Oriented Assembled TiO2 Hierarchical Nanowire Arrays with Fast

Mar 14, 2014 - Developing high surface area nanostructured electrodes with rapid charge transport is essential for artificial photosynthesis, solar ce...
5 downloads 19 Views 6MB Size
Download PDF
Letter pubs.acs.org/NanoLett

Oriented Assembled TiO2 Hierarchical Nanowire Arrays with Fast Electron Transport Properties Xia Sheng,† Dongqing He,† Jie Yang,† Kai Zhu,*,‡ and Xinjian Feng*,† †

Suzhou Institute of Nano-Tech and Nano-Bionics, Chinese Academy of Sciences, Suzhou 215123, Jiangsu, China National Renewable Energy Laboratory, Golden, Colorado 80401, United States

‡

S Supporting Information *

ABSTRACT: Developing high surface area nanostructured electrodes with rapid charge transport is essential for artificial photosynthesis, solar cells, photocatalysis, and energy storage devices. Substantial research efforts have been recently focused on building one-dimensional (1D) nanoblocks with fast charge transport into three-dimensional (3D) hierarchical architectures. However, except for the enlargement in surface area, there is little experimental evidence of fast electron transport in these 3D nanostructure-based solar cells. In this communication, we report single-crystal-like 3D TiO 2 branched nanowire arrays consisting of 1D branch epitaxially grown from the primary trunk. These 3D branched nanoarrays not only demonstrate 71% enlargement in large surface area (compared with 1D nanowire arrays) but also exhibit fast charge transport property (comparable to that in 1D single crystal nanoarrays), leading to 52% improvement in solar conversion efficiency. The orientated 3D assembly strategy reported here can be extended to assemble other metal oxides with one or multiple components and thus represents a critical avenue toward high-performance optoelectronics. KEYWORDS: Three dimensional, nanowire, semiconductor, charge transport, solar cells

D

that would limit charge transport within these hierarchical structures. Currently there is little experimental evidence of fast electron transport in these 3D nanostructure-based solar cells. Herein, we describe the fabrication and fast charge transport properties of oriented assembled TiO2 hierarchical nanoarrays consisting of 1D branches epitaxially grown from the primary trunk. We find that charge transport in these single-crystal-like branched nanowire (B-NW) arrays is about 200 times faster than that in the nanoparticle (NP) films. We further demonstrate that the 3D B-NW arrays have a larger surface area (71% higher) relative to the 1D NW arrays, leading to 52% improvement in solar conversion efficiency without affecting the electron collection. In light of these results, we conclude that photoelectrodes composed of single-crystal-like 3D hierarchical NW arrays are attractive for artificial photosynthesis, solar cells, and other optoelectronic applications. Figure 1a,b show field emission scanning electron microscopy (FE-SEM) top views of the as-assembled 3D B-NW arrays on fluorine-doped tin oxide (FTO) coated substrates. The needlelike nanobranches grow in four symmetrical directions with a length of about 70 nm (Figure 1b). The length of the nanobranch can be controlled between 20 and 100 nm by adjusting the growth time (Figure S1, Supporting Information).

eveloping high surface area nanostructured electrodes with rapid charge transport is essential for effective charge collection in solar cells,1,2 artificial photosynthesis,3 photocatalysis,4 and energy storage devices.5 While the large surface area of electrode materials is normally associated with the high porosity and small size of building blocks2 (e.g., nanoparticles), the rapid charge transport property usually relies on straight conducting pathways.6 As an example for achieving the latter goal, one-dimensional (1D) single-crystal rutile TiO2 nanowire (NW) arrays have been reported. Also, 1D rutile TiO2 NW electrodes have been shown to markedly improve the electron transport property relative to nanoparticle (NP) counterparts.6d However, 1D NW arrays usually have relatively low surface area because of the large volume of free space between the NWs. Substantial research efforts have been recently focused on building one-dimensional (1D) nanoblocks with fast charge transport into three-dimensional (3D) hierarchical architectures.7−10 Various techniques including pulsed laser deposition,8 vapor deposition,9 and hydrothermal and solution methods10 have been used to prepare 3D hierarchical TiO2 NW arrays. For example, Lee10d and Wu et al10g, respectively, have reported a seed-induced multisteps and a one-step hydrothermal approaches for the fabrication of branched NWs. These assembly strategies have demonstrated enhanced light harvesting by enlarging the electrode surface area. However, they could in turn lead to the formation of grain boundaries and lattice defects between the branches and trunks © 2014 American Chemical Society

