- Email: [email protected]
Anodized TiO2 Nanotubes: Effect of anodizing time on film length, morphology and photoelectrochemical properties
Materials Letters ∎ (∎∎∎∎) ∎∎∎–∎∎∎
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Q1 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66
Contents lists available at ScienceDirect
Materials Letters journal homepage: www.elsevier.com/locate/matlet
Anodized TiO2 Nanotubes: Effect of anodizing time on film length, morphology and photoelectrochemical properties D. Regonini n, F.J. Clemens nn Laboratory for High Performance Ceramics, EMPA-Swiss Federal Laboratories for Materials Science & Technology, Überlandstrasse 129, 8600 Dübendorf, Switzerland
art ic l e i nf o
a b s t r a c t
Article history: Received 13 October 2014 Accepted 27 November 2014
The influence of anodizing time on the length, morphology and photoelectrochemical properties of TiO2 Nanotubes (NTs) has been investigated. An optimum anodizing time of 20 min at 30 V leads to 1.1 mm long NTs films, generating a photocurrent density (J photo ) of 460 mA/cm2. Anodized films grown for a time shorter than 10 min show only a thin (100 nm) porous layer, instead of well-defined NTs, and exhibit lower photocurrent densities. Similarly, J photo is lower in NTs grown for 1–2 h, partly because the NTs are damaged after extended anodization and partly because their length (l) already exceeds the electron diffusion length (Ln ), found to be 0.87 0.1 mm. The Incident Photon to Current Efficiency (IPCE) study reveals how the microstructural changes (morphology and length) affect the behavior of the anodic films over the photoaction spectrum. IPCE analysis confirms the superiority of 1.1 mm long and well-defined NTs, with a maximum value of 18% recorded at 340 nm. In contrast, the ability to collect electrons decreases for longer and partially damaged NTs, especially in the region 360–300 nm. A small response above 400 nm is observed in all the photoelectrodes, due to oxygen vacancies induced during annealing. & 2014 Elsevier B.V. All rights reserved.
Keywords: Anodization TiO2 Nanotubes IPCE Photoelectrochemical Water Splitting
1. Introduction Anodized TiO2 Nanotubes (NTs) [1,2] have received wide attention from the research community over the last decade, as they are relevant in dye sensitized solar cells [3], photocatalysis [4], biomedicine [5] and gas-sensing [6]. The (photo) electrochemistry of TiO2 NTs is a fascinating and key topic to study in order to optimize the performances in the aforementioned fields [7]. Concerning the solar water splitting process, an electron diffusion length (Ln ) of 24 mm is reported for smooth NTs, although in the presence of a sacrificial agent [8]. The presence of long-lived electrons in TiO2 NTs, indicating excellent charge carrier lifetime and potential for efficient electron transfer and transport, is also confirmed by time resolved spectroscopy studies [9]. Beranek [10] reported that the highest photocurrent gain with the NTs is achieved at a relatively low bias because the maximum space charge layer (W) generated within such nanostructure is limited by the NTs' wall thickness. In contrast for compact semiconductors the photoresponse follows the Gärtner model [11] and keeps increasing proportionally to the applied bias. Despite the intensive research efforts on TiO2 NTs, Paramasivam [4] argued that a
n
Corresponding author. Tel.: þ 41 58 765 4490. Corresponding author. Tel.: þ 41 58 765 4821. E-mail addresses: [email protected] (D. Regonini), [email protected] (F.J. Clemens). nn
more quantitative survey on the influence of tube morphology, length, wall thickness and tube diameter is needed to better understand their photoelectrochemical properties. Likewise, Augustynski [12] underlined that the assessment of the photoelectrochemistry of TiO2 NTs is made difficult by the lack of systematic data, particularly those showing how their length affect the Incident Photon to Current Efficiency (IPCE) over the photoaction spectrum. Aim of the work: The aim of our work is to evaluate the effect of anodizing time on the length, film morphology/microstructure and photoelectrochemistry of undoped TiO2 NTs. The investigation helps to understand what is required to develop an anodic film with an optimized photoresponse. An often neglected aspect is the significant impact of the anodizing potential on the semiconducting properties of TiO2 NTs [13,14]; to minimize changes in the electronic properties of the anodic films, a constant anodizing potential of 30 V has been utilized within our study. The solar water splitting process [15] is chosen as a model reaction.
2. Experimental methods (i) Anodization. Prior to anodizing, the Ti specimens (Ti 99.6%, Goodfellow, Cambridge Limited, 0.5 mm thick) were mechanically grinded with SiC paper and ultrasonically cleaned for 10 min in isopropyl alcohol. Anodization was carried out in a two
http://dx.doi.org/10.1016/j.matlet.2014.11.145 0167-577X/& 2014 Elsevier B.V. All rights reserved.
