Journal of ELECTRONIC MATERIALS, Vol. 44, No.

4, 2015
DOI: 10.1007/s11664-015-3636-5
 2015 The Minerals, Metals & Materials Society

A Comparative Study of Nanostructured TiO2, ZnO and Bilayer

TiO2/ZnO Dye-Sensitized Solar Cells

MAMTA RANI1,2 and S.K. TRIPATHI1,3,4

1.—Department of Physics, Centre of Advanced Study in Physics, Panjab University, Chandigarh,

Chandigarh 160014, India. 2.—Department of Physics, DAV University, Jalandhar, Punjab 144012,
India. 3.—e-mail: surya@pu.ac.in. 4.—e-mail:surya_tr@yahoo.com

Titanium dioxide (TiO2), Zinc oxide (ZnO) and bilayer TiO2/ZnO (TZO) based
cells have been developed and sensitized with five organic dyes and one cocktail
dye composed of five dyes. Photovoltaic performance of TiO2 and ZnO solar cell
sensitized with six dyes is compared to that of bilayer TZO cells. The forward
current is found to increase with applied voltage in the range V £ 0.4 V, which
is dominated by thermionic emission, whereas in 0.4 £ V £ 0.7 V, the current
transport is due to space charge-limited current controlled by exponential trap
distribution in all devices. The combined properties of the materials enhance
the efficiency of composite TZO cells. TiO2 permits the formation of an energy
barrier at the ZnO electrode/electrolyte interface, which reduces the back
electron transfer from the conduction band of ZnO to I3 in the electrolyte.
Also, due to the TiO2 layer on the ZnO, the latter forms a compact layer be-
tween flourine-doped tin oxide (FTO)/TiO2 which benefits the fast electron
transfer from TiO2 to ZnO to FTO glass. This reduces the charge recombination
occurring at the ZnO/FTO interface leading to higher open circuit voltage (Voc),
higher short circuit current (Jsc), lower series resistance (Rs), and in turn
higher efficiency in TZO solar cells as compared to ZnO cells. Among the six
dyes, Eosin-Y and Rose Bengal dye gave the best performance as sensitizers
with TZO.

Key words: TZO, dye-sensitized solar cell, conversion efficiency, organic dye

INTRODUCTION and suitable energy band positioning ensure the

regeneration of the dye as well as diffusion of the
Dye-sensitized solar cells (DSSCs) have attracted
charge. The oxide material plays a fundamental role
great attention as a promising photovoltaic tech-
in device functioning. It acts as a substrate for the
nologies alternative to thin films solar cells, owing
dye, controls the energy alignment of the excited
to their low cost production, usage of medium purity
state of the dye, which eventually controls the
materials and impressive incident photons to elec-
charge transfer, and it is the main source of electron
trical current conversion efficiency and long-term
recombination. Binary metal oxides such as TiO2,
stability.1,2 A typical DSSC is an electrochemical
ZnO, Nb2O5, Fe2O3, ZrO2, Al2O3, CeO2 and ternary
device, which is composed of a nanocrystalline
compounds such as SrTiO3, Zn2SnO4 have been
semiconductor film, sensitizing dye, electrolyte, and
tested for their use as photoelectrodes in DSSC.3
counter electrode. The dye is promoted into an
Among them, the d-block binary metal oxides, TiO2,
excited state upon light absorption from where an
ZnO, and Nb2O5 are the best candidates to be pho-
ultrafast electron injection process takes place into
toelectrodes due to the dissimilarity in orbitals con-
the conduction band of the semiconductor. The
stituting their conduction band and valence band.3
presence of the redox couple in the electrolyte
The high efficiency of these DSSCs depends upon
(Received May 27, 2014; accepted January 8, 2015;
the effective separation and transportation of the
published online February 7, 2015) charges generated in the dye molecules over the

