Solution-derived 40 µm vertically aligned ZnO nanowire arrays as photoelectrodes in dye-sensitized solar cells

Home Search Collections Journals About Contact us My IOPscience Solution-derived 40 µm vertically aligned ZnO nanowire arrays as photoelectrodes in dyesensitized solar cells This content has been downloaded from IOPscience. Please scroll down to see the full text. 2010 Nanotechnology 21 195602 (http://iopscience.iop.org/0957-4484/21/19/195602) View the table of contents for this issue, or go to the journal homepage for more Download details: IP Address: 133.1.248.56 This content was downloaded on 22/10/2014 at 04:34 Please note that terms and conditions apply. IOP PUBLISHING NANOTECHNOLOGY Nanotechnology 21 (2010) 195602 (9pp) doi:10.1088/0957-4484/21/19/195602 Solution-derived 40 µm vertically aligned ZnO nanowire arrays as photoelectrodes in dye-sensitized solar cells Jijun Qiu1,2 , Xiaomin Li1,4 , Fuwei Zhuge1 , Xiaoyan Gan1 , Xiangdong Gao1 , Weizhen He3 , Se-Jeong Park3 , Hyung-Kook Kim3 and Yoon-Hwae Hwang3,4 1 State Key Laboratory of High Performance Ceramics and Superfine Microstructures, Shanghai Institute of Ceramics, Chinese Academy of Sciences, Shanghai 200050, People’s Republic of China 2 RCDAMP, Pusan National University, Busan 609-735, Korea 3 Department of Nano-Materials Engineering & BK 21 Nano Fusion Technology Division, Pusan National University, Miryang 627-706, Korea E-mail: jj qiu@163.com, Lixm@mail.sic.ac.cn and yhwang@pusan.ac.kr Received 19 November 2009, in final form 20 November 2009 Published 21 April 2010 Online at stacks.iop.org/Nano/21/195602 Abstract Well-aligned ZnO nanowire arrays with a long length of more than 40 μm were prepared successfully by using the polyethylenimine (PEI)-assisted preheating hydrothermal method (PAPHT). Several important synthetic parameters such as PEI content, growth time, preheating time and zinc salt concentration were found to determine the growth of ultralong ZnO nanowire arrays, including length, diameter, density and alignment degree. The photoluminescence (PL) spectrum of as-grown ultralong ZnO nanowire arrays revealed a UV emission and a yellow emission, which was attributed to the absorbed hydroxyl group based on the peak shift after annealing in various atmospheres. The performance of dye-sensitized solar cells (DSSCs) increased with increasing length of ZnO nanowire arrays, which was mainly ascribed to the aggrandized photocurrent and reduced recombination loss according to electrochemical impedance spectroscopy (EIS). A maximum efficiency of 1.3% for a cell with a short-circuit current density (Jsc ) = 4.26 mA cm2 , open-circuit voltage (Voc ) = 0.69 V and (fill factor) FF = 0.42 was achieved with a length of 40 μm. S Online supplementary data available from stacks.iop.org/Nano/21/195602/mmedia (Some figures in this article are in colour only in the electronic version) recombination rate, which can be achieved in the following seven ways. First, is to maximize the area of the dye-TiO2 interface by using nanocrystalline-mesoporous architecture to increase the dye loading [1]. Second is to dope nanocrystalline TiO2 with special elements to increase the open-circuit voltage of the cells [2]. Third is to replace the nanocrystalline-mesoporous structure by one-dimensional single-crystalline array structures to decrease the electron traps in the nanocrystalline boundaries [3]. Fourth is to coat the nanocrystallines or one-dimensional nanostructures in a conformal metal oxide shell to suppress the carrier recombination by passivating recombination centers and introducing an energy barrier [4]. Fifth is to synthesize new 1. Introduction As a third-generation photovoltaic cell, dye-sensitized solar cells (DSSC) are promising devices for inexpensive, stable, large-scale soar energy conversion. Although the solar power conversion efficiencies of DSSC have been improved in the recent two decades, they are still lower than 20% for conventional Si-based solar cells. In principle, the improvement in DSSC performance has been made by increasing the light-harvesting capability of the photoelectrodes and by reducing the electron–hole 4 Authors to whom any correspondence should be addressed. 0957-4484/10/195602+09$30.00 1 © 2010 IOP Publishing Ltd Printed in the UK & the USA Nanotechnology 21 (2010) 195602 J Qiu et al as 20 μm were successfully fabricated by the PEI-assisted hydrothermal method (PAHT) at 92 ◦ C for 50 h. However, without the additive the length of the nanorods was no more than 5 μm. In this synthesis process, the growth solution needs to be repeatedly refreshed every 2.5 h to complement the depletion of reagents for a rapid growth rate of 0.5 μm h−1 . At the same time, acid must be added to the growth solution to decrease the pH value in the range from 10.5 to 7.2, which ensures the PEI can be adsorbed on the lateral plane of the ZnO nanorods by electrostatic attraction to achieve a large aspect ratio of 100 [20]. However, the introduction of fresh solution and the acid not only increases the fabrication cost, but also decreases the reproducibility, because ZnO seeds are easily destroyed by fresh growth solution involving hydrogen ions (H+ ) in the initial growth stage due to low chemical stability. Moreover, ZnO nanowire arrays with more than 30 μm length have not been successfully fabricated by the PAHT method. Herein, an improved PAHT, called preheating PEI-assisted hydrothermal synthesis (PPAHT), was developed to fabricate longer ZnO nanowire arrays with high reproducibility. Firstly, a preheating process is employed to avoid adding acid to the growth solution and the ZnO dissolution, giving rise to a higher reproducibility. At the same time, the population density of ZnO nanowires can be easily tuned