Enhanced photovoltaic performance with co-sensitization of quantum dots and an organic dye in dye-sensitized solar cells

Article abstract of DOI:10.1039/c4ta03098b

CdSe quantum dots of two different sizes exhibiting a maximum emission at 495 nm (CdSe₄₉₅) and 545 nm (CdSe₅₄₅) were combined with di-tetrabutylammonium cis-bis(isothiocyanato)bis(2,2-bipyridyl-4,4-dicarboxylato)ruthenium(ii) (N719)

Full text of DOI:10.1039/c4ta03098b

Journal of
Materials Chemistry A
View Article Online
View Journal
PAPER
Enhanced photovoltaic performance with
co-sensitization of quantum dots and an organic
Cite this: DOI: 10.1039/c4ta03098b
dye in dye-sensitized solar cells†
a
b
c
Camilla Lelii, Moungi G. Bawendi, Paolo Biagini, Po-Yen Chen,d
*
*
a
d
a
d
*
Marcello Crucianelli, Julio M. D'Arcy, Francesco De Angelis, Paula T. Hammond
and Riccardo Poc
CdSe quantum dots of two different sizes exhibiting a maximum emission at 495 nm (CdSe495) and 545 nm
(CdSe545 were combined with di-tetrabutylammonium cis-bis(isothiocyanato)bis(2,2-bipyridyl-4,4-
)
dicarboxylato)ruthenium(II) (N719) or 2-cyano-3-{5-[7-(4-diphenylamino-phenyl)benzo[1,2,5]thiadiazol-
4-yl]-thiophen-2-yl}-acrylic acid (TBTCA) resulting in four novel hybrid organic–inorganic sensitizers,
which were used in the fabrication of dye-sensitized solar cells. The results showed that with N719, both
the CdSe dots decreased the power conversion efficiencies when compared to a standard device
consisting only of N719 as the sensitizer. With the organic dye TBTCA, CdSe545 showed no significant
effect, while CdSe495 interacted favorably, leading to a 25% increase in power conversion efficiency
compared to a device sensitized solely by TBTCA. Studies on excited-state lifetimes of N719 in the
presence of CdSe did not distinguish between energy and/or charge transfer mechanisms. On the other
hand, time correlated single photon counting experiments on the photoelectrodes suggest that the
advantages due to the CdSe495–TBTCA combination could be ascribed to FRET from quantum dots to
the organic dye and to a further contribution, as suggested by IPCE spectra, consisting of electron
transfer via cascade from the LUMO level of TBTCA to CdSe495 to TiO2, which produces a higher flux of
electrons in the external circuit.
Received 18th June 2014
Accepted 28th August 2014
DOI: 10.1039/c4ta03098b
www.rsc.org/MaterialsA
and copper indium gallium selenide, as well as via organic
technologies such as bulk heterojunction, bilayered, and dye-
Introduction
sensitized solar cells (DSSCs). Among these photovoltaics,
DSSCs have attracted considerable attention due to an ideal
compromise between efficiency and cost-performance.4–6 The
photoanode of a DSSC is comprised of a nanoporous TiO2 layer
sensitized by photosensitizers which is commonly considered
as a crucial part in this kind of devices. Ideally the photosen-
sitizer might absorb energy both in the whole visible region and
in the near-infrared (NIR) region; moreover its excited state
should be at a higher energy level than the conduction band of
the semiconductor to facilitate electron injection. However, the
traditional dyes generally used in DSSCs suffer from narrow
absorption spectra and/or low molar extinction coefficients,5 so
a new approach has emerged, which further improves the effi-
ciency, by co-sensitizing dyes with small band gap semi-
conductor nanocrystals (QDs). In fact, QDs have advantages,
such as high molar extinction coefficients and large intrinsic
dipole moment; furthermore quantum connement and a
tunable band gap, dependent on the size and shape of QDs,
enable designing light-absorbing materials with tailored optical
properties.7–9 These peculiar characteristics can be exploited in
co-sensitized devices, where QDs play an important role in
charge and/or energy transfer processes.10 The advantages of the
The worldwide energy demand is dramatically increasing, and
primary energy sources are derived from carbon-based fossil
fuels. 90% of greenhouse gas emissions come from the
combustion of fossil fuels, leading to serious climate change.
The lack of clean energy motivates scientists to nd alternative
and eco-friendly energy sources and therefore solar energy is an
ideal candidate because it is clean and freely available.1–3 Har-
vesting energy from the sun can be accomplished via inorganic
solar cells such as single crystal silicon, polycrystalline silicon,
aDepartment of Physical and Chemical Sciences, University of L'Aquila, via Vetoio
(Coppito 1) – 67100, L'Aquila, Italy
bDepartment of Chemistry, Massachusetts Institute of Technology, 77 Massachusetts
Avenue, Cambridge, MA 02139, USA. E-mail: mgb@mit.edu
cResearch Center for Non Conventional Energy, Istituto eni-Donegani, eni S.p.A., via
Fauser, 4 – 28100, Novara, Italy. E-mail: paolo.biagini@eni.com
dDepartment of Chemical Engineering, Massachusetts Institute of Technology, 77
Massachusetts Avenue, Cambridge, MA 02139, USA. E-mail: hammond@MIT.edu
† Electronic supplementary information (ESI) available: Synthetic procedures; 1H
NMR and CV characterization of TBTCA; absorption and emission spectra of
TBTCA, N719, CdSe495 and CdSe545; TEM-EDS of the QD/TiO2 lm; detailed
results of XPS measurements on the QD/TiO2 lm; photovoltaic measurements
and statistical analysis of all the devices prepared. See DOI: 10.1039/c4ta03098b
This journal is © The Royal Society of Chemistry 2014
J. Mater. Chem. A

View Article Online
Journal of Materials Chemistry A
Paper
co-sensitization are several: (i) the problems arising from the
Here we describe devices using iodide-based electrolytes and
internal recombination of QDs can be overcome thanks to the shed light on the conditions under which the co-sensitization
capability of the dye that works as the hole scavenger, leading to between the organic dye (TBTCA) and CdSe QDs can realize. The
better charge separation and higher electron injection yields;9 achieved increase around 25% represents, to the best of our
moreover (ii) tuning the optical properties of the dots according knowledge, the top augment for a device based on a push–pull
to the dye affords the possibility for increasing light absorption organic dye, QD and I2/I3ꢀ redox couple; by comparison we also
in order to match the solar spectrum;10 nally (iii) by combining report the attempts to obtain, under the same conditions, co-
QDs and the dye the interaction between them can lead to FRET sensitization between the same CdSe QDs and the N719 dye.
