J. Phys. Chem. B 2005, 109, 9651-9655
9651
Photocurrent Generation from Semiconducting Manganese Oxide Nanosheets in Response to Visible Light Nobuyuki Sakai, Yasuo Ebina, Kazunori Takada, and Takayoshi Sasaki* AdVanced Materials Laboratory, National Institute for Materials Science, 1-1 Namiki, Tsukuba, Ibaraki 305-0044, Japan ReceiVed: January 4, 2005; In Final Form: March 6, 2005
Unilamellar nanosheet crystallites of manganese oxide generated the anodic photocurrent under visible light irradiation (λ < 500 nm), while the nanosheets themselves were stable as revealed by in-plane XRD and UV-visible absorption spectra. The band gap energy was estimated to be 2.23 eV on the basis of the photocurrent action spectrum. The molecular thickness of ∼0.5 nm may facilitate the charge separation of excited electrons and holes, which is generally very difficult for strongly localized d-d transitions. The monolayer film of MnO2 nanosheets exhibited the incident photon-to-electron conversion efficiency of 0.16% in response to the monochromatic light irradiation (λ ) 400 nm), which is comparable to those for sensitization of monolayer dyes adsorbed on a flat single-crystal surface. The efficiency declined with increasing the layer number of MnO2 nanosheets, although the optical absorption was enhanced. The recombination of the excited electron-hole pairs may become dominant when the carriers need to migrate a longer distance than 1 layer through multilayered nanosheets.
Introduction The transition metal oxides such as TiO2 and ZnO are wellknown semiconductors that can generate photocurrent under UV light irradiation.1-6 These oxides are characterized by their electronic structure with empty d-shell (d ) 0) or with fully occupied orbitals (d ) 10). In these oxides, charge separation takes place via photon absorption associated with a gap energy between valence and conduction bands, which are mainly generated from oxygen 2p and metal 3d or 4s orbitals, respectively. This electronic transition involving two different atoms may stabilize the charge separation and allow the migration of carriers having a relatively longer lifetime (10-100 ns),7-10 which is the origin of photocatalytic activities and photovoltaic conversion capabilities of these oxides. However, their large band gap (>3 eV), which requires UV light to excite, leads to a disadvantage in solar energy utilization. Extensive efforts have been made to overcome this problem by dye-sensitization11,12 and introduction of impurity level by anion doping.13 Transition metal oxides with partially filled d-levels may be attractive because they absorb visible light involving d-d transitions. However, because the d electrons are confined in central transition metal ions, migration of the excited carriers into surfaces and/or interfaces is difficult, and the excited electrons are rapidly recombined with the holes. This may be the main reason there have been no reports on photocurrent generation from oxides with d ) 1-9 configurations except for a few exceptional cases.14 These difficulties can be overcome with a new modification of nanosized oxides as described below. Recently, several groups including our laboratories have reported the synthesis of functional oxide nanosheets, for example, Ti0.91O2,15,16 MnO2,17,18 and Ca2Nb3O10,19 by delaminating precursor crystals of a layered oxide into its elementary layers. The crystallites have an ultrathin thickness of less than * Corresponding author. Phone: +81-29-860-4313. Fax: +81-29-8549061. E-mail:
[email protected].
1 nm and a lateral size ranging from submicrometers to several tens of micrometers. These unusual structural features with extremely high two-dimensional anisotropy are expected to evolve novel chemical and physical properties that differ from those for granular bulk oxides. In practice, titania nanosheets exhibit a larger band gap energy than that for bulk TiO2 due to quantum size effects.20 MnO2 nanosheets have an absorption peak centered at around 380 nm, which is tailed to the visible regime,17,21 while the parent layered manganese oxide has featureless absorption over an entire wavelength range of 200-800 nm. In the present work, we fabricated the MnO2 nanosheet electrodes and examined the photoelectrochemical properties of these electrodes. Experimental Section All chemicals were of >99.9% purity or of analytical grade. Polyethylenimine (PEI, MW ) ∼7.5 × 105) and polydiallyldimethylammonium (PDDA, MW ) ∼1 × 105 to 2 × 105) chloride were obtained from Aldrich Co. and used without further purification. Ultrapure water, filtered by a Milli-Q reagent water system to a resistivity of >17 MΩ cm, was used throughout the experiments. Propylene carbonate (Aldrich, H2O content 375 nm, Q ) 7.4 mC cm-2). The data for the film on a quartz glass chip are also shown as a broken trace.
