Crystallographic phase-mediated dye-sensitized solar cell performance of ZnO nanostructures

Available online at www.sciencedirect.com

Scripta Materialia 69 (2013) 291–294 www.elsevier.com/locate/scriptamat

Crystallographic phase-mediated dye-sensitized solar cell performance of ZnO nanostructures T. Ganesh,a Sambhaji S. Bhande,b Rajaram S. Manea,b,⇑ and Sung-Hwan Hana,⇑ a

Inorganic Nanomaterials Laboratory, Department of Chemistry, Hanyang University, Sungdong-Ku, Haengdang-dong 17, Seoul 133-791, Republic of Korea b Center for Nanomaterials and Energy Devices, School of Physical Sciences, Swami Ramanand Teerth Marathwada University, Dnyanteerth, Vishnupuri, Nanded 4316006, India Received 1 March 2013; accepted 25 April 2013 Available online 7 May 2013

A facile fabrication of ZnO nanostructures with different crystallographic phases on indium tin oxide glass substrates by using an aqueous solution growth method at low temperature is explored. Electron charge transportation in ZnO crystallographic phases is probed for dye-sensitized solar cells. Due to the decrease in the electron lifetime from 5.5 to 1.36 ms the ZnO electrode primed with nanoparticles and sensitized with ruthenium dye (N3) delivers light conversion efficiency of up to 1.33%, which is higher than electrodes made up of nanoplates (0.81%), nanorods (0.50%) and microrods (0.28%). Ó 2013 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. Keywords: ZnO crystallographic phases; Dye-sensitized solar cells (DSSCs); Impedance analysis; J–V; IPCE

Dye-sensitized solar cells (DSSCs) using sensitizing dye molecules linked to wide band gap mesoporous metal oxides have attracted a great deal of attention [1]. In general, DSSCs are composed mainly of abundantly available and non-toxic metal oxides, a monolayer of dye molecules and an electrolyte. Light harvesting is a key factor for improving the efficiency of DSSCs. The molar extinction coefficient, surface coverage of the sensitizing dye and the total surface area of the metal oxide electrode are the main factors determining the light absorption and, thereby, the power conversion efficiency of DSSCs [2,3]. In addition to the development of suitable dyes, electron transport through mesoporous oxides needs to be probed in detail, with the specific intention of finding materials of higher electron mobility ensuring longer electron lifetime, and it is also necessary to determine the intrinsic material composition and the properties of the loading dye. ZnO is a uniquely sensitive material that belongs to the group IIb–VI compound semiconductors whose intrinsic properties

⇑ Corresponding authors. Address: Inorganic Nanomaterials Laboratory, Department of Chemistry, Hanyang University, Sungdong-Ku, Haengdang-dong 17, Seoul 133-791, Republic of Korea (R.S. Mane); e-mail addresses: [email protected] (R.S. Mane); [email protected] (S.-H. Han)

depend strongly on their crystallographic phase, orientation, surface treatment and nanostructure form. Apart from micro- and nanorods of the same orientations, a variety of nanostructures including nanoparticles (NPs), nanowires (NWs), tetrapods, hierarchical structures, nanoflowers, nanotips, nanosheets, nanotubes and branched nanostructures have been widely employed in recent years. The growth of ZnO can be deduced from its crystal lattice structure or the nucleation sites of the substrate surface, and the concentration and electrostatic interaction between the cations (Zn2+ and ZnOH+) and anions (OH) influences growth along polar or non-polar facets [4]. The most prominent ZnO rods are composed of a polar metastable phase and a stable non-polar Zn-terminated (Zn-rich) phase. These phases influence the material surface characteristics, including the charge transfer kinetics and dye-loading properties, in addition to surface defects which are affected by the preparation conditions [5]. The electron transport of ZnO NPs can be increased by about two orders of magnitudes and the electron mobility by several orders of magnitudes when a thin film of ZnO NPs is deposited on vertically aligned ZnO NWs [6]. Furthermore, the unidirectional growth of ZnO nanostructures with fewer surface traps enables a faster and more direct electrical pathway for the