1151
1152 Rani and Tripathi

semiconducting nanoparticles. However, the energy ratio of Ti(OC4H9)4:H2O:C2H5OH was about 1.3:
conversion efficiency of DSSCs is still inferior to 320.6:3.9. The above solution was aged at 353 K for
conventional Si and thin films solar cells. This is a 30 min. A clear solution was obtained after 7 days.
hindrance to their commercialization. In this con- The filtered residue was dried at 353 K for 6 h and
text, the development of a new sensitizer has played then ground to a fine powder to get a dried gel. The
a decisive role in enhancing overall efficiency and dried gel was calcined in air for 6 h at 873 K to
device stability, but the theoretical limit of 26.8%4 obtain nanosized TiO2 powder. Selection of tem-
for DSSC remains far from reach. An additional perature range from 873 K is based on all organic
factor that influences the efficiency of a DSSC is the residue burning up at 673 K, while above 873 K,
open circuit voltage (Voc). To enhance efficiency, anatase starts to convert into rutile phase, having
some researchers have attempted to improve the less porosity, as reported in literature. The high
properties of the TiO2 electrode5–7 or used a bilayer porosity is a crucial factor for adsorption of dye. In
system.8,9 This bilayer electrode is obtained by our previous study,28 the x-ray diffraction (XRD)
coating another metal oxide on a base metal oxide and transmission electron microscopy (TEM) of
electrode. This reduces the recombination rate of TiO2 at 673 K, 773 K and 873 K was mentioned.
photo-injected electrons by efficient spatial separa- The XRD pattern of TiO2 nanoparticles annealed at
tion of photo-injected electrons, while some authors those temperatures confirmed the anatase phase of
have reported improvements by appropriate modi- TiO2. Transformation was confirmed by the
fications of the electrolyte or dye.10–12 appearance of a detectable rutile peak by XRD at
As a result of a considerable review of experi- 873 K. The TEM image of anatase phase of TiO2
mental data, it has been suggested that the electron obtained at 673 K consisted of well-dispersed
transport in nanocrystalline TiO2 films may be the nanorods with an aspect ratio of about 7:1. The TEM
factor that limits the overall efficiency of TiO2 based image of TiO2 nanorods calcined at 773 K revealed
solar cells in spite of the high porosity of TiO2.13,14 the gradual increase in both diameter and length.28
Many attempts have been made to improve the ori- Further increasing the calcination temperature to
ginal DSSC design and circumvent its inherent 873 K, small nanorods agglomerated to form large
limitations, especially the low conductivity of the nanorods, which showed more porous irregular
TiO2 nanoparticle film and potential charge recom- networks as is clear from the TEM image at 100 nm
bination at the TiO2 surface.15 ZnO is an attractive scale at 873 K. Also, for ZnO at 673 K to 873 K, we
material and has been widely investigated for its obtained the wurtzite phase. TEM and XRD of ZnO
optical, catalytic, conductivity, and photoelectro- at 673 K are reported in Ref. 29. So, the selected
chemical properties and for applications in piezo- calcination temperature was 873 K for the applica-
electric devices,16 ultraviolet (UV) light-emitting tion of TiO2, ZnO and TZO films in DSSC.
diodes,17 in UV sensors18 and in solar cells.19 The A 0.1 M ZnO solution was prepared by dissolving
electron mobility is much higher in ZnO than in zinc acetate dehydrate in methanol at room tem-
TiO2.20 To get the benefits of both constituent prop- perature. The pH value of the solution was adjusted
erties, i.e. high porosity of TiO2 and high electrical to 10.5 using sodium hydroxide (1 M) solution. The
conductivity of ZnO, bilayer film has been prepared. resulting filtrate was kept for 48 h to complete the
Apart from this, the CB edge of both oxides at the gelation and hydrolysis process. The ZnO gel was
same level21 acts as an added advantage in a bilayer first dried at 373 K for 2 h and then annealed in air
structure. In our report on the fabrication, charac- at 723 K for 30 min. PEG (polyethylene glycol) was
terization and photoconducting behavior of a ZnO, added to the TiO2 and ZnO nanopowder separately
TiO2 and TZO bilayer system, it has been demon- to prevent the aggregation of the particles and to
strated that the TZO is highly photoconducting and improve the porosity of the films. TZO film was
photosensitive at room temperature, which can lead prepared by coating TiO2 nanorods on prepared
to high Jsc in solar devices.22 The composite ZnO/ ZnO film and sintering at 723 K. The thickness of
TiO2 films have been developed with the aim of the sintered TiO2, ZnO films was 5 lm and that of
producing solar cells with improved photovoltaic TZO film was 10 lm. Six organic dyes, Eosin Y (EY),
properties compared to those displayed by either Fast Green (FGF), Rose Bengal (Rose), Rhodamine
ZnO or TiO2 cells. All cells presented here have been B (RhB), Acridine orange (AO), and Cocktail dye (C)
sensitized with the six low-cost organic dyes, which were used for sensitizing the metal oxide films. The
are commonly used with TiO223,24 and ZnO.25–27 dyes were dissolved in ethanol having a concentra-
tion of 3.0 9 104 M. Cocktail dye solution was
prepared by mixing different dyes at specific con-
EXPERIMENTAL DETAILS
centrations as reported by Seema et al.30 Dye-sen-
ZnO, TiO2 and bilayer TZO films were prepared sitized photoelectrodes were prepared by dipping
as by using the doctor blade method. For this, the the film in the different dye solutions for 48 h. The
colloidal solution is made by dissolving 16 mL of films were then rinsed with ethanol. A sandwich-
titanium butoxide (Ti(OC4H9)4) in 8 mL ethyl alco- type DSSC was fabricated with the dye-sensitized
hol (EtOH). The solution was added dropwise to a oxide electrode and platinum-coated FTO as counter
solution of H2O and concentrated H2SO4. The mole electrode and electrolyte. The counter electrode was
A Comparative Study of Nanostructured TiO2, ZnO and Bilayer TiO2/ZnO 1153
Dye-Sensitized Solar Cells

