dark field TEM, and (d) HRTEM images of
Pt/f-MWNTs depicting the uniform decoration of Pt nanoparticles on f-MWNTs.

A. Kaniyoor, S. Ramaprabhu / Electrochimica Acta 72 (2012) 199–206
203
Table 2
Equivalent circuit parameters. Charge transfer resistance and other parameters
obtained for the various counter electrodes measured in symmetric sandwich cell
configuration.
Material
Rs (ꢀ)
Rct (ꢀ cm2)
Q:Yo (S)
Q:n
Rpore (ꢀ cm2)
Pt
19.1
15.4
16.2
1.8
360
0.4
1.7 × 10−5 0.93
5.2 × 10−4 0.95
4.4 × 10−5 0.90
–
10.9
–
f-MWNTs
Pt/f-MWNTs
increases by ∼15% and 32% with respect to that of f-MWNTs and
Pt based cells. The FF increases from 0.3 in f-MWNTs based cells
to 0.64 in Pt/f-MWNTs based cells, which is comparable to that
of Pt. As a consequence of these changes, the power conversion
efficiency (ꢂ) of Pt/f-MWNTs based cells is ∼5.4%, which is greater
than that of f-MWNTs based cells by ∼170% and Pt based cells by
20%.
This remarkable improvement in the efficiency of Pt/f-MWNTs
based DSSC is clearly due to an increase in the Jsc, though it has a
slightly lower FF. This behavior is also seen in f-MWNTs based cells
wherein, though the FF is too less, Jsc comparable to that of Pt is
obtained. In recent reports on similar systems [39,40], enhanced
efficiency was also obtained due to increase in Jsc, although the
FFs were lower than those of their Pt based reference DSSCs. The
increase in Jsc was attributed mainly to increased surface rough-
ness and catalytic activity. However, the lower FFs observed in
these systems have not been addressed. Generally, an increase in
Jsc is accompanied by an increase in FF. While this is true for pla-
nar electrodes such as Pt, the same need not be true for rough
electrodes. Therefore, the increase in Jsc could indeed be due to
increased surface area, improved catalytic activity, or conductivity
effects in these rough systems. The lower FF, on the other hand,
could be due to porosity in the electrodes. Note that the porosity
being referred to is not that of the material but of the electrode. We
propose that in carbon based counter electrodes fabricated with-
out binders (or removed due to sintering), significant amount of
pores could be present. Consequently, the electrolyte can slowly
percolate through the pores and interact with the FTO electrode.
Hence, a combined effect of FTO and carbon material would be
observed, resulting in low FF. Other carbon nanostructures such
as graphene also exhibit comparable or increased current densities
at low FF [41]. However, if the pure carbon nanostructures are mod-
ified by metal decoration, etc., the behavior can become complex.
Surface area and catalytic effects arise due to the nanostructur-
ing of Pt. To verify this, DSSCs were fabricated with Pt/f-MWNTs
counter electrode containing 40 wt.% Pt (20 ␮g cm−2). This cell
showed a lower efficiency due to the decrease in Jsc. In this sam-
ple, the Pt nanoparticles agglomerate due to the high loading,
resulting in an increase in crystallite size. Consequently, surface
area and catalytic activity reduce which decrease the current den-
sity. Interestingly, the FF improved by a small extent. To further
understand this behavior, electrochemical studies were carried
out.
Fig. 6. I–V characteristics. Current voltage characteristics of DSSCs with f-MWNTs
(green circles), Pt (red squares) and Pt/f-MWNTs (20 wt.% Pt – blue triangles; 40 wt.%
Pt – purple diamonds) based counter electrodes, measured under AM 1.5 simulated
solar radiation. (For interpretation of the references to color in this figure legend,
the reader is referred to the web version of this article.)
Also the outer few tubes were found to be damaged due to the
harsh oxidation treatment given. Pt/f-MWNTs (Fig. 5c) exhibit mor-
phology similar to that of MWNTs but in addition consist of Pt
nanoparticles attached to MWNTs. Even at a scale of 100 nm (mag-
nification ∼60,000) the particles were not visible clearly in normal
bright field imaging in TEM. Hence, a dark field TEM image of
Pt/f-MWNTs is presented. The image shows uniform coating of
f-MWNTs with Pt nanoparticles, which appear as shiny dots on
f-MWNT surface. The HRTEM image of Pt/f-MWNTs presented in
Fig. 5d shows Pt particles of very small size attached on the surface
of MWNTs. Majority of the particles had a size of ∼2 nm. Some larger
particles of ∼4 nm in size were also observed. Since, larger crystal-
lites generally dominate in XRD patterns, the Scherrer formula gave
a crystallite size of ∼3.7 nm, instead of 2 nm.