Please cite this article as: Regonini D, Clemens FJ. Anodized TiO2 Nanotubes: Effect of anodizing time on film length, morphology and photoelectrochemical properties. Mater Lett (2014), http://dx.doi.org/10.1016/j.matlet.2014.11.145i
67 68 69 70 71 72 73 74 75 76 Q2 77 78 79 80 81 82 83 84 85 Q3 86 87 88 89 90 91 92 93 94 95 96 97
2
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66
D. Regonini, F.J. Clemens / Materials Letters ∎ (∎∎∎∎) ∎∎∎–∎∎∎
67 length being respectively 0.870.1 mm and 1.170.1 mm; the top of 68 the NTs is mostly covered by an initiation porous layer [19], a 69 common scenario in organic electrolytes (Fig. 2b, c). However the 70 mouth of the NTs is free of this porous layer in some parts of the film, 71 as shown in the inset of Fig. 2c. The anodic NTs can grow longer 72 (3 mm) by extending the anodizing time further, as shown in 73 Fig. 2d, e, although the top of the NTs is gradually destroyed. Indeed, 74 the partial collapse of the NTs due to over-dissolution of the top of 75 the tubes can already be appreciated after 1 h (Fig. 2d) and becomes 76 very significant after 2 h (Fig. 2e), leading to the formation of spikes 77 [20] or clusters of NTs and therefore to a substantial change in the 78 morphology of the anodic film. 79 The GAXRD spectra of the TiO2 NTs are shown in Fig. 3. As80 prepared NTs are amorphous and converted to anatase following 81 the thermal treatment at 450 1C [16]. Clearly, as the NTs grow 82 longer, TiO2 peaks become more significant than peaks belonging 83 to the Ti substrate underneath the oxide. As the crystal phase 84 (anatase) does not depend on the anodizing time (Fig. 3), any 85 difference in the photoelectrochemical response of the films can 86 be ascribed to the different lengths (l) of the NTs and changes 87 within their microstructure/morphology. 88 The photoelectrochemical analysis of the anodic films is summar89 ized in Fig.4a, b. The thin porous layer obtained after 5 min anodization 90 behaves differently than films grown for a longer time (10–20 min) 91 and having well-defined NTs. Its poor J Photo in comparison to NTs 92 grown for 10 min, particularly evident at negative bias (0.8 to 93 0.4 V vs Ag/AgCl), suggests that after 5 min the anodic film is far 94 from an optimal photoresponse, both in terms of morphology and 95 length. Such a thin film resembles the behavior of a compact layer and 96 its J Photo does not saturate, but increases proportionally with the applied bias [10,11]. 97 A different scenario is observed once well-defined NTs are formed; 98 J Photo at 0.0 V vs Ag/AgCl (1.0 V vs RHE) raises from 360 to 460 mA/cm2 99 100 by increasing the anodization time from 10 to 20 min. More impor101 tantly, the J Photo at negative bias ( 0.8 to 0.4 V vs Ag/AgCl) is 102 significantly improved in comparison to 5 min growth, providing a 3. Results and discussion 103 direct evidence for the lower charge carrier recombination occurring 104 in well-formed NTs, which may be ascribed to a directional charge Fig. 1 illustrates how the morphology of the anodic layer evolves 105 transport (i.e. 1D transport) within the tubular structure. The photoand the length of the NTs film increases with the anodizing time. 106 response then decreases to 320 and 300 mA/cm2 (at 0.0 V vs Ag/AgCl) After 5 min the anodic film is 100720 nm thick and rather than 107 well-defined NTs, only a porous layer can be seen (Fig. 2a). NTs are when extending the anodization time to 1 or 2 h, respectively (Fig.4a). instead clearly visible after 10 and 20 minutes (Fig. 2b and c), their This suggests that the optimal length has been exceeded (l is 2.3 mm at Q4108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 Fig. 1. The NTs thickness vs anodizing time plot shows that the NTs are 1 mm after 20 min at 30 V. Within the explored time range (up to 2 h), the maximum thickness was 3 mm. electrodes cell, applying a potential of 30 V (power supply: Keithley 2450) between the Ti anode and a Pt cathode and varying the anodizing time (5 min–2 h). The electrolyte consisted of 75 ml of anhydrous Ethylene Glycol (Sigma-Aldrich), 3 ml of distilled H2O and 0.3 g of NH4F (Sigma Aldrich). As-prepared anodized samples were rinsed with water, ultrasonically cleaned with ethanol and stored in ethanol overnight. The amorphous specimens were converted to anatase at 450 1C [16,17] in air (heating rate: 100 1C/h, dwelling time: 1 h, furnace type: LHT 04/ 17, Nabertherm). (ii) Morphology, Microstructure and Crystallinity. The photoelectrodes were analyzed by a Field Emission Nova NanoSEM 230 (Nova FEI). Crystal phases were measured by Grazing Angle X-Ray Diffraction (GAXRD, geometry, ω¼11), using a Panalytical, X'Pert Pro instrument (Cu-Kα1, λ¼1.5406 Å). (iii) Photoelectrochemical Characterization. The photocurrent measurements were performed with a Voltalab80 PGZ-402 (Radiometer Analytical) potentiostat, with a TiO2 NTs Working Electrode (WE), Pt plate (XM120, Radiometer Analytical) Counter Electrode (CE) and Ag/AgCl/3M-KCl (XR300, Radiometer Analytical) Reference Electrode (RE) fitted in a Cappuccino Cell [18] filled with 10 ml of 1 M KOH. A potential sweep from 900 to þ700 mV was applied to WE vs RE (scan rate of 20mV/sec). The WE was irradiated by a Xe lamp (solar simulator L.O.T–Oriel AG), at an intensity of 80 mW/cm2 and the resulting photocurrent density (J Photo ) was recorded. Incident Photon to Current Efficiency (IPCE) analysis was performed using a Xe lamp (L.O.T– Oriel AG) and a monochromator (Omni-λ 300, L.O.T–Oriel AG). The TiO2 NTs photoanodes, immersed in 1 M KOH, were irradiated by different wavelengths (λ) ranging from 550 to 280 nm, while a potential bias of 0.23 V (Keithley 2450) was applied between WE and CE. The resulting current was measured as a function of λ, J Photo ðλÞ, and used to calculate the IPCE.
Please cite this article as: Regonini D, Clemens FJ. Anodized TiO2 Nanotubes: Effect of anodizing time on film length, morphology and photoelectrochemical properties. Mater Lett (2014), http://dx.doi.org/10.1016/j.matlet.2014.11.145i
D. Regonini, F.J. Clemens / Materials Letters ∎ (∎∎∎∎) ∎∎∎–∎∎∎
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 Q9 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66
3
Fig. 2. (a)–(e) Micro and nanostructure of TiO2 NTs films grown at different anodizing times, investigated by SEM: 5 min (a), 10 min (b), 20 min (c), 1 h (d) and 2 h (e).