Figure 2. In-plane XRD patterns for (a) ITO and monolayer film of MnO2 nanosheets (b) before and (c) after the photocurrent generation (λ > 375 nm, Q ) 7.4 mC cm-2). The circles in the figure indicate the diffraction from the two-dimensional hexagonal architecture of MnO2 nanosheets. The wavelength of the incident X-ray is 0.13988 nm.
AFM image (not shown) revealed a dense monolayer coverage of the substrate surface with the nanosheets. The in-plane X-ray diffraction (XRD) technique could detect diffraction peaks from the monolayer film of MnO2 nanosheets on the ITO coated quartz glass substrate. Upon self-assembly of MnO2 nanosheets, two new peaks appeared at 2θ ) 32.9° and 58.6° (indicated by circles in Figure 2b) in addition to many reflections from ITO substrate (Figure 2a). These two peaks with d-values of 2.47 and 1.43 Å are attributable to 10 and 11 reflections from the two-dimensional hexagonal architecture of MnO2 nanosheet. Photocurrent Generation from Monolayer Films of MnO2 Nanosheet Electrodes. We detected the anodic photocurrent from MnO2 nanosheets, which were deposited as a monolayer on ITO coated quartz glass as shown in the inset of Figure 3. The generation of anodic photocurrent indicates the n-type semiconducting nature of this material.10 A full understanding of the redox reactions of excited electrons and holes associated with photocurrent generation is not available at the present stage. We speculate that the electrons may reduce water and/or molecular oxygen dissolved in the electrolyte, while the holes may oxidize water and/or propylene carbonate used as electrolyte. The photocurrent action spectrum (Figure 3), in which the incident photon-to-electron conversion efficiency (IPCE) is plotted as a function of excitation wavelength, shows the onset wavelength for the photocurrent generation of around 500 nm, indicating that the photoelectric conversion occurred in response to visible light. The IPCE for the monolayer film of MnO2
Photocurrent Generation from MnO2 Nanosheets
Figure 3. Photocurrent action spectra for monolayer film of MnO2 nanosheet/ITO electrode at 0.4 V vs Ag/Ag+ in a propylene carbonate solution containing 0.1 mol dm-3 LiClO4. The solid line shows the optical absorption spectra of MnO2 nanosheet normalized to the photocurrent profile. Inset shows periodic anodic photocurrent generation from the monolayer film of MnO2 nanosheet/ITO electrode under monochromatic light centered at 450 nm.
nanosheet electrode was approximately 0.16% under a monochromatic light (λ ) 400 nm) at 0.4 V, which is comparable to that from dyes, for example, ruthenium complexes, adsorbed in monolayer on the flat single-crystal surface.12 The spectral profile in a wavelength range of 350-500 nm closely matches the optical absorption feature of the MnO2 nanosheets, confirming that the observed photocurrent arises from MnO2 nanosheets. To the best of our knowledge, this is the first report on photocurrent generation from manganese oxides, although a wide variety of compositions, crystal structures, and morphologies have been reported so far in this system.22 It is worth considering why the photocurrent involving d-d transitions is generated in the present system. The twodimensional structure with an ultrathin thickness of less than 1 nm may play an essential role in the separation of the excited electron-hole pairs. The MnO2 nanosheet is composed only of one plane of Mn ions sandwiched by two hexagonal planes of oxygen atoms. The two oxygen planes are stacked in a relation to provide octahedral sites for Mn ions, producing a coplanar oxide nanosheet. The atomic architecture normal to the sheet is comparable to molecules or metal complexes, while that parallel to the sheet corresponds to bulk crystals. Such an ultrathin MnO2 nanosheet (∼0.5 nm) may facilitate the excited electrons and holes to migrate into ITO and interface with electrolyte, respectively, before their loss via recombination. To estimate the band gap energy for the MnO2 nanosheet electrode, the square root of the IPCE (η) times the photon energy, (ηhν)0.5, is plotted against photon energy hν (Figure 4).23 This plot gives a straight line in a photon energy range close to the absorption threshold, suggesting that the MnO2 nanosheets have an indirect electronic transition near the band gap. The band gap energy for the monolayer film of the MnO2 nanosheet electrode is derived as 2.23 eV from the intercept of the linear portion with the abscissa. We also examined the dependence of the photocurrent generation on the applied potential of the MnO2 nanosheet electrodes (Figure 5). Under each applied potential, the monolayer film of MnO2 nanosheet electrode was irradiated with monochromatic light centered at 400 nm. The behavior of photocurrent generation was dependent on the direction of potential change. When the potential was swept from -0.5 to 1.0 V by a 0.1-V difference, the photocurrent appeared to flow at -0.3 V and increased with increasing the potential until 0.4 V and then decreased in more positive potentials. When the potential was changed from 1.0 to -0.5 V, the photocurrent
J. Phys. Chem. B, Vol. 109, No. 19, 2005 9653
Figure 4. Variation of the square root of IPCE (η) times hν with photon energy for the monolayer film of MnO2 nanosheet electrode at 0.4 V. Inset shows the band gap energy for the MnO2 nanosheet electrode with various numbers of layer pairs.