Received: December 13, 2013 Revised: February 18, 2014 Published: March 14, 2014 1848

dx.doi.org/10.1021/nl4046262 | Nano Lett. 2014, 14, 1848−1852

Nano Letters

Letter

Figure 1. FE-SEM images of the as-synthesized TiO2 3D B-NW array on a transparent FTO-coated glass substrate. (a,b) Top views at low and high magnifications, respectively; (c,d) cross-sectional views with 45° titling angle at low and high magnifications, respectively.

The crystal phase of the primary 1D NW and 3D B-NW arrays were both identified as tetragonal rutile according to their X-ray diffraction (XRD) patterns (Supporting Information Figure S2). Figure 1c shows that the well-separated 3D B-NW arrays grow vertically from the FTO substrate with an average thickness of about 3 μm. The nanobranches grow uniformly and with a clear orientation along specific side crystal facets of the trunk (Figure 1d). Figure 2a shows a typical transmission electron microscopy (TEM) image of a single 3D branched nanowire. The central nanowire core has uniform, needlelike side branches. The branches grow either upward (Region 1) or downward (Region 2) and are both oriented at a specific angle away from the trunk (Figure 2a,b (schematic illustration)). As shown in Figure 2d,e, the (110) lattice crystal plane with a fringe spacing of 0.325 nm in the trunk and branch are both well-developed indicating that they are both single crystal and grow along the crystallographic c-axis. The crystallographic relationship between the trunk and branch was further analyzed by TEM and high-resolution transmission electron microscopy (HR-TEM) along the [010] zone axis (Figure 2c,f). Figure 2c is a typical high-magnification TEM image of a trunk with one upward branch. The angle between the branch and the trunk is calculated to be about 65°, which is consistent with the angle between the crystallographic c-axis of the trunk and branch. The HR-TEM image taken from region f in Figure 2c is highlighted in Figure 2f. The (1̅01) lattice plane of the trunk and the (101) lattice plane of the branch both have a fringe spacing of 0.249 nm perfectly linked and form a {101} twinned structure. On the basis of the above TEM and HR-TEM analysis, the twinned structural relationship between the trunk and branch is illustrated schematically by the atomic model (Figure 2g,h). Thus, the nanobranches can be considered as epitaxial growth of the trunk along the (100) and (010) crystalline planes at four symmetrical directions with a typical angle of 65° between the c-axis of the trunk and

Figure 2. TEM and HR-TEM images of the as-synthesized 3D B-NW. (a,b) TEM image and schematic illustration, respectively, of a single 3D B-NW; (c) TEM image of a part of the 3D B-NW structure showing the trunk and one upward branch; (d,e) the HR-TEM images of a 1D single crystal trunk and a branch, respectively; (f) HR-TEM of the region f in panel (c) showing the interface between the trunk and the branch; (g) schematic illustration of a profile atomic model along [010] zone axes of the trunk corresponding to the panel (f); (h) a 3D atomic model along the [001] zone axes of the trunk. The gray and red balls denote Ti and O, respectively.