by controlling the preheating time. Secondly, the original growth solution has been utilized until the whole growth process is over, which significantly decreases the fabrication cost. In this paper, we investigated the effect of the synthetic parameters, involving PEI content, growth solution concentration and preheating time on the length and aspect ratio of ZnO nanowire arrays obtained from the PPAHT method. In addition, we also used synthesized ZnO nanowires to assemble DSSC, and to see if the further increase of nanowire length can improve the device performance. dye sensitizers with a higher molar extinction coefficient and broader spectral response [5]. Sixth is to maximize the area of the interface between the transparent conducting substrate and the oxide layer to increase the electron capture [6]. The last way is to use a double light-scattering layer TiO2 [7] or ZnO [8, 9] film to enhance the light-scattering capability by significantly extending the traveling distance of light within the photoelectrode films. Among these, replacing TiO2 nanocrystalline-mesoporous networks with one-dimensional well-aligned semiconductor arrays has demonstrated a great potential to boost the DSSC performance [3, 10–14]. In 2005, Baxter [10] and co-workers first fabricated ZnO-nanowire-based DSSC using metal–organic chemical vapor deposition (MOCVD) and the hydrothermal method, and an overall light conversion efficiency as high as 0.5% was achieved. Law et al [3] also fabricated ZnOnanowire-based DSSC using a polyethylenimine (PEI)-assisted hydrothermal method, obtaining an overall efficiency as high as 1.5%. Subsequently, they further studied core– shell structures combining their fabricated ZnO nanowire system with layers of TiO2 , resulting in a further increase in the overall light conversion efficiency to 2.25% [4]. Henceforth, the one-dimensional ZnO nanowire and its derivative hierarchical nanostructures, including core–shell nanostructures [4], nanotubes [11], flower-like nanowires [12], dendritic nanowires [13] and branched nanowires [14], have been intensively investigated as alternative photoelectrodes to improve the DSSC performance. Although there are many available methods to fabricate one-dimensional ZnO nanostructures, it is also difficult to establish an optimal synthesis methodology for DSSC application because their performances depend to a great extent on the preparation conditions. To date, the most attractive synthesis technique for obtaining well-aligned ZnO nanowires and hierarchical or branched nanowires for DSSC is the hydrothermal method. Compared with other hightemperature physical or chemical vapor deposition methods, including metal–organic vapor deposition (MOCVD), atomic layer deposition (ALD), sputtering, thermal decomposition, thermal evaporation and condensation, the low-temperature hydrothermal method offers the potential for much lower cost because of eliminating the expense associated with high-temperature manufacturing and vacuum processing. In addition, hydrothermal synthesis is easily scalable to large areas and is compatible with roll-to-roll processing of soft plastic substrates. At the same time, DSSC using the hydrothermal-route-synthesized nanowires was found to yield a better efficiency compared to cells prepared using ZnO nanowires fabricated by the MOCVD method. Although a lot of effort has been made to investigate the effect of synthesis parameters on the dimension and quality of ZnO nanowires [3, 15–19], it is still difficult to fabricate ultralong ZnO nanowire arrays with a high aspect ratio by the hydrothermal method, which blocks a further rise of DSSC performance. Recently, it is found that the introduction of polyethylenimine (PEI) [3, 15] into the growth solution as an additive could efficiently increase the length and decrease the aspect ratio of ZnO nanowires. ZnO nanowire arrays as long 2. Experimental details 2.1. Synthesis of ultralong ZnO nanowire arrays 2.1.1. Deposition of ZnO seed layer by sol–gel dipcoating. Firstly, the transparent conducting F:SnO2 (FTO) substrates were seeded with a oriented ZnO thin film by using the sol–gel dip-coating method. Prior to the seeding, the FTO substrates were ultrasonically cleaned in ethanol and acetone for 15 min, respectively. For the oriented seeds, 0.075 mol Zn(CH3 COO)2 ·2H2 O was dissolved in a 100 ml 2-methoxyethanol-monoethanolamine solution at room temperature. The resultant solution was stirred at 60 ◦ C for 0.5 h to yield a clear and homogeneous sol. Then cleaned FTO substrates were slowly immersed into the ZnO sol for 10 s and then withdrawn at 3.0 cm min−1 . Subsequently they were preheated at 300 ◦ C for 10 min to form a ZnO gel seed layer. The desired thickness of ZnO seed layers was controlled by repeating the dip-coating and preheating processes. Finally, the preheated substrates were heated from 300 to 550 ◦ C at 2 ◦ C min−1 and held at 550 ◦ C for 1 h to obtain the oriented ZnO seed layers. 2 Nanotechnology 21 (2010) 195602 J Qiu et al 2.1.2. Synthesis of ultralong ZnO nanowire arrays by PPAHT. The growth solutions for the hydrothermal route were prepared by dissolving Zn(NO3 )2 ·6H2 O, (CH2 )6 N4 and PEI (branched, low molecular weight) in distilled water. The equivalent concentrations of Zn(NO3 )2 ·4H2 O and (CH2 )6 N4 were tuned in the range from 0.01 to 0.1 M to investigate the effect of the reagent concentration on the length and aspect ratio of ZnO nanowires. At the same time, the PEI content was changed from 0 to 0.015 M corresponding to the change of reagent concentration. Then 150 ml growth solution was first sealed in a glass bottle of maximum volume 250 ml and was then heated to 95 ◦ C for 2–12 h, namely a preheating process. After the preheating process, the ZnO-seeded FTO substrates were quickly immersed in the hot preheated growth solution and tilted against the wall of the bottle with ZnO seed layers facing down. Subsequently, the bottle was sealed again and heated to 95 ◦ C for various times from 12 to 150 h without any stirring. These various synthesis conditions were given in supporting information (available at stacks.iop.org/Nano/21/ 195602/mmedia) to study the evolution of the morphology as a function of the preheating time and the growth time. After growth, the resultant samples were removed from the vials, rinsed thoroughly with ethanol to remove any residual reactants and dried in air at 80 ◦ C. Figure 1. General transmission line model of ZnO-nanowire-based solar cells. 