¨
(Forster Resonance Energy Transfer) processes, where the
former can act as donors and the latter as the acceptor thereby
enhancing the cell efficiency. In this way common dyes for
DSSCs with good charge injection ability but with narrow
absorption spectra or low molar extinction coefficients now can
serve as efficient sensitizers.5,11
Experimental
Measurements
The cyclovoltammetric (CV) characterization was carried
out using an Autolab PGSTAT 12 potentiostat/galvanostat
(EcoChemie, The Netherlands) run by a PC with GPES soware.
XPS measurements were performed on a TiO2 lm (5 mm
thick) with a PHI VersaProbe II XPS instrument. TCSPC
measurements were performed with an apparatus composed as
follow. Excitation source: Picoquant PDL-800B driver with a
LDH-C 400 module; 2.5 MHz pulse train; 400 nm center wave-
length; ꢁ30 ps pulse duration (estimated); 450 nm short-pass
lter (Thor Labs FES0450). Beamsplitter: Pellicle beamsplitter
(Thor Labs BP208); roughly 92% transmission over the visible
spectrum. Sample objective: 4ꢂ Nikon air objective. Emission
lter: 500 nm and 600 nm long-pass (Thor Labs FEL0500 and
FEL0600). Detection: Perkin-Elmer SPCM (single-photon APD);
PicoHarp 300 (timing module). TEM imaging analysis was
performed using a JEOL JEM-2100F Field Emission Electron
Microscope at 200 kV accelerating voltage and illumination
angle a ¼ 1. The CdSe-sensitized TiO2 lm was scratched from
FTO glass and transferred into a vial containing ethanol (1–2 ml)
and sonicated for 5 min. Few drops of this solution were applied
on a formvar/carbon coated copper grid 400 mesh and dried in
air. Electron Probe Micro-Analysis (EPMA) was performed on
the QD-sensitized TiO2 lm cross-sections with a JXA-8200
Superprobe (JEOL), with an operating voltage of 15 kV. The lm
was fabricated by the doctor blade technique on a 1 mm soda-
lime glass substrate. The samples were coated with carbon,
using a high vacuum carbon evaporator and all measurements
were referred to pure element standards.
Previous studies have focused on the interactions between
semiconductor nanocrystals and dyes12–21 as well as the charge
transfer processes.9,22–33 Wang and Liu26 reported co-sensitization
with N719 and PbS quantum dots and they obtained an increasing
efficiency from 5.95%, of device sensitized only with N719, to
6.35% in the co-sensitized device. Moreover, Song et al.28 built the
photoanode of the solar cells using ZnO nanowires coated with
CdS quantum dots and the cis-bis(isothiocyanato)bis(2,2-bipyridyl-
4,40-dicarboxylic acid)ruthenium(II) (N3) dye reporting a conversion
efficiency enhancement from 0.5% of the N3 sensitizer to 1.9% of
the co-sensitizer and attributing this improvement to efficient hole
transfer from CdS to N3, which avoids the exciton recombination
inside the QDs and ensures the electron injection into ZnO
nanowires. Many advantages have been obtained using CdSe or
CdS nanocrystals and squaraine dyes that absorb selectively in the
NIR region34 and can work as an acceptor of both holes and energy
via FRET deriving from QDs. For example Kamat et al.20 proved a
protable energy transfer from CdSe to the squaraine dye that
enhanced the efficiency of the devices from 3.05% in DSSCs to
3.65% in QDDSSCs. Another important result was obtained by
Etgar et al.,21 which enhanced the power conversion efficiency by
47% due to FRET from CdSe quantum dots to the squaraine dye.
Despite considerable improvement, the overall efficiency of
devices sensitized with squaraine dyes and QDs is still rather
low,20,21,30 and these devices use electrolytes based on cobalt or
sulde/polysulde redox couples. Even though the mentioned
redox couples have shownꢀto be, in QD-based devices, more suit-
able than traditional Iꢀ/I3 based electrolytes, the iodide couple
remains a good choice for DSSCs because it is a good regenerator
of the oxidized state of the dye and minimizes the recombination
of the electrons of TiO2 with electrolyte acceptors.5 Only in few
cases traditional push–pull organic dyes, perceived as more effi-
cient than squaraine, have been employed together with QDs.27,29,33
Fan and co-workers29 obtained an improvement of performances
using QDs together with 2-cyano-3-(5-(9-(4-(2,2-diphenylvinyl)
phenyl)carbazol-6-yl)thiophen-2-yl)-acrylic acid (JK28) passing
from efficiency of 0.41% in the devices sensitized with JK28
only to 1.03% in a co-sensitized device. Recently Guo33 obtained
an efficiency up to 7.20% in a co-sensitized device based on
CuInS2 QDs and 2-cyano-3-(5-(50-(5-(4-(diphenylamino)phenyl)
Synthesis
All reactions were performed using standard glass vessels under
an inert nitrogen atmosphere. All the solvents were purchased
from Sigma Aldrich and, unless otherwise specied, they have
been used without further purication. K2CO3 (Carlo Erba),
MgSO4 (Carlo Erba), TiO2 paste, T/SP – 20 nm (Solaronix), N719
(Solaronix), and TBTCA were prepared according to a modi-
cation of the literature procedures;35,36 all the other chemicals
employed here were acquired from Aldrich and they were all
used as received.