Fig. 3. GAXRD of TiO2 NTs grown at 30 V with anodizing time ranging from 5 min to 2 h and annealed at 450 1C. The spectra of as-prepared TiO2 NTs and Ti substrate annealed at 450 1C are given as reference. As-prepared NTs are amorphous and the formation of anatase crystals is induced by the thermal treatment at 450 1C [16].
1 h and 2.9 mm at 2 h). Based on SEM analysis, the partial collapse of NTs grown for 1–2 h and the formation of spikes/clusters of tubes (Fig. 2d, e), are also responsible for the lower J Photo observed. It has been previously shown that such clusters are detrimental to the 1D transport, as they create additional three dimensional pathways, increasing recombination (in the case of dye sensitized solar cells also decreasing dye loading) [21]. In contrast, it is difficult to establish whether the initial thin porous layer covering the top of the NTs (Fig. 2b, c), has a negative effect on the electron transport (concerning the solar water splitting process), as proposed by Albu et al. [19]. Considering it is only few nm thick and does not seem to prevent penetration of the electrolyte within the entire NTs film, we expect it has a minor effect only at very low λ (i.e. 300 nm), as such light is likely to be absorbed by this porous layer rather than by the NTs underneath. As a guideline, according to Eagles [22], the light penetration depth in TiO2 is 10 nm at λ ¼300 nm and 1 mm at 380 nm. More insights are provided by the IPCE studies (Fig. 4b). All the NTs show a weak response above 400 nm, due to oxygen vacancies induced during annealing by the significant presence of C within the as-prepared films grown in organic media [23]. Similarly, it was reported by Park [24] that a carbon thermal treatment triggered a visible response in the NTs. Each IPCE curve has a maximum (IPCEmax ), which slightly shifts towards longer λ as the length of the film
Please cite this article as: Regonini D, Clemens FJ. Anodized TiO2 Nanotubes: Effect of anodizing time on film length, morphology and photoelectrochemical properties. Mater Lett (2014), http://dx.doi.org/10.1016/j.matlet.2014.11.145i
67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132
4
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 Q5 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Q7 Q6 61 62 63 64 65 66
D. Regonini, F.J. Clemens / Materials Letters ∎ (∎∎∎∎) ∎∎∎–∎∎∎
67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 Fig. 4. Photocurrent density (a) and IPCE response (b) of TiO2 NTs anodized for a time ranging from 5 min to 2 h. All measurements were performed in 1 M KOH and indicate 84 the higher photoactivity of 1.1 mm thick NTs grown at 30 V for 20 min. 85 86 87 (EPFL). Dr. Fabio La Mattina and A. Kupferschmid (Empa) are also increases (IPCEmax is at 330 nm for l¼ 125 nm and at 350 nm when 88 gratefully acknowledged for their contribution to the IPCE set-up. l ¼2.9 mm). Such red shift is also observed in colloidal nanoparticles 89 under front illumination [25]. To evaluate the electron transport 90 properties, a useful parameter is the electron diffusion length, Ln , References 91 describing the competition between the diffusive transport and the 92 recombination of the charge carriers [26]. For photoanodes with [1] Roy P, Berger S, Schmuki P. TiO2 nanotubes: synthesis and applications. Angew 93 l=Ln 41, Ln can be quantified accordingly to Söedergren [27], where Chem Int Ed Engl 2011;50:2904–39. [2] Regonini D, Bowen CR, Jaroenworaluck A, Stevens R. A review of growth 94 the short wavelength side of the action spectra is simplified as mechanism, structure and crystallinity of anodized TiO2 nanotubes. Mater Sci 95 ln ðIPCEÞ ¼ l=Ln . The drop in the IPCE at λ¼300 nm is particularly Eng R 2013;74:377–406. 96 evident for an anodizing time of 1–2 h (Fig. 4b), therefore we assume [3] Roy P, Kim D, Lee K, Spiecker E, Schmuki P. TiO2 nanotubes and their application in dye-sensitized solar cells. Nanoscale 2010;2:45–59. 97 that the condition l=Ln 41 is satisfied for NTs grown 1 h or longer. This [4] Paramasivam I, Jha H, Liu N, Schmuki P. A review of photocatalysis using self98 is in line with the decrease observed in the J Photo of NTs anodized for organized TiO2 nanotubes and other ordered oxide nanostructures. Small 99 1–2 h (Fig. 4a). Using the IPCE values at 300 nm of 5.5 and 3% at 1 h 2012;8:3073–103. 100 and 2 h respectively, we estimate Ln to be 0.870.1 mm. This indicates [5] Kummer KM, Taylor E, Webster TJ. Biological applications of anodized TiO2 nanostructures: a review from orthopedic to stent applications. Nanosci 101 that despite the 1D transport within the NTs further studies are Nanotechnol Lett 2012;4:483–93. required to improve their electron transport and photoelectrochemical 102 [6] Galstyan V, Comini E, Faglia G, Sberveglieri G. TiO2 nanotubes: recent advances properties. 103 in synthesis and gas sensing properties. Sensors 2013;13:14813–38. [7] Shrestha NK, Schmuki P. Electrochemistry at TiO2 nanotubes and other Q8 104 semiconductor nanostructures. Electrochemistry, vol. 12. The Royal Society 105 of Chemistry; 87–131 (Chapter 3). 106 [8] Lynch RP, Ghicov A, Schmuki P. A photo-electrochemical investigation of self4. Conclusions organized TiO2 nanotubes. J Electrochem Soc 2010;157:G76–84. 