1359-6462/$ - see front matter Ó 2013 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.scriptamat.2013.04.021

292

T. Ganesh et al. / Scripta Materialia 69 (2013) 291–294

collection of the electrons [7]. To investigate the relationship between the ZnO phase and DSSC performance, electron lifetime values for NRs, microrods (MRs), nanoplates (NPls) and NPs were calculated. Ruthenium(II) cis-di(thiocyano) bis(2,20 -bypyridyl-4,40 dicarboxylic acid) (N3) dye was applied for a fixed, i.e. 45 min, period of time [8]. Various ZnO nanostructures, including NPs, NRs, NPls and MRs, were produced via different procedures [9–11]. In brief, a commercially available ZnO NPs dispersion, electrodeposited NPls, chemically deposited NRs and MRs electrodes were presented as the dyeloading agents. The film thickness of the NPs electrode was varied by dilution with distilled water in the ZnO dispersion and by dip coating, i.e. qualitative dilution of ZnO dispersion can form a thin film via a dip-coating process. It is worth mentioning here that the growth conditions for NPls, MRs and NRs were stringent and the film thickness was 1–1.5 lm. The nanostructural evolution of ZnO NPs, NPls, NRs and MRs was monitored by scanning electron microscopy (SEM) with a Cambridge Stereoscan 250-MK unit. The Zn:O ratio in the ZnO nanostructures was confirmed by energy-dispersive X-ray analysis using an attachment on an EM912 transmission electron microscope. Film thickness was measured with a surface profilometer (DEKTAK 3) by electromechanically moving the samples beneath a diamond-tipped stylus. For structural elucidation, Xray diffraction (XRD) patterns were recorded. The electrochemical impedance spectroscopy (EIS) measurements were performed using a BAS-Zahner IM6 impedance analyzer. For the solar cells, an electrolyte solution consisting of 15 ml methoxyacetonitrile (98%) containing 0.6 M 1-hexyl-2-3-dimethylimidazolium iodide (C6DMI), 0.1 M lithium iodide (LiI), 0.05 M iodide (I2) and 0.5 M 4-tert-butylpyridine (t-BPy) was used. The soaking time for all the electrodes was 45 min and the DSSCs measurements were performed at a photointensity of 100 mW cm2 (0.25 cm2) using a solar simulator. The incident photon-to-current conversion efficiency (IPCE) was measured using a PolaronixÒ K3100 spectral IPCE measurement system (South Korea); prior to the sample measurement, the equipment was calibrated with a standard silicon photodiode. The crystallographic phases of ZnO nanostructures are shown in Figure 1. The (0 0 1) plane observed in the peripheral phase of the hexagonal rods appears the region more likely to absorb the dye molecules. The ZnO nanostructures for DSSCs consist of a non-stoichiometric excess of oxygen ions compared to zinc ions. In the growth of NPls, the Cl ions from the KCl electro-

lyte used in the electrodeposition process prevent growth from the basal plane and, consequently, the vertical stacking of ZnO hexagonal platelets is promoted. These platelets are more horizontally stacked. Figure 2a–d presents different ZnO nanostructures employed in this study. The NPs are not exactly spherical as seen in the inset of Figure 2a, but are the agglomeration of many small spherical crystallites with diameters ranging from 20 to 30 nm. These irregular crystallites are important for light scattering and dye absorption to enhance DSSCs performance. The NPls morphology of the ZnO grown against the indium tin oxide (ITO) substrate surface is shown in Figure 2b. The horizontal in-plane top width of the individual ZnO NPls (inset) is 100– 150 nm. The vertical placement of the individual plate makes it difficult to visualize at the bottom even under high-resolution scanning [12]. The ZnO NRs of 100– 120 nm in diameter and MRs of 500–600 nm in diameter, respectively, are clearly distinguished from one another (Fig. 2c and d). Uniformly distributed ZnO NWs are grown randomly with the evolution of a neck-like structure at their tips (Fig. 2c, inset). MRs composed of horizontally stacked hexagonal plates are perpendicular to the substrate surface (Fig. 2d, inset). Under the influence of non-covalently bonded intermolecular forces, such as hydrogen bonding, van der Waals interactions or electrostatic forces, the nuclei will become clustered and possibly rearrange into crystallographically organized configurations, ensuring the formation of various nanostructures [13]. The intensity of the XRD for the (0 0 2) reflection plane varies (Fig. 3), indicating that there must be a change in the structural, optical and electrical properties of ZnO phases. The intensity of the (0 0 2) plane is systematically increased from NPs to the MRs. Due to fewer oxygen ions, and a lower ionic diffusion resistance, i.e. a higher charge transfer resistance, the ZnO NPs phase was confirmed to have a weaker (0 0 2) reflection plane intensity compared to that of other phases. Wurtzitetype hexagonal ZnO rods have six prismatic faces in addition to the polar basal ð0 0 1Þ and peripheral polar metastable (0 0 1) plane (c-axis growth) [5,14]. In the