2.3.2. Characterization of ZnO-based DSSC photovoltaic properties. The photovoltaic properties of the fabricated solar cells under a simulated AM 1.5G illumination with a light intensity of 100 mW cm−2 (300 W, model 91160B, Oriel) were measured with the aid of a potentiostat (CHI-660B, CH Instruments). The electrochemical impedance spectra (EIS) measurements were carried out by applying bias of the opencircuit voltage under illumination of AM 1.5G 100 mW cm−2 over the frequency range of 10−1 –105 Hz with a 10 mV ac signal. Zview equivalent circuit modeling software was used to fit the data by utilizing a built-in extended element (DX type 11-Bisquert #2), which allows for transmission line modeling (see figure 1). In this equivalent circuit modeling, the fitted parameters denote the electron transport resistance (Rw ), recombination resistance (Rk ), the charge transfer constant (K eff ), capacitance of the electrical double layer (C p ), electron density (n s ) and the effective diffusion coefficient (Deff ). Although experimental error is inevitable in the determination of the parameters of each sample, it is too small (5%) to have any influence on the change tendency of the parameters. 2.2. Fabrication of ZnO-nanowire-based DSSCs The cis-bis(isothiocyanato)bis(2,2 -bipyridyl-4,4 -dicarboxylato)-ruthenium(II)bistetrabutylammonium (N719) (Solaronix S. A.) was used as a sensitizer. ZnO nanowire array photoelectrodes were prepared by immersing in a 0.0005 M ethanolic solution of N719 for 2 h. The counter electrode was an F:SnO2 substrate with a 20 nm Pt layer deposited by e-beam evaporation. The photoelectrode was sandwiched between the counter electrodes separated by a 100 μm hot-melt polypropylene spacer. Then the electrolyte, which consisted of 0.5 M tetrabutylammonium iodide, 0.05 M I2 and 0.5 M 4tertbutylpyridine in acetonitrile, was introduced between the electrodes by capillary forces. The active electrode area was typically 0.27 cm2 . Colloidal liquid silver (Ted Pella Inc., Redding, CA) was placed at the electrical contacts to improve the contact points and allowed to cure for 30 min at room temperature. 3. Results and discussion 3.1. Structure and morphology A well-aligned ultralong ZnO nanowire array was synthesized from optimizing the synthetic parameters (given in the supporting information available at stacks.iop.org/Nano/21/ 195602/mmedia), and its typical side-view FESEM images under different magnifications were shown in figures 2(a)– (c). Low- and medium-magnification FESEM images shown in figures 2(a) and (b) indicate that well-aligned ZnO nanowire arrays with high density are observed on the surface of the substrate. Almost all ZnO nanowires are perpendicular to the substrate and have a uniform length of 40 μm. The high-magnification FESEM image in figure 2(c) shows the diameter of the nanowires is distributed in the range of 100– 150 nm with an average value of 120 nm, and the aspect ratio of nanowires reaches up to around 330. A typical HRTEM image of a nanowire in figure 2(d) clearly shows the distances between two parallel lattice fringes, which is about 0.52 nm and corresponds to the (0002) planes of the wurtzite hexagonal ZnO, indicating that the nanowires are single crystalline with 2.3. Characterization 2.3.1. Characterization of structures and morphologies of ZnO nanowire arrays. The crystalline structure of the samples was characterized by using an x-ray diffraction (XRD) system (Bruker D8 ADVANCE) with Cu Kα of 1.5406 Å. The morphology and chemical composition of the samples were taken with a field-emission scanning electron microscopy (FESEM, Hitachi S-4700). Transmission electron microscope (TEM) and high-resolution (HRTEM) images and selected-area electron diffraction (SAED) patterns were obtained on a transmission electron microscope (JEOL JEM2100F). Photoluminescence spectra were measured by using a photoluminescence spectrometer (SLM48000DSCF), using an He–Cd laser (325 nm) as the excitation source. 3 Nanotechnology 21 (2010) 195602 J Qiu et al Figure 2. Characterizations of a typical ultralong ZnO nanowire array with a length of 40 μm. (a) Low-magnification, (b) medium-magnification and (c) high-magnification side-view SEM images of well-aligned ZnO nanowire arrays, respectively. (d) HRTEM image of an individual ZnO nanowire and (e) its corresponding SAED pattern. (f) EDS profile of ZnO nanowire arrays; the inset is the XRD pattern. are dominant, but there are still nanowire bundles on this ultralong nanowire array, as indicated by the arrowhead in figure 2(b). The occurrence of nanowire bundles is presumably due to the following three reasons: (1) tilted nanowire growth direction caused by a thin ZnO seed layer, which exhibits a polycrystalline structure with weak crystallinity; (2) the crowding effect resulting from the collision between neighboring nanowires in the solution growth process and (3) the bending and bundling of nanowires caused by surface tension effects of water evaporation during the drying process. 