3.5. I–V characteristics of fabricated solar cells
To evaluate the I–V characteristics of solar cells prepared with
different counter electrodes, DSSCs were fabricated with similar
photoanodes, dye and electrolyte. Keeping all other parameters
constant, it is possible to evaluate the performance of counter
electrode alone. Three types of counter electrodes were prepared
namely, f-MWNTs, Pt, and Pt/f-MWNTs. 5 cells were fabricated for
each counter electrode type and the I–V characteristic were mea-
sured under 1 Sun AM 1.5 simulated solar illumination. The I–V
curves are plotted in Fig. 6 and average values are summarized in
Table 1. It can be seen that Pt counter electrode based DSSC has a
short circuit current density (Jsc) of ∼9.3 mA cm−2 and a fill factor
(FF) of ∼0.66, resulting in a high efficiency of 4.5%. With f-MWNTs
as counter electrode, the FF reduces to 0.32 and Jsc to 8.9 mA cm−2
.
This suggests that undecorated f-MWNTs based counter electrode
has a poor catalytic activity toward tri-iodide reduction. In the case
of Pt/f-MWNTs (20 wt.% Pt) counter electrode based DSSC, the Jsc
3.6. Electrochemical impedance spectroscopy
To determine the catalytic activity of the electrodes toward tri-
iodide reduction, electrochemical impedance spectroscopy (EIS)
was carried out. Fig. 7a shows the impedance spectra for Pt, f-
MWNTs and Pt/f-MWNTs electrodes measured in a symmetric
sandwich cell configuration. Zꢀ and Zꢀꢀ are real and imaginary parts
of impedance, following IUPAC convention. The inset, Fig. 7b and
c shows the magnified view of the spectra. Fig. 7d and e gives the
equivalent circuits which are used to model the impedance behav-
ior in these electrodes. The values have been summarized in Table 2.
The Pt counter electrode shows a large semicircle in the high
Table 1
Solar cell parameters. The characteristic parameters of the fabricated DSSCs such
as open circuit voltage (Voc), short circuit current density (Jsc), fill factor (FF) and
efficiency (ꢂ).
Material
Voc (V)
Jsc (mA cm−2
)
FF
ꢂ (%)
Pt
0.74
0.69
0.70
0.71
9.3
8.9
12.2
9.96
0.66
0.33
0.64
0.63
4.5
2.1
5.4
4.7
f-MWNTs
Pt/f-MWNTs (20 wt.%)
Pt/f-MWNTs (40 wt.%)

204
A. Kaniyoor, S. Ramaprabhu / Electrochimica Acta 72 (2012) 199–206
Fig. 7. Catalytic activity. (a) Electrochemical impedance spectra of Pt (red squares), f-MWNTs (green circles) and Pt/f-MWNTs (blue triangles) in a symmetric sandwich cell
configuration. (b) and (c) are the magnified data of (a). (d) and (e) are the equivalent circuits used to model the impedance spectra of Pt, Pt/f-MWNTs and f-MWNTs. (For
interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
frequency region corresponding to the Pt/electrolyte interface and
a small semicircle in the low frequency region corresponding to the
diffusion of ionic moieties. This kind of behavior is modeled accord-
ing to Randles type equivalent circuit with minor modifications as
shown in Fig. 7d. The ideal capacitance Cdl is replaced by a constant
phase element, Q which represents the capacitance at rough sur-
faces. The admittance in this case is represented by YQ = YQo(iω)n,
where YQo and n are frequency-independent parameters (0 ≤ n ≤ 1;
for n = 1, Q = Cdl) [41,42]. In the case of Pt counter electrode, n = 0.93.
