Synergistic Interaction of Dyes and Semiconductor Quantum Dots for Advanced Cascade Cosensitized Solar Cells

A new procedure for the cosensitization with quantum dots (QDs) and dyes for sensitized solar cells is reported here. Cascade cosensitization of TiO2 electrodes is obtained by the sensitization with CdS QDs and zinc phthalocyanines (ZnPcs), in which ZnPcs containing a sulfur atom are specially designed to produce a cascade injection by direct attachment to QDs. This strategy causes a double synergetic interaction. This is the differentiating point of cascade cosensitization in comparison with other approaches in which dyes with conventional functionalization are anchored to TiO2 electrodes. Cosensitization produces a panchromatic response from the visible to near‐IR region already observed with other sensitization strategies. However, cascade cosensitization produces in addition a synergistic interaction between QDs and dye, that it is not merely limited to the complementary light absorption, but dye enhances the efficiency of QD sensitization acting as a passivating agent. The cascade cosensitization concept is demonstrated with using [Co(phen)3]3+/2+ redox electrolyte. The TiO2/CdS QD‐ZnPc/[Co(phen)3]3+/2+ sensitized solar cell shows a large improvement of short‐circuit photocurrent and open‐circuit voltage in comparison with samples just sensitized with QDs. The advent of such cosensitized QD‐ZnPc solar cells paves the way to extend the absorbance region of the promising QD‐based solar cells and the development of a new family of molecules designed for this purpose.

phthalocyanines (Pcs) in order to prepare sensitized solar cells with enhanced effi ciency. QDs are among potential key players in the next generation of photovoltaic devices [ 1 ] due to their low-cost solution-phase processability, large absorption cross sections, a spectrally tunable absorption onset (achieved via the quantum size effect), and enhanced multiple exciton generation or carrier multiplication. [ 2 ] After the fi rst report on a certifi ed QD solar cell, [ 3 ] a remarkable progress has been reached just in a couple of years obtaining certifi ed power conversion efficiencies (PCEs) approaching 9%. [ 4 ] On the contrary, Pcs are outstanding dye candidates in dye-sensitized solar cells (DSSCs) due to their high extinction coeffi cient in the infrared spectral region and to their high thermal and chemical stabilities. [ 5 ] Pcs incorporated in DSSCs have achieved PCEs as high as 6.4 %, [ 6 ] still far away from the 12.75% PCE obtained by porphyrins, [ 7 ] their closest relatives. Light harvesting ability of QDs and Pcs can be enhanced by binding them, either covalently or supramolecularly, producing a system capable to charge injection from both chromophores. However, the chemical combination of Pc rings and QDs has hardly been explored probably due to the diffi culties to synthesize Pcs with the adequate anchoring groups. Until now, just a few articles have been published, where Pcs are covalently linked to QDs showing in most of the cases Förster resonance energy transfer (FRET) from QDs to Pcs. SiPc has been connected to CdSe QDs through axial ligation, [ 8 ] and also tetraminoZnPc [ 9 ] and unsymmetrically tristert -butyl-imidoZnPc [ 10 ] have been linked to CdSe showing FRET and quenching the QDs emission. But few complete photovoltaic devices taking benefi t of this interaction are reported. Nazeeruddin and co-workers published a very interesting seminal work on cosensitization, where zinc carboxyphthalocyanine dye (TT1) is added on top of a TiO 2 /PbSsensitized electrode, prepared by successive ionic layer absorption and reaction (SILAR), to obtain a panchromatic response, which results in an improved effi ciency versus the TiO 2 /TT1 system. [ 11 ] TT1, as most of the dyes, was designed with a carboxylic group to attach to TiO 2 . These authors obtained cosensitized cells with enhanced effi ciency due to the broadening of

Synergistic Interaction of Dyes and Semiconductor Quantum Dots for Advanced Cascade Cosensitized Solar Cells
Vicente M. Blas-Ferrando , Javier Ortiz , Victoria González-Pedro , Rafael S. Sánchez , Iván Mora-Seró , * Fernando Fernández-Lázaro , and Ángela Sastre-Santos * A new procedure for the cosensitization with quantum dots (QDs) and dyes for sensitized solar cells is reported here. Cascade cosensitization of TiO 2 electrodes is obtained by the sensitization with CdS QDs and zinc phthalocyanines (ZnPcs), in which ZnPcs containing a sulfur atom are specially designed to produce a cascade injection by direct attachment to QDs. This strategy causes a double synergetic interaction. This is the differentiating point of cascade cosensitization in comparison with other approaches in which dyes with conventional functionalization are anchored to TiO 2 electrodes. Cosensitization produces a panchromatic response from the visible to near-IR region already observed with other sensitization strategies. However, cascade cosensitization produces in addition a synergistic interaction between QDs and dye, that it is not merely limited to the complementary light absorption, but dye enhances the effi ciency of QD sensitization acting as a passivating agent. The cascade cosensitization concept is demonstrated with using [Co(phen) 3 ] 3+