3.2. Synthetic parameters A series of experiments indicated that synthetic parameters also dominate the diameter, length, degree of alignment and growth rate of ZnO nanowire arrays, including the PEI content, preheating time, growth time, seed layer thickness and zinc salt concentration. Figure 3. The length and the diameter of ZnO nanowire arrays were plotted as a function of the PEI content in the precursor solution. the wurtzite hexagonal phase and grow preferentially along the [0001] (c axis) direction, which was further confirmed by the corresponding SAED dot pattern in figure 2(e). An energy dispersive spectroscopy (EDS) image of ZnO nanowire arrays, as shown in figure 2(f), contains only Zn and O elements, without any other impurity contamination. The XRD pattern in the scan range of 2θ = 20◦ –80◦ was illustrated in the inset of figure 2(f). No other diffraction peaks were detected, excluding peaks assigning to the FTO conducting layer and ZnO nanowires. All ZnO diffraction peaks are in good agreement with JCPDS card no. 36-1451 for a typical wurtzite hexagonal ZnO (P 63mc). A significantly higher (0002) diffraction peak intensity indicates that ZnO nanowires are preferentially oriented along the direction perpendicular to the (0001) crystallographic face (c-axis direction). It should be noted that ZnO nanowires perpendicular to the substrate 3.2.1. PEI content. Based on the reports by Yang [3] and Wu [20], the effect of the PEI content on the length of nanowires was first investigated. Figure 3 shows the plot of the length and diameter of ZnO nanowire arrays versus the PEI content in solution. The typical FESEM images of the ZnO nanowire arrays synthesized at 95 ◦ C for 24 h with various PEI contents in the range from 0 to 0.0015 M was shown in supporting information (figure S1 available at stacks.iop.org/ Nano/21/195602/mmedia). Without PEI, ZnO nanowire arrays with a length of 4 μm were obtained after growing for 24 h. A magnified FESEM image shown in the inset of figure S1 (available at stacks.iop.org/Nano/21/195602/mmedia) clearly shows that every nanowire has a syringe-like morphology, exhibiting an abrupt decrease in diameter along the c-axis direction of ZnO nanorods, which is similar to the results in papers reported by Sun [19, 21, 22]. With increasing PEI 4 Nanotechnology 21 (2010) 195602 J Qiu et al Figure 4. The length of ZnO nanowire arrays was plotted as a function of the growth time. The morphology change of ZnO nanowire arrays was shown in the inset side-view SEM images. Figure 5. The length of ZnO nanowire arrays was plotted as a function of the preheating time. The morphology change of ZnO nanowire arrays was shown in the inset side-view SEM images. content, the length of nanowires sharply increases and reaches a maximum of 12 μm when the PEI content was up to 0.010 M. As the PEI content increases from 0.010 to 0.012 M, the nanowire length decreases from 10 to 6.5 μm. Moreover, after the PEI concentration exceeds 0.012 M, nothing was found both on the substrates and the growth solution. At the same time, the ZnO seed layer was also etched off from the FTO substrate. Additionally, the average diameter of the nanowires linearly decreases with increasing PEI content, as shown in figure 3. Therefore, the optimum PEI content should be located in the range from 0.008 to 0.010 M to obtain ZnO nanowire arrays with the maximum length and aspect ratio. It is well known that ZnO has a polar hexagonal wurtzite structure consisting of two polar {0001} planes and six crystallographic non-polar {011̄0} planes. The positively charged Zn-terminated (0001) plane has higher surface energy, resulting in the faster growth rate along the [0001] direction. PEI is a non-polar polymer with a large amount of side amino groups (–NH2 ), which can be protonated over a wide range of pH values from 3 to 11, resulting in the increase of the solution’s pH value and positively charged PEI molecules [22]. The initial pH value of the precursor solution of Zn(NO3 )2 and HMT is about 6.2, falling in the range of protonation of PEI. Therefore, after introducing PEI, the pH value increases and is proportional to the PEI content. After the pH value of the growth solution is higher than the isoelectric point (IEP) of ZnO [23], the six lateral facets of the ZnO nanowires will be negatively charged due to the dissolution of ZnO into [Zn(OH)4 ]2− . Subsequently, the protonated positively charged PEI will be adsorbed onto the lateral facets of ZnO nanorods due to the electrostatic affinity, restraining the nanowire growth in the radial direction. At the same time, the dissolution of ZnO, especially ZnO particle precipitation in solution, can complement the depletion of Zn2+ , maintaining a high Zn2+ concentration and a high c-axis growth rate during the whole growth duration. However, as the pH value is higher than 13, the dissolution rate is much larger than the growth rate, resulting in the decrease of length and diameter with increasing PEI concentration. The dissolution can be confirmed by the reduced ZnO particle precipitants in solution with the increase in PEI content. 