Electrochemically synthesized mesoporous thin fi lms of ZnO for highly ef fi cient dye sensitized solar cells

In this work, nanostructured thin fi lms of ZnO were electrochemically grown on FTO substrates. The morphology was tuned by modifying the synthesis parameters. The synthesis was carried out by applying Zn(NO 3 )6H 2 O as the sole component of the aqueous electrolyte, avoiding the use of capping agents. The composition and morphology of the prepared ZnO were characterized by energy-dispersive X-ray spectroscopy (EDX) and scanning electron microscopy (SEM), respectively. The as-deposited fi lms were applied as electrodes in dye sensitized solar cells (DSCs). The performance of the cells was investigated by JV curves and IPCE (incident photon to charge carrier ef fi ciency) measurements. The SEM analysis demonstrated a direct relationship between ZnO morphology and Zn precursor concentration. It has been shown that the lower the concentration is, the more porous the morphology is. Increasing the amount of dye adsorbed on the ZnO decreased the power conversion ef fi ciency of the fi nal DSCs. The best cell presented the following parameters: open circuit voltage V OC ¼ 0 : 59 V, short circuit current J SC ¼ 7 : 64 mA = cm 2 , fi ll factor FF ¼ 50.41%, and power conversion ef fi ciency PCE ¼ 2.27%.


Introduction
ZnO has recently emerged as a promising candidate for its application in excitonic solar cells (XSC). Different ZnO nanostructures like nanowires, nanotubes, nanobelts, nanotetrapods, mesoporous thin films, among others have been successfully obtained by low-cost techniques such as the hydrothermal and the electrochemical synthesis routes [1][2][3][4][5][6]. TiO2 is still the most commonly used semiconductor oxide in XSC applications. However, interest at ZnO XSC has increased in the same rate as it has been investigated and successfully applied in this technology [7][8][9]. ZnO presents unique properties that are very similar to those of TiO2. In fact, ZnO presents some advantages in comparison with TiO2 like faster charge transport due to electron mobility and conductivity of several orders of magnitude higher than that n Corresponding author. E-mail address: anderson@metalmat.ufc.br (F.A.S. Lima). observed for TiO2 [10,11]. Among the various synthesis methods developed to produce ZnO nanostructured thin films, the electrodeposition is a very attractive technique due to its low cost, scalability, and low temperature processing. Direct control of film morphology is possible by manipulating the electrodeposition parameters like applied current, applied potential, electrolyte concentration, temperature, and pH [12][13][14][15]. Moreover, electrochemical deposition of ZnO permits the fabrication of novel nanostructured thin films with superior advantages like large surface area and electron mobility, and more efficient charge transport [16][17][18][19][20]. Zhu et al. [21] have proved that a combination of nanorod/nanoparticle composite architecture can improve the performance of TiO2 DSCs. Over the last few years, different ZnO nanostructures have been successfully applied in DSCs [22][23][24][25].
In this work, thin films of ZnO were obtained by electrodeposition under different synthesis parameters. These films were morphologically, compositionally, and structurally characterized. The photoanodes were finally applied to assemble DSCs.

Synthesis of nanostructured ZnO thin films
Nanostructured ZnO thin films were fabricated by an electrochemical deposition process. The film deposition was carried out by applying the Zn(NO3)6H2O as the sole component of the aqueous electrolyte avoiding the use of any capping agent. A Pt foil and a FTO-coated glass from SOLEMS s with sheet resistance of 50-70 Ω=square were used as a counter electrode and a working electrode, respectively. The potential was applied with respect to an Ag/AgCl (KCl-Sat) reference electrode. Three different concentrations of zinc nitrate on the aqueous solution were used: 12.5 mM, 25 mM and 50 mM. The temperature was always kept at 80 1C, through the use of a hot plate and the applied potential for all procedures was 1.0 V. The reaction time ranged from 0.5 h to 2 h. The crystal structure, morphology and composition of the as-deposited ZnO nanostructured thin films were characterized by X-ray diffraction (XRD), scanning electron microscopy (SEM) and energy-dispersive X-ray spectroscopy (EDX) respectively.

Solar cell fabrication and characterization
As-deposited ZnO thin films were immersed in a 0.5 mM/L ethanolic solution of N719 dye for 1 h in order to provide dye loading at the ZnO. After dye loading, the working electrodes were used to assemble DSSC. The platinized FTO counter electrode was prepared by electron beam physical vapor deposition (EBPVD), with a Pt thickness of 50 nm. The work and counter electrodes were bonded together with a hot melt sealing foil. The commercial liquid iodide-iodine electrolyte was introduced through a small hole on the Pt counter-electrode by capillary forces applying vacuum. Finally, the DSSC were hermetically sealed with a small piece of glass to close the filling hole. Photoelectrochemical tests were carried out by measuring the JV characteristic curves under simulated AM 1.5 solar illumination at 100 mW/cm 2 from a sun simulator Steuernagel Solarkonstant s KHS1200. JV curves were measured using a Keithley s 2601 multimeter. Incident Photon to Charge Carrier Efficiency (IPCE) measurements were carried out with a QE/IPCE measurement system from Oriel s at 10 nm intervals between 300 and 700 nm. The results were not corrected for intensity losses due to light absorption and reflection by the glass support.