Synthesis of CdSe495 and CdSe545 QDs
thiophen-2-yl)-4,40-dihexyl-2,20-bithiazol-5-yl)furan-2-yl)acrylic acid Trioctylphosphine oxide, TOPO (3.0 g, 7.8 mmol), octadecyl-
(BTF), with an increase around 7% with respect to the corre- phosphonic acid, ODPA (0.28 g, 0.84 mmol) and CdO (0.06 g,
ꢃ
sponding device without QDs.
0.47 mmol) were mixed in a 50 ml ask, heated at 150 C and
J. Mater. Chem. A
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
Journal of Materials Chemistry A
exposed to vacuum for ca. 1 hour. Then, under nitrogen, the mismatch between the simulated light and AM 1.5. J–V curves
solution was heated to about 300 ꢃC to dissolve the CdO until it were obtained by applying an external bias to the cell and
turns optically clear and colorless. At this point, 1.5 g (4.0 mmol) measuring the generated photocurrent with a Keithley model
of trioctylphosphine, TOP, was injected into the ask and the 2400 digital source meter. The voltage step and delay time of the
temperature was allowed to recover to the value required for the photocurrent were 10 mV and 40 ms, respectively. IPCE spectra
injection of the Se–TOP solution [Se (0.058 g, 0.73 mmol) + TOP were obtained with a computer-controlled system (Model QEX7,
(0.36 g, 0.97 mmol)]. The injection temperature and the reaction PV Measurements, Inc.) consisting of a 150 W xenon lamp light
time were modied in order to synthesize CdSe dots of different source and a monochromator with two 1200 g mmꢀ1 diffraction
sizes. As an example, for the synthesis of green uorescent CdSe gratings. The monochromator was incremented through the
dots the Se–TOP solution was injected at 380 ꢃC and the heating wavelength range from 350 nm to 900 nm to generate the
mantle was removed immediately aer the injection. On the spectral response of IPCE with a spectral resolution of 10 nm.
other hand, red uorescent CdSe dots were synthesized by The incident photon ux was determined using a calibrated
injecting the Se–TOP at 370 ꢃC and stopping the reaction silicon photodiode (calibrated by PV Measurements, Inc.).
heating only aer 3 minutes. Aer the synthesis, the nano- Measurements were performed under short-circuit conditions,
crystals were precipitated with methanol/butanol; they were while the cell was under background illumination from a bias
puried by repeated redissolution in hexane and precipitation light of 50 mW cmꢀ2. Bias illumination was from the same
with the same mixture of methanol/butanol, and they were direction as the monochromatic light, which was from the FTO
nally dissolved in hexane.37
side. The monochromatic beam was chopped using a computer-
controlled shutter at a frequency of 4 Hz, and averaging of up to
40 shutter cycles was employed.
Fabrication and characterization of solar devices
FTO glass (TEC15; thickness, 2.2 mm; 15 U per square; Pil-
kington, USA) was rst cleaned in a detergent solution using an
ultrasonic bath for 15 min, and then rinsed with water and
Results and discussion
Evaluation of coverage of the TiO2 surface by QDs
ethanol. The FTO glass plates were immersed in a 40 mM A crucial point of QD-sensitized solar cells is the sensitization of
aqueous TiCl4 solution at 80 ꢃC for 30 min and washed with TiO2 with QDs. QDs do not present an anchoring functional
water and ethanol. A layer of TiO2 paste was deposited on the group to bind the TiO2 surface; in addition they are passivated
FTO glass plates ꢃby the doctor blade technique and then dried with organic ligands, which make the sensitization on TiO2 and
for 5 min at 120 C. This procedure was repeated, if necessary, charge transfer processes across the interfaces complicated.
to achieve an appropriate thickness of 10 mm. Aer drying at Several approaches to deposit QDs on the TiO2 lm are
120 ꢃC the lm thus produced was sintered at 500 ꢃC for 30 min. described in the literature:38 Chemical-Bath Deposition (CBD)
ꢃ
Aer cooling at 80 C the lm (about 10 mm thick, measured and Successive Ionic-Layer Adsorption and Reaction (SILAR) are
using a VEECO Dektak 150 prolometer) was immediately in situ techniques, which are able to provide high surface
dipped into a 0.5 mM dye solution in the appropriate solvent coverage and intimate contact with a wide bandgap semi-
(N719 in ethanol and TBTCA in dichloromethane) for 18 h. For conductor, facilitating charge injection processes. However, the
the co-sensitization, the lm was rst immersed in a 1 mM CBD and SILAR processes lead to a polydisperse size distribu-
hexane solution of CdSe QDs for 18 h and then in the afore- tion, which affects the band gap of QDs. Otherwise pre-
mentioned dye solutions.