3.2.2. Growth time. Figure 4 shows the plot of the length of ZnO nanowire arrays versus the growth times. Clearly, under the optimum PEI content of 0.009 M, the growth time is the key parameter to dominate the length of ZnO nanowire arrays. On extending the growth time, the color of the growth solution changes from colorless transparent to turbid yellow, then to light brown and finally to dark red. The length linearly increases from 10 to 42 μm with an addition of about 0.3 μm h−1 . The steady and continuous growth is mainly derived from a continuous reactive ion supply, resulting from the accelerated dissolution of ZnO particle precipitation in the solution due to the introduction of PEI. At the same time, the diameter of every ZnO nanowire shows a slightly decreasing tendency along the c axis with increasing growth time, which is due to the decrease of the concentrations of the supplied reactive ions resulting from the depletion of the reactive mass. It can also be found from the side-view SEM image (inset in figure 4) that the alignment of ZnO nanowire arrays gets worse with increasing length, and various numbers of neighboring nanowires are slightly bent and cling to others, forming pyramid-like bundles after growing for 48 h. The bending and bundling of nanowires is a result of the balance between the mechanical flexibility and the capillary interaction, derived from the increased length (or larger aspect ratio) of nanowires and the larger surface tension of water during the drying process, respectively. 3.2.3. Preheating time. Figure 5 shows the plot of the length of ZnO nanowire arrays versus the preheating time. Clearly, ZnO nanowires synthesized from different preheating times exhibit obvious differences in length, diameter, density and alignment degree, as shown in the inset side-view SEM images. Firstly, the length linearly decreases from 20 to 7 μm as the preheating time increases from 1 to 12 h, as shown in figure 5. At the same time, the average diameter and density of ZnO nanowires decreases with increasing preheating 5 Nanotechnology 21 (2010) 195602 J Qiu et al Figure 6. The length of ZnO nanowire arrays was plotted as a function of the thickness of the ZnO seed layer. The morphology change of ZnO nanowire arrays was shown in the inset side-view SEM images. Figure 7. The length of ZnO nanowire arrays was plotted as a function of the concentration of zinc salt. The morphology change of ZnO nanowire arrays was shown in the inset side-view SEM images. layer thickness. Firstly, the surface roughness of the seed layer increases with increasing thickness, which increases the active nuclei site density and restricts the migration and coalescence of ZnO nuclei formed in the initial growth stage, finally resulting in an increase of nanowire density. Because of the polycrystalline feature of seeds, the crystallinity and c-axis orientation of ZnO seeds on the layer surface will be improved due to the decrease of lattice mismatch with the FTO substrate on increasing thickness, resulting in the improvement of the alignment of the nanowire arrays. At the same time, the increased nanowire density accelerates the effect of space confinement and selective competition between nanowires, further enhancing the alignment degree. This is because the increased nanowire density pronouncedly reduces the free growth space of an individual nanowire both in radial and axial directions. Therefore, the inclined nanowires grown from the seeds with tilted c axis to the substrate will collide with neighboring nanowires and stop growing, whereas,only the vertical nanowires grown from the nucleation seeds along the c axis perpendicular to the substrate can continuously grow unimpeded indefinitely in the axis direction, and selfassemble into vertical well-aligned nanowire arrays. The inverse relationship between the nanowire length and the density is attributed to the faster precursor consumption over dense areas, leading to lower local reactive concentration and growth rate. The thicker the seed layer, the higher the nanowire density. Therefore the greater nanowire surface area will lead to a larger precursor depletion near the surface and hence a lower surface reactive concentration, consequently resulting in decreasing the growth rate. time. Following the decrease in diameter, the alignment degree becomes poor, and some neighboring nanowires in an area with a diameter of 1–3 μm also bend towards the center, forming pyramid-like bundles at their tips, as shown in the inset SEM images. With the prolongation of preheating, the initial reagent concentration continuously decreases due to the formation of ZnO particle precipitation in the growth solution, resulting not only in the decrease of nucleation density on the seed layer, and consequently decreased nanowire density and diameter, but also in lowering the reagent supplying concentration and transport rate, and consequently a decreased growth rate of nanowires. At the same time, the mechanical flexibility of nanowires increases with the decrease in diameter, and slender nanowires facilitate bending and aggregation under the effect of centripetal surface tension induced by the rapid evaporation of water drops on the top of nanowires in the drying process. 3.2.4. ZnO seed layer thickness. Figure 6 shows the plot of the length of ZnO nanowire arrays versus the ZnO seed layer thickness. The side-view SEM images of the ZnO nanowires on seed layers with thicknesses of 30, 60 and 90 nm were shown in the inset of figure 6. Clearly, on increasing the seed layer thickness, the nanowire density increases, the degree of vertical alignment becomes better and the length sharply decreases. Although the nanowires with the longest length of 30 μm were obtained on the thinnest seed layer of 30 nm thickness, they exhibit the lowest density and the worst degree of vertical alignment. Vertically well-aligned nanowire arrays with a relative high density were formed on a seed layer with 50 nm thickness. However, the nanowire length shows a decrease of 10 μm compared with the length of nanowires grown from the thinner seed layer. No significant density or alignment changes were found after further increasing the seed layer thickness, except for a further decreased length. The reason for the morphology change can be explained by the diversity of surface roughness, crystal quality and reactive ion concentration resulting