Structural, morphological and compositional analysis
Special attention was given to the optimization of the reaction time and to the concentration of zinc nitrate in the aqueous electrolyte. Samples 1, 2 and 3 were deposited from a 50 mM zinc nitrate electrolyte for 0.5 h (sample 1) and 2 h in 4 steps of 0.5 h each (samples 2 and 3). Sample 2 was cleaned with deionized water and dried with nitrogen between two consecutive 0.5 h steps left standing in air between cycles. Sample 3 was maintained in the electrolyte solution between cycles. For samples 4 and 5, the total reaction time was 2 h in one single step with concentration of zinc nitrate in the aqueous electrolyte of 25 mM and 12.5 mM respectively. Table 1 summarizes the deposition conditions for the preparation of each electrode. Fig. 1 shows XRD diffractograms for Samples 3, 4 and 5. The diffractogram corresponding to sample 3 electrode (electrolyte concentration of 50 mM) reveals the existence of a single crystalline phase with hexagonal wurtzite structure (CPDS card file number 36-1451, zincite phase). The maximum intensity for the ZnO corresponds to the (1011) diffraction peak. Therefore, the observation of a strong and narrow (0002) peak suggests that the nanostructure is highly oriented along the c-axis perpendicular to the plane of the substrate ([0001] direction). The diffractograms of samples 4 and 5 (electrolyte concentrations of 25 mM and 12.5 mM) also show a preferential orientation along the (0002) plane but with increasingly lower relative intensity of the (0002) peak. The diffractograms of samples 4 and 5 agree with the presence of ZnO as described before. Extra peaks at 2θ equal to 26.51, 36.31 and 56.61 correspond to the FTO substrate.
The influence of deposition reaction time and contact with air on film morphology was studied. Films grown from a 50 nM zinc nitrate electrolyte (samples 1, 2, and 3) were used. Fig. 2(a) shows a SEM micrograph (top view) of a sample 1 where it is possible to observe a very dense array of vertically aligned nanorods. A different microstructure was formed in the region between the nanorods as can be seen more clearly on a micrograph taken with a larger magnification and shown in  Fig. 2(b). The deposition of vertically aligned nanorods confirms the XRD results and is explained by the fact that ZnO crystallizes with the wurtzite crystal structure, which presents hexagonal symmetry and consists of tetrahedrally coordinated Zn and O atoms in alternating layers along the caxis. These layers are composed uniquely of either Zn 2þ or O 2 ions and are therefore polar [9]. The polar (0001) plane has higher surface energy and leads to the fast crystal growth rate along the [0001] direction resulting in the formation of onedimensional ZnO nanostructures [26].
Although the internal structure of ZnO favors the anisotropic growth along the [0001] direction, the surrounding conditions during the electrodeposition process may affect the growth habit of ZnO [27]. The later could explain the different morphology observed between the nanorods. This growth mechanism can be also responsible for the coalescence between adjacent nanorods observed on the microstructure of the sample 2 as can be clearly seen in the micrograph in Fig. 2(c). Moreover, the growth of a porous nanostructure between the nanorods of sample 3 is observed in Fig. 2(d). According to Ludwig et al. [28] there is a strong correlation between the amount of electrochemically deposited ZnO and the morphology of the thin film. The total amount of electrochemically deposited ZnO can be estimated from the total charge deposited per area (integrated from the monitored current density versus time). In this study, the average current density for all the deposited films was 1.3 mA/cm 2 . This high current density probably led to a fast growth rate of the ZnO nuclei, and the growth of ZnO along other directions. As a result, the creation of different nanostructures between the nanorods and their coalescence was observed. Fig. 2(e) shows the cross sectional micrograph for the sample 2. The effect of air contact between consecutive deposition steps is observed by discontinued layers corresponding to each of four 0.5 h steps. As sample 3 was also grown in a four step deposition cycle but maintaining the sample inside the electrolyte (no air contact) between consecutive steps. Thus, the sample does not show discontinued layers as can be seen in the SEM micrograph shown in Fig. 2(f). Air contact influences film composition as well as microstructure as can be seen in Table 2 which shows EDX results on the films deposited with and without air contact (samples 2 and 3, respectively). Films grown with air contact between deposition steps show an excess of oxygen content with respect to zinc while films grown without air contact show stoichiometric ZnO composition. The excess of oxygen can be explained by the reaction of the ZnO surface with oxygen from the air forming an insulating layer of zinc hydroxide (ZnOH) which is an insulator.
The influence of the concentration of zinc precursor on the morphology of the films was also investigated. For this study, films were grown from 25 nM and 12.5 mM electrolytes in a single step of 2 h (samples 4 and 5) with all other parameters kept the same as before. The high density current during the electrochemical reaction together with the low concentrations of the zinc precursor in the aqueous electrolyte led to novel nanostructures as can be seen in the SEM micrographs shown in Fig. 3. Fig. 3(a) and (b) shows top view images of the electrodes grown from 25 mM and 12.5 mM electrolytes, respectively. In these images, porous films are seen instead of the vertically aligned nanorods found for samples grown from 50 mM electrolytes. The cross sectional views of these same films that are shown in Fig. 3(c) and (d) do not show vertically aligned structures, which means that the films grown in other directions which not along the c-axis. The late is in well agreement with the XRD results that showed a lower relative intensity of the (0002) peaks for these samples. In fact, high resolution SEM images of sample 5 (12.5 mM) presented in Fig. 4 show the formation of a mesoporous nanostructure (see Fig. 4).

