Spatial Charge Separation in Asymmetric Structure of Au Nanoparticle

Letter pubs.acs.org/NanoLett

Spatial Charge Separation in Asymmetric Structure of Au Nanoparticle on TiO2 Nanotube by Light-Induced Surface Potential Imaging Hyunjun Yoo,†,∥ Changdeuck Bae,†,‡,∥ Yunjeong Yang,† Seonhee Lee,† Myungjun Kim,† Hyunchul Kim,† Yunseok Kim,*,§ and Hyunjung Shin*,† †

Department of Energy Science, ‡Integrated Energy Center for Fostering Global Creative Researchers (BK 21 plus), and §School of Advanced Materials Science and Engineering, Sungkyunkwan University, Suwon 440-746, South Korea S Supporting Information *

ABSTRACT: Both enhancing the excitons’ lifetime and ingeniously controlling the spatial charge transfer are the key to the realization of efficiently photocatalytic and artificially photosynthetic devices. Nanostructured metal/metal-oxide interfaces often exhibit improved energy conversion efficiency. Understanding the surface potential changes of nano-objects under light illumination is crucial in photoelectrochemical cells. Under ultraviolet (UV) illumination, here, we directly observed the charge separation phenomena at the Au-nanoparticle/TiO2-nanotube interfaces by using Kelvin probe force microscopy. The surface potential maps of TiO2 nanotubes with and without Au nanoparticles were compared on the effect of different substrates. We observed that in a steady state, approximately 0.3 electron per Au particle of about 4 nm in diameter is effectively charged and consequently screens the surface potential of the underlying TiO2 nanotubes. Our observations should help design improved photoelectrochemical devices for energy conversion applications. KEYWORDS: Charge separation, Kelvin probe force microscopy, TiO2 nanotubes, Au nanoparticles, surface potential map

P

systems have been suggested by inducing electrostatic anisotropy such as metal−semiconductor interfaces.2,3,11−14 Au nanoparticles (NPs)-decorated TiO2 system is a wellestablished example where electrons can stay longer in the Au NPs upon charge transfer through the junction interfaces. Previously, charge separation/transfer phenomena have been studied by spectroscopic methods such as photoluminescence quenching, photoinduced absorption spectroscopy, and surface photovoltaic spectroscopy.15−18 These methods provide the dynamics, yet are limited in terms of the spatial information. Complimentarily, local probe techniques allow for direct observation of not only the spatially resolved surface properties, but also the local manifestation of physical/chemical characteristics in nanostructures.19−23 Kelvin probe force microscopy (KPFM) employed in this study has the capability of directly

hotoinduced electron−hole pairs in semiconductors allow for solar conversion into electrical or chemical energies. To improve the overall conversion efficiency, both enhancing the excitons’ lifetime and ingeniously controlling the spatial charge transfer are essential, as in natural photosynthesis. Rationally formed metal/semiconductor interfaces at the nanoscale often exhibit improved device/cell performance. Investigating and tailoring the surface potential changes of the system at around the interfaces during light illumination is of significance in the fundamental understanding and application of photoelectrochemical cells.1−3 TiO2 is a well-known material for many photoactive applications including photocatalysts,4,5 dye-sensitized solar cells,6 and photodetectors7 due to its chemical stability,8 high electron mobility,9 and high photocatalytic activity. On its own, it shows a limited performance because the recombination is fast and most of the excitons cannot be involved in the surface reactions upon traveling in TiO2, that is, bulk recombination.5,10 To increase the excitons’ lifetime, asymmetric material © XXXX American Chemical Society