107 [9] Kahnt A, Oelsner C, Werner F, Guldi DM, Albu SP, Kirchgeorg R, et al. Excited 108 state properties of anodic TiO2 nanotubes. Appl Phys Lett 2013;102:233109. Our study reveals that an optimal tube length of 1.1 mm is 109 [10] Beranek R, Tsuchiya H, Sugishima T, Macak JM, Taveira L, Fujimoto S, et al. required to maximize the photoresponse of undoped TiO2 NTs. The Enhancement and limits of the photoelectrochemical response from anodic 110 morphology of the anodic layer is very important and there is TiO2 nanotubes. Appl Phys Lett 2005;87:243114. 111 [11] Gärtner WW. Depletion-layer photoeffects in semiconductors. Phys Rev clearly a change in behavior once the transition from a thin porous 112 1959;116:84–7. layer to well-defined NTs is completed. For example, the [12] Augustynski J, Alexander BD, Solarska R. Metal oxide photoanodes for water 113 J Photo under mild bias ( 0.8 to 0.4 V vs Ag/AgCl) is much higher splitting. Top Curr Chem 2011;303:1–38. 114 for well-defined NTs (grown for 10–20 minutes). Extended ano[13] Ali Yahia SA, Hamadou L, Kadri A, Benbrahim N, Sutter EMM. Effect of 115 anodizing potential on the formation and EIS characteristics of TiO2 nanotube dizing time (1–2 h) at 30 V causes some damage to the NTs and 116 arrays. J Electrochem Soc 2012;159:K83–92. their photoresponse decreases, whereas the impact of the thin [14] Acevedo-Pena P, Gonzalez I. TiO2 nanotubes formed in aqueous media: 117 porous layer covering the top of the NTs is currently not clear. The relationship between morphology, electrochemical properties and photoelec118 trochemical performance for water oxidation. J Electrochem Soc 2013;160: electron diffusion length of undoped NTs is estimated to be 119 H452–H458. 0.8 70.1 mm and the study of slightly doped anodic TiO2 films [15] Fujishima A, Honda K. Electrochemical photolysis of water at a semiconductor 120 with improved electron transport properties [28,29] is ongoing. electrode. Nature 1972;238:37–8. 121 [16] Regonini D, Jaroenworaluck A, Stevens R, Bowen CR. Effect of heat treatment 122 on the properties and structure of TiO2 nanotubes: phase composition and chemical composition. Surf Interface Anal 2010;42:139–44. 123 [17] Jaroenworaluck A, Regonini D, Bowen CR, Stevens R. A microscopy study of the Acknowledgments 124 effect of heat treatment on the structure and properties of anodised TiO2 125 nanotubes. Appl Surf Sci 2010;256:2672–9. [18] Van de Krol R, Gratzel M. Photoelectrochemical hydrogen production. New 126 The authors gratefully thank the financial support received by the York: Springer; 2012; 102. 127 EU COFUND Marie Curie Actions Program (NANOSPLIT) and by the [19] Albu SP, Schmuki P. Highly defined and ordered top-openings in TiO2 128 Swiss National Science Foundation (project n. R'Equip 206021nanotube arrays. Phys Status Solidi (RRL) 2010;4:151–3. [20] Berger S, Hahn R, Roy P, Schmuki P. Self-organized TiO2 nanotubes: factors 129 121306, Fundamental Aspects of Photocatalysis and Photoelectroaffecting their morphology and properties. Phys Status Solidi B 2010;247:2424–35. 130 chemistry/Basic Research Instrumentation for Functional Character[21] Zhu K, Vinzant TB, Neale NR, Frank AJ. Removing structural disorder from 131 ization). The Cappuccino cell was machined at Empa based on the oriented TiO2 nanotube arrays: reducing the dimensionality of transport and 132 design provided by the Laboratory for Photonics and Interfaces recombination in dye-sensitized solar cells. Nano Lett 2007;7:3739–46. Please cite this article as: Regonini D, Clemens FJ. Anodized TiO2 Nanotubes: Effect of anodizing time on film length, morphology and photoelectrochemical properties. Mater Lett (2014), http://dx.doi.org/10.1016/j.matlet.2014.11.145i
D. Regonini, F.J. Clemens / Materials Letters ∎ (∎∎∎∎) ∎∎∎–∎∎∎
1 2 3 4 5 6 7 8 9
[22] Eagles DM. Polar modes of lattice vibration and polaron coupling constants in rutile (TiO2). J Phys Chem Solids 1964;25:1243–51. [23] Regonini D, Satka A, Jaroenworaluck A, Allsopp DWE, Bowen CR, Stevens R. Factors influencing surface morphology of anodized TiO2 nanotubes. Electrochim Acta 2012;74:244–53. [24] Park JH, Kim S, Bard AJ. Novel carbon-doped TiO2 nanotube arrays with high aspect ratios for efficient solar water splitting. Nano Lett 2005;6:24–8. [25] Hagfeldt A, Björkstén U, Lindquist S-E. Photoelectrochemical studies of colloidal TiO2-films: the charge separation process studied by means of action spectra in the UV region. Sol Energy Mater Sol Cells 1992;27:293–304.
5
[26] Leng WH, Barnes PRF, Juozapavicius M, O'Regan BC, Durrant JR. Electron diffusion length in mesoporous nanocrystalline TiO2 photoelectrodes during water oxidation. J Phys Chem Lett 2010;1:967–72. [27] Soedergren S, Hagfeldt A, Olsson J, Lindquist S-E. Theoretical models for the action spectrum and the current–voltage characteristics of microporous semiconductor films in photoelectrochemical cells. J Phys Chem 1994;98:5552–6. [28] Das C, Roy P, Yang M, Jha H, Schmuki P. Nb doped TiO2 nanotubes for enhanced photoelectrochemical water-splitting. Nanoscale 2011;3:3094–6. [29] Yang M, Jha H, Liu N, Schmuki P. Increased photocurrent response in Nbdoped TiO2 nanotubes. J Mater Chem 2011;21:15205–8.