branch. Such orientated assembled 3D branched nanowire arrays with a twin-structured coherent interface should have few grain boundaries and lattice defects. Consequently, they should facilitate charge transport along the 3D TiO2 hierarchical nanowire arrays. Figure 3 shows the dependence of the electron diffusion coefficients (D) on the photoelectron density (n) for 1D NW, 3D B-NW, and NP rutile TiO2 films based dye sensitized solar cells (DSSCs). The values of D and n are determined from the transport time constants (τc) and film thickness (d) using procedures described elsewhere.11 The diffusion coefficient in the 3D B-NW arrays is about 2 orders of magnitude higher than that of randomly packed NP films at the same photoelectron density (e.g., 1 × 1017 cm−3).12 The fast charge transport property of 3D B-NW arrays can be attributed to the elimination of grain boundaries and lattice defects between the 1D NW trunks and the nanobranches. A closer examination of Figure 3 indicates that the D value of the 3D B-NW arrays is about a factor of 3 lower than that of the 1D single crystal NW arrays. Further increase in branch length results in a slower electron transport in the B-NW film (Supporting Information Figure S3). This could be due to the higher density of surface 1849

dx.doi.org/10.1021/nl4046262 | Nano Lett. 2014, 14, 1848−1852

Nano Letters

Letter

Figure 4. Comparison of recombination lifetimes as a function of the photoelectron density for 3D B-NW and 1D NW rutile TiO2 based DSSCs.

Figure 3. Comparison of electron diffusion coefficients as a function of the photoelectron density for 1D NW, 3D B-NW, and NP rutile TiO2 films based DSSCs.

states13 associated with the larger surface area of the 3D B-NW film. The dependence of photoelectron density on voltage is usually used to measure the distribution of sub-bandgap trap states of the photoelectrodes. Supporting Information Figure S4 shows that for a given voltage, the photoelectron density of the 3D B-NW-based cell is about 40% higher than that of the 1D NW-based cells. This suggests that the total density of trap states (Ntot) in the 3D B-NW sample is about 40% higher than that in the 1D NW one. We have shown previously that D is related to Ntot by the expression D ∝ (Ntot)−1/α, where α is related to the shape of the distribution of the sub-bandgap trap states.13a Best fits to the data show that α = 0.43 for both the 3D B-NW and 1D NW films. On the basis of this α value, we estimate that the 40% larger Ntot should lead to about 2.2-fold slower transport for the 3D B-NW film than the 1D NW arrays. This could account for the observed electron transport properties of the 3D B-NW and 1D single crystal NW arrays shown in Figure 3. Figure 4 shows the recombination lifetimes (τr) for DSSCs based on the 3D B-NW and 1D NW arrays as a function of the photoelectron density. Recombination is an interfacial charge transfer process that occurs mainly at the electrode/electrolyte interface14,15 and is thus influenced by the number of surface states.16 At a given photoelectron density, the lifetime of 3D BNW array-based cells is about a factor of 2 longer than that of 1D NW array-based cells because of its larger number of surface states as discussed above. The charge collection property of the DSSCs can be measured by the electron diffusion length (Ln) given by Ln = (Dτr)1/2. A longer Ln usually leads to a higher charge-collection efficiency. Ln is usually several times (e.g., more than three times) greater than the film thickness in order to obtain optimum cell performances.15 Analyses of the data shown in Figures 3 and 4 yield an comparable electron diffusion length for 3D B-NW (Ln = 30 μm) and 1D NW (Ln = 38 μm) based cells. Therefore, constructing orientated assembled hierarchical 3D B-NW arrays (up to 10 μm) can increase the surface area of the 1D NW arrays without negatively affecting the electron collection efficiency. The J−V characteristics of DSSCs based on 3D B-NW and 1D NW of similar NW thickness (3 μm) are shown in Figure 5.

Figure 5. Comparison of the J−V characteristics of the 3D B-NW and 1D NW array-based DSSCs under simulated AM 1.5 sunlight and in the dark.