synthesized monodisperse QDs are usually prepared ex situ and
The counter-electrode was 100 nm-thick platinum, sputtered deposited directly or through a molecular bifunctional linker
on an ITO glass (Delta Technologies). A solution of 0.6 M of 1- onto a semiconductor lm. In this case, the functional molec-
butyl-3-methylimidazolium iodide, 0.03 M of I2, 0.10 M of ular linkers provide a better coverage and they can interfere in
guanidinium thiocyanate and 0.5 M of 4-tert-butyl pyridine charge transfer processes.39–42 So we thought that depositing
acetonitrile/valeronitrile (85 : 15, v/v) has been used as the monodisperse QDs directly onto the TiO2 lm, without any
electrolyte. The dye-adsorbed TiO2 photoanodes and platinum molecular linker, although it is not the optimal method to
counter electrodes were assembled into a sandwich-type cell obtain a complete coverage of the TiO2 layer, is the best method
and sealed with a hot-melt Surlyn resin with a thickness of to study their role in the devices, and in fact it avoids any
25 mm (Solaronix). The TiO2 electrodes had an area of about interference due to the molecular linker in charge separation
0.16 cm2. The aperture of the Surlyn frame was larger than that processes and to the heterogeneity of the properties of QDs
of the TiO2 area by 2 mm. Copper tape was adhered onto the caused by the polydisperse size distribution obtained when
edge of the FTO outside the cell. The position of the tape was other in situ methods (CBD and SILAR) are employed.38,43
1 mm away from the edge of the Surlyn gasket and 4 mm away
from the edge of the TiO2 layer.
We have prepared CdSe QDs according to a standard litera-
ture procedure37 which consists of the reaction of a CdO solu-
Photovoltaic measurements were performed using an AM 1.5 tion, tri-octylphosphine oxide (TOPO) and tri-octylphosphine
solar simulator (Photo Emission Tech.). The power of the (TOP) in octadecylphosphonic acid (ODPA) and by means of an
simulated light was calibrated to 100 mW cmꢀ2 by using a accurate control of the temperature and of the reaction time
reference silicon photodiode with a power meter (1835-C, with a Se–TOP solution, we were able to easily regulate the
Newport) and a reference silicon solar cell to reduce the dimension of the CdSe QDs, which also keep an elevated
This journal is © The Royal Society of Chemistry 2014
J. Mater. Chem. A

View Article Online
Journal of Materials Chemistry A
Paper
uniformity in their size distribution as conrmed by TEM
analysis (Fig. 1), while the determination of the corresponding
amount in solution was determined by the spectroscopic
method.44
QD/TiO2 lms were obtained by prolonged dipping of a TiO2
lm (10 mm thick, sintered on FTO glass) into CdSe QD solution
in hexane (see the Experimental method section for more
details).
To preliminarily assess the uniformity of the CdSe disper-
sion on TiO2 lms, X-ray Photoelectron Spectroscopy (XPS)
characterization of CdSe–TiO2 composite lms (5 mm thick) on
the surface and the depth prole were carried out. The surface
survey, performed on two different spots of the surface, showed
an average atomic percentage of cadmium of 5.9 ꢄ 0.4 versus a
titanium percentage of 15.1 ꢄ 0.1. Depth prole characteriza-
tion has been performed by sputtering the sample with high
energy Ar+ ions. To correlate the sputtering time with the
thickness of the lm, we have controlled the disappearance of
the Ti peak over the sputtering time. The result is that the Ti
peak disappears aer 220 minutes of sputtering and this time
corresponds to 5 mm thickness. On the other hand the analysis
on the CdSe–TiO2 composite lms reveals that the Cd peak
disappears aer just 5 minutes so we can reasonably assume
that CdSe QDs are located in a very thin top layer and consid-
ering that 220 min were needed to remove a 5 mm layer of TiO2
we can assume that CdSe QDs penetrate up only to around 100
nm (see the ESI† for more details).
Fig. 2 SEM image, performed with an electron probe micro-analyzer,
of the cross-section of the CdSe-sensitized TiO2 film (10 mm thick)
deposited on FTO glass.
electron beam used can detect a sample sphere of 2–3 mm in
diameter. Complete results of EMPA are reported in Table 1.
Ultimately, considering the results obtained by XPS and
EMPA we recovered the following information: (i) the amount of
atomic Cd on the top of the TiO2 layer is 5.9%; (ii) QDs are
located within around 100 nm thickness of the TiO2 layer; (iii)
the mean value of atomic Cd in the upper 2 mm is 0.14%. So we
can conclude that under these conditions CdSe QDs cannot
penetrate homogeneously into the whole mesoporous titania
structure.
Evaluation of CdSe QD effects in solar cells
These results can be compared with electron probe micro-
analysis performed directly on the cross-section of the QD/TiO2
composite lms (10 mm thick) (Fig. 2).
The analyzer works similar to a scanning electron micro-
scope (SEM): the sample is bombarded with an electron beam,
emitting X-rays at wavelengths characteristic to the elements
being analyzed. This enables us to determine the abundances
of elements present within small sample volumes (typically 10–
30 mm3 or less).45
As can be seen from the SEM image of the Electron Micro-
Probe Analyzer (Fig. 2) the maximum amount of Cd is on the top
of the photoanode, and its atomic percentage is estimated to be
around 0.14%. This small value may seem apparently incon-
sistent with the XPS results that indicated 5.9% of Cd atoms,
but we condently consider the value obtained with EMPA, an
average in the rst 2 mm of the overall thickness, since the
The fabrication of the devices was performed by conventional
techniques46 and is exhaustively described in the experimental
methods. We tested two different CdSe QDs: CdSe495 and
CdSe545 (the number refers to the wavelength of maximum of
emission), each one with two different dyes: the commercially
available ruthenium complex N719 and an already described35
organic dye (TBTCA), which we prepared according to a partially
modied method allowing us to avoid the use of stannyl
derivatives36 (Fig. 3).
The co-sensitized devices (QDDSSCs) were compared with
the respective cells assembled with dye alone (DSSCs). Table 2
summarizes the main photovoltaic parameters measured in the
presence of Iꢀ/I3ꢀ based electrolyte solutions (see Experimental
methods for more details). Under these conditions we did not
observe any decrease in the efficiency for all the devices during
the characterization (J–V curves and IPCE), which normally was
completed within 3–4 h.