Received: April 15, 2014 Revised: June 25, 2014

A

dx.doi.org/10.1021/nl501381a | Nano Lett. XXXX, XXX, XXX−XXX

Nano Letters

Letter

imaging the charge separation/transfer phenomena because of the nanometer scale spatial resolution and sub-millivolt sensitivity in the surface potential. Other researchers have investigated various nanomaterials/structures such as nanowires,24,25 nanocrystals,26 and nanodumbbell structures27 by using KPFM. In this Letter, we investigate the charge transfer/separation of TiO2 NTs (TNTs) with and without NPs of Au on their outer surfaces at the Fermi energy equilibrium under the UV on/off using KPFM. The surface potential of TiO2 nanotubes is monitored before and after attachment of Au nanoparticles on the different substrates (i.e., Si/SiO2 and Pt substrates) under the UV on/off using KPFM. From direct imaging, we found that in a steady state, approximately 0.3 electron per Au particle of about 4 nm in diameter is effectively charged and consequently screens the surface potential of the underlying TiO2 nanotubes. Our observations should help design improved photoelectrochemical devices for energy conversion applications. TNTs were prepared by template-directed atomic layer deposition techniques, followed by annealing that had been originally developed by some of the authors.28,29 Indeed, TiO2 with a nanotubular geometry is promising for photoactive applications such as photocatalytic and artificial photosynthetic devices. It has a wide band gap (∼3.2 eV) and thus the UV light absorbed generates excitons everywhere within TiO2.30 Because the exciton diffusion length is very short (∼ nm),1,31 only small amounts of photogenerated charge carriers are involved in the surface reactions. In other words, the bulk recombination is predominant even in nanostructured TiO2 possessing a characteristic length of more than tens of nanometers. Therefore, the ultrathin wall layers of TiO2 nanotubes should be beneficial in maximizing the charge separation efficacy. Figure 1a shows the resulting TNTs, having an average diameter of ∼60 nm and a wall thickness of about 8 nm. Au NPs were coated by soaking TNTs in an aqueous solution of HAuCl4 (5 mM, pH 6) at 70 °C for 2 h, as shown in Figure 1b (an average diameter of about 4 nm). Such a controlled precipitation of Au NPs on the outer surfaces of TNTs with a very high areal density should serve as many junction interfaces across which photoexcited electrons are likely to be separated by an energy barrier. Crystalline Au NPs were found asprepared and no annealing was carried out. The nanotubular geometry also ensures a large specific surface area.32 The number density is obtained to be ∼7.0 × 1011 per cm2 based on the SEM image as shown in Figure 1b. It is also noted that the clear interfaces between Au NPs and TNTs are believed to be critical for the electron injection efficiency from TNTs to Au NPs.33 KPFM measurements were performed on smooth substrates with a tunnel oxide, that is, SiO2 (5 nm)/n++ Si (see Figure 2). Under the UV illumination (that is, in a steady state), we measured the topography and surface potential of the samples under ambient conditions, simultaneously. It is noted that the environment of the TiO2 as oxygen-rich or oxygen-poor influences the surface potential imaging of charge-injected regions in TiO2 as reported by others.34 A commercial atomic force microscope (SPA-400, SII, Japan) was employed using both Au-coated tips (SI-DF3-A, SII, a resonant frequency, 27 kHz; a spring constant, 1.6 N/m) and Pt/Ir coated tips (EFM, Nanoworld, 75 kHz, 2.8 N/m). Images were acquired at scan rates of 0.1−0.2 Hz with applied voltages of 2−3 V and ac frequencies near the resonant frequency of

Figure 1. Scanning electron microscopy image of (a) templatedirected ALD grown TNTs and (b) Au NPs-attached TNT. (c) Highresolution TEM image of an Au NPs-attached TNT showing the most intensive diffracting plane of anatase {101}. Single crystalline Au dots with the average size of ∼4 nm in diameter showing clear lattice images of {111} planes deposited on the surface of anatase TNTs. Moiré fringes (indicated by red circles) are clearly observed between two different crystalline lattices of Au and, anatase, TiO2.