Figure 1. Crystallographic phases of ZnO: (a) NPs, (b) NPls, (c) NRs and (d) MRs.

Figure 2. SEM images of ZnO: (a) NPs, (b) NPls, (c) NRs and (d) MRs; inset shows a magnified version.

T. Ganesh et al. / Scripta Materialia 69 (2013) 291–294

Figure 3. (a) XRD, (b) Nyquist, (c) J–V and (d) IPCE measurements of different ZnO crystallographic phases discussed in Scheme 1. Arrow indicates the NPs–NPls–NRs–MRs direction.

growth phase, the ammonium salts along with Zn2+ precursor in solution form Zn2+–amino complexes, thereby attaching to the six prismatic side phases of the ZnO, and preventing their lateral growth. Energy-dispersive X-ray analysis shows a higher content of oxygen ions compared to zinc ions in ZnO NPs. The zinc/oxygen ratio is decreased from NPs to MRs. We presume that the presence of excess oxygen ions in the MRs phase might have restricted the flow of photoexcited electrons from the metal oxide to a charge-collecting electrode, i.e. ITO. For example, in the formation of nanotubes by etching NRs using an alkaline KCl solution, the polar metastable (0 0 1) axis surface composition differs from that of the non-polar stable six prismatic Zn-terminated surface [15], enabling faster etching on the polar metastable (0 0 1) phase. Hence, the surface composition of NRs/MRs consists of excess zinc ions, which could be a probable reason for the lower dye adsorption, and increased amount of oxygen ions in NPs and NPls. To obtain more qualitative evidence for the charge transport, recombination and lifetime of the electrons, EIS measurements were performed under open circuit potentials in 1 M KCl electrolyte. As observed from the semicircle in the high-frequency range, the charge transfer resistance of the ZnO NPs electrode is the lowest of all the nanostructures examined here. The straight line in the low-frequency region is due to the diffusion of Cl ions from the KCl electrolyte into the ZnO electrode matrix as seen in Figure 3b. The initial lag in the highfrequency region is considered to correspond to the ohmic resistance of the electrolyte solution. The smaller semicircular loop of the ZnO NPs electrode is increased