As shown in Table 2, CdSe QDs have different effects on N719
and TBTCA. When CdSe QDs are employed together with N719,
Table 1 Atomic concentration at different depths. Analysis performed
on the cross-section of the 10 micron thick TiO2 film sensitized with
CdSe by Electron Probe Micro-Analysis
Depth (mm) Ti%
O%
Cd%
Se%
2
4
6
8
27.8 ꢄ 0.2 71.8 ꢄ 0.8 0.141 ꢄ 0.005 0.198 ꢄ 0.006
29.2 ꢄ 0.3 70.6 ꢄ 0.8 0.041 ꢄ 0.004 0.081 ꢄ 0.005
29.7 ꢄ 0.3 70.2 ꢄ 0.8 0.025 ꢄ 0.004 0.061 ꢄ 0.005
29.1 ꢄ 0.3 70.9 ꢄ 0.8 0.018 ꢄ 0.004 0.057 ꢄ 0.005
Fig. 1 TEM image of CdSe solution in n-hexane.
J. Mater. Chem. A
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
Journal of Materials Chemistry A
Fig. 3 Structure of dyes employed.
Table 2 Photovoltaic parameters of DSSCs and QDDSSCs
Fig. 4 (a) Comparison of J–V curves of devices sensitized with N719
alone (dark dotted line), N719–CdSe495 (green dotted line), and N719–
CdSe545 (orange dotted line); (b) comparison of IPCE spectra of
devices sensitized with N719 alone (dark dotted line), N719–CdSe495
(green dotted line), and N719–CdSe545 (orange dotted line); (c)
comparison of J–V curves of devices sensitized with TBTCA alone
(dark bold line), TBTCA–CdSe495 (green bold line), and TBTCA–
CdSe545 (orange bold line); (d) comparison of IPCE spectra of devices
sensitized with TBTCA alone (dark bold line), TBTCA–CdSe495 (green
bold line), and TBTCA–CdSe545 (orange bold line).
Entry
Sensitizer
JSC mA cmꢀ2
VOC
V
FF %
h %
1
2
3
4
5
6
7
8
CdSe495
CdSe545
N719
N719–CdSe495
N719–CdSe545
TBTCA
TBTCA–CdSe495
TBTCA–CdSe545
0.3
0.3
13.5
11.6
11.1
8.4
0.62
0.65
0.79
0.76
0.68
0.63
0.67
0.64
61
57
71
75
74
76
77
74
0.1
0.1
7.6
6.6
5.6
4.0
5.0
3.9
9.6
8.3
Finally it should be noted that, in the examined conditions,
the efficiency of the devices with QDs alone was quite low and in
all the considered combinations QD/dye we observed a signi-
cant improvement of the device efficiency. This behaviour had
already been reported.20,21,33,49,50 In particular Guo33 ascribed the
enhancement to the interaction between QDs and dyes (QDs
formed a barrier layer to suppress the recombination from
injection electron to the electrolyte and improve the open-
circuit voltage). On the other hand the <0.1% efficiency
measured for the devices containing only the QD was explained
they indifferently produced a decrease in the device efficiency.
In the literature there are several reports demonstrating that
Ru-dye/QD devices work better than Ru dye devices,47,48 but
there are also a few exceptions. For example, Lin et al.31 attrib-
uted the degradation of the performances of a N719/CIGS QD
solar cell, compared to a pristine N719 device, to an energy level
10
´
misalignment; Mora-Sero et al. ascribed the performance
degradation to the existence of new charge recombination
channels. When CdSe QDs are used with TBTCA, the efficiency
of the co-sensitized device with TBTCA–CdSe495 was increased
to about 25% (h rises from 4.0% to 5.0%). The enhanced effi-
ciency was mainly derived from the improved photocurrent
density (JSC). The open circuit voltage which corresponds to the
difference between the Fermi level of TiO2 and the redox
on the basis of a very low light harvesting capability of the QD.
9
´
However we believe, according to Mora-Sero's observations,
that the issue arises from the effective internal recombination
of QDs, which could be bypassed thanks to the capability of the
dye to work as the hole scavenger, producing better charge
separation and higher electron injection yields.
ꢀ
potential of the Iꢀ/I3 couple5,11 and it can be signicantly
affected when the wide bandgap semiconductor is in contact
with a dye or other materials. In the cases considered here, the
VOC is mainly altered by passing from N719 (0.79 V) to TBTCA
Evaluation of excited state lifetimes
(0.63 V), while the addition of CdSe QDs to the corresponding To study energy transfer kinetics in our devices, we performed
DSSCs generated only small variations, negative for N719 and time correlated single photon counting (TCSPC) experiments on
positive in the case of TBTCA. Fig. 4 shows the J–V curves and the photoanode. The resulting PL lifetime traces are shown in
the IPCE spectra of all the combinations, between N719 and Fig. 5.
TBTCA with both types of CdSe QDs investigated.
We cannot observe marked differences between N719 and
The improved efficiency obtained with the combination N719 + CdSe and in both cases the decay prole clearly shows a
TBTCA–CdSe495 which, as explained above, was essentially due multiexponential behavior (Fig. 5a) meaning that several
to an increase in the photocurrent density (Fig. 3c), conse- competitive decay processes may happen at the same time. In
10
´
quently the corresponding IPCE curve (Fig. 3d) get to 65%, close fact, as explained by Mora-Sero et al., there are different
to the 500 nm region, starting from about 50% referred to the pathways for the photogenerated charge, as shown in Fig. 6a,
device sensitized with organic dye TBTCA only. All these which schematically proposed the most likely situations for
evidences suggest that on adding CdSe495 to TBTCA a surplus of QDDSSCs in terms of energy level arrangement.
electrons ows through the circuit resulting in an improved
The best case, to produce improvement in the device effi-
performance of the device.
ciency, is the type II alignment where there is directional charge
This journal is © The Royal Society of Chemistry 2014
J. Mater. Chem. A

View Article Online
Journal of Materials Chemistry A
Paper
Fig. 7 (a) Absorption spectra of N719 (dark dotted line); emission
spectrum of CdSe495 (green dotted line); emission spectrum of
CdSe545 (orange dotted line); (b) absorption spectra of TCBCA (dark
bold line); emission spectrum of CdSe495 (green bold line); emission
spectrum of CdSe545 (orange bold line).