from variation of the seed 3.2.5. Zinc salt concentration. Figure 7 shows the plot of the length of ZnO nanowire arrays versus the zinc salt concentration in the growth solution and their corresponding side-view SEM images (inset). The ZnO nanowires were fabricated from various zinc salt concentrations ([Zn2+ ]) of 0.01, 0.025 and 0.06 M, which all have a fixed concentration ratio of Zn(NO3 )2 to HMT and PEI ([Zn2+ ]/[HMT] = 1, [Zn2+ ]/[PEI] = 4.5). Clearly, the length, the density and 6 Nanotechnology 21 (2010) 195602 J Qiu et al the corresponding alignment degree of nanowires increase with increasing [Zn2+ ]. For a relatively low [Zn2+ ], ZnO nanowires randomly grow from seed-layer-coated substrates with low density, not exhibiting an oriented growth feature perpendicular to the substrates. The length is about 30 μm after growing for 140 h. Although no evident improvement in the degree of alignment was found after increasing the amount of [Zn2+ ] to 0.025 M, the corresponding length of nanowires increases to 35 μm. As the amount of [Zn2+ ] increased to 0.06 M, vertically well-aligned nanowire arrays with a length of 44 μm were obtained. Obviously, the improvement in the degree of alignment can only be attributed to the further increased nanowire density, because there is no obvious diameter change with increasing [Zn2+ ]. It should be noticed that a larger concentration of [Zn2+ ] was not investigated in our case, because the plentiful emulsion precipitation of ZnO particles emerges during the preparation procedure of the precursor solution at room temperature. The increased nanowire density and the improvement of the degree of alignment are related to the increase of nucleation density in the initial growth stage, which results from increased [Zn2+ ]. The higher the precursor concentration, the larger the supersaturation. The more the nucleation density, the higher the nanowire density, and the better the alignment degree. The increase of the reagent concentration not only increases the transport rate of active ions, resulting in a relatively high nanowire growth rate, but also extends the virtual feeding time of active ions, maintaining the relatively high growth rate through the whole growth duration. Figure 8. Room temperature PL spectra of ultralong ZnO nanowire arrays after annealing in different atmospheres. for 1 h. At the same time, the ratio of UV to red emission further decreased to 0.05. The reason for the redshift of the yellow emission and the decreased intensity ratio of the UV to red emission is due to an excessively oxidized layer formed on the surface of ZnO nanorods by annealing in oxygen-rich conditions at high temperatures, resulting in increasing the defect concentrations related to excess oxygen [27]. Therefore, it is believed that the origin of the yellow emission is not related to excess oxygen defects. Otherwise, an increased intensity, rather than a redshift, could be observed in PL spectra after annealing in air due to the increase of oxygen defects. However, after annealing in vacuum at 550 ◦ C for 1 h, a blueshift from yellow emission to green emission centered at 515 nm (2.40 eV) was observed from the corresponding PL spectra shown in figure 8. The green emission is often attributed to the recombination of electrons and holes in singly ionized oxygen vacancies, and could be quenched or redshifted after annealing in an oxygen-rich atmosphere. In addition, the corresponding intensity ratio of UV to green emission sharply increased to 1.35, indicating a decrease of the point defect concentration and an improvement of crystallinity. This result indicates that the yellow emission in as-grown ZnO nanowire arrays is attributed to OH groups, instead of the commonly assumed interstitial oxygen defect. The reason for the blueshift and the enhancement of UV emission is due to the desorption of hydroxyl groups (150 ◦ C) and hydrogen (420 ◦ C) on the hydrothermally grown ZnO nanowires. Therefore, vacuum annealing at 550 ◦ C is an effective way to improve the optical properties of ultralong ZnO nanowire arrays resulting from desorbing the hydroxyl group absorbed from the growth precursor solution, whereas annealing in air will result in crystal degradation resulting from increasing oxygen defects. 3.3. Optical properties Figure 8 shows the room temperature (RM) PL spectra of ultralong ZnO nanowire arrays with a length of 42 μm. The PL spectrum of the as-grown sample is shown by a weak UV peak at 378 nm (3.28 eV) and a strong and broad yellow emission centered at 550 nm (2.25 eV). The PL intensity ratio of the UV emission to the yellow emission is very low (0.07), indicating the presence of high density point defects in the ultralong nanowire arrays grown at low temperatures. The UV emission originates the free excitonic recombination, which can be observed at room temperature due to the large exciton binding energy of ZnO (about 60 meV). It has been reported that thermal energy at room temperature may be enough to release bound excitons because the binding energy of bound excitons is only a few millielectronvolts [24]. The yellow emission is commonly reported in the PL spectrum of ZnO nanostructures and represents a common feature in samples prepared from aqueous solutions of Zn(NO3 )2 and HMT [25]. This emission is typically attributed to oxygen interstitials, although Li impurity can be another possible candidate [26]. Recently, hydroxyl groups are also proposed to explain the origin of the yellow emission in ZnO nanorods fabricated by the hydrothermal method [25]. In order to address the origin of the yellow emissions, post-annealing in various atmospheres was carried out and the resultant PL spectra were also shown in figure 8. An obvious redshift from the yellow