Please cite this article as: Regonini D, Clemens FJ. Anodized TiO2 Nanotubes: Effect of anodizing time on film length, morphology and photoelectrochemical properties. Mater Lett (2014), http://dx.doi.org/10.1016/j.matlet.2014.11.145i
10 11 12 13 14 15 16 17
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Q1 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66
Contents lists available at ScienceDirect
Materials Letters journal homepage: www.elsevier.com/locate/matlet
Anodized TiO2 Nanotubes: Effect of anodizing time on film length, morphology and photoelectrochemical properties D. Regonini n, F.J. Clemens nn Laboratory for High Performance Ceramics, EMPA-Swiss Federal Laboratories for Materials Science & Technology, Überlandstrasse 129, 8600 Dübendorf, Switzerland
art ic l e i nf o
a b s t r a c t
Article history: Received 13 October 2014 Accepted 27 November 2014
The influence of anodizing time on the length, morphology and photoelectrochemical properties of TiO2 Nanotubes (NTs) has been investigated. An optimum anodizing time of 20 min at 30 V leads to 1.1 mm long NTs films, generating a photocurrent density (J photo ) of 460 mA/cm2. Anodized films grown for a time shorter than 10 min show only a thin (100 nm) porous layer, instead of well-defined NTs, and exhibit lower photocurrent densities. Similarly, J photo is lower in NTs grown for 1–2 h, partly because the NTs are damaged after extended anodization and partly because their length (l) already exceeds the electron diffusion length (Ln ), found to be 0.87 0.1 mm. The Incident Photon to Current Efficiency (IPCE) study reveals how the microstructural changes (morphology and length) affect the behavior of the anodic films over the photoaction spectrum. IPCE analysis confirms the superiority of 1.1 mm long and well-defined NTs, with a maximum value of 18% recorded at 340 nm. In contrast, the ability to collect electrons decreases for longer and partially damaged NTs, especially in the region 360–300 nm. A small response above 400 nm is observed in all the photoelectrodes, due to oxygen vacancies induced during annealing. & 2014 Elsevier B.V. All rights reserved.
Keywords: Anodization TiO2 Nanotubes IPCE Photoelectrochemical Water Splitting
1. Introduction Anodized TiO2 Nanotubes (NTs) [1,2] have received wide attention from the research community over the last decade, as they are relevant in dye sensitized solar cells [3], photocatalysis [4], biomedicine [5] and gas-sensing [6]. The (photo) electrochemistry of TiO2 NTs is a fascinating and key topic to study in order to optimize the performances in the aforementioned fields [7]. Concerning the solar water splitting process, an electron diffusion length (Ln ) of 24 mm is reported for smooth NTs, although in the presence of a sacrificial agent [8]. The presence of long-lived electrons in TiO2 NTs, indicating excellent charge carrier lifetime and potential for efficient electron transfer and transport, is also confirmed by time resolved spectroscopy studies [9]. Beranek [10] reported that the highest photocurrent gain with the NTs is achieved at a relatively low bias because the maximum space charge layer (W) generated within such nanostructure is limited by the NTs' wall thickness. In contrast for compact semiconductors the photoresponse follows the Gärtner model [11] and keeps increasing proportionally to the applied bias. Despite the intensive research efforts on TiO2 NTs, Paramasivam [4] argued that a
n
Corresponding author. Tel.: þ 41 58 765 4490. Corresponding author. Tel.: þ 41 58 765 4821. E-mail addresses: [email protected] (D. Regonini), [email protected] (F.J. Clemens). nn
more quantitative survey on the influence of tube morphology, length, wall thickness and tube diameter is needed to better understand their photoelectrochemical properties. Likewise, Augustynski [12] underlined that the assessment of the photoelectrochemistry of TiO2 NTs is made difficult by the lack of systematic data, particularly those showing how their length affect the Incident Photon to Current Efficiency (IPCE) over the photoaction spectrum. Aim of the work: The aim of our work is to evaluate the effect of anodizing time on the length, film morphology/microstructure and photoelectrochemistry of undoped TiO2 NTs. The investigation helps to understand what is required to develop an anodic film with an optimized photoresponse. An often neglected aspect is the significant impact of the anodizing potential on the semiconducting properties of TiO2 NTs [13,14]; to minimize changes in the electronic properties of the anodic films, a constant anodizing potential of 30 V has been utilized within our study. The solar water splitting process [15] is chosen as a model reaction.
2. Experimental methods (i) Anodization. Prior to anodizing, the Ti specimens (Ti 99.6%, Goodfellow, Cambridge Limited, 0.5 mm thick) were mechanically grinded with SiC paper and ultrasonically cleaned for 10 min in isopropyl alcohol. Anodization was carried out in a two
http://dx.doi.org/10.1016/j.matlet.2014.11.145 0167-577X/& 2014 Elsevier B.V. All rights reserved.