Fig. 5 (a) Comparison of the decay profile between N719 (dark filled
squares), N719–CdSe495 (green filled squares) and N719–CdSe545
(orange filed squares) @TiO2 film deposited on FTO glass, and the PL
wavelength detected for the decay profile is around 700 nm; (b)
comparison of the decay profile between TBTCA (dark circles),
TBTCA–CdSe495 (green circles) and TBTCA–CdSe545 (orange circles)
@TiO2 film deposited on FTO glass, and the PL wavelength detected
for the decay profile is around 680 nm.
dened slower decay than TBTCA@TiO2 one; moreover it is
possible to roughly t the decay curve (Fig. 5b) with a mon-
oexponential function, so we can suppose that in this case
only one process could preferably operate. These observa-
tions are consistent with FRET from CdSe495 to TBTCA, which
could be the main reason for the increased efficiency in the
corresponding devices (Table 2, entry 7). Spectroscopic
investigation (Fig. 7b) also suggests the real possibility to
have FRET between CdSe QDs and TBTCA. It is worth noting
that only CdSe495 are characterized by signicant trap emis-
sions (Fig. 7a and b) clearly evidenced by the presence of a
broadened red-shied band centered at 580 nm which is
closely related to the presence of intra-band states.53 More-
over, if we look carefully at the IPCE spectrum (Fig. 4d), the
increased photocurrent realized around 400 nm can be
reasonably ascribed to FRET from CdSe495 to TBTCA, which
might be more efficient owing to the optimal superimposi-
tion between dye absorption and QD emission bands also
ensured by the presence of trap emissions, but the
improvement evidenced around 600 nm could be attributed
to charge transfer via cascade of electrons from the LUMO of
TBTCA to the CB of CdSe495 and nally into the TiO2 semi-
conductor (Fig. 8).
Fig. 8 shows the energy level scheme of the sensitizers. In
this case the energy levels of TBTCA are experimentally
determined by cyclovoltammetric measurements on a lm
generated from 1,2-dichlorobenzene solution, whereas TiO2,
CdSe545 and CdSe495 levels are recovered from the litera-
ture.51,52 The experimental data of devices suggest that only
CdSe495 has the right energy levels to ensure the proper
alignment between TiO2 and TBTCA that allows further
driving force to the electrons to be injected into the CB of the
TiO2 and on the basis of these observations it is possible to
explain the different behavior of the two CdSe QD samples
observed in the corresponding QDDSSC devices. Lastly,
although TCSPC experiments reveal some physical–chemical
processes involved between the constituents of the photo-
anode under open circuit conditions, the IPCE measure-
separation for both electrons and holes. Otherwise, if the LUMO
of the dye is located below the Conduction Band (CB) of the QD,
as in the case of the type I alignment, directional charge sepa-
ration will be weakened; the dye can only act as the hole scav-
enger and it could be an increased photoluminescence
emission of the dye. Moreover many new recombination chan-
nels, internal to the QD, could appear when QDs are combined
with other materials.
In our specic case (Fig. 6b) the energy level values were
not determined rigorously, but were recovered on the basis of
previous studies,21,51,52 readjusting the band gap of QDs
according to their size. The LUMO of N719 is very close in
energy to the CB of both CdSe495 and CdSe545 QDs, it could be
a little above or below the CB resulting in similar scenery to
10
´
that proposed by Mora-Sero can t even into this case.
Furthermore, analyzing the UV-visible spectra (Fig. 7a), FRET
¨
(Forster Resonance Energy Transfer) energy transfer may
occur too, since the overlap between the absorption
spectrum of N719 and emission spectra of CdSe QDs has
been clearly observed. This complicated framework makes
interpreting the interaction between CdSe QDs and N719 very
difficult.
The combination of TBTCA with CdSe545 did not lead to
any remarkable changes in the decay time of the dye, whereas
the mixture of CdSe495–TBTCA@TiO2 brings about a well
ments provide the evidence of
a more complicated
system and by intersecting all the experimental results, it is
possible to attribute the enhancing efficiency of TBTCA–
CdSe495-sensitized solar cells to either energy and charge
transfers.
Fig. 6 (a) Schematic illustration, adapted from the study of Mora-Sero
`
et al.,10 of possible energy level alignments and related photo-
generated-charge-carrier dynamics adapted from ref. 10; (b) energy
level alignment of N719 and CdSe.
J. Mater. Chem. A
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
Journal of Materials Chemistry A
possibility to have QDs smaller than CdSe495 might improve the
mechanism of the electron transfer via cascade owing to better
separation between the CB of the TiO2 and the LUMO of the QD.
Acknowledgements
The authors acknowledge eni S.p.A for nancial support and are
grateful to Ou Chen (MIT) for providing the QDs and Thomas
Bischof (MIT) for TCSPC experiments. The authors also thank
Andrea Alessi and Mario Salvalaggio (eni S.p.A.) for UV-Vis
Spectroscopy, Nilanjan Chatterjee (MIT) for Electron Probe
Micro-Analysis, Stefano Carlo Chiaberge (eni S.p.A.) for Mass
Spectrometry, Elizabeth L. Shaw (MIT) for XPS experiments,
Alessandra Tacca (eni S.p.A.) for Cyclovoltammetry, Silvia Spera
(eni S.p.A.) for NMR and Dong Soo Yun (MIT) for TEM
characterization.