emission to red emission centered at 600 nm (2.06 eV) was observed after annealing in air at 550 ◦ C 3.4. ZnO nanowire dye-sensitized solar cells Figure 9 shows the photocurrent–voltage ( J –V ) characteristics of ZnO nanowire dye-sensitized solar cells with various lengths. With increasing length of the nanowire array to 40 μm, the efficiency of the solar cell increases gradually. A maximum efficiency of 1.31% was achieved in the cell with a 7 Nanotechnology 21 (2010) 195602 J Qiu et al Table 1. Performances and electron transport properties of N719-sensitized ZnO nanowire DSSCs with various thicknesses. Length (μm) Jsc (mA cm−2 ) Voc (V) FF η (%) Rw () Rk () K err (s−1 ) C p (F) n s (cm−3 ) Deff (cm2 s−1 ) 10 20 30 40 2.95 3.49 4.15 4.26 0.71 0.72 0.70 0.69 0.35 0.38 0.38 0.42 0.73 0.97 1.13 1.31 15.6 14.6 12.7 2.44 199 180 138 90 11.9 11.9 8.07 8.07 1.2 × 10−4 1.4 × 10−4 1.3 × 10−4 1.8 × 10−4 3.69 × 1017 2.04 × 1017 1.21 × 1017 1.38 × 1017 1.02 × 10−4 3.98 × 10−4 1.16 × 10−3 7.02 × 10−3 Figure 9. J –V curves of the ZnO nanowire DSSCs fabricated from different lengths. Figure 10. Nyquist plots of the impedance data of the ZnO nanowire DSSCs constructed from different lengths. The solid lines are the fitting results based on the equivalent circuit model shown in figure 1. 40 μm ZnO nanowire array, with short-circuit current density (Jsc ) = 4.26 mA cm2 , open-circuit voltage (Voc ) = 0.69 V and (fill factor) FF = 0.42. The improvement of the conversion efficiency comes from the increased short-circuit current, since the open-circuit voltage decreases slightly in the lengthened nanowire array. The enlarged surface area in the lengthened nanowire array results in improved dye loading: besides, the multi-scattering effect between the nanowires enhances the optical path of the incident light in the nanowire array. Both of these effects contribute to the improved energy conversion of solar cells with longer nanowire arrays. The electron transfer processes in the nanowire array are further investigated by electrochemical impedance spectroscopy at the open-circuit condition in the frequency range of 10−1 –105 Hz. The results are fitted with the equivalent circuit in figure 1 and some basic parameters of rw , rk , Ck are obtained. The Nyquist plots of both the measured data points and fitted curves are shown in figure 10. The electron density n s and diffusion coefficient Deff in the ZnO nanowire are calculated according to the process proposed by [28]. The overall characteristics of the ZnO nanowire solar cells are summarized in table 1. The largest Helmholtz capacitance C p , which corresponds to the capacitance of the electrical double layer, is observed for the nanowire array of 40 μm, which is inconsistent with the enlarged surface area and dye loading capability. On the other hand, the recombination resistance Rk at the photoanode/electrolyte interface decreases gradually with increasing length of the nanowire array, indicating more recombination loss in the solar cells with longer nanowire arrays. However, by considering the enlarged surface area by rk = Rk /area, the recombination process is indeed slowed in the solar cells with longer nanowire arrays, as indicated by the recombination coefficient K eff = 1/τeff , in which τeff is the effective electron lifetime. Therefore, the enhanced overall efficiency in the DSSC with long ZnO nanowire arrays is therefore mainly caused by the aggrandized surface area and slowed recombination process, as is verified by the increased fill factor. Surprisingly, decreased electron transport resistance in the nanowire was obtained with decreasing electron density, which was in contrast with the trap/detrap transport model, in which electron transport was limited by the detrapping process of trapped electrons in the defects. This phenomenon could be attributed to locally concentrated electrons, probably at the bottom of the nanowire, or mutual scattering transfer of electrons in the nanowire, which are beyond the discussions of the current work. Consistently, the efficiency of the DSSC composed of ZnO nanowire arrays increases with increasing length of the nanowires, probably due to the enlarged dye loading and slowed recombination process. However, it should be noticed that the efficiency of ZnO nanowire solar cells is still much lower than that of porous TiO2 nanocrystalline solar cells, due to the unfavorable formation of the Zn2+ /dye-complex layer on the ZnO nanowire surface, which inhibited electron injection from dye molecules into the ZnO nanowire [29–31]. The efficiency of the dye-sensitized ZnO nanowire solar cells could be further improved by better engineering the ZnO nanowire and the interface of dye molecules and the nanowire, e.g. by modifying the ZnO nanowire with surface coatings (SiO2 , Al2 O3 and TiO2 ) to form various core/shell structures. 8 Nanotechnology 21 (2010) 195602 J Qiu et al 4. Conclusion [11] Martinson A B F, Elam J W, Hupp J T and Pellin M J 2007 ZnO nanotube based dye-sensitized solar cells Nano Lett. 7 2183 [12] Jiang C Y, Sun X W, Lo G Q and Kwong D L 2007 Improved dye-sensitized solar cells with a ZnO-nanoflower photoanode Appl. Phys. Lett. 90 263501 [13] Baxter J B and Aydil E S 2006 Dye-sensitized solar cells based on semiconductor morphologies with ZnO nanowires Sol. Energy Mater. Sol. Cells 90 607 [14] Cheng H M, Chiu W H, Lee C H, Tsai S Y and Hsieh W F 2008 Formation of branched ZnO nanowires from solvothermal method and dye-sensitized solar cells applications J. Phys. Chem. C 112 16359 [15] Tian Z R, Voigt J A, Liu J, Mckenzie B, Mcdermott M J, Rodriguez M A, Konishi H and Xu H 2003 Complex and oriented ZnO nanostructures Nat. Mater. 