The DSSC based on the 3D B-NW arrays exhibits an opencircuit voltage (Voc) of 0.71 V, a fill factor (FF) of 0.64, and a short-circuit photocurrent density (Jsc) of 10.14 mA/cm2 to give a solar conversion efficiency of 4.61%. In contrast, the cell based on 1D NW arrays shows a Voc of 0.73 V, a FF of 0.63, and a Jsc of 6.57 mA/cm2 to yield an efficiency of 3.02%. The Jsc value of the 3D B-NW based cell is about 54% higher than that of the 1D NW based cell, which is attributed to the larger surface area of the 3D B-NW film than the 1D NW film. These results are in agreement with the quantum efficiency (QE) spectra (Supporting Information Figure S5). According to dye desorption measurements as shown in Supporting Information Figure S6, the surface area of 3D B-NW arrays is found to be about 71% larger than that of the 1D NW arrays. Consistent with their difference in surface areas, the dark current of 3D BNW array-based DSSC is slightly higher than that of 1D NW array-based cells (Figure 5). It is also worth noting that the FF and Voc of 3D B-NW is comparable to that of 1D NW based cells, an observation consistent with their similar transport and recombination properties. Overall, the 3D B-NW arrays based 1850

dx.doi.org/10.1021/nl4046262 | Nano Lett. 2014, 14, 1848−1852

Nano Letters

Letter

were measured. The photocurrent density and photovoltage of the DSSCs were measured with active sample areas of 0.25 cm2 using AM-1.5 simulated sunlight (Oriel Sol3A Class AAA Solar Simulator). Electron transport and recombination properties of DSSCs were measured by intensity modulated photocurrent and photovoltage spectroscopies as described previously.16

cells demonstrate about 52% higher solar conversion efficiency than the 1D NW-based devices and thus should be attractive for organic/inorganic hybrid, QDs and perovskite-based solar cells, and other applications. In conclusion, we have reported oriented assembled singlecrystal-like TiO2 hierarchical nanowire arrays grown on transparent conductive substrates. The 1D nanobranches are formed via epitaxial growth from the trunk with few grain boundaries and provide a direct photoexcited carrier collection pathway in the 3D architecture. In comparison to the 1D NW arrays, the as-synthesized 3D B-NW arrays not only demonstrate higher surface area but also exhibit comparable fast charge transport properties, leading to significantly improved photovoltaic performance. Such 3D nanostructurebased electrodes have great potential in optoelectronic devices, particularly in solid state solar cells where recombination is very fast. The orientated assembly strategy reported here can be extended to assemble other metal oxides with one or multicomponents into 3D hierarchical nanostructures, and thus represents a critical avenue toward high-performance optoelectronics. Experimental Section. A two-step procedure was employed to synthesize 3D B-NW arrays on FTO glass (TEC-8) substrate. In the first step, aligned 1D single crystal TiO2 NW arrays were grown on FTO substrate through a solvothermal method adapted from a previous study.6d In a typical process, FTO substrates coated with a thin TiO2 layer were loaded into a sealed Teflon-lined stainless steel reactor filled with 6 mL of 2-butanone, 6 mL of 37% hydrochloric acid, and 0.4 mL of tetrabutyl titanate, and then kept at 200 °C for 60 min. For the growth of 3D B-NW arrays, the as-obtained 1D NW arrays were loaded into a sealed Teflon-lined stainless steel reactor filled with 10 mL of DI water, 0.1−0.3 mL 37% hydrochloric acid and 0.1−0.2 mL TiCl3 solution (20 wt % of TiCl3 in 2 M HCl), and then kept at 80 °C for 1−2 h. The obtained 3D B-NW samples were exposed to O2 plasma at 50 W for 10 min with an oxygen flow rate of 0.6 L/min and then annealed at 450 °C for 30 min in O2. For comparison purpose, aligned 1D TiO2 NW arrays on FTO substrate were treated using the same method, that is, exposed to O2 plasma at 50 W for 10 min and then annealed at 450 °C for 30 min in O2. Oxygen plasma treatment was carried out by using a plasma cleaner (Ming Heng, PDC-MG). The morphologies and microstructures of the samples were characterization using a FE-SEM (S4800, Hitachi, Tokyo, Japan). Average film thickness was determined using a surface profiler (KLA Tencor Alpha-Step 500). The TEM and HR-TEM images were taken using a Tecnai F20 (FEI, Hillsboro, OR, USA) microscope at an accelerating voltage of 200 kV. The crystal phase structures of the samples were investigated using an X-ray powder diffractometer (X′Pert PRO, PANalytical, Almelo, The Netherlands). For use in DSSCs, NWs array samples were coated with dye by immersion overnight at ambient temperature in a 0.5 mM ethanolic solution of commercial N719 dye. The electrolyte was composed of 0.8 M 1-hexyl-2,3-dimethylimidazolium iodide and 50 mM iodine in methoxypropionitrile. The Pt counter electrode was prepared by spreading a thin layer of 5-mM H2PtCl6 isopropanol solution on a FTO substrate, which was followed by drying in air and then heating at 400 °C for 20 min. The thickness of the electrolyte layer between the NW array and counter-electrode was fixed by the use of a 25 μm thick Surlyn as the spacer. Three samples for each electrode type