Figure 5. Dependence of the IPCE values on applied potential for the monolayer film of MnO2 nanosheet electrode. The incident light source was monochromatic light centered at 400 nm.
was almost constant from 0.8 to 0.4 V and gradually decreased as the potential becomes more negative, and became undetectable at -0.4 V. In both cases, the onset potential is at around -0.35 V, and the anodic photocurrent was observed under more positive potential than the onset potential. This value is consistent with the redox potential of Mn3+/Mn4+ of the MnO2 nanosheet observed in the cyclic voltammogram, which will be published elsewhere. The hysteresis observed in the potential range between 0.0 and 0.9 V may suggest that the Mn3+ state is rather superior for the photocurrent generation to Mn4+ state. The Mn3+ state, which is produced by partial reduction of MnO2 nanosheets due to polarization at -0.5 V, is gradually oxidized to the Mn4+ state during the potential sweep in a positive direction. The decrease of the Mn3+/Mn4+ ratio brings about the attenuation of the photocurrent in a potential range of 0.4-0.9 V. In contrast, the potential change in a negative direction from 1.0 V does not change the oxidation state of Mn4+ ions, showing that the photocurrent magnitude enhances with larger polarization from the onset potential and becomes saturated in a potential range of 0.4-0.8 V. This behavior is similar to that for well-known n-type semiconductors such as TiO2.10 On the basis of the results shown above, the energy level diagram of the MnO2 nanosheets can be described as shown in Figure 6. Mn 3d energy level is split into higher (eg) and lower (t2g) energy levels due to the ligand field of MnO6 octahedra. The optical absorption of the MnO2 nanosheets in the visible light range principally originates from d-d transitions of Mn ions. The energy level of the Mn 3d eg state should be equal to the redox potential of Mn3+/Mn4+ of MnO2 nanosheets because
9654 J. Phys. Chem. B, Vol. 109, No. 19, 2005
Sakai et al.
Figure 7. Dependence of the IPCE value (λ ) 400 nm) on the layer number of MnO2 nanosheets.
Figure 6. Schematic illustration of energy level diagram of MnO2 nanosheet in comparison with titania nanosheet.
the Mn3+ state in the MnO6 octahedra takes a high spin state, whose electrons occupy the eg state, due to its low ligand-field strength. The agreement between the redox potential of Mn3+/Mn4+ of MnO2 nanosheets and the onset potential of photocurrent generation strongly suggests that the excited energy level of MnO2 nanosheets corresponds to the energy level of the Mn 3d eg state. The band gap energy corresponds to the energy difference between Mn 3d t2g and eg states. Therefore, the photocurrent can be generated by an excitation of electrons from the Mn 3d t2g level to the Mn 3d eg level. Mn 3d t2g states and eg states serve as the “valence band” and the “conduction band” as semiconductor, respectively. It is interesting to compare the energy diagram with that for the titania nanosheet (Figure 6), which shows a large band gap due to size quantization effects.20 The energy levels of Mn 3d t2g and eg are higher and lower than the valence and the conduction bands of titania nanosheet, respectively. Alternate layer-by-layer assembly of titania and MnO2 nanosheets may create two-dimensional multiple-quantum-well structures.10,24 Stability of MnO2 Nanosheet Electrodes toward Photocurrent Generation. A serious problem of some semiconductors in practical applications is the instability against photochemical processes including photocurrent generation, which is typically found for CdS and CdSe despite their sensitivity to visible light.25-27 The MnO2 nanosheets were found to be stable against the photocurrent generation. Because the hexagonal two-dimensional unit cell of the nanosheet crystallite (a ≈ 0.285 nm) contains one formula unit, an ideal monolayer film without gaps and overlaps of the nanosheets has 1.4 × 1015 Mn atom cm-2. If the MnO2 nanosheet is assumed to undergo photodecomposition as eq 1 upon light irradiation, 0.92 mC cm-2 of the photocurrent should be consumed to decompose the monolayer of MnO2 nanosheet.