prepared by thermal decomposition of hexacloro- 615 nm. When the dyes are attached to the oxide
platinic acid solution in isopropanol on FTO. The films, the absorption peaks are shifted as compared
electrolyte solution was a mixture of 0.5 M KI and to the spectra in solution. The device sensitized with
0.05 M iodine in an ethylene carbonate and aceto- cocktail dye has a much stronger optical absorption
notrile mixed solvent (60:40 by volume). compared with the five dyes in the broad visible
UV–Visible absorption spectra of all the dye-sen- region. The shifts of the absorption spectra by
sitized metal oxide photoelectrodes were recorded adsorption on the oxide surface have also been
on a Perkin Elmer LS-35 spectrometer. Solar cell observed in other organic dyes31 and may be
performance was studied in the standard solar cell attributed to a strong interaction between the
laboratory of the Solar Energy Centre (SEC), Min- anchoring group of the dye and the semiconductor
istry of New and Renewable Energy, Gurgaon, surface.2 This kind of interaction also leads to an
using a class AAA solar simulator model SP1000- aggregated state of the dye on the film. Although
4966-based Solar Cell Tester of M.S Oriel, USA, to p–p stacking of organic dyes is advantageous for
study under standard test conditions (STC). Photo- light harvesting because of its broad features in the
electrochemical data were measured using a 1600- UV–Vis absorption spectrum, it is important to
W Xenon light source that was focused to give emphasize that only the dye covalently attached to
1000 Wm2 equivalent to 1 sun at AM 1.5 at the the oxide surface by the anchoring group will con-
surface of the test cell. The working distance of the tribute to the change in the conductivity of the
lamp was 12.00 ± 0.5 cm. Current–voltage (I–V) device. UV–Vis absorption spectra of all dye-sensi-
characteristics of the cell were recorded using a tized oxide films demonstrate that the dyes are com-
four-point probe Keithley source meter (Model pletely anchored with all the ZnO and TZO films and
2400). The active area was typically 1 cm2. that these compounds absorb light of appropriate
wavelengths which indicates that they can be used as
RESULTS AND DISCUSSION photosensitizers for wide band gap semiconductors
which on their own cannot adsorb visible light.
Fourier transform infrared (FTIR) results confirm
Figures 3, 4 and 5 represent the dark J–V char-
the formation of TZO film.22 Direct band gap energy
acterstics on a semi-logarithmic scale of the solar
for ZnO, TiO2 and TZO film was 3.2 eV, 3.59 eV and
cells made with metal oxide/dye/KI using TiO2, ZnO
3.0 eV, respectively, as reported earlier.22,28,29 TZO
and TZO (TiO2/ZnO) and sensitized with six dyes,
had broader absorption as compared to ZnO. ZnO
EY, FGF, Rose, RhB, AO and C. The current den-
and TZO nanopowder films were sensitized with the
sity–volatge (J–V) characteristics of the prepared
five types of commercially available organic dyes,
DSSC were found to be non-linear. The deviation
EY, FGF, Rose, RhB, and AO, and the C dye com-
from linearity can be attributed to the series resis-
posed of these five dyes. The absorption spectra of
tance, bulk resistance and the interface state prop-
these dyes in solution and adsorbed on the ZnO and
erties of the DSSC device. Resistance is the
TZO films are displayed in Figs. 1 and 2. In solu-
important factor for scaling the performance of
tion, the absorption peaks are observed at 529 nm
metal-oxide/dye/KI soalr cells. The existence of an
for EY, 617 nm for FGF, 558 nm for Rose, 544 nm
interfacial layer and series resistance affects the
for RhB, and 490 nm for AO. The Cocktail dye has
ideality factor of the diode. From the dark charac-
three absorption maxima at 495 nm, 539 nm and
terstics of solar cells, the values of built-in voltage

Fig. 1. Normalized UV–Vis absorbance spectra of six dyes in etha- Fig. 2. Normalized UV–Vis absorbance spectra of six dyes in etha-
nol solution and adsorbed on a ZnO electrode. nol solution and adsorbed on a TZO electrode.
1154 Rani and Tripathi

Fig. 5. Semi-logarithmic plot of the dark I–V characteristics of the

prepared TZO dye-sensitized solar cells.
Fig. 3. Semi-logarithmic plot of the dark I–V characteristics of the
prepared TiO2 dye.

and A is the device area. There is a lot of variation in

the electron effective mass value of TiO2. Monticone
et al. take the anatase TiO2 electron effective mass
me = 10 mo.34 On the one hand, Pascual and co-
workers obtained me = 3 m0,35 whereas Stamate
reported a value of me = (0.71–1.26) mo.36 We con-
sidered the intermediate value (me = 3 mo). Effec-
tive mass of ZnO is 0.29 mo and TZO is 1.65 mo. The
values of the Richardson constant using the above
value of effective mass for TiO2, ZnO and TZO are
35.12, 363.33 and 199.23, respectively.
On taking the log on both side of Eq. 1, we get
q/B
ln Is ¼ ln AA T 2  ; (2)
kB T
which gives
kB T AA T 2
/B ¼ ln : (3)
Fig. 4. Semi-logarithmic plot of the dark I–V characteristics of the q Is
prepared ZnO dye.