Please cite this article as: Regonini D, Clemens FJ. Anodized TiO2 Nanotubes: Effect of anodizing time on film length, morphology and photoelectrochemical properties. Mater Lett (2014), http://dx.doi.org/10.1016/j.matlet.2014.11.145i
67 68 69 70 71 72 73 74 75 76 Q2 77 78 79 80 81 82 83 84 85 Q3 86 87 88 89 90 91 92 93 94 95 96 97
2
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66
D. Regonini, F.J. Clemens / Materials Letters ∎ (∎∎∎∎) ∎∎∎–∎∎∎
67 length being respectively 0.870.1 mm and 1.170.1 mm; the top of 68 the NTs is mostly covered by an initiation porous layer [19], a 69 common scenario in organic electrolytes (Fig. 2b, c). However the 70 mouth of the NTs is free of this porous layer in some parts of the film, 71 as shown in the inset of Fig. 2c. The anodic NTs can grow longer 72 (3 mm) by extending the anodizing time further, as shown in 73 Fig. 2d, e, although the top of the NTs is gradually destroyed. Indeed, 74 the partial collapse of the NTs due to over-dissolution of the top of 75 the tubes can already be appreciated after 1 h (Fig. 2d) and becomes 76 very significant after 2 h (Fig. 2e), leading to the formation of spikes 77 [20] or clusters of NTs and therefore to a substantial change in the 78 morphology of the anodic film. 79 The GAXRD spectra of the TiO2 NTs are shown in Fig. 3. As80 prepared NTs are amorphous and converted to anatase following 81 the thermal treatment at 450 1C [16]. Clearly, as the NTs grow 82 longer, TiO2 peaks become more significant than peaks belonging 83 to the Ti substrate underneath the oxide. As the crystal phase 84 (anatase) does not depend on the anodizing time (Fig. 3), any 85 difference in the photoelectrochemical response of the films can 86 be ascribed to the different lengths (l) of the NTs and changes 87 within their microstructure/morphology. 88 The photoelectrochemical analysis of the anodic films is summar89 ized in Fig.4a, b. The thin porous layer obtained after 5 min anodization 90 behaves differently than films grown for a longer time (10–20 min) 91 and having well-defined NTs. Its poor J Photo in comparison to NTs 92 grown for 10 min, particularly evident at negative bias (0.8 to 93 0.4 V vs Ag/AgCl), suggests that after 5 min the anodic film is far 94 from an optimal photoresponse, both in terms of morphology and 95 length. Such a thin film resembles the behavior of a compact layer and 96 its J Photo does not saturate, but increases proportionally with the applied bias [10,11]. 97 A different scenario is observed once well-defined NTs are formed; 98 J Photo at 0.0 V vs Ag/AgCl (1.0 V vs RHE) raises from 360 to 460 mA/cm2 99 100 by increasing the anodization time from 10 to 20 min. More impor101 tantly, the J Photo at negative bias ( 0.8 to 0.4 V vs Ag/AgCl) is 102 significantly improved in comparison to 5 min growth, providing a 3. Results and discussion 103 direct evidence for the lower charge carrier recombination occurring 104 in well-formed NTs, which may be ascribed to a directional charge Fig. 1 illustrates how the morphology of the anodic layer evolves 105 transport (i.e. 1D transport) within the tubular structure. The photoand the length of the NTs film increases with the anodizing time. 106 response then decreases to 320 and 300 mA/cm2 (at 0.0 V vs Ag/AgCl) After 5 min the anodic film is 100720 nm thick and rather than 107 well-defined NTs, only a porous layer can be seen (Fig. 2a). NTs are when extending the anodization time to 1 or 2 h, respectively (Fig.4a). instead clearly visible after 10 and 20 minutes (Fig. 2b and c), their This suggests that the optimal length has been exceeded (l is 2.3 mm at Q4108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 Fig. 1. The NTs thickness vs anodizing time plot shows that the NTs are 1 mm after 20 min at 30 V. Within the explored time range (up to 2 h), the maximum thickness was 3 mm. electrodes cell, applying a potential of 30 V (power supply: Keithley 2450) between the Ti anode and a Pt cathode and varying the anodizing time (5 min–2 h). The electrolyte consisted of 75 ml of anhydrous Ethylene Glycol (Sigma-Aldrich), 3 ml of distilled H2O and 0.3 g of NH4F (Sigma Aldrich). As-prepared anodized samples were rinsed with water, ultrasonically cleaned with ethanol and stored in ethanol overnight. The amorphous specimens were converted to anatase at 450 1C [16,17] in air (heating rate: 100 1C/h, dwelling time: 1 h, furnace type: LHT 04/ 17, Nabertherm). (ii) Morphology, Microstructure and Crystallinity. The photoelectrodes were analyzed by a Field Emission Nova NanoSEM 230 (Nova FEI). Crystal phases were measured by Grazing Angle X-Ray Diffraction (GAXRD, geometry, ω¼11), using a Panalytical, X'Pert Pro instrument (Cu-Kα1, λ¼1.5406 Å). (iii) Photoelectrochemical Characterization. The photocurrent measurements were performed with a Voltalab80 PGZ-402 (Radiometer Analytical) potentiostat, with a TiO2 NTs Working Electrode (WE), Pt plate (XM120, Radiometer Analytical) Counter Electrode (CE) and Ag/AgCl/3M-KCl (XR300, Radiometer Analytical) Reference Electrode (RE) fitted in a Cappuccino Cell [18] filled with 10 ml of 1 M KOH. A potential sweep from 900 to þ700 mV was applied to WE vs RE (scan rate of 20mV/sec). The WE was irradiated by a Xe lamp (solar simulator L.O.T–Oriel AG), at an intensity of 80 mW/cm2 and the resulting photocurrent density (J Photo ) was recorded. Incident Photon to Current Efficiency (IPCE) analysis was performed using a Xe lamp (L.O.T– Oriel AG) and a monochromator (Omni-λ 300, L.O.T–Oriel AG). The TiO2 NTs photoanodes, immersed in 1 M KOH, were irradiated by different wavelengths (λ) ranging from 550 to 280 nm, while a potential bias of 0.23 V (Keithley 2450) was applied between WE and CE. The resulting current was measured as a function of λ, J Photo ðλÞ, and used to calculate the IPCE.
Please cite this article as: Regonini D, Clemens FJ. Anodized TiO2 Nanotubes: Effect of anodizing time on film length, morphology and photoelectrochemical properties. Mater Lett (2014), http://dx.doi.org/10.1016/j.matlet.2014.11.145i
D. Regonini, F.J. Clemens / Materials Letters ∎ (∎∎∎∎) ∎∎∎–∎∎∎
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 Q9 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66
3
Fig. 2. (a)–(e) Micro and nanostructure of TiO2 NTs films grown at different anodizing times, investigated by SEM: 5 min (a), 10 min (b), 20 min (c), 1 h (d) and 2 h (e).