Photovoltaic performance
The electrodes described above were used without any further treatment as photoanodes to construct dye sensitized solar cells (DSCs). The solar cells were tested under simulated AM 1.5 solar illumination at 100 mW/cm 2 . Fig. 5 shows JV curves and IPCE spectra for the samples using different ZnO layers, while Table 3 shows the photovoltaic parameters for these cells. The data show that all DSCs presented short circuit current density JSC greater than 2.89 mA/cm 2 , open circuit potential VOC greater than 0.59 V, fill factor FF greater than 46.58% and power conversion efficiency PCE greater than 0.97%. These values are in good agreement with the ones recently reported for ZnO thin films deposited in a similar way and without any post-deposition treatment [30,31].
The data show that VOC values are approximately the same for all the cells. As the VOC for a DSCs is defined as the difference between the redox potential of the mediator and the Fermi level of the nanocrystalline film [29], it is possible to state that Fermi level energies for all thin films are approximately the same. The data also reveals that JSC values tend to increase with decreasing electrolyte concentration. According to SEM and XRD results, a decrease in electrolyte concentration leads to more porous films which is desired in order to get the best photovoltaic performance. In fact, the solar cell built with sample 4 (12.5 mM), that presented a mesoporous nanostructure, reached the maximum JSC value. Moreover, considering the absence of post-deposition treatment, a remarkably high PCE value of 2.27% was obtained for this cell. The improvement on JSC can be attributed to mesoporous structure of the thin film, which allows the absorption of larger amount of dye, as monolayers, on the ZnO nanostructures which provides the generation of more photoelectrons and consequently higher JSC. The series and shunt resistance were determined from the slopes of the I-V curves. The photovoltaic parameters for the DSCs analyzed in this study are presented in Table 3.
The amount of dye adsorbed on the oxide was measured by the dye desorption method, which consists of the separation of the two electrodes, followed by the immersion of the ZnO-dye electrode in an aqueous basic solution of 0.1 mM KOH. The dye desorption took place after 1 h, with the solution turning into a pink color due to the presence of the N719 dye. At the same time, the ZnO electrode turned colorless as was found by Neale et al. [32]. The quantity of dye adsorbed on the ZnO electrode was determined by UV-vis analysis of the desorbed dye solution applying the Lambert-Beer law. The results were corrected by the active area of the DSCs. The relationship between the amount of adsorbed dye and efficiency, for the electrodes grown on the electrolyte containing 50 mM of zinc nitrate, is shown in Fig. 6. The data revealed higher power conversion efficiency for samples with the    lowest amount of dye adsorbed. This result can be explained by the dye precipitation on the ZnO nanostructure. According to Keis et al. [33] dye precipitation can take place on the ZnO nanostructure and, in this case, some dissolution of the ZnO by the acidic carboxylic groups of the dye can occur. The resulting Zn 2þ ions form insoluble complexes with the N719 dye, causing precipitation of these complexes in the pores of the film. This gives rise to a filter effect (inactive dye molecules) [34].

Conclusions
ZnO thin films were synthesized by a low cost, electrodeposition technique, and used without any postdeposition treatment, to assembled DSCs. All the photoelectrodes were successfully used in DSCs. The morphological, structural and electrical characterization of the samples revealed the best performance for the more porous electrodes. The analysis of the amount of adsorbed dye reveals that a big amount of adsorbed dye decreases the PCE, probably due to the precipitation of the dye on the porous of the thin film. The devices present the feature to improve the PCE with the aging under ambient conditions. A remarkably high PCE 2.27% was obtained with the cell present with a mesoporous nanostructure. For this device the photovoltaic parameters were VOC ¼ 0:59 V, the JSC ¼ 7:64 mA=cm 2 , FF ¼ 50:41%, and the PCE ¼ 2:27%.