The value of /B calculated using the above equa-

(Vbi), breakdown voltage (VBd) and breakdown cur- tion is tabulated in Table I. The ideality factor
rent (IBd) were found and are listed in Table I. higher than unity represents the non-ideal behavior
Different regions of the J–V curve are dominated of J-V characteristics at lower voltages. The non-
by different loss mechanisms. So, Eq. 1 is fitted with ideality decreases the photocurrent in the solar cell
the lower part of the dark J–V curve in the region by diffusion of the iodide or tri-iodide ions within the
(V < 0.4 V). The plot of ln (I) versus V is a straight nano-structured oxide film.37 The higher value of n
line. The value of slope = q/nkBT gives the value of may be due to the potential drop in the interfacial
ideality factor and the y-intercept gives the value of layer and the presence of excess current, and to the
Is.32 The values of Is and n for all the diodes are recombination current through the interfacial
listed in Table I. states between the metal insulator layers. There are
The saturation current is given by33 different possible factors which can affect the ide-
  ality values such as: series resistance, image forces,
q/ tunnelling, generation-recombination, interface
Is ¼ AA T 2 exp  B ; (1) impurities and interfacial oxide layer.38,39
kB T
The double logarithmic plots of forward J–V
where A* = 4pqm*k2B/h3 is the Richardson constant, charateristics in voltage range (0.4 < V £ 0.7) are
m* is the effective mass of the carriers, /B is the shown for TiO2, ZnO and TZO in Figs. 6, 7, and 8,
potential barrier height, T is the device temperature respectively.
A Comparative Study of Nanostructured TiO2, ZnO and Bilayer TiO2/ZnO 1155
Dye-Sensitized Solar Cells

Table I. Current–voltage (J–V) parameters: ideality factor (n), potential barrier height (/B ), reverse
saturation current (Is), built-in voltage (Vbi), breakdown volatge (VBd), SPLC exponent (m), and
characteristic temperature (Tt)
Samples n /B (eV) Is (A) Vbi (V) VBd (V) m Tt (K)

TiO2/EY 4.42 0.58 7.37 9 103 0.22 0.43 2.52 1518

TiO2/FGF 1.85 0.73 1.73 9 105 0.32 0.32 4.43 1029
TiO2/Rose 1.49 0.77 3.06 9 106 0.28 0.63 6.06 454
TiO2/RhB 1.84 0.69 8.87 9 105 0.37 0.40 5.10 1231
TiO2/AO 2.60 0.60 3.27 9 103 0.30 0.30 4.16 949
TiO2/C 1.98 0.69 8.96 9 105 0.42 0.50 5.72 1415
ZnO/EY 1.78 0.79 1.90 9 106 0.35 0.53 6.57 1671
ZnO/FGF 1.26 0.81 1.05 9 106 0.36 0.60 7.38 1914
ZnO/Rose 2.33 0.75 9.15 9 106 0.26 0.63 3.21 663
ZnO/RhB 1.39 0.75 1.05 9 105 0.24 0.61 6.03 1509
ZnO/AO 1.26 0.76 7.25 9 106 0.30 0.61 5.89 1467
ZnO/C 1.10 0.80 1.97 9 106 0.34 0.65 8.08 2124
TZO/EY 1.66 0.76 2.88 9 106 0.24 0.58 3.28 685
TZO/FGF 2.03 0.73 9.38 9 106 0.20 0.55 1.45 136
TZO/Rose 1.35 0.74 8.08 9 106 0.30 0.66 6.15 1545
TZO/RhB 1.36 0.73 1.33 9 105 0.29 0.58 4.98 1196
TZO/AO 2.03 0.75 4.18 9 106 0.20 0.64 2.79 539
TZO/C 1.46 0.77 2.44 9 106 0.25 0.59 4.07 921

Fig. 6. Double-logarithmic plot of the dark forward I–V characteris- Fig. 7. Double-logarithmic plot of the dark forward I–V characteris-
tics of the prepared TiO2 dye-sensitized solar cells. tics of the prepared ZnO dye-sensitized solar cells.