The slight deviation from the ideal capacitance is probably due
to the roughness introduced by brush coating method used for Pt
deposition and hence the use of Q instead of Cdl. The charge trans-
fer resistance (Rct) was determined to be 1.8 ꢀ cm2. As we do not
expect any surface area effects in pure Pt based films, we conclude
that the low Rct of Pt is solely due to its catalytic activity toward tri-
iodide reduction. This catalytic activity is responsible for the high
FF, generally observed in Pt electrodes. For modeling the diffusion
of ionic species, the diffusion parameter O is used. This is different
from Warburg impedance and appears at finite length conditions.
Its admittance is given as YO = YOo(iω)1/2coth[B(iω)1/2], where YOo is
related to conductance and B is inversely related to time constant of
the regeneration of diffusion process [42]. This bounded diffusion
is responsible for second semicircle in the impedance spectrum. If
the infinite Warburg term (W) is used, a 45◦ line is obtained instead
of a semicircle.
In the case of the present f-MWNTs based electrodes, the
traditional equivalent circuit could not satisfactorily explain the
observed behavior. It was recently reported by Roy-Mayhew et al.
[41] that in porous carbon based electrodes, an additional dif-
fusion impedance term arises in the high frequency region due
to the diffusion through the electrode pores. However, in the
present case it was found that a simple Opore did not satisfacto-
rily fit the high frequency semicircle. The traditional circuit shown
in Fig. 7d, also does not simulate the observed spectrum. The
equivalent circuit shown in Fig. 7e fits the obtained impedance
spectra more accurately. Further modeling/experiments may be
required to understand the interaction at porous electrodes. Nev-
ertheless, the first semicircle does not represent the charge transfer
resistance of the electrode, as can be confirmed from the poor per-
formance of our f-MWNTs based DSSC. The second larger semicircle
now represents the charge transfer resistance and the constant
phase element. The Rct for f-MWNTs was found to be 360 ꢀ cm2,
which is reason for its poor catalytic activity and hence lower
FF.
The Pt/f-MWNTs counter electrode behaves similar to Pt, as can
be seen from Fig. 7c. The first semicircle represents the electrode
electrolyte interface and the second semicircle represents the dif-
fusion of ionic species between the two electrodes. Hence the same
equivalent circuit was used to fit the impedance spectrum (Fig. 7d).
Pt/f-MWNTs electrode has a very low Rct of 0.4 ꢀ cm2. This very low
Rct is responsible for the high catalytic activity and therefore high
current density of Pt/f-MWNTs based counter electrode. Interest-
ingly, the FF of Pt/f-MWNTs based DSSC is lower than that of Pt, even
though it has a lower Rct. This kind of behavior is also observed in
recent reports on similar systems [39,40,43]. The reason for this
is not yet clear. Jeon et al. have attributed this to surface resistiv-
ity of the electrode. Similarly, we determined the bulk electrical
conductivity of the samples by linear four probe method. While
purified f-MWNTs have a conductivity of 60 S m−1, Pt/f-MWNTs
have 3.2 × 103 S m−1. Pure Pt has been reported to have a conduc-
tivity of 17.1 × 103 S m−1 [40]. Thus, Pt has a higher FF than other
counter electrodes. The porosity effects described in the case of
f-MWNTs is not significant in the case of Pt/f-MWNTs. The pres-
ence of Pt nanoparticles alters the kinetics of tri-iodide reduction
considerably. Moreover, the uniform distribution of Pt on MWNTs
ensures that there are a large number of catalytic sites available for
reduction which is not the case with f-MWNTs.
It can be noted that the series resistance of f-MWNTs is lower
than that of Pt/f-MWNTs and Pt. The trend is in fact, opposite to
the changes in bulk electrical conductivity. Therefore, this behavior
can be explained only in terms of contact/adhesion between the
material and the FTO counter electrode. The f-MWNTs make better
contact due to their hydrophilic nature. After Pt decoration, only
some functional groups remain free making them less hydrophilic.
This results in poorer contact with the FTO substrate and hence the
Rs is higher. However, it is not clear if these changes affect the DSSC
behavior since dominant role is generally played by Rct or the total
series resistance (Rs + Rct) [44].