■

ASSOCIATED CONTENT

S Supporting Information *

The time-dependent experimental growth process of nanobranches on the trunk, XRD patterns of 3D B-NW and 1D NW arrays, comparison of photoelectron density as a function of the voltage for 1D NW and 3D B-NW array-based cells, comparison of quantum efficiency (IPCE) spectra for 1D NW and 3D B-NW array-based cells, and UV−vis spectra of dye solution desorbed from 1D and 3D nanowire arrays. This material is available free of charge via the Internet at http:// pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail: (X.F.) [email protected]. *E-mail: (K.Z.) [email protected]. Author Contributions

X.S. is responsible for the experiments and preparation of the paper; D.H. performed the TEM and HRTEM analysis; J.Y. performed the SEM analysis; K.Z. performed the IMPS, IMVS, and IV measurements; X.F. provided guidance for the experiments and for editing the paper. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was financially supported by grants of Chinese Thousand Talents Program (YZBQF11001) and the National Natural Science Foundation of China (21371178). K.Z. acknowledges the support by the Division of Chemical Sciences, Geosciences, and Biosciences, Office of Basic Energy Sciences, U.S. Department of Energy, under contract No. DEAC36-08GO28308 with the National Renewable Energy Laboratory.

■

REFERENCES

(1) (a) Cai, N.; Moon, S. J.; Cevey-Ha, L.; Moehl, T.; HumphryBaker, R.; Wang, P.; Zakeeruddin, S. M.; Grätzel, M. Nano Lett. 2011, 11, 1452. (b) Jang, S. R.; Zhu, K.; Ko, M. J.; Kim, K.; Kim, K. C.; Park, N. G.; Frank, A. J. ACS Nano 2011, 5, 8267. (2) (a) Crossland, E. J. W.; Noel, N.; Sivaram, V.; Leijtens, T.; Alexander-Webber, J. A.; Snaith, H. J. Nature 2013, 495, 215. (b) O’Regan, B.; Grätzel, M. Nature 1991, 353, 737. (3) (a) Habisreutinger, S. N.; Schmidt-Mende, L.; Stolarczyk, J. K. Angew. Chem., Int. Ed. 2013, 52, 7372. (b) Walter, M. G.; Warren, E. L.; Mckone, J. R.; Boettcher, S. W.; Mi, Q. X.; Santori, E. A.; Lewis, N. S. Chem. Rev. 2010, 110, 6446. (4) (a) Kubacka, A.; Fernández-García, M.; Colón, G. Chem. Rev. 2012, 112, 1555. (b) Hoffmann, M. R.; Martin, S. T.; Choi, W.; Bahnemann, D. W. Chem. Rev. 1995, 95, 69. (5) Xia, X. H.; Tu, J. P.; Zhang, Y. Q.; Wang, X. L.; Gu, C. D.; Zhao, X. B.; Fan, H. J. ACS Nano 2012, 6, 5531. (6) (a) Law, M.; Greene, L. E.; Johnson, J. C.; Saykally, R.; Yang, P. D. Nat. Mater. 2005, 4, 455. (b) Zhu, K.; Vinzant, T. B.; Neale, N. R.; Frank, A. J. Nano Lett. 2007, 7, 3739. (c) Varghese, O. K.; Paulose, M.; Grimes, C. A. Nat. Nanotechnol. 2009, 4, 592. (d) Feng, X. J.; Zhu, K.;