Figure 2. A comparative KPFM study between TNT and Au/TNT on n++ Si substrates with a tunnelling oxide (thermal SiO2, 5 nm thick). From left to right, AFM height images (a,d) and surface potential images in the dark (b,e) and under UV illumination (c,f). Each pair of topography and surface potential images was collected simultaneously with/without light illumination, respectively.

cantilevers. Different light-emitting diodes (LEDs) were used as UV (λ = 365 nm), red (λ = 620 nm), green (λ = 525 nm), and blue (λ = 400 nm) light sources. Surface potential differences, ΔV = (Vsample − Vsubstrate)UV − (Vsample − Vsubstrate)dark, are displayed by area-averaging each micrograph. In the dark, the surface potential of TNTs was about 7.5 mV higher than that of B

dx.doi.org/10.1021/nl501381a | Nano Lett. XXXX, XXX, XXX−XXX

Nano Letters

Letter

Figure 3. Photoresponse results of TNT and Au/TNT on Pt substrates with various light sources. The representative surface potential imaging and line profiles of (a−d) TNT and (e−h) Au/TNT on Pt substrates, respectively. (i) Energy band diagrams of the Au/TNT junctions, highlighting the ultrathin wall layer of TNTs.44 (j,k) Schematic illustrations of cross-sectional view of our nanotubular structures with/without Au NPs (as yellow dots) for charge separation behaviors with and without Au NPs under UV illumination.

underlying TNTs in which holes were accumulated in accordance with the above control experiments, leading to a smaller ΔV (Figure 2d−f). The above scenario is further confirmed as follows, by employing Pt as substrate. Now, the surface potential measurements on Au-decorated TNTs were carried out with different light sources and intensities. In this case, we expect that the depletion layers inside our n-type TiO2 are to be more visible (i.e., positive sign in our KPFM for the tip is grounded) due to the excess electrons in Au and the direct electrical contact to the Pt substrate. As shown in Figure 3a−h, ΔV of TNT was 88.6 mV while ΔV of Au/TNT was 33.4 mV. Moreover, ΔV values under the red (2.00 eV), green (2.36 eV), and blue (3.10 eV) LED illumination show no changes within our detectable voltage resolution (few millivolts) (Figure 3g,h). This indicates that the majority of the photogenerated electrons are coming from the valence band edge of TNTs rather than possible defect states or hot carriers by the surface plasmon resonances of Au NPs. Such explanations are indeed reasonable when considering the length scale of exciton diffusion for the electrically asymmetric nanotubular geometry. (i) TNTs on planar Pt substrate: The majority of excitons are recombined within the body of TNTs before traveling until the TNT/Pt interfaces through the TNT walls. Some of the electrons near

the substrate, mainly due to the work function difference between them. After UV illumination, ΔV was only 6.8 mV, indicative of a change in the effective Fermi level. The increase in the positive sign possibly implies the hole accumulation within TNTs by injecting excitonic electrons into the substrate with an electrical ground through tunneling in a steady state. Such an argument is rational in that the outer surfaces of TNTs are highly curved, working like a sharp tip in contact with the substrate. The tips’ work function was calibrated against the freshly cleavaged highly ordered pyrolytic graphite (HOPG) [see Supporting Information Figure S1 and Table S1]. As the results, the work function values of Pt, TiO2, and Au by KPFM were 5.08, 4.97, and 5.22 eV, respectively. Those are comparable with the literature values for Au (5.10−5.47 eV), Pt (5.12−5.93 eV), and TiO2 (4−6 eV).35−38 For the case of Au NPs-attached TNTs, remarkably, ΔV was 0.4 mV (that is, Vsample − Vsubstrate was from 9.5 to 9.9 mV). It is noteworthy that ΔV was consistent with that without Au NPs, showing that our measurements are robust and reproducible. Although the electron transfer through the interfaces between Au/TNT and SiO2(5 nm)/n++ Si happens in a similar way, the photogenerated electrons transferred to the Au NPs are also expected under the UV illumination (∼3.4 eV).39 The separated electrons to the Au side were found to screen the C

dx.doi.org/10.1021/nl501381a | Nano Lett. XXXX, XXX, XXX−XXX

Nano Letters

Letter

Figure 4. Investigation of ΔV as a function of illumination intensity: Line profiles of potential difference between the Pt substrate and (a) TNTs and (b) Au/TNTs. Plots of (c) the average ΔV and (d) the estimated number of electrons by the illumination intensity.