2 821 [16] Greene L E, Yuhas B D, Law M, Zitoun D and Yang P 2006 Solution-grown zinc oxide nanowires Inorg. Chem. 45 7535 [17] Wen B, Huang Y and Boland J J 2008 Controllable growth of ZnO nanostructures by a simple solvothermal process J. Phys. Chem. C 112 106 [18] Vayssieres L 2003 Synthesis and optical properties of white-light-emitting alumina/ZnO nanotubes Adv. Mater. 15 464 [19] Sun Y, Fuge G M, Fox N A, Riley D J and Ashfold M N R 2005 Synthesis of aligned arrays of ultrathin ZnO nanotubes on a Si wafer coated with a thin ZnO film Adv. Mater. 17 2477 [20] Wu W, Hu G, Cui S, Zhou Y and Wu H 2008 Epitaxy of vertical ZnO nanorod arrays on highly (001)-oriented ZnO seed monolayer by a hydrothermal route Cryst. Growth Des. 8 4014 [21] Sun Y, Riley D J and Ashfold M N R 2006 Mechanism of ZnO nanotube growth by hydrothermal methods on ZnO film-coated Si substrates J. Phys. Chem. B 110 15186 [22] Zhou Y, Wu W, Hu G, Wu H and Cui S 2008 Hydrothermal synthesis of ZnO nanorod arrays with the addition of polyethyleneimine Mater. Res. Bull. 43 2113 [23] Kosmulski M 2001 Chemical Properties of Material Surfaces (New York: Dekker) p 102 [24] Liu Y X, Liu Y C, Shao C L and Mu R 2004 Excitonic properties of ZnO nanocrystalline films prepared by oxidation of zinc-implanted silica J. Phys. D: Appl. Phys. 37 3025 [25] Djurišić A B et al 2007 Defect emissions in ZnO nanostructures Nanotechnology 18 095702 [26] Li D, Leung Y H, Djurišić A B, Liu Z T, Xie M H, Shi S L, Xu S J and Chan W K 2004 Different origins of visible luminescence in ZnO nanostructures fabricated by the chemical and evaporation methods Appl. Phys. Lett. 85 1601 [27] Kang H S, Kang J S, Kim J W and Lee S Y 2004 Annealing effect on the property of ultraviolet and green emissions of ZnO thin films J. Appl. Phys. 95 1246 [28] Bisquert J, Garcia-Belmonte G, Fabregat-Santiago F, Ferriols N, Bogdanoff P and Pereira E C 2000 Doubling exponent models for the analysis of porous film electrodes by impedance relaxation of TiO2 nanoporous in aqueous solution J. Phys. Chem. B 104 2287 [29] Horiuchi H, Katoh R, Hara K, Yanagida M, Murata S, Arakawa H and Tachiya M 2003 Electron injection efficiency from excited N3 into nanocrystalline ZnO films: effect of (N3-Zn2+) aggregate formation J. Phys. Chem. B 107 2570 [30] Westermark K, Rensmo H, Siegbahn H, Keis K, Hagfeldt A, Ojamae L and Persson P 2002 PES studies of Ru(dcbpyH2 )2 (NCS)2 adsorption on nanostructured ZnO for solar cell applications J. Phys. Chem. B 106 10102 [31] Chou T P, Zhang Q F and Cao G Z 2007 Effects of dye loading conditions on the energy conversion efficiency of ZnO and TiO2 dye-sensitized solar cells J. Phys. Chem. C 111 18804 Vertically aligned ZnO nanowire arrays with a long length and high aspect ratio were prepared successfully by using the polyethylenimine (PEI)-assisted preheating hydrothermal method. Several important synthetic parameters were found to determine the growth of ultralong ZnO nanowire arrays. The lengths of ZnO nanowire arrays are mainly determined by PEI content and growth time, whereas the preheating time, the thickness of the ZnO seed layer and the concentration of zinc salt have a large influence on the density and the alignment degree of ZnO nanowire arrays. The room temperature photoluminescence (PL) spectrum of as-grown ultralong ZnO nanowire arrays reveals a UV emission and a yellow emission. The yellow emission was attributed to the absorbed hydroxyl group based on the peak shift after annealing at various atmospheres: a redshift after annealing in air and a blueshift after annealing in vacuum. The performance of dye-sensitized solar cells (DSSCs) increased with increasing length of the ZnO nanowire arrays, indicating that the ultralong ZnO nanowire arrays have great potential in improving the performance of dye-sensitized solar cells. References [1] O’Regan B and Grätzel M 1991 A low-cost, high-efficiency solar cell based on dye-sensitized colloidal TiO2 films Nature 353 737 [2] Wijayarathna T R C K, Aponsu G M L P, Ariyasinghe Y P Y P, Premalal E V A, Kumara G K R and Tennakone K 2008 A high efficiency indoline-sensitized solar cell based on a nanocrystalline TiO2 surface doped with copper Nanotechnology 19 485703 [3] Law M, Greene L E, Johnson J C, Saykally R and Yang P D 2005 Nanowire dye-sensitized solar cells Nat. Mater. 4 455 [4] Law M, Greene L E, Radenovic A, Kuykendall T, Liphardt J and Yang P 2006 Core–shell nanowire dye-sensitized solar cells J. Phys. Chem. B 110 22652 [5] Chen K S, Liu W H, Wang Y H, Lai C H, Chou P T, Lee G H, Chen K, Chen H Y, Chi Y and Tung F C 2007 New family of ruthenium-dye-sensitized nanocrystalline TiO2 solar cells with a high solar-energy-conversion efficiency Adv. Funct. Mater. 17 2964 [6] Wang H W, Ting C F, Hung M K, Chiou C H, Liu Y L, Liu Z, Ratinac K R and Ringer S P 2009 Three-dimensional electrodes for dye-sensitized solar cells: synthesis of indium–tin-oxide nanowire arrays and ITO/TiO2 core–shell nanowire arrays by electrophoretic deposition Nanotechnology 20 055601 [7] Ito S, Nazeeruddin M K, Zakeeruddin S M, Péchy P, Comte P, Grätzel M, Mizuno T, Tanaka A and Koyanagi T 2009 Study of dye-sensitized solar cells by scanning electron micrograph observation and thickness optimization of porous TiO2 electrodes Int. J. Photoenergy 51 7609 [8] Zhang Q F, Chou T P, Russo B, Fryxell G E, Jenekhe S A and Cao G Z 2008 High conversion efficiency in dye-sensitized solar cells through controlled aggregation of ZnO nanocrystallites Angew. Chem. Int. Edn 47 2402 [9] Zhen Y Z, Tao X, Wang L X, Xu H, Hou Q, Zhou W L and Chen J F 2009 Novel ZnO-based film with double light-scattering layers as photoelectrodes for enhanced efficiency in dye-sensitized solar cells Chem. Mater. 22 928 [10] Baxter J B and Aydil E S 2005 Nanowire-based dye-sensitized solar cells Appl. Phys. Lett. 86 053114 9