Fig. 3. GAXRD of TiO2 NTs grown at 30 V with anodizing time ranging from 5 min to 2 h and annealed at 450 1C. The spectra of as-prepared TiO2 NTs and Ti substrate annealed at 450 1C are given as reference. As-prepared NTs are amorphous and the formation of anatase crystals is induced by the thermal treatment at 450 1C [16].
1 h and 2.9 mm at 2 h). Based on SEM analysis, the partial collapse of NTs grown for 1–2 h and the formation of spikes/clusters of tubes (Fig. 2d, e), are also responsible for the lower J Photo observed. It has been previously shown that such clusters are detrimental to the 1D transport, as they create additional three dimensional pathways, increasing recombination (in the case of dye sensitized solar cells also decreasing dye loading) [21]. In contrast, it is difficult to establish whether the initial thin porous layer covering the top of the NTs (Fig. 2b, c), has a negative effect on the electron transport (concerning the solar water splitting process), as proposed by Albu et al. [19]. Considering it is only few nm thick and does not seem to prevent penetration of the electrolyte within the entire NTs film, we expect it has a minor effect only at very low λ (i.e. 300 nm), as such light is likely to be absorbed by this porous layer rather than by the NTs underneath. As a guideline, according to Eagles [22], the light penetration depth in TiO2 is 10 nm at λ ¼300 nm and 1 mm at 380 nm. More insights are provided by the IPCE studies (Fig. 4b). All the NTs show a weak response above 400 nm, due to oxygen vacancies induced during annealing by the significant presence of C within the as-prepared films grown in organic media [23]. Similarly, it was reported by Park [24] that a carbon thermal treatment triggered a visible response in the NTs. Each IPCE curve has a maximum (IPCEmax ), which slightly shifts towards longer λ as the length of the film
Please cite this article as: Regonini D, Clemens FJ. Anodized TiO2 Nanotubes: Effect of anodizing time on film length, morphology and photoelectrochemical properties. Mater Lett (2014), http://dx.doi.org/10.1016/j.matlet.2014.11.145i
67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132
4
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 Q5 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Q7 Q6 61 62 63 64 65 66
D. Regonini, F.J. Clemens / Materials Letters ∎ (∎∎∎∎) ∎∎∎–∎∎∎
67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 Fig. 4. Photocurrent density (a) and IPCE response (b) of TiO2 NTs anodized for a time ranging from 5 min to 2 h. All measurements were performed in 1 M KOH and indicate 84 the higher photoactivity of 1.1 mm thick NTs grown at 30 V for 20 min. 85 86 87 (EPFL). Dr. Fabio La Mattina and A. Kupferschmid (Empa) are also increases (IPCEmax is at 330 nm for l¼ 125 nm and at 350 nm when 88 gratefully acknowledged for their contribution to the IPCE set-up. l ¼2.9 mm). Such red shift is also observed in colloidal nanoparticles 89 under front illumination [25]. To evaluate the electron transport 90 properties, a useful parameter is the electron diffusion length, Ln , References 91 describing the competition between the diffusive transport and the 92 recombination of the charge carriers [26]. For photoanodes with [1] Roy P, Berger S, Schmuki P. TiO2 nanotubes: synthesis and applications. Angew 93 l=Ln 41, Ln can be quantified accordingly to Söedergren [27], where Chem Int Ed Engl 2011;50:2904–39. [2] Regonini D, Bowen CR, Jaroenworaluck A, Stevens R. A review of growth 94 the short wavelength side of the action spectra is simplified as mechanism, structure and crystallinity of anodized TiO2 nanotubes. Mater Sci 95 ln ðIPCEÞ ¼ l=Ln . The drop in the IPCE at λ¼300 nm is particularly Eng R 2013;74:377–406. 96 evident for an anodizing time of 1–2 h (Fig. 4b), therefore we assume [3] Roy P, Kim D, Lee K, Spiecker E, Schmuki P. TiO2 nanotubes and their application in dye-sensitized solar cells. Nanoscale 2010;2:45–59. 97 that the condition l=Ln 41 is satisfied for NTs grown 1 h or longer. This [4] Paramasivam I, Jha H, Liu N, Schmuki P. A review of photocatalysis using self98 is in line with the decrease observed in the J Photo of NTs anodized for organized TiO2 nanotubes and other ordered oxide nanostructures. Small 99 1–2 h (Fig. 4a). Using the IPCE values at 300 nm of 5.5 and 3% at 1 h 2012;8:3073–103. 100 and 2 h respectively, we estimate Ln to be 0.870.1 mm. This indicates [5] Kummer KM, Taylor E, Webster TJ. Biological applications of anodized TiO2 nanostructures: a review from orthopedic to stent applications. Nanosci 101 that despite the 1D transport within the NTs further studies are Nanotechnol Lett 2012;4:483–93. required to improve their electron transport and photoelectrochemical 102 [6] Galstyan V, Comini E, Faglia G, Sberveglieri G. TiO2 nanotubes: recent advances properties. 103 in synthesis and gas sensing properties. Sensors 2013;13:14813–38. [7] Shrestha NK, Schmuki P. Electrochemistry at TiO2 nanotubes and other Q8 104 semiconductor nanostructures. Electrochemistry, vol. 12. The Royal Society 105 of Chemistry; 87–131 (Chapter 3). 106 [8] Lynch RP, Ghicov A, Schmuki P. A photo-electrochemical investigation of self4. Conclusions organized TiO2 nanotubes. J Electrochem Soc 2010;157:G76–84. 