The results exhibit a linear relationship up to a

certain voltage range and the current follows the gap. Semiconductor localized states act as carrier
voltage dependence I a Vm. m  1 corresponds to the trapping centers. Carriers from electrodes trapped
Ohmic conduction, while the m ‡ 1 case is inter- in these centers and these localized states become
preted as an indication of the space charge-limited charged. Subsequently, a space charge region would
conduction mechanism (SCLC). The value of expo- be built up in the material, which has a strong effect
nent m for each sample is reported in Table I. The on the J–V characteristics of a device, especially at
value of m is greater than 2 for all devices except for higher bias levels. So, at high voltage, the trapped
the TZO/FGF-based solar cell. This means that the carriers contribute to the conduction paths and
current in the device is controlled by the SCLC reliable information about the density of states can
mechanism dominated by exponential trap distri- be deduced from the J–V characteristics. In this
bution. The semiconductors may contain a large model, the relationship between voltage and current
number of localized defect states in their forbidden is expressed as.40,41
1156 Rani and Tripathi

Fig. 8. Double-logarithmic plot of the dark forward I–V characteris-

tics of the prepared TZO dye-sensitized solar cell. Fig. 9. Light I–V curve of the TiO2 dye-sensitized solar cells.

 l
elN 22o
I¼ V lþ1 ; (4)
d2lþ1 ePo kB Tt
where e is the dielectric constant of the effective
layer, e is the electronic charge, l is the mobility of
carrier charges, N is the effective density of states in
electronic band edge, d is the thickness, and l is the
ratio Tt/T where Ti and Tt represents the room and
characteristic temperature of the exponential trap
distribution, respectively. Tt is given by
 
E
Pð EÞ ¼ Po exp ; (5)
kB Tt
where P(E) is the concentration of traps per unit
energy range above the valence band edge. The total
concentration of traps Nt is given as follows: Fig. 10. Light I–V curve of the ZnO dye-ensitized solar cells.

Nt ¼ Po kB Tt (6)
The value of parameters Nt, Tt, l and Po can be
found by studying the variation of LnI versus 1/T at
different voltages.
The J–V curves for all dye-sensitized TiO2, ZnO
solar cells and dye-sensitized TZO under illumina-
tion are shown in Figs. 9, 10 and 11, respectively.
The figures of merit (Voc, Jsc, FF, Jmax, Pmax, g, Rs,
Rsh) for these solar cells are listed in Table II. Sev-
eral conclusions can be drawn from these J–V
characteristics. The addition of TiO2 to ZnO layer
increases the Voc and Jsc as compared to pure TiO2-
and ZnO-based solar cells with all organic dyes
except the cocktail dye.
Figure 12 shows different processes involved in
charge injection from metal oxide into liquid elec-
trolyte in DSSCs. Electrons can be injected from
excited dye molecules to the conduction band of ZnO
and from the FTO to trap states of ZnO. For solar
cell applications, the injection of these electrons into Fig. 11. Light I–V curve of the TZO dye-sensitized solar cells.
A Comparative Study of Nanostructured TiO2, ZnO and Bilayer TiO2/ZnO 1157
Dye-Sensitized Solar Cells

Table II. The figures of merit for all prepared DSSCs

Sample Voc (Mv) Jsc (mA/cm2) Jmax (mA/cm2) Pmax (mW/cm2) F.F (%) g (%) Rs (X) Rsh (X)

TiO2/EY 66 0.241 0.150 0.005 31 0.005 665 1845

TiO2/FGF 182 0.145 0.099 0.010 38 0.010 325 2178
TiO2/Rose 116 0.096 0.056 0.004 36 0.004 552 1650
TiO2/RhB 223 0.301 0.272 0.036 53 0.036 250 2560
TiO2/AO 55 0.248 0.150 0.005 37 0.005 455 1877
TiO2/C 121 0.028 0.026 0.002 30 0.002 670 1478
ZnO/EY 319 0.431 0.337 0.074 54 0.08 178 8862
ZnO/FGF 203 0.120 0.076 0.010 39 0.01 731 4388
ZnO/Rose 296 0.343 0.237 0.041 40 0.04 391 7162
ZnO/RhB 133 0.157 0.091 0.010 33 0.01 490 1490
ZnO/AO 218 0.165 0.102 0.014 39 0.02 524 2410
ZnO/C 357 0.476 0.320 0.073 43 0.07 295 2618
TZO/EY 379 1.027 0.803 0.213 55 0.21 85 3532
TZO/FGF 239 0.256 0.160 0.024 39 0.02 363 2540
TZO/Rose 298 0.688 0.459 0.089 44 0.09 139 1579
TZO/RhB 143 0.205 0.104 0.008 28 0.01 438 721
TZO/AO 283 0.177 0.106 0.019 39 0.02 484 2632
TZO/C 261 0.264 0.160 0.026 38 0.03 405 2003