they are not recombined in TNTs. Subsequently, they are involved in the water oxidation reactions under ambient conditions in a steady state. The back-side transfers are probably limited by a Schottky barrier. In our system, the Schottky barrier height (ΦSBH) and width at the Au/TNT interfaces are modified because of the environmental Fermi level pinning by the small size of Au and the thin wall layers of TNTs, respectively. The modified band diagram is displayed in Figure 3i (see the solid and dotted lines). Indeed, the reduced ΦSBH was extracted to be 0.25−0.5 eV from the conductive atomic force microscopy even though the ideality factor (η) was found to deviate toward higher values (∼5 to 10). Nonetheless, the Fermi level pinning by the Au NPs should extend the lifetime of excitonic electrons, enhancing the charge separations through the Au/TNT interfaces. Currently, the enhancement mechanism of photocatalytic activities by attachment of metal nanoparticles is still under debate. There are three complementary contributions: (i) Excitons are selectively transferred from the semiconductor into metal nanoparticles, disturbing the recombination. (ii) Metal nanoparticles on the semiconductor cause the shift of the quasiFermi level of the semiconductor to more negative potential. (iii) During photocatalytic reactions in the aqueous media, electrons are accumulated at the metal−semiconductor interfaces under the UV illumination.1−3,14,41−43 The present observation provides more insight into the enhancement mechanisms by demonstrating the screening effects by the presence of the metal−semiconductor interfaces. In conclusion, we have observed the charge separation/ transfer of the photogenerated electrons from TNTs into Au NPs by using KPFM. We found that the transferred electrons act to screen the underlying semiconductor, reducing the surface potential differences under the UV illumination. By varying the illumination intensity, we estimated the electron number in an Au NP to be ∼0.3 in a steady state. The present

the interfaces should transfer to the Pt substrate, accumulating holes behind, as shown in Figure 3j. Consequently, the effective Fermi level of TNTs is lowered, resulting in the increased work function differences by KPFM. (ii) Au NPs/TNTs on planar Pt substrate: Near the Pt side, similar processes to those in the TNT/Pt case are expected. Because of the presence of the very thin TiO2 wall layers, however, the photoexcited electrons, even with a short diffusion length, can be transferred into the Au NPs on them. Considering the contact areas between the TNT/Pt and the TNT/Au, we conclude that the electron separation toward the Au NPs is the principal process. The effective number of electrons in the Au NPs is quantitatively estimated, as will be discussed later. Qualitatively, the electrons in Au could screen the holes accumulated in TNTs in a steady state under UV, resulting in a small ΔV (see, Figure 3k). Note that such a small ΔV found in metal nanoparticle-decorated semiconductors shows much higher photocatalytic activities.12,14 ΔV of both TNT and Au/TNT as a function of the UV illumination intensity gives us further insight into the charge separation behaviors (see Figure 4). Figure 4c shows that ΔV is saturated with increasing light intensity. As we assumed that the surface potential on the tubes is screened by the electrons in Au, ΔV should be related to the number of electrons in/on the Au NPs. On the basis of a simple capacitor model,40 we estimated the effective surface charge density on the individual Au particles assuming that the exciton production per volume in the tubes is the same for both cases (i.e., TNT and Au/ TNT). As shown in Figure 4d, it is saturated to be ∼0.3 with increasing UV intensity. At the current stage, the details of the microscopic reactions at the three phase contact lines of Au/ TNT/air are not clear. Most probably, the photocatalytic oxidation of ambient water molecules under ambient conditions will happen. Therefore, it might be rate-determining in our systems: Excitonic electrons are injected into Au NPs unless D