Introduction
An elegant strategy to improve the light-harvesting in photovoltaic devices is to use complementary light harvesters capable to produce panchromatic absorption. Here, we have combined semiconductor quantum dots (QDs) and ad hoc-designed the light absorption region. Another example of cosensitization of this kind has been reported with CdS and SQ1. [ 11 ] Nevertheless a close inspection of the external quantum effi ciency (EQE) showed a decrease of the performance in the light absorption region of QDs after cosensitization. Conversely, very recently we have shown that it is possible to enhance the performance in the light absorption region of QDs with an appropriate QD passivation. Disulphide bisphthalocyanine was covalently bonded to CdSe and CdS QDs improving, by a passivation mechanism, the effi ciency of the QD solar cell compared with the QD without phthalocyanine. [ 12 ] In our previous work, despite the improvement in cell effi ciency with the cosensitization, no contribution from the ZnPc in the EQE was observed, probably due to the nonconjugated bridge between the sulphur anchoring group and the ZnPc. Here, we show a cascade cosensitization by the use of especially designed ZnPcs that can produce both benefi cial effects, broadening of light absorption and QD passivation.
Here, we will like to highlight the synthesis of Pcs designed for anchoring selectively on semiconductor QDs in TiO 2 /QDs electrodes through a thiol (SH) group. These two new mercaptophthalocyanine dyes, where SH is either directly connected to the macrocycle or through a conjugated linker, exhibit the capability of injecting electrons into the QD. Therefore, we have proved in this work a cascade injection from the ZnPcs into the TiO 2 electrode though the CdS QD, with contributions of both sensitizers to the EQE. In addition, ZnPc passivates the surface of QDs enhancing the EQE of the QD-sensitizer in comparison with devices prepared without cosensitization. Several trials to synthesize thiol-substituted Pcs were unsuccessful in our hands, mostly attributable to the SH oxidation catalyzed by the own phthalocyanine ring. [ 13 ] For this reason, we have synthetized two unsymmetrically substituted Pcs, CNC 2 H 4 SZnPc 1 , and AcSPhC 2 ZnPc 2 ( Scheme 1 , 2 ), with different protecting groups in the sulphur, ethylennitrile, and thioester, respectively, to avoid oxidation. Moreover, these Pcs present donor substituents, such as tert -octylphenoxy and tert -butyl groups, to increase the energy of the LUMO levels over the conduction band of the CdS QD in order to facilitate the electron injection, and place the HOMO levels lower in energy than that of the electrolyte for effi cient charge regeneration ( Figure 1 ).
UV-vis spectra in THF as solvent show typical Q and B Pc absorption bands with high extinction coeffi cients corresponding to nonaggregated Pcs centered respectively at 682 and 361 nm for CNC 2 H 4 SZnPc and 686, 671 (Q-band splitting is due probably to the presence of 4 different regioisomers) and 351 nm for AcSPhC 2 ZnPc ( Figure 2 a). Figure 2 b confi rms by UV-vis in solid state of the QDSC devices that the two Pcs have been attached to the CdS-QD, observing the Q-bands of the ZnPcs at the same wavelength that in THF solution. However, the UV-vis band of QD in the hybrid CdS-ZnPcs increased in intensity and appears at 475, 28 nm batochromically shiftted compared with the pristine QD-band centered at 447 nm. This fact could be explained by the reduction of quantum confi nement after being covered by the large macrocyclic ring. Moreover, comparing the Q-band intensity of the two ZnPcs, a larger amount of load is observed for -SZnPc than for -SPhC 2 ZnPc. This factor is also observable by the naked eye comparing the green color intensity of the devices (see Figure S14, Supporting Information). Figure 3 a shows the current-potential ( J-V ) curves comparing the Pc-untreated CdS device as a reference, with our CdS-SZnPc hybrid systems using polysulfi de S 2− /S n 2− as electrolyte. Higher open-circuit voltage ( V oc ) and short-circuit current ( J sc ) were obtained when ZnPcs were attached to the CdS QD ( Table 1 ). It is very important to note that the new ZnPc dyes here reported do not attach directly to TiO 2 as we have verifi ed experimentally. Figure 3b shows the EQE versus wavelength for these cells. A signifi cant increase of the EQE in the area between 300 and 500 nm is observed, however, it does not occur the same in the area of 600-700 nm where the phthalocyanine absorbs, concluding that the main effect of QD-dye interaction is the passivation of the QD surface, [ 12,17 ] but no evidence of cascade injection from dye into TiO 2 electrode, through the QD is observed. The introduction of the Pc dyes increases the effi ciency around 50%, from 1.1% up to 1.7% in CdS QD-SZnPc and 1.5% in CdS QD-SPhC 2 ZnPc 2 (Table 1 ).