the electrolyte must be suppressed under dark or ZnO cells is indicative of faster electron recombi-
illuminated conditions. Under normal DSSC oper- nation with the electrolyte than the bilayer struc-
ation, photo-generated carriers ideally collect at the ture.42 The observed decrease in dark current is
FTO, but carrier loss can occur when they are essentially due to the suppression of I3 reduction
injected into the electrolyte and collect at the plat- at the dye-sensitized TZO electrode/electrolyte
inum-coated FTO. As discussed in Ref. 28, surface interface. Also, it is reported in the literature that
states of metal oxide material have played a large bilayer films have higher IPCE values due to better
role in the charge transport process. So, we believe charge separation by a fast electron transfer process
that the injection of carriers from the conduction using two semiconductors with different conduction
band of metal oxide into the electrolyte happens band edges and energy positions. The use of nano-
through surface states at the metal oxide/dye/elec- crystalline TiO2 permits the formation of an energy
trolyte interface. Due to the large surface area of the barrier at the ZnO electrode/electrolyte interface
device and also very small photo-absorption in oxide due to the fact that TiO2 conduction band edge is
film, carrier loss due to recombination in the bulk of more negative than ZnO. This energy barrier
oxide is much less than that of the carrier loss reduces the back electron transfer from the con-
through the interface. Recombination at the inter- duction band of ZnO to I3 in the electrolyte, thus
face mainly occurs through surface defects in two reducing the recombination rate and improving the
ways. Firstly, this is by the transition of an electron cell performance. Also, from XRD analysis,22 we
from the conduction band of oxide to the highest found that TZO has a large crystallization as com-
occupied molecular orbitals (HOMO) state of an pared to ZnO and TiO2, which is in favor of reducing
oxidized dye molecule through diffusion and surface the energy losses of electron travelling in the
states, as discussed in our earlier report.28 Sec- semiconductor photoelectrodes film. TZO film has
ondly, an electron is trapped at oxide surface states, broader UV absorption range.22 From photocon-
which then react with I3 ions and form I ions ductivity measurements, we found that TZO has
(reduction process in the electrolyte). Recombina- higher photosensitivity as compared to TiO2 and
tion of electrons at the oxide/electrolyte occurs ZnO.22 So, another explanation is that, due to the
through surface states. In both cases, carrier loss is TiO2 layer on ZnO, ZnO forms a compact layer
due to surface states. between FTO/TiO2 which benefits the fast electron
Series resistance is less in the case of TZO as is transfer from TiO2 to ZnO to FTO glass and reduces
clear from Table II. From a lower charge transfer the charge recombination occurring at the ZnO/FTO
resistance, a large electron injection driving force interface. ZnO cells have higher efficiency than
and, as a consequence, a larger Jsc can be obtained. TiO2-based cells. In general, xanthene dyes such as
The J–V performance of the dye-sensitized TZO Rose or EY have a better performance when used to
electrode shows a decrease in the recombination sensitize ZnO rather than TiO2 cells.43 From I–V
rate, with a large increase in device power conver- measurements, it has been found that EY dye is the
sion efficiency. The dark current of TZO film is less most photosensitive dye for TZO and ZnO-based
than TiO2 and ZnO, when compared in dark J–V solar cells. Rose dye is the next most photosensitive
graphs (Figs. 3, 4, 5). The higher dark current in the dye that provides good photocurrent value. C dye is
1158 Rani and Tripathi

Fig. 12. Schematic representation of energy levels of FTO/ZnO/dye/electrolyte for significant (1) photo excitation of electrons into surface states,
(2) photo excitation of electrons from HOMO to LUMO (lowest unoccupied molecular orbital) of the dyes, (3) injection from LUMO of dyes to ZnO
CB ,(4) transport of electrons through electron traps to HOMO of dyes, (5) subsequent electron transfer between distributed surface states and
recombination with holes in the valence band, (6) hole capture and thermal detrapping processes, (7) injection carrier from surface states of
metal oxide to the electrolyte, and (8) electron injection from FTO to trap states of ZnO.

more photosensitivite for ZnO-based solar cells, but iodide (KI), but the solubility of KI in organic eth-
photocurrent decreases for TZO-based cells. The ylene carbonate (EC) was quite low. So, a higher
photocurrent of AO-sensitized ZnO is insufficient concentration of KI in EC could not be achieved.
due to insufficient bonding between ZnO and AO, Due to the low solubility of the KI, there are fewer
because of the absence of any anchoring group,44 iodide ions in the electrolyte resulting in a low
although the efficiency of cells is not high as these concentration of tri-iodide in the cell and, conse-
photoelectrodes were prepared 2 months before quently, in the production of a relatively low cur-
their characterization. It is well known that most of rent. So, an interesting step forward from this
the organic dyes undergo degradation when they developed work, to improve the solar cell efficiency,
are adsorbed on the surface of a semiconductor, will be to replace KI with organic tetrapropyl-
such as TiO2 or ZnO, which can cause catalytic ammonium iodide (TPAI) and to incorporate Ag and
photodegradation. In addition, reactive oxygen Au nanoparticles in highly efficient TZO film, which
species (mainly oxygen ions) may be formed with will definitely improve the value of Jsc and Voc.
time, which react with oxidized dye molecules under
visible or ultraviolet light irradiation in the pre-
CONCLUSIONS
sence of TiO2 particles.45,46 As a result of this
reaction, the number of oxidized dye molecules may In summary, the photoelectrochemical perfor-
reduce. Dye degradation also depends on the mance of porous ZnO, TiO2 and bilayer TiO2/ZnO
adsorption sites47 and constitutes a major problem, (TZO) solar cells have been investigated by
not only for the simple organic dyes but also for the employing five organic dyes, plus a cocktail dye
highly efficient metal complex dyes.48 composed of the five dyes, as sensitizers. UV–Vis
The Voc is related to the interfacial charge absorption spectra of all dye-sensitized oxide films
recombination process between the dye-sensitized confirms the anchoring of the dyes with all the ZnO,
heterojunction and electrolyte. In open circuit con- TiO2 and TZO films. The device mechanism is con-
ditions, all photo-generated carriers recombine trolled by thermionic emission at lower voltages,
within the solar cell diode.49 Thus, if recombination followed by space charge-limited current at higher
can be minimized, Voc can approach more closely to voltages. It was observed that TZO films have a
the maximum values. Also, we used potassium higher efficiency with the five organic dyes than
A Comparative Study of Nanostructured TiO2, ZnO and Bilayer TiO2/ZnO 1159
Dye-Sensitized Solar Cells