3.7. Cyclic voltammetry studies
Cyclic voltammetry (CV) studies were carried out to further
understand the catalytic activity of the electrode materials. Fig. 8
shows the CV curves for f-MWNTs, Pt and Pt/f-MWNTs in the
potential window of −0.4 to 1.2 V vs. Ag/AgCl taken at scan rate
of 50 mV s−1. Two pairs of redox waves are expected, according
to literature [41,45]. A pair of anodic and cathodic currents corre-
sponding to the oxidation (Aox) and reduction (Ared) of I2/I3−, as

A. Kaniyoor, S. Ramaprabhu / Electrochimica Acta 72 (2012) 199–206
205
Fig. 8. Catalytic surface area. Cyclic voltammograms for f-MWNTs (green dotted
line), Pt (red dashed line) and Pt/f-MWNTs (blue solid line) electrodes, obtained
at scan rate of 50 mV s−1. The counter electrode was a Pt wire and the reference
electrode was Ag/AgCl. The electrolyte consisted of 5 mM LiI, 0.5 mM I2 and 0.1 M
LiClO4 in acetonitrile. (For interpretation of the references to color in this figure
legend, the reader is referred to the web version of this article.)
given by Eq. (1), is expected at higher potentials. Another pair, Bo−x
and Bred, corresponding to the oxidation and reduction of I−/I3
couple (Eq. (2)) is expected at lower potentials.
−
3I2 + 2e− ↔ 2I3
(1)
(2)
I3− + 2e− ↔ 3I−
From Fig. 8, it can be seen that Pt and Pt/f-MWNTs exhibit both the
redox waves whereas f-MWNTs do not display any peak other than
Box (the current is too low to be resolved from the background for
the other peaks). This directly suggests that f-MWNTs based elec-
trode has very low catalytic activity. Furthermore, the peak anodic
and cathodic currents in Pt/f-MWNTs are higher than in Pt. This sug-
gests that Pt/f-MWNTs is more catalytic than Pt. However, since
f-MWNTs do not show any catalytic activity, the higher activity
of Pt/f-MWNTs is mainly due to the presence of nanostructured
Pt. It is known that the peak separation between the anodic and
cathodic peaks, Epp, varies inversely as the rate of a reaction [41].
From Fig. 8, it can be seen that Pt and Pt/f-MWNTs have Epp of ∼470
and 380 mV, respectively. This immediately suggests that tri-iodide
reduction happens rapidly at Pt/f-MWNTs surface rather than on
pure Pt which can be again attributed to the presence of large
number of highly catalytic nanosized Pt clusters. This increased
catalytic activity is responsible for higher exchange current den-
sity in Pt/f-MWNTs based electrodes. The Bred peak in Pt/f-MWNTs
is shifted towards positive potential compared to that of Pt which
suggests that there are lower overpotential losses in Pt/f-MWNTs.
The total series resistance values of the electrodes, Pt/f-MWNTs
∼16.6 ꢀ and Pt ∼20.9 ꢀ reaffirm that there is lower overpotential
loss in Pt/f-MWNTs.
Further cyclic voltammetric experiments were carried out by
changing the scan rates. Fig. 9a and b depicts the variation of Epp
and Ip (peak cathodic, Ipc and peak anodic, Ipa currents) vs. square
root of the scan rate (v). It can be seen from Fig. 9a that, as the
scan rate is reduced, Epp decreases. As the scan rate is reduced the
oxidized species have sufficient time to diffuse from the bulk of the
solution toward the electrode and interact with the catalyst leading
to an increase in the catalytic activity. The decrease in Epp with scan
rate is steeper for Pt than for Pt/f-MWNTs, which could be under-
stood on the basis of the following argument. The particle size of Pt
loaded on MWNTs is very less compared to that of the bulk film. As
a result, the Pt nanoparticles have a higher catalytic activity than in
Fig. 9. Rapid electron transfer. (a) Variation of Epp with square root of scan rate and
(b) variation of peak cathodic (Ipc) and anodic (Ipa) currents with square root of scan
rate for Pt (red squares) and Pt/f-MWNTs (blue triangles) suggesting rapid electron
transfer kinetics at Pt/f-MWNTs surface. (For interpretation of the references to color
in this figure legend, the reader is referred to the web version of this article.)
the pure film form. Therefore, even at high scan rates Pt/f-MWNTs
can reduce all the tri-iodide species present near the electrode.