However, these results were signifi cantly improved by changing the electrolyte to [Co  as reference to 6.02 mA cm −2 /750 mV in CdS QD-SZnPc and 4.52 mA cm -2 /734 mV in CdS QD-SPhC 2 ZnPc ( Figure 4 a). These enhancements for the cosensitized systems are understandable paying attention to the EQE spectra (Figure 4 b); the contribution of the ZnPcs within the 600-800 nm region can be observed in addition to the contribution of QDs in the 300-500 nm region. Furthermore, an increase of FF is also reported for cosensitized samples ( Table 1). As a result, the overall performance of the cosensitized device is increased from 0.8% using CdS QD as sensitizer to 2.5% and 1.9% with cosensitization for CdS QD-SZnPc and CdS QD-SPhC 2 ZnP, respectively. These increases represent a threefold and twofold enhancements, respectively, in the PCE. From the J-V curve and EQE spectra, we can infer that the TiO 2 /CdS-ZnPc/[Co(phen) 3 ] 3+/2+ solar cell devices deliver a photovoltaic response from both materials. In addition, resonant Förster energy transfer can be ruled out due to the low overlapping between dye emission and QD absorption. The enhanced device effi ciencies are attributed to the charge injection occurring from both sensitizers, with electron transfer from high energy ZnPc to the lower energy CdS QDs and from QD into TiO 2 demonstrating a cascade cosensitization. The use of a cobalt electrolyte, with a redox potential closer in energy to the HOMO of the ZnPcs compared to the polysulfi de, seems to be a crucial parameter for the regeneration of the dye. Despite it has been experimentally verifi ed that TiO 2 was not able to be sensitized using ZnPcs 1 and 2 , due to the presence of an inadequate anchoring group, a small contribution from direct injection from ZnPc to the TiO 2 cannot be ruled out. The energetic diagram provided in Figure 1 indicates that the direct electron transfer from dye to TiO 2 is energetically possible. Since QD sensitization is conducted using the SILAR method, leading to a large particle distribution of QDs, it is possible that dye molecules coordinated with small QD particles could be in contact with TiO 2 , thus allowing the electron transfer. For this reason, the electron transfer mechanism will be analyzed using femtosecond transient fl ash photolysis measurements and the conclusions will be published elsewhere.
To unveil the origin of this enhanced performance cells plotted in Figure 4 a,b have been characterized by impedance spectroscopy and analyzed with the standard models for QDSCs. [ 18,19 ] The compared values obtained for recombination resistance, R rec , and chemical capacitance, C µ , are depicted in Figure 5 a,b, respectively. The similar values obtained for both magnitudes, independently of the dye used indicate: i) The TiO 2 conduction band does not change depending on the employed dye, because the same C µ is observed; ii) Recombination rate does not vary, as the same R rec has been obtained for both systems. Consequently, both solar cells are similar in terms of IS characterization, although their photovoltaic behavior is clearly different, particularly in terms of J sc and V oc . The origin of the different photocurrents relies on a process which is not accessible by IS. The higher photovoltaic values obtained for CdS QD-SZnPc versus the CdS QD-SPhC 2 ZnPc device can be explained by the closer distance between both QD and ZnPc in the fi rst case that could make the electron transfer more favorable and/or by the higher dye loading. Higher dye loading with SZnPc is confi rmed by taking into account the extinction coeffi cient, Figure 2 a and the absorption measurements Figure 2b. Different injection rate could also contribute to this effect but this point is beyond the scope of the present paper.