Fig. 8 Energy level alignment of TBTCA and CdSe QDs.
Conclusions
Notes and references
In this work we studied the effects of the co-sensitization of TiO2
with CdSe QDs and the dye. First we characterized qualitatively
and quantitatively the CdSe-sensitized titania lm showing that
QDs do not penetrate completely into the mesoporous TiO2 lm
but they are just conned to a very thin top layer. We also
evaluated the effect on the device efficiency using both CdSe495
and CdSe545 QDs combined with two different dyes: the ruth-
enium(II) complex N719 and metal-free TBTCA. We found a
successful combination between TBTCA and CdSe495 that
resulted in an increase in the cell efficiency of 25%. Consider-
ations on TCSPC experiments performed on the photoelectrode
revealed FRET from CdSe495 to TBTCA as one of the presumable
mechanisms, as long as IPCE curves attested that could be also
a good matching of the energy levels of the entire system
contributing to a better injection via electron cascade. In the
case of N719 the exploited combinations are all pejorative with
1 K. W. J. Barnham, M. Mazzer and B. Clive, Nat. Mater., 2006,
5, 161.
2 M. Jacoby, Chem. Eng. News, 2007, 85(35), 16.
3 J.-L. Bredas and J. R. Durrant, Acc. Chem. Res., 2009, 42, 1689.
4 N. Robertson, Angew. Chem., Int. Ed., 2006, 45, 2338.
5 A. Hagfeldt, G. Boschloo, L. Sun, L. Kloo and H. Pettersson,
Chem. Rev., 2010, 110, 6595.
6 M. K. Nazeeruddin, C. Klein, P. Liska and M. Graetzel, Coord.
Chem. Rev., 2005, 249, 1460.
7 P. Yu, K. Zhu, A. G. Norman, S. Ferrere, A. J. Frank and
A. J. Nozik, J. Phys. Chem. B, 2006, 110, 25451.
8 R. Vogel, P. Hoyer and H. Weller, J. Phys. Chem. B, 1994, 98,
3183.
´
´
`
9 I. Mora-Sero, V. Likodimos, S. Gimenez, E. Martınez-Ferrero,
J. Albero, E. Palomares, A. G. Kontos, P. Falaras and
J. Bisquert, J. Phys. Chem. C, 2010, 114, 6755.
´
respect to the dye alone and the causes are not completely 10 I. Mora-Sero, D. Gross, T. Mittereder, A. A. Lutich, S. Susha,
attributed yet.
T. Dittrich, A. Belaidi, R. Caballero, F. Langa, J. Bisquert and
Generally co-sensitization had been studied with the main
A. L. Rogach, Small, 2010, 2, 221.
¨
aim to obtain an enhancement in the efficiency of the device 11 A. Mishra, M. K. Fischer and P. Bauerle, Angew. Chem., Int.
through improved combinations of materials. Indeed in view of Ed., 2009, 48, 2474.
an actual application of this photovoltaic technology, co-sensi- 12 S. Itzhakov, S. Buhbut, E. Tauber, T. Geiger, A. Zaban and
tization would provide improvements in efficiency and stability D. Oron, Adv. Energy Mater., 2011, 1, 626.
with respect to a single-component system and the use of 13 K. Becker, J. M. Lupton, J. Muller, A. L. Rogach, D. V. Tapalin,
different couples of QDs and more efficient organic dyes would
H. Weller and J. Feldmann, Nat. Mater., 2006, 5, 777.
¨
be a very stimulating argument for future studies. Taking into 14 A. J. Nozik, Phys. E, 2002, 14, 115.
account the outcomes of this work we may think that it would 15 J. A. McGuire, J. Joo, J. M. Pietryga, R. D. Schaller and
be possible to obtain further improvement in QDDSSCs by
V. I. Klimov, Acc. Chem. Res., 2008, 41, 1810.
using organic dyes with red shied absorption bands, that, in 16 M. T. Trinh, A. J. Houtepen, J. M. Schins, T. Hanrath, J. Piris,
the case of the TBTCA type structure, could be obtained – for
example – introducing one or more thiophene units in the
W. Knulst, A. P. L. M. Goossens and L. D. A. Siebbeles, Nano
Lett., 2008, 8, 1713.
p-bridge and/or adding one or more electron-donating groups 17 H. Lee, H. C. Leventis, S.-J. Moon, P. Chen, S. Ito,
¨
S. A. Haque, T. Torres, F. Nuesch, T. Geiger,
on the terminal phenyls of the triphenylamine unit. Additional
enhancements might be obtained by using QDs with modied
shell having better compatibility with the photo-anode in order
¨
S. M. Zakeeruddin, M. Gratzel and M. K. Nazeerruddin,
Adv. Funct. Mater., 2009, 19, 2735.
to increase the amount of adsorbed particles and achieve a 18 A. M. Funston, J. J. Jasienak and P. Mulvaney, Adv. Mater.,
more effective penetration inside of the TiO2 lm. Finally the
2008, 20, 4274.
This journal is © The Royal Society of Chemistry 2014
J. Mater. Chem. A

View Article Online
Journal of Materials Chemistry A
Paper
19 A. Y. Koposov, P. Szymanski, T. Cardolaccia, T. J. Meyer, 36 D. J. Schipper and K. Fagnou, Chem. Mater., 2011, 23, 1594.
V. I. Klimov and V. I. Sykora, Adv. Funct. Mater., 2011, 21, 37 L. Carbone, C. Nobile, M. De Giorgi, F. Della Sala, G. Morello,
3159.
P. Pompa, M. Hytch, E. Snoeck, A. Fiore, I. R. Franchini,
M. Nadasan, A. F. Silvestre, L. Chiodo, S. Kudera,
R. Cingolani, R. Krahne and L. Manna, Nano Lett., 2007, 7,
2942.