293

to a maximum in the ZnO MRs electrode. A smaller semicircular loop indicates a lower Rct [16]. Therefore, the ZnO electrode with high (0 0 2) reflection plane intensity has significantly higher charge transfer resistance (Rct) and thereby limits the charge injection kinetics when used in DSSCs. Thus, by lessening the (0 0 2) plane intensity, better charge transfer kinetics, and hence enhanced performance, can be obtained due to the low Rct value. The electron transport and recombination in dye-sensitized ZnO NR solar cells using EIS is widely reported. The electron transport and the recombination rates are exponentially distributed by the surface traps, thereby limiting the fill factor (ff) and hence the power efficiency of the device [16]. The frequency of the peak in the middle-frequency region (fmid) of the Bode plote (not shown) is 28.7 Hz for the ZnO NPs electrode, less than for the NPl (36.8 Hz), NR (36.8 Hz) and MRs (117 Hz) electrodes. The electron lifetime can be calculated from the equation se = 1/2pfmid. The electron lifetime (se) values are 5.5, 4.32, 4.32 and 1.36 ms for ZnO NPs, NPls, NRs and MRs electrodes, respectively. The similarity of the electron lifetimes for NPls and NRs electrodes suggests that the ZnO NRs are seeded with 100 nm NPs film, possibly enhancing their lifetime. The longer electron lifetime for NPs electrode indicates more effective suppression of the reverse reaction between photoelectrons in the conduction band of ZnO and I 3 in the electrolyte, which is reflected in the improvement of the open circuit voltage (Voc), photocurrent and energy conversion efficiency of the device [17]. Figure 3c presents the current–voltage (J–V) curves of the ZnO NPs, NPls, NRs and MRs electrodes loaded with N3 dye for 45 min. Instead of thick films, thin ZnO films (11.5 lm) were preferred to elucidate the material characteristics, as uniform dye percolation and a pit-free surface can be obtained. Table 1 depicts the electronic parameters obtained from the J–V curves, as well as the quantity of dye loaded on the ZnO nanostructures. The ZnO NPs electrode efficiency is higher than that of the corresponding NPls, NRs and MRs electrodes. The ZnO electrode composed of NPs show a current density (Jsc) of 4.11 mA cm2, a Voc of 0.69 V, a fill factor (ff) of 0.46 and a power conversion efficiency (g%) of 1.13%. The Jsc values are decreased from 4.11 to 1.06 mA cm2 when the (0 0 2) reflection plane intensity is increased from NPs to MRs. For the ZnO NPls and NRs electrodes, the electronic parameters are in between those of NPs and MRs (NPl: Jsc = 3.45 mA cm2, ff = 0.36 and g = 0.81%; NRs: Voc=0.64 V, Jsc = 2.15 mA cm2, Voc = 0.61 V, ff = 0.38 and g = 0.50%). The higher Jsc of the ZnO NP electrode originates from the higher dye loading, associated with its reduced (0 0 2) peak intensity. This is due to its crystallographic phases, limiting the surface composition of

Table 1. The electronic parameters related to ZnO nanostructure-based DSSCs. ZnO phase type

Jsc (mA cm2)

Voc (V)

ff

g (%)

Dye amount (mol cm2)

Nanoparticles (NPs) Nanoplatelets (NPls) Nanorods (NRs) Microrods (MRs)

4.11 3.45 2.15 1.06

0.69 0.64 0.61 0.58

0.46 0.36 0.38 0.46

1.33 0.81 0.50 0.28

1.17  106 8.57  107 7.14  107 5.00  107

294

T. Ganesh et al. / Scripta Materialia 69 (2013) 291–294

zinc and oxygen ions, as well as the dye loading. The recombination process involves the recombination of photo-induced electrons from the dye-sensitized photoanodes with the I 3 ions in the electrolyte, which limits the charge collection on the conducting substrate surface side. Due to the lower Rct value, the electron–hole pairs in the ZnO MRs and NRs recombine faster than those in the NPs and NPls, leading to a decrease in the Voc value as well as a reduction in the device conversion efficiency [18]. Disadvantageously, with NRs/MRs electrodes, the effective surface area with respect to the NPs film is significantly less, thus rendering it less suitable for light harvesting from monolayer coverings of dye molecules [19]. It is worthwhile noting that the amounts of N3 dye loaded in thin NPs, NPls, NRs and MRs electrodes are 1.17  106, 8.57  107, 7.14  107 and 5  107 mol cm2. Dye adsorption is lowest for ZnO MRs and highest for ZnO NP electrodes. Figure 3d shows the IPCE measurements of the ZnO nanostructures. The IPCE of the ZnO NPs electrode is close to 12.2%, which is far higher than for NPls (9.4%), NRs (8.5%) and MRs (4.4%) electrodes, respectively. This implies that not only are the one-dimensional nanostructures of ZnO uniquely required for fast electron transportation but also the porosity and charge transfer resistance of ZnO, originating from its crystallographic phases influenced by its zinc and oxygen contents, are equally critically important when designing efficient ZnO-based DSSCs. In summary, the impact of the crystallographic phases of ZnO on DSSCs performance is investigated. The intensity of the (0 0 2) reflection plane in the XRD spectra is lower for ZnO NPs and higher for MRs. The surface composition of zinc and oxygen ions critically influences the dye loading amount, charge transfer resistance and device conversion efficiency, causing the thin NPs electrode to deliver an efficiency of 1.33%. However, due to the higher charge transfer resistance, the ZnO electrode composed of MRs shows a lower conversion efficiency (0.05%) which opens a new avenue in designing efficient ZnO-based DSSCs, in particular, in solid-state DSSCs where charge collection and interfacial contacts are of prime importance.