In other words, Pt/f-MWNTs based electrode has rapid electron
transfer kinetics due to the surface area effect ensuing from the
nanoparticle distribution. From Fig. 9b, we see that Ip versus v0.5 is
nearly linear in the case of Pt/f-MWNTs whereas some deviation can
be observed in the case of Pt. Further, from the inset in Fig. 9b we can
see that Ipa/Ipc is nearly constant and closer to 1 in the case of Pt/f-
MWNTs whereas significant deviation from reversibility (Ipa/Ipc = 1)
can be observed in the case of pure Pt electrode. Deviation from
reversibility suggests slower electron transfer kinetics in the case
of Pt. Pt/f-MWNTs based electrode allows rapid electron transfer
kinetics and is therefore reversible. Thus, Pt/f-MWNTs is a bet-
ter electrode for the reduction of tri-iodide in DSSCs due to the
presence of Pt nanoparticles.
4. Conclusions
Nanostructured Pt particles were successfully and uniformly
decorated on f-MWNTs by a rapid and facile method. The DSSC
fabricated with Pt/f-MWNTs as counter electrodes with a Pt
loading of ∼10 ␮g cm−2 showed an efficiency of 5.4%, an improve-
ment by 20% and 135% over the conventional Pt and f-MWNTs

206
A. Kaniyoor, S. Ramaprabhu / Electrochimica Acta 72 (2012) 199–206
counter electrodes based DSSCs. Based on cyclic voltammetric and
electrochemical impedance measurements, the superior perfor-
mance of Pt/f-MWNTs could be attributed to higher catalytic
activity and large accessible surface area of nanosized Pt.
[17] G. Khelashvili, S. Behrens, C. Weidenthaler, C. Vetter, A. Hinsch, R.
Kern, K. Skupien, E. Dinjus, H. Bönnemann, Thin Solid Films 342 (2006)
511–512.
[18] F. Cai, J. Liang, Z. Tao, J. Chen, R. Xu, Journal of Power Sources 177 (2008) 631.
[19] P. Li, J. Wu, J. Lin, M. Huang, Y. Huang, Q. Li, Solar Energy 83 (2009) 845.
[20] S. Iijima, Nature 354 (1991) 56.
[21] P.M. Ajayan, Chemical Reviews 99 (1999) 1787.
Acknowledgments
[22] A. Kaniyoor, R.I. Jafri, T. Arockiadoss, S. Ramaprabhu, Nanoscale 1 (2009) 382.
[23] A. Kaniyoor, S. Ramaprabhu, Carbon 49 (2011) 227.
[24] N. Jha, S. Ramaprabhu, Nanoscale 2 (2010) 806.
The authors thank Indian Institute of Technology Madras (IITM),
Sophisticated Analytical Instrument Facility (SAIF, IITM), Depart-
ment of Science and Technology (DST, India) and Defense Research
Development Organization (DRDO, India) for financial and infras-
tructural support.
[25] T.T. Baby, S. Ramaprabhu, Talanta 80 (2010) 2016.
[26] A.K. Mishra, S. Ramaprabhu, Energy & Environmental Science 4 (2011) 889.
[27] W.J. Lee, E. Ramasamy, D.Y. Lee, J.S. Song, ACS Applied Material & Interfaces 1
(2009) 1145.
[28] M.M. Shaijumon, S. Ramaprabhu, Applied Physics Letters 88 (2006) 253105.
[29] Z. Guo, Y. Chen, L. Li, X. Wang, G.L. Haller, Y. Yang, Journal of Catalysis 276
(2010) 314.
[30] S. Song, Y. Wang, P.K. Shen, Journal of Power Sources 170 (2007) 46.
[31] M.K. Nazeeruddin, A. Kay, I. Rodicio, R. Humphry-Baker, E. Mueller, P. Liska,
N. Vlachopoulos, M. Graetzel, Journal of the American Chemical Society 115
(1993) 6382.
References
[1] B. O’Regan, M. Grätzel, Nature 353 (1991) 737.
[2] N. Papageorgiou, W.F. Maier, M. Gra¨tzel, Journal of the Electrochemical Society
[32] H. Hirua, T.W. Ebbesen, K. Tanigaki, Advanced Materials 7 (1995) 275.
[33] A.A. Mikhaylova, E.K. Tusseeva, N.A. Mayorova, A.Yu. Rychagov, Yu.M.
Volfkovich, A.V. Krestinin, O.A. Khazova, Electrochimica Acta 10 (2011) 3656.
[34] N. Rajalakshmi, H. Ryu, M.M. Shaijumon, S. Ramaprabhu, Journal of Power
Sources 140 (2005) 250.