107 [9] Kahnt A, Oelsner C, Werner F, Guldi DM, Albu SP, Kirchgeorg R, et al. Excited 108 state properties of anodic TiO2 nanotubes. Appl Phys Lett 2013;102:233109. Our study reveals that an optimal tube length of 1.1 mm is 109 [10] Beranek R, Tsuchiya H, Sugishima T, Macak JM, Taveira L, Fujimoto S, et al. required to maximize the photoresponse of undoped TiO2 NTs. The Enhancement and limits of the photoelectrochemical response from anodic 110 morphology of the anodic layer is very important and there is TiO2 nanotubes. Appl Phys Lett 2005;87:243114. 111 [11] Gärtner WW. Depletion-layer photoeffects in semiconductors. Phys Rev clearly a change in behavior once the transition from a thin porous 112 1959;116:84–7. layer to well-defined NTs is completed. For example, the [12] Augustynski J, Alexander BD, Solarska R. Metal oxide photoanodes for water 113 J Photo under mild bias ( 0.8 to 0.4 V vs Ag/AgCl) is much higher splitting. Top Curr Chem 2011;303:1–38. 114 for well-defined NTs (grown for 10–20 minutes). Extended ano[13] Ali Yahia SA, Hamadou L, Kadri A, Benbrahim N, Sutter EMM. Effect of 115 anodizing potential on the formation and EIS characteristics of TiO2 nanotube dizing time (1–2 h) at 30 V causes some damage to the NTs and 116 arrays. J Electrochem Soc 2012;159:K83–92. their photoresponse decreases, whereas the impact of the thin [14] Acevedo-Pena P, Gonzalez I. TiO2 nanotubes formed in aqueous media: 117 porous layer covering the top of the NTs is currently not clear. The relationship between morphology, electrochemical properties and photoelec118 trochemical performance for water oxidation. J Electrochem Soc 2013;160: electron diffusion length of undoped NTs is estimated to be 119 H452–H458. 0.8 70.1 mm and the study of slightly doped anodic TiO2 films [15] Fujishima A, Honda K. Electrochemical photolysis of water at a semiconductor 120 with improved electron transport properties [28,29] is ongoing. electrode. Nature 1972;238:37–8. 121 [16] Regonini D, Jaroenworaluck A, Stevens R, Bowen CR. Effect of heat treatment 122 on the properties and structure of TiO2 nanotubes: phase composition and chemical composition. Surf Interface Anal 2010;42:139–44. 123 [17] Jaroenworaluck A, Regonini D, Bowen CR, Stevens R. A microscopy study of the Acknowledgments 124 effect of heat treatment on the structure and properties of anodised TiO2 125 nanotubes. Appl Surf Sci 2010;256:2672–9. [18] Van de Krol R, Gratzel M. Photoelectrochemical hydrogen production. New 126 The authors gratefully thank the financial support received by the York: Springer; 2012; 102. 127 EU COFUND Marie Curie Actions Program (NANOSPLIT) and by the [19] Albu SP, Schmuki P. Highly defined and ordered top-openings in TiO2 128 Swiss National Science Foundation (project n. R'Equip 206021nanotube arrays. Phys Status Solidi (RRL) 2010;4:151–3. [20] Berger S, Hahn R, Roy P, Schmuki P. Self-organized TiO2 nanotubes: factors 129 121306, Fundamental Aspects of Photocatalysis and Photoelectroaffecting their morphology and properties. Phys Status Solidi B 2010;247:2424–35. 130 chemistry/Basic Research Instrumentation for Functional Character[21] Zhu K, Vinzant TB, Neale NR, Frank AJ. Removing structural disorder from 131 ization). The Cappuccino cell was machined at Empa based on the oriented TiO2 nanotube arrays: reducing the dimensionality of transport and 132 design provided by the Laboratory for Photonics and Interfaces recombination in dye-sensitized solar cells. Nano Lett 2007;7:3739–46. Please cite this article as: Regonini D, Clemens FJ. Anodized TiO2 Nanotubes: Effect of anodizing time on film length, morphology and photoelectrochemical properties. Mater Lett (2014), http://dx.doi.org/10.1016/j.matlet.2014.11.145i
D. Regonini, F.J. Clemens / Materials Letters ∎ (∎∎∎∎) ∎∎∎–∎∎∎
1 2 3 4 5 6 7 8 9
[22] Eagles DM. Polar modes of lattice vibration and polaron coupling constants in rutile (TiO2). J Phys Chem Solids 1964;25:1243–51. [23] Regonini D, Satka A, Jaroenworaluck A, Allsopp DWE, Bowen CR, Stevens R. Factors influencing surface morphology of anodized TiO2 nanotubes. Electrochim Acta 2012;74:244–53. [24] Park JH, Kim S, Bard AJ. Novel carbon-doped TiO2 nanotube arrays with high aspect ratios for efficient solar water splitting. Nano Lett 2005;6:24–8. [25] Hagfeldt A, Björkstén U, Lindquist S-E. Photoelectrochemical studies of colloidal TiO2-films: the charge separation process studied by means of action spectra in the UV region. Sol Energy Mater Sol Cells 1992;27:293–304.
5
[26] Leng WH, Barnes PRF, Juozapavicius M, O'Regan BC, Durrant JR. Electron diffusion length in mesoporous nanocrystalline TiO2 photoelectrodes during water oxidation. J Phys Chem Lett 2010;1:967–72. [27] Soedergren S, Hagfeldt A, Olsson J, Lindquist S-E. Theoretical models for the action spectrum and the current–voltage characteristics of microporous semiconductor films in photoelectrochemical cells. J Phys Chem 1994;98:5552–6. [28] Das C, Roy P, Yang M, Jha H, Schmuki P. Nb doped TiO2 nanotubes for enhanced photoelectrochemical water-splitting. Nanoscale 2011;3:3094–6. [29] Yang M, Jha H, Liu N, Schmuki P. Increased photocurrent response in Nbdoped TiO2 nanotubes. J Mater Chem 2011;21:15205–8.
Please cite this article as: Regonini D, Clemens FJ. Anodized TiO2 Nanotubes: Effect of anodizing time on film length, morphology and photoelectrochemical properties. Mater Lett (2014), http://dx.doi.org/10.1016/j.matlet.2014.11.145i
10 11 12 13 14 15 16 17