This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MEST) (No.2012-0008678). [1] H.J. Snaith, L. Schmidt-Mende, Adv. Mater. 19 (2007) 3187. [2] T.P. Chou, Q. Zhang, G.E. Fryxell, G. Cao, Adv. Mater. 19 (2007) 2588. [3] H.J. Snaith, G.L. Whiting, B. Sun, N.C. Greenham, W.T.S. Huck, R.H. Friend, Nano Lett. 5 (2005) 1653. [4] Y. Tong, Y. Liu, C. Shao, Y. Liu, C. Xu, J. Zhang, Y. Lu, D. Shen, X. Fan, J. Phys. Chem. B 110 (2006) 14714. [5] L. Vayssieres, K. Keis, A. Hagfeldt, S.E. Lindquist, Chem. Mater. 13 (2001) 4395. [6] A.B.F. Martinson, J.E. Mcgarrah, M.O.K. Parpia, J.T. Hupp, Phys. Chem. Chem. Phys. 8 (2006) 4655. [7] S.B. Ambade, R.S. Mane, A.V. Ghule, M.G. Takwale, A. Abhyankar, B. Cho, S.H. Han, Scr. Mater. 61 (2009) 12. [8] M. Law, L.E. Greene, J.C. Jonhson, P. Yang, Nat. Mater. 4 (2005) 455. [9] H.Y. Park, H.Y. Go, S. Kalme, R.S. Mane, S.H. Han, M.Y. Yoon, Anal. Chem. 81 (2009) 4280. [10] T. Ganesh, R.S. Mane, G. Cai, J.H. Chang, S.H. Han, J. Phys. Chem. C 113 (2009) 7666. [11] R.S. Mane, S.J. Roh, S.H. Han, Chem. Lett. 34 (2005) 536. [12] R.S. Mane, H.M. Nguyen, T. Ganesh, S.B. Ambade, N. Kim, S.H. Han, Electrochem. Commun. 11 (2009) 752. [13] C.D. Lokhande, A. Barkschat, H. Hributsch, Sol. Energy Mater. Sol. Cells 79 (2003) 293. [14] D. Polsongkrama, P. Chamninok, S. Pukird, L. Chow, O. Lupan, G. Chai, H. Khallaf, S. Park, A. Schulte, Physica B 403 (2008) 3713. [15] J. Elias, R. Tena-Zaera, G.-Y. Wang, C. Levy-Clement, Chem. Mater. 20 (2008) 6633. [16] M. Wang, P. Chen, R. Humphry-Baker, M. Zakeeruddin, M. Gratzel, Chem. Phys. Chem. 10 (2009) 290. [17] W. Zhang, R. Zhu, X. Liu, B. Liu, S. Ramakrishna, Appl. Phys. Lett. 95 (2009) 43304. [18] R.S. Mane, D.V. Shinde, S.J. Yoon, S.B. Ambade, J.K. Lee, S.H. Han, Appl. Phys. Lett. 101 (2012) 31906. [19] I. Gonzalez-Valls, M. Lira-Cantu, Energy Environ. Sci. 2 (2009) 19.