[35] R. Yu, L. Chen, Q. Liu, K.L. Tan, S.C. Ng, H.S.O. Chan, G.Q. Xiu, T.S.A. Hor, Chemistry
of Materials 10 (1998) 718.
[36] K. Lee, J. Zhang, H. Wang, D.P. Wilkinson, Journal of Applied Electrochemistry
36 (2006) 507.
[37] X. Li, W.X. Chen, J. Zhao, W. Xing, Z.D. Xu, Carbon 10 (2005) 2168.
[38] J. Yang, T.C. Deivaraj, H.P. Too, J.Y. Lee, Langmuir 20 (2004) 4241.
[39] A. Mathew, G.M. Rao, N. Munichandraiah, Materials Research Bulletin 46 (2011)
2045.
[40] K.-C. Huang, Y.-C. Wang, R.-X. Dong, W.-C. Tsai, K.-W. Tsai, C.-C. Wang, Y.-
H. Chen, R. Vittal, J.-J. Lin, K.-C. Ho, Journal of Materials Chemistry 20 (2010)
4067.
[41] J.D. Roy-Mayhew, D.J. Bozym, C. Punckt, I.A. Aksay, ACS Nano 4 (2010) 6203.
[42] S. Gagliardi, L. Giorgi, R. Giorgi, N. Lisi, Th.D. Makris, E. Salernitano, A. Rufoloni,
Superlattices and Microstructures 46 (2009) 205.
144 (1997) 876.
[3] N. Papageorgiou, Coordination Chemistry Reviews 248 (2004) 1421.
[4] A. Hauch, A. Georg, Electrochimica Acta 46 (2001) 3457.
[5] L. Kavan, J.H. Yum, M. Grätzel, ACS Nano 5 (2011) 165.
[6] J.-L. Lan, Y.-Y. Wang, C.-C. Wana, T.-C. Wei, H.-P. Feng, C. Peng, H.-P. Cheng, Y.-H.
Chang, W.-C. Hsu, Current Applied Physics 10 (2010) S168.
[7] P. Calandra, G. Calogero, A. Sinopoli, P.G. Gucciardi, International Journal of
Photoenergy 2010 (2010) 109495.
[8] S. Ito, T.N. Murakami, P. Comte, P. Liska, C. Grätzel, M.K. Nazeeruddin, M. Grätzel,
Thin Solid Films 516 (2008) 4613.
[9] M. Wei, Y. Konishi, H. Zhou, M. Yanagida, H. Sugihara, H. Arakawa, Journal of
Materials Chemistry 16 (2006) 1287.
[10] C. Barb, F. Arendse, P. Comte, M. Jirousek, F. Lenzmann, V. Shklover, M. Gratzel,
Journal of the American Ceramic Society 80 (1997) 3157.
[11] M. Wang, A.M. Anghel, B. Marsan, N.C. Ha, N. Pootrakulchote, S.M. Zakeeruddin,
M. Grätzel, Journal of the American Chemical Society 131 (2009) 15976.
[12] T.N. Murakami, M. Grätzel, Inorganica Chimica Acta 361 (2008) 572.
[13] M. Wu, X. Lin, T. Wang, J. Qiu, T. Ma, Energy & Environmental Science 4 (2011)
2308.
[43] S.S. Jeon, C. Kim, J. Ko, S.S. Im, Journal of Physical Chemistry C 115 (2011) 22035.
[44] R. Trevisan, M. Döbbelin, P.P. Boix, E.M. Barea, R.T. Zaera, I.M. Seró, J. Bisquert,
Advanced Engineering Materials 1 (2011) 781.
[45] B. Fang, S.-Q. Fan, J.H. Kim, M.-S. Kim, M. Kim, N.K. Chaudhari, J. Ko, J.-S. Yu,
Langmuir 26 (2010) 11238.
[14] P. Joshi, Y. Xie, M. Ropp, D. Galipeau, S. Bailey, Q. Qiao, Energy & Environmental
Science 2 (2009) 426.
[15] Q. Qin, J. Tao, Y. Yang, Synthetic Metals 160 (2010) 1167.
[16] S.S. Kim, K.W. Park, J.H. Yum, Y.E. Sung, Solar Energy Materials and Solar Cells
90 (2006) 283.