20 H. Choi, P. K. Santra and P. V. Kamat, ACS Nano, 2012, 6,
5718.
21 L. Etgar, J. Park, C. Barolo, V. Lesnyak, S. K. Panda,
P. Quagliotto, S. G. Hickey, M. K. Nazeeruddin, 38 Y. J. Shen and Y. L. Lee, Nanotechnology, 2008, 19, 045602.
¨
´
´
¨
A. A. Eychmuller, G. Viscardi and M. Gratzel, RSC Adv., 39 I. Mora-Sero, S. Gimenez, T. Moehl, F. Fabregat-Santiago,
´
2012, 2, 2748.
T. Lana-Villareal, R. Gomez and J. Bisquert,
22 M. Shalom, J. Albero, Z. Tachan, E. Martinez-Ferrero,
Nanotechnology, 2008, 19, 424007.
A. Zaban and E. Palomares, J. Phys. Chem. Lett., 2010, 1, 1134. 40 J. Chen, J. L. Song, X. W. Sun, W. Q. Deng, C. Y. Jiang, W. Lei,
23 J. Huang, D. Stockwell, Z. Huang, D. L. Mohler and T. Lian, J.
Am. Chem. Soc., 2008, 130, 5632.
24 S. N. Sharma, Z. S. Pillai and P. V. Kamat, J. Phys. Chem. B,
2003, 107, 10088.
J. H. Huang and R. S. Liu, Appl. Phys. Lett., 2009, 94, 153115.
41 I. Robel, V. Subramanian, M. Kuno and P. V. Kamat, J. Am.
Chem. Soc., 2385, 128.
´
42 N. Guijarro, T. Lana-Villareal, I. Mora-Sero, J. Bisquert and
´
R. Gomez, J. Phys. Chem. C, 2009, 113, 4208.
25 M. Sykora, M. A. Petruska, J. Alstrum-Acevedo, I. Bezel,
¨
T. J. Meyer and V. I. Klimov, J. Am. Chem. Soc., 2006, 128, 43 S. Ruhle, M. Shalom and A. Zaban, Chem. Phys. Chem., 2010,
9984.
11, 2290.
26 Y. Liu and J. Wang, Thin Solid Films, 2010, 518, e54.
27 S. So, S.-Q. Fan, H. Choi, C. Kim, N. Cho, K. Song and J. Ko,
Appl. Phys. Lett., 2010, 97, 263506.
28 X. Song, X.-L. Yu, Y. Xie, J. Sun, T. Ling and X.-W. Du,
Semicond. Sci. Technol., 2010, 25, 095014.
29 S.-Q. Fan, R.-J. Cao, Y.-X. Xi, M. Gao, M.-D. Wang, D.-H. Kim
and J.-J. Ko, Optoelectron. Adv. Mater., Rapid Commun., 2009,
3, 1027.
44 W. W. Yu, L. Qu, W. Guo and X. Peng, Chem. Mater., 2003, 15,
2854.
45 S. J. B. Reed, Electron microprobe analysis, 2nd ed.,
Cambridge University Press, Cambridge, 1993.
¨
46 S. Ito, T. N. Murakami, P. Comte, P. Liska, C. Gratzel,
¨
M. K. Nazeeruddin and M. Gratzel, Thin Solid Films, 2008,
516, 4613.
´
´
47 V. Gonzalez-Pedro, Q. Shen, V. Jovanovski, S. Gimenez,
´
30 H. Choi, R. Nicolaescu, S. Paek, J. Ko and P. V. Kamat, ACS
Nano, 2011, 5, 9238.
R. Tena-Zaera, T. Toyoda and I. Mora-Sero, Electrochim.
Acta, 2013, 100, 35.
31 C.-A. Lin, K. P. Huang, S. T. Ho, M.-W. Huang and J. -H. He, 48 Y. Gao, X. Zou, Z. Sun, Z. Huang and H. Zhou, J. Nanomater.,
Appl. Phys. Lett., 2012, 101, 123901. 2012, 415370.
32 V. Gonzalez-Pedro, Q. Shen, V. Jovanovski, S. Gimenez, 49 H. Shen, H. Lin, Y. Liu, J. Li and D. Oron, J. Phys. Chem. C,
´
´
´
R. Tena-Zaera, T. Toyoda and I. Mora-Sero, Electrochim.
2012, 116, 15185.
Acta, 2013, 100, 35.
33 F. Guo, J. He, J. Li, W. Wu, Y. Hang and J. Hua, J. Colloid
Interface Sci., 2013, 408, 59.
50 H. Choi and P. V. Kamat, J. Phys. Chem. Lett., 2013, 4,
3983.
51 C.-F. Chi, H.-W. Cho, H. Teng, C.-Y. Chuang, Y. -M. Chang,
Y.-J. Hsu and Y. -L. Lee, Appl. Phys. Lett., 2011, 98, 012101.
34 S. Paek, H. Choi, C. Kim, N. Cho, S. So, K. Song,
¨
M. K. Nazeeruddin and J. Ko, Chem. Commun., 2011, 47, 52 F. Lenzmann, J. Krueger, S. Burnside, K. Brooks, M. Gratzel,
¨
2874.
D. Gal, S. Ruhle and D. Cahen, J. Phys. Chem. B, 2001, 105,
35 M. Velusamy, K. R. J. Thomas, J. T. Lin, Y.-C. Hsu and
K.-C. Ho, Org. Lett., 2005, 7, 1899.
6347.
53 C. J. Murphy and J. L. Coffer, Appl. Spectrosc., 2002, 56, 16A.
J. Mater. Chem. A
This journal is © The Royal Society of Chemistry 2014