dx.doi.org/10.1021/nl501381a | Nano Lett. XXXX, XXX, XXX−XXX

Nano Letters

Letter

(18) Tachibana, Y.; Moser, J. E.; Grätzel, M.; Klug, D. R.; Durrant, J. R. J. Phys. Chem. 1996, 100, 20056−20062. (19) Hinterdorfer, P.; Dufrêne, Y. F. Nat. Methods 2006, 3, 347−355. (20) Yoo, H.; Bae, C.; Kim, M.; Hong, S.; No, K.; Kim, Y.; Shin, H. Appl. Phys. Lett. 2013, 103, 022902. (21) Lee, J.-S.; Cho, J.; Lee, C.; Kim, I.; Park, J.; Kim, Y.-M.; Shin, H.; Lee, J.; Caruso, F. Nat. Nanotechnol. 2007, 2, 790−795. (22) Baik, S. J.; Lim, K. S.; Choi, W.; Yoo, H.; Lee, J.-S.; Shin, H. Nanoscale 2011, 3, 2560−2565. (23) Liu, Z.; Zhu, H.; Song, N.; Lian, T. Nano Lett. 2013, 13, 5563− 5569. (24) Schäfer, S.; Wang, Z.; Zierold, R.; Kipp, T.; Mews, A. Nano Lett. 2011, 11, 2672−2677. (25) Liu, C.; Hwang, Y. J.; Jeong, H. E.; Yang, P. Nano Lett. 2011, 11, 3755−3758. (26) Krauss, T.; Brus, L. Phys. Rev. Lett. 1999, 83, 4840−4843. (27) Costi, R.; Cohen, G.; Salant, A.; Rabani, E.; Banin, U. Nano Lett. 2009, 9, 2031−2039. (28) Shin, H.; Jeong, D.-K.; Lee, J.; Sung, M. M.; Kim, J. Adv. Mater. 2004, 16, 1197−1200. (29) Bae, C.; Yoo, H.; Kim, S.; Lee, K.; Kim, J.; Sung, M. M.; Shin, H. Chem. Mater. 2008, 20, 756−767. (30) Mor, G. K.; Kim, S.; Paulose, M.; Varghese, O. K.; Shankar, K.; Basham, J.; Grimes, C. A. Nano Lett. 2009, 9, 4250−4257. (31) Warren, S. C.; Thimsen, E. Energy Environ. Sci. 2012, 5, 5133− 5146. (32) Kim, M.; Bae, C.; Kim, H.; Yoo, H.; Montero Moreno, J. M.; Jung, H. S.; Bachmann, J.; Nielsch, K.; Shin, H. J. Mater. Chem. A 2013, 1, 14080−14088. (33) Pu, Y.-C.; Wang, G.; Chang, K.-D.; Ling, Y.; Lin, Y.-K.; Fitzmorris, B. C.; Liu, C.-M.; Lu, X.; Tong, Y.; Zhang, J. Z.; Hsu, Y.-J.; Li, Y. Nano Lett. 2013, 13, 3817−3823. (34) Kim, H.; Hong, S.; Kim, D.-W. Appl. Phys. Lett. 2012, 100, 022901. (35) CRC Handbook of Chemistry and Physics, version 20.; Lide, D. R., Ed.; CRC Press/Taylor and Francis Group: Boca Raton, FL, 2008; pp 12−114. (36) Liu, G.; Jaegermann, W.; He, J.; Sundström, V.; Sun, L. J. Phys. Chem. B 2002, 106, 5814−5819. (37) Onda, K.; Li, B.; Petek, H. Phys. Rev. B 2004, 70, 045415. (38) Xiong, G.; Shao, R.; Droubay, T. C.; Joly, A. G.; Beck, K. M.; Chambers, S. A.; Hess, W. P. Adv. Funct. Mater. 2007, 17, 2133−2138. (39) Williams, R. Phys. Rev. 1965, 140, A569−A575. (40) Chen, X. Q.; Yamada, H.; Horiuchi, T.; Matsushige, K.; Watanabe, S.; Kawai, M.; Weiss, P. S. J. Vac. Sci. Technol., B 1999, 17, 1930−1934. (41) Wood, A.; Giersig, M.; Mulvaney, P. J. Phys. Chem. B 2001, 105, 8810−8815. (42) Jakob, M.; Levanon, H.; Kamat, P. V. Nano Lett. 2003, 3, 353− 358. (43) Adams, D. M.; Brus, L.; Chidsey, C. E. D.; Creager, S.; Creutz, C.; Kagan, C. R.; Kamat, P. V.; Lieberman, M.; Lindsay, S.; Marcus, R. A.; Metzger, R. M.; Michel-Beyerle, M. E.; Miller, J. R.; Newton, M. D.; Rolison, D. R.; Sankey, O.; Schanze, K. S.; Yardley, J.; Zhu, X. J. Phys. Chem. B 2003, 107, 6668−6697. (44) Park, H.; Beresford, R.; Hong, S.; Xu, J. J. Appl. Phys. 2010, 108, 094308.