Conclusion
This work reports, for the fi rst time, on the cascade cosensitization using QD and dyes. We have demonstrated this concept with CdS QDs and ad hoc designed ZnPcs with thiol groups in a sensitized solar cell where the dye is selectively anchored to the QDs. As a result of this cosensitization, the photocurrent, Adv. Funct. Mater. 2015, 25, 3220-3226 www.afm-journal.de www.MaterialsViews.com  open circuit voltage and FF are increased. Effi ciency of cosensitized CdS QD-SZnPc, with Co(phen) 3 ] 3+ /[Co(phen) 3 ] 2+ electrolyte, is 212% higher than that of the cell produced just with CdS QD sensitization. The cascade cosensitization opens the possibility to achieve devices with a high response in the visible region due to the outstanding QDs properties and to extend their functionality to the NIR by the contribution of ZnPcs. This works establishes the potentiality of a new family of dyes especially designed for anchoring to QDs instead to high bandgap material, i.e., TiO 2 . The fact that individual QDs already demonstrated high effi ciency opens up new avenues for future improving these effi ciencies by cosensitization with properly designed dyes. Cascade cosensitization allows a synergistic panchromatic light absorption, on one side, with recombination reduction by means of surface passivation in the presence of an appropriate electrolyte, on the other side. This work could have important implications on the future development of sensitized devices.

Experimental Section
Materials and Methods for the Synthesis of New Compounds : All chemicals were reagent-grade, purchased from commercial sources, and used as received. Column chromatography was performed using SiO 2 (40-63 µm). TLC plates coated with SiO 2 60F254 were used and visualized by UV light. NMR spectra were measured with a Bruker AC 300. Fluorescence spectra were recorded with Perkin-Elmer LS 55 Luminescence Spectrometer, UV/Vis spectra were recorded with a Helios Gamma spectrophotometer and IR spectra with Nicolet Impact 400D Spectrometer. Mass spectra were obtained from Bruker Microfl ex matrixassisted laser desorption/ionization time of fl ight (MALDI-TOF). CV measurements were performed in a conventional three-electrode cell using a µ-AUTOLAB type III potentiostat/galvanostat at 298 K, over benzonitrile and deaerated sample solutions (≈0.5 × 10 −3 M ), containing 0.10 M tetrabutylammonium hexafl uorophosphate (TBAPF 6 ) as supporting electrolyte. A glassy carbon (GC) working electrode, Ag/AgNO 3 reference electrode and a platinum wire counter electrode were employed. Ferrocene/ferrocenium was used as an internal standard for all measurements. Impedance spectroscopy measurements were carried out under dark conditions at different forward biases, by applying a 20 mV AC sinusoidal signal over the constant applied bias with the frequency ranging between 400 kHz and 0.1 Hz.
Preparation on Sensitized TiO 2 : The electrode confi guration was a 9-nm-thicked transparent layer DSL 18NR-T (20 nm average particle size) and a 6-nm-thicked scattering layer WERO-4 (300-400 nm particle size distribution). The FTO (SnO 2 :F) coated glass was previously covered by a compact layer of TiO 2 deposited by spray pyrolysis of titanium(IV) bis(acetoacetonato) di(isopropanoxylate). These electrodes were sintered at 450 °C for 30 min. The mesoporous TiO 2 electrodes were in situ sensitized with CdS QDs grown by SILAR. For CdS growth the electrodes were successively immersed in two different solutions for 1 min each: one consisting of 0.05 M Cd(OAc) 2 dissolved in ethanol, another of 0.05 M Na 2 S in methanol/Milli-Q ultrapure (1:1). Following each immersion, rinsing and drying was undertaken using a solution without the precursor in order to rinse the excess of precursor. All these processes constitute one SILAR cycle. The SILAR process was carried out using SILAR equipment from ISTest at room temperature under an air atmosphere. All the samples analyzed in this study were coated with fi ve SILAR cycles General Procedure for Anchoring ZnPcs to QD : 0.002 mmol of the ZnPc and 0.006 mmol of CsOH were diluted in 500 µl of dried THF in argon atmosphere at room temperature and stirring during 2 h to eliminate the protecting group. After that time, an additional quantity of dried THF was added to obtain 5 × 10 −3 M ZnPc solution. In inert conditions, the mesoporous TiO 2 electrodes sensitized by CdS QDs grown by SILAR were immersed in the 5 × 10 −3 M ZnPc solution for 4 h. Then, the device was washed with dried THF to remove the unattached Zn to the QD.   www.afm-journal.de www.MaterialsViews.com NREL-calibrated Si solar cell to 1 sun intensity (100 mW cm −2 ). The IPCE measurements were performed employing a 150 W Xe lamp coupled with a monochromator controlled by a computer; the photocurrent was measured using an optical power meter 70310 from Oriel Instruments, using a Si photodiode to calibrate the system. The absorption spectra were registered using an AndoriDus DV-420A intensifi ed CCD with thermoelectric cooling coupled with a Newport 77400 MS125 TM spectrograph. Light absorption of electrodes sensitized with CdS was carried out on electrodes without TiO 2 scattering layer. A nonsensitized electrode of TiO 2 prepared in the same way has used as reference sample. The absorption from reference samples has been subtracted in the absorption measurements.