ZnO and TiO2 solar cells, which is due to improved 17. H. Guo, J. Zhou, and Z. Lin, Electrochem. Commun. 10, 146
values of Voc and Jsc. This may be due to the for- (2008).
18. W. Wu, S. Bai, N. Cui, F. Ma, Z. Wei, Y. Qin, and E. Xie, Sci.
mation of energy barriers at the ZnO electrode/ Adv. Mater. 2, 402 (2010).
electrolyte interface because of the presence of the 19. A. Umar, M.S. Akhtar, S.H. Kim, A. Al-Hajry, M.S.
TiO2 layer. This reduces the back electron transfer Chauhan, and S. Chauhan, Sci. Adv. Mater. 3, 695 (2011).
from CB of ZnO to I3 in the electrolyte, thus 20. T.P. Chou, Q.F. Zhang, and G. Cao, J. Phys. Chem. C 111,
18804 (2007).
reducing the recombination rate and improving the 21. Y. Xu and M.A.A. Schoonen, Am. Mineral. 85, 543 (2000).
cell performance. Also, TZO film has high crystal- 22. M. Rani and S.K. Tripathi, Energy Environ. Focus 2, 227
linity, broad UV absorption and higher photosensi- (2013).
tivity as compared to ZnO film. ZnO forms a 23. S.S. Mali, C.A. Betty, P.N. Bhosale, and P.S. Patil, Elec-
compact layer between FTO/TiO2 which benefits the trochim. Acta 59, 113 (2012).
24. Z.F. Fang, L.X. Ping, Z.J. Bo, Z.X. Wen, and L. Yuan, Chin.
fast electron transfer from TiO2 to ZnO to FTO glass Sci. Bull. 54, 2633 (2009).
and reduces the charge recombination occurring at 25. N. Khongchareon, S. Choopuna, N. Hongsith, A.
the ZnO/FTO interface. Eosin-Y dye gave the best Gardchareon, S. Phadungdhitidhada, and D. Wongratana-
performance as a sensitizer with both ZnO and TZO. phisan, Electrochim. Acta 106, 195 (2013).
26. K. Hongsith, N. Hongsith, D. Wongratanaphisan, A. Gard-
The lowest efficiency of TiO2-based cells was with chareon, S. Phadungdhitidhada, P. Singjai, and S. Choopun,
Xanthene dye which gives the best results with ZnO Thin Solid Films 539, 260 (2013).
and photocatalytic degradation of dye by TiO2. 27. P.K. Baviskar, J.B. Zhang, V. Gupta, S. Chand, and B.R.
Sankapal, J. Alloy Compd. 510, 33 (2012).
ACKNOWLEDGEMENT 28. M. Rani and S.K. Tripathi, Mater. Sci. Eng. B 187, 214
(2014).
This work is financially supported by University 29. M. Rani and S.K. Tripathi, J. Electron. Mater. 43, 426
Grant Commission (U.G.C.) New Delhi (Major (2014).
Research Project). Mamta Rani would like to 30. S. Rani, P.K. Shishodia, and R.M. Mehra, J. Renew. Sus-
acknowledge the Council of Scientific and Industrial tain. Energy 2, 0431031 (2010).
31. K. Hara, Z.S. Wang, T. Sato, A. Furube, R. Katoh, H.
Research (CSIR), New Delhi, for providing fellow- Sugihara, Y. Dan-oh, C. Kasada, A. Shinpo, and S. Suga,
ship. The authors are grateful to Hemant Singh, J. Phys. Chem. B 109, 15476 (2005).
Indraprastha University, New Delhi, for sharing his 32. J.J. Liou and J.S. Yuan, Semiconductor Device Physics and
research experience, fruitful discussions and sup- Simulations (New York: Plenum Press, 1998), p. 8.
port for solar cell measurements. The authors are 33. S.M. Sze, Physics of Semiconductor Devices, 1st ed. (New
York: Wiley, 1981), p. 146.
also grateful to Solar Energy Centre, MNRE, Gov- 34. S. Monticone, R. Tufeu, A.V. Kanaev, E. Scolan, and C.