MnO2 + 4h+ f Mn4+ + O2v
(1)
Even after the photocurrent of 7.4 mC cm-2, which is 8 times that required to decompose the entire monolayer film of MnO2 nanosheet, was generated, the in-plane XRD peaks corresponding to MnO2 nanosheets did not appear to change as shown in Figure 2c. UV-vis absorption spectra (Figure 1) also showed a negligible difference before and after the photocurrent generation. These data indicate that the MnO2 nanosheet was stable without dissolution itself against photocurrent generation.
Photoelectrochemical Behaviors on Multilayer Films of MnO2 Nanosheets. We have also examined the photocurrent generation from multilayer films of MnO2 nanosheet electrodes. The success in multilayer formation was confirmed by UVvis absorption spectra. The band gap energies for MnO2 nanosheet electrode with various layer numbers estimated from the photocurrent action spectra are plotted in the inset of Figure 4. They were nearly independent of the layer number of MnO2 nanosheets. Dependence of the photocurrent generation on the applied potential was also similar to that for the monolayer film of MnO2 nanosheet electrodes, showing the onset potential of -0.35 V for photocurrent generation. These results indicate that the nanosheet is electronically isolated in the multilayer assemblies without affecting the electronic state of neighboring nanosheets. The quantum efficiency declined with increasing the layer number. Even though the optical absorption was enhanced, the IPCE value reduced contrarily (Figure 7). This tendency may reflect the essential nature of the phenomenon. The recombination of the excited electron-hole pairs rather than the charge transfer via some redox reactions at the interface predominantly proceeds when the carriers need to migrate a longer distance through the stacked multiple numbers of nanosheets. This phenomenon raises a presumption that only a topmost layer of multilayer films of MnO2 nanosheets plays a role in generating the photocurrent. To confirm this hypothesis, we examined the dependence of photocurrent generation on the direction of incident light. When the light was irradiated from the back face of the electrode, the incident intensity of irradiated light to the topmost layer in contact to the electrolyte decreases to ca. 40% at λ ) 400 nm, because the 9 layers of MnO2 nanosheets behind the topmost layer absorb ca. 60% of incident photon. The IPCE value at λ ) 400 nm was actually damped down to approximately 40% in comparison with the case of the front-face illumination (Figure 8). This strongly supports the photosensitivity of only the topmost layer. The photoexcited carriers within the other layers are extinct via recombination. Moreover, the increase of the distance between the topmost layer and ITO electrode decreased the IPCE value, as shown in Figure 7. These behaviors may be compatible with the process of the photocurrent generation discussed above for the monolayer film. This may also provide a reasonable explanation why no photocurrent has been observed from the bulk materials of manganese oxide.28 Conclusions The MnO2 nanosheet electrode generates anodic photocurrent under visible light irradiation without dissolution itself. The twodimensional ultrathin oxide system may account for the charge
Photocurrent Generation from MnO2 Nanosheets
Figure 8. Photocurrent action spectra for 10-layer film of MnO2 nanosheet electrodes. The direction of the incident light is from (a) the front and (b) the back face of the electrode.
separation of excited electrons and holes involving strongly localized d-d transitions. Combination with electron donors or acceptors sandwiched by the layers of MnO2 nanosheets may enhance the quantum efficiency of photocurrent generation in multilayer film of MnO2 nanosheet electrodes. The findings here should be of significant interest from a fundamental viewpoint as a new effect of nanoscale materials, which may lead to utilization of localized d electrons for photoinduced electrontransfer reactions, and be of great importance from technology as a new possibility in photoelectric conversion involving visible light harvesting, which may lead to achievement of extremely thin film solar cells with two-dimensional ultrathin nanosheets. Furthermore, the electronic diagram obtained here may shed light on the mechanism of photocatalytic reactions with manganese oxides.29,30 Acknowledgment. This work was supported by CREST of Japan Science and Technology Agency (JST). We thank Dr. K. Fukuda and Prof. I. Nakai of Science University of Tokyo for the in-plane XRD analysis. References and Notes (1) Fujishima, A.; Honda, K. Nature 1972, 238, 37. (2) Bard, A. J. Science 1980, 207, 139.