works reported here may help to stimulate study of the efficient electron−hole separation in photoactive devices using asymmetric, one-dimensional metal/semiconductor hybrid structures.

■

ASSOCIATED CONTENT

S Supporting Information *

Fabrication of TiO2 nanotubes (NTs), work function measurement by Kelvin probe force spectroscopy, calculation of charge density, band diagrams, and table of CPD and work function of the Pt, Au, and TiO2 thin film. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected]. Author Contributions ∥

H.Y. and C.B. contributed equally to this work.

Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors acknowledge financial support from the National Research Foundation of Korea Grant funded by the Korean Government (2010-0028972, 2012M3A7B4049986, and 2013R1A2A2A01068499). This work was supported by the Human Resources Development program (No. 20124010203270) of KETEP grant funded by the Korean Government Ministry of Knowledge Economy. The authors also acknowledge financial support in part from the Agency for Defense Development (ADD) in Korea.

■

REFERENCES

(1) Linic, S.; Christopher, P.; Ingram, D. B. Nat. Mater. 2011, 10, 911−921. (2) Tada, H.; Mitsui, T.; Kiyonaga, T.; Akita, T.; Tanaka, K. Nat. Mater. 2006, 5, 782−786. (3) Kamat, P. J. Phys. Chem. B 2002, 7729−7744. (4) Fujishima, A.; Honda, K. Nature 1972, 238, 37−38. (5) Linsebigler, A. L.; Lu, G.; Yates, J. T. Chem. Rev. 1995, 95, 735− 758. (6) O’Regan, B.; Grätzel, M. Nature 1991, 353, 737−740. (7) Zou, J.; Zhang, Q.; Huang, K.; Marzari, N. J. Phys. Chem. C 2010, 114, 10725−10729. (8) Jones, D. A. Principles and prevention of corrosion, 2nd ed.; Prentice Hall: New York, 1995. (9) Forro, L.; Chauvet, O.; Emin, D.; Zuppiroli, L.; Berger, H.; Lévy, F. J. Appl. Phys. 1994, 75, 633. (10) Frank, A. J. Coord. Chem. Rev. 2004, 248, 1165−1179. (11) Hirakawa, T.; Kamat, P. V. V J. Am. Chem. Soc. 2005, 127, 3928−3934. (12) Subramanian, V.; Wolf, E. E.; Kamat, P. V. J. Am. Chem. Soc. 2004, 126, 4943−4950. (13) Zheng, Y.; Zheng, L.; Zhan, Y.; Lin, X.; Zheng, Q.; Wei, K. Inorg. Chem. 2007, 46, 6980−6986. (14) Subramanian, V.; Wolf, E.; Kamat, P. V. V J. Phys. Chem. B 2001, 105, 11439−11446. (15) Greenham, N.; Peng, X.; Alivisatos, A. Phys. Rev. B 1996, 54, 17628−17637. (16) Yu, G.; Heeger, A. J. J. Appl. Phys. 1995, 78, 4510−4515. (17) Jailaubekov, A. E.; Willard, A. P.; Tritsch, J. R.; Chan, W.-L.; Sai, N.; Gearba, R.; Kaake, L. G.; Williams, K. J.; Leung, K.; Rossky, P. J.; Zhu, X.-Y. Nat. Mater. 2013, 12, 66−73. E

dx.doi.org/10.1021/nl501381a | Nano Lett. XXXX, XXX, XXX−XXX