ernment of India, for providing the facility for pho- Sanchez, Appl. Surf. Sci. 162, 565 (2000).
tovoltaic measurements. 35. J. Pascual, J. Camassel, and H. Mathieu, Phys. Rev. Lett. 39,
1490 (1977).
REFERENCES 36. M.D. Stamate, Appl. Surf. Sci. 205, 353 (2003).
37. M.K. Nazeeruddin, F. De Angelis, S. Fantacci, A. Selloni, G.
1. B.O. Regan and M. Grätzel, Nature 35, 737 (1991).
Viscardi, P. Liska, S. Ito, B. Takeru, and M. Grätzel, J. Am.
2. A. Hagfeldt, G. Boschloo, L. Sun, L. Kloo, and H. Pettersson,
Chem. Soc. 127, 16835 (2005).
Chem. Rev. 110, 6595 (2010).
38. S. Aydogan, M. Saglam, and A. Türüt, Appl. Surf. Sci. 250,
3. R. Jose, V. Thavasi, and S. Ramakrishna, J. Am. Ceram.
43 (2005).
Soc. 92, 289 (2009).
39. M. Pattabi, S. Krishna, N. Ganesh, and X. Mathew, Sol.
4. B.P. Fingerhut, W. Zinth, and R.D. Vivie-Riedle, Phys.
Energy Mater. Sol. Cells 81, 111 (2007).
Chem. Chem. Phys. 12, 422 (2010).
40. T.G. Abdel Malik and R.M. Abdel-Latif, Thin Solid Films
5. P. Cheng, C.S. Deng, D.N. Liu, and X.M. Dai, Appl. Surf.
305, 336 (1997).
Sci. 254, 3391 (2008).
41. H.M. Zeyada, M.M. El-Nahass, and E.M. El-Menyyawy, Sol.
6. S.S. Kim, J.H. Yum, and Y.E. Sung, Sol. Energy Mater. Sol.
Energy Mater. Sol. C 92, 1586 (2008).
C 79, 495 (2003).
42. M.K.I. Senevirathne, P.K.D.D.P. Pitigala, V. Sivakumar,
7. S. Kushwaha and L. Bahadur, Int. J. Hydrog. Energy 36,
P.V.V. Jayaweera, A.G.U. Perera, and K. Tennakone,
11620 (2011).
J. Photochem. Photobiol. A 195, 364 (2008).
8. J.Y. Liao and K.C. Ho, Sol. Energy Mater. Sol. C 86, 229 (2005).
43. E. Guillén, F. Casanueva, J.A. Anta, A. Vega-Poot, G.
9. A. Kitiyanan and S. Yoshikawa, Mater. Lett. 59, 4038 (2005).
Oskam, R. Alcántara, C. Fernández-Lorenzo, and J. Martı́n-
10. A. Fukui, R. Komiya, R. Yamanaka, A. Islam, and L. Han,
Calleja, J. Photochem. Photobiology A 200, 364 (2008).
Sol. Energy Mater. Sol. C 90, 649 (2006).
44. F. Labat, I. Ciofini, H.P. Hratchian, M.J. Frisch, K.
11. H.K. Singh, S. Aggarwal, D.C. Agrawal, P. Kulria, S.K.
Raghavachari, and C. Adamo, J. Am. Chem. Soc. 131, 14290
Tripathi, and D.K. Avasthi, Vacuum 87, 21 (2013).
(2009).
12. G.C. Vougioukalakis, A.I. Philippoulos, T. Stergiopoulos,
45. F. Zhang, J. Zhao, L. Zang, T. Shen, H. Hidaka, E. Pelizzetti,
and P. Falaras, Coord. Chem. Rev. 255, 2602 (2011).
and N. Serpone, J. Mol. Catal. A 120, 173 (1997).
13. S.A. Haque, Y. Tachibana, R.L. Willis, J.E. Moser, M.
46. T. Wu, G. Liu, J. Zhao, H. Hidaka, and N. Serpone, J. Phys.
Grätzel, D.R. Klug, and J.R. Durrant, J. Phys. Chem. B 104,
Chem. B 102, 5845 (1998).
538 (2000).
47. H. Tributsch, Coord. Chem. Rev. 248, 1511 (2004).
14. J. Nelson, S.A. Haque, D.R. Klug, and J.R. Durrant, Phys.
48. B. Macht, M. Turrion, A. Barkschat, P. Salvador, K. Ellmer,
Rev. B. 63, 205321 (2001).
and H. Tributsch, Sol. Energy Mater. Sol. C 73, 163 (2002).
15. M. Grätzel, J. Photochem. Photobiol. C 4, 145 (2003).
49. G. Boschloo and A. Hagfeldt, Acc. Chem. Res. 42, 1819
16. S.J. Kang, J. Korean Phys. Soc. 47, 589 (2005).
(2009).