J. Phys. Chem. B, Vol. 109, No. 19, 2005 9655 (3) Bard, A. J. J. Phys. Chem. 1982, 86, 172. (4) Gerischer, H. Electrochim. Acta 1990, 35, 1677. (5) Tryk, D. A.; Fujishima, A.; Honda, K. Electrochim. Acta 2000, 45, 2363. (6) Memming, R. Semiconductor Electrochemistry; Wiley-VCH: Weinheim, 2001. (7) Martin, S. T.; Herrmann, H.; Choi, W.; Hoffmann, M. R. J. Chem. Soc., Faraday Trans. 1994, 90, 3315. (8) Martin, S. T.; Herrmann, H.; Hoffmann, M. R. J. Chem. Soc., Faraday Trans. 1994, 90, 3323. (9) Bahnemann, D. W.; Hilgendorff, M.; Memming, R. J. Phys. Chem. B 1997, 101, 4265. (10) Ohko, Y.; Hashimoto, K.; Fujishima, A. J. Phys. Chem. A 1997, 101, 8057. (11) O’Regan, B.; Gra¨tzel, M. Nature 1991, 353, 737. (12) Gra¨tzel, M. Nature 2001, 414, 338. (13) Asahi, R.; Morikawa, T.; Ohwaki, T.; Aoki, K.; Taga, Y. Science 2001, 293, 269. (14) Dare-Edwards, M. P.; Goodenough, J. B.; Hamnett, A.; Nicholson, N. D. J. Chem. Soc., Faraday Trans. 2 1981, 77, 643. (15) Sasaki, T.; Watanabe, M.; Hashizume, H.; Yamada, H.; Nakazawa, H. J. Am. Chem. Soc. 1996, 118, 8329. (16) Sasaki, T.; Watanabe, M. J. Am. Chem. Soc. 1998, 120, 4682. (17) Omomo, Y.; Sasaki, T.; Wang, L. Z.; Watanabe, M. J. Am. Chem. Soc. 2003, 125, 3568. (18) Yang, X.; Makita, Y.; Liu, Z.-h.; Sakane, K.; Ooi, K. Chem. Mater. 2004, 16, 5581. (19) Schaak, R. E.; Mallouk, T. E. Chem. Mater. 2002, 14, 1455. (20) Sakai, N.; Ebina, Y.; Takada, K.; Sasaki, T. J. Am. Chem. Soc. 2004, 126, 5851. (21) Wang, L. Z.; Omomo, Y.; Sakai, N.; Fukuda, K.; Nakai, I.; Ebina, Y.; Takada, K.; Watanabe, M.; Sasaki, T. Chem. Mater. 2003, 15, 2873. (22) Feng, Q.; Kanoh, H.; Ooi, K. J. Mater. Chem. 1999, 9, 319. (23) Kavan, L.; Gra¨tzel, M.; Gilbert, S. E.; Klemenz, C.; Scheel, H. J. J. Am. Chem. Soc. 1996, 118, 6716. (24) Nozik, A. J.; Memming, R. J. Phys. Chem. 1996, 100, 13061. (25) Miller, B.; Heller, A. Nature 1976, 262, 680. (26) Ellis, A. B.; Kaiser, S. W.; Bolts, J. M.; Wrighton, M. S. J. Am. Chem. Soc. 1977, 99, 2839. (27) Ellis, A. B.; Bolts, J. M.; Kaiser, S. W.; Wrighton, M. S. J. Am. Chem. Soc. 1977, 99, 2848. (28) Hardee, K. L.; Bard, A. J. J. Electrochem. Soc. 1977, 124, 215. (29) Cao, H.; Suib, S. L. J. Am. Chem. Soc. 1994, 116, 5334. (30) Segal, S. R.; Suib, S. L.; Tang, X.; Satyapal, S. Chem. Mater. 1999, 11, 1687.