1851

dx.doi.org/10.1021/nl4046262 | Nano Lett. 2014, 14, 1848−1852

Nano Letters

Letter

Frank, A. J.; Grimes, C. A.; Mallouk, T. E. Angew. Chem., Int. Ed. 2012, 51, 2727. (7) Wang, D.; Qian, F.; Yang, C.; Zhong, Z. H.; Lieber, C. M. Nano Lett. 2004, 4, 871. (8) Sauvage, F.; Fonzo, F. D.; Bassi, A. L.; Casari, C. S.; Russo, S.; Divitini, G.; Ducati, C.; Bottani, C. E.; Comte, P.; Grätzel, M. Nano Lett. 2010, 10, 2562. (9) (a) Shi, J.; Hara, Y.; Sun, C. L.; Anderson, M. A.; Wang, X. D. Nano Lett. 2011, 11, 3413. (b) Yang, X. F.; Zhuang, J. L.; Li, X. Y.; Chen, D. H.; Ouyang, G. F.; Mao, Z. Q.; Han, Y. X.; He, Z. H.; Liang, C. L.; Wu, M. M.; Yu, J. C. ACS Nano 2009, 3, 1212. (10) (a) Wang, H. G.; Ma, D. L.; Huang, X. L.; Huang, Y.; Zhang, X. B. Sci. Rep. 2012, 2, 701. (b) Wu, W. Q.; Rao, H. S.; Xu, Y. F.; Wang, Y. F.; Su, C. Y.; Kuang, D. B. Sci. Rep. 2013, 3, 1892. (c) Cho, I. S.; Chen, Z. B.; Forman, A. J.; Kim, D. R.; Rao, P. M.; Jaramillo, T. F.; Zheng, X. L. Nano Lett. 2011, 11, 4978. (d) Lee, D. H.; Rho, Y.; Allen, F. I.; Minor, A. M.; Ko, S. H.; Grigoropoulos, C. P. Nanoscale 2013, 5, 11147. (e) Zhu, H.; Tao, J.; Wang, T.; Deng, J. Appl. Surf. Sci. 2011, 257, 10494. (f) Yin, Z. Y.; Wang, Z.; Du, Y. P.; Qi, X. Y.; Huang, Y. Z.; Xue, C.; Zhang, H. Adv. Mater. 2012, 24, 5374. (g) Wu, W. Q.; Lei, B. X.; Rao, H. S.; Xu, Y. F.; Wang, Y. F.; Su, C. Y.; Kuang, D. B. Sci. Rep. 2013, 3, 1352. (11) Zhao, Y.; Zhu, K. J. Phys. Chem. Lett. 2013, 4, 2880. (12) Park, N. G.; Schlichthörl, G.; van de Lagemaat, J.; Cheong, H. M.; Mascarenhas, A.; Frank, A. J. J. Phys. Chem. B 1999, 103, 3308. (13) (a) Kopidakis, N.; Neale, N. R.; Zhu, K.; van de Lagemaat, J.; Frank, A. J. Appl. Phys. Lett. 2005, 87, 202106. (b) Mohammadpour, A.; Farsinezhad, S.; Wiltshire, B. D.; Shankar, K. Phys. Status Solidi RRL 2014, DOI: 10.1002/pssr.201308296. (14) Hamann, T. W.; Ondersma, J. W. Energy Environ. Sci. 2011, 4, 370. (15) Peter, L. M. J. Phys. Chem. C 2007, 111, 6601. (16) Zhu, K.; Kopidakis, N.; Neale, N. R.; van de Lagemaat, J.; Frank, A. J. J. Phys. Chem. B 2006, 110, 25174.

1852

dx.doi.org/10.1021/nl4046262 | Nano Lett. 2014, 14, 1848−1852