Hydrogen generation through direct photoelectrolysis of water was studied using photoelectrochemical cells made of different facets of free-standing polar GaN system. To build the fundamental understanding at the differences of surface photochemistry afforded by the GaN and polar surfaces, we correlated the relationship between the surface structure and photoelectrochemical performance on the different polar facets. The photoelectrochemical measurements clearly revealed that the Ga-polar surface had a more negative onset potential relative to the N-polar surface due to the much negative flat-band potential. At more positive applied voltages, however, the N-polar surface yielded much higher photocurrent with conversion efficiency of 0.61% compared to that of 0.55% by using the Ga-polar surface. The reason could be attributed to the variation in the band structure of the different polar facets via Mott-Schottky analyses. Based on this work, understanding the facet effect on photoelectrochemical activity can provide a blueprint for the design of materials in solar hydrogen applications.
© 2013 Optical Society of America
In the course of developing new strategies to convert and store the energy from sunlight to chemical fuels, solar water splitting has recently attracted a lot of attention [1,2]. Over the past 4 decades, the development of photocatalysis has primarily focused upon large band gap metal oxides involving ions with filled or empty d-shell bonding configurations [3–5] and oxynitrides . Recently, the use of group III nitride semiconductors for water splitting has attracted considerable attention [7,8]. Due to the more negative potential of the nitrogen 2p orbital, compared to that of the oxygen 2p orbital, metal nitrides often possess a narrow band gap and can potentially encompass nearly the entire solar spectrum. Moreover, the inherent chemical stability of nitrides also favors their use in the harsh photocatalysis reaction environment . Particularly, GaN demonstrates considerable resistance to corrosion in many aqueous solutions  and its band edge potentials are situated in positions that allow for zero-bias hydrogen generation . Indeed, recent ab initio molecular dynamic simulations further showed that the overall water oxidation reaction at GaN surfaces can be energetically driven by photogenerated holes .
Considering that photocatalytic reactions take place on the surfaces of semiconductors, the exposed crystal facets play a critical role in determining the photocatalytic reactivity and efficiency. For example, Hsu et al. reported the polar facets of ZnO yielded three times increase in conversion efficiency than nonpolar-facets because of the high surface energy, spontaneous polarization and negative flat-band potential of a polar-orientated surface . Similarly, the polar GaN surfaces have also been proposed as being sufficiently accessible for photoelectrochemistry compared to the semipolar and nonpolar facets . Although the polarity-dependent photocatalytic activities in the GaN case has been intensively studied, the differences of surface photochemistry between Ga-polar and N-polar GaN is still not understood. For the wurtzite structure of GaN, there are two typical faces of the polar surface, which give rise to a microscopic spontaneous polarization as shown in Fig. 1(a) . The presence of the spontaneous polarization may induce the adsorption of positive or negative ions from electrolyte at polar surface, causing a partial compensating of polarization charges and altering the charge-transfer behavior. Owing to their peculiar nature, in this study, comparative studies to correlate the surface structure and photoelectrochemical performance on Ga-polar and N-polar surface are systematically investigated.
A free-standing GaN film was obtained commercially from Everlight Electronics Co., Ltd. (Taipei, Taiwan). The thickness of both GaN films is approximately 100 µm. Our Ga-face and N-face free-standing GaN films are all commercial bulk products, which was fabricated through hydride vapor phase epitaxy (HVPE) growth method. The HVPE system used in this study was the horizontal home-built quartz reactor with the rotating quartz susceptor. GaCl was supplied vertically, just over the surface of the susceptor. The growth temperature of 1030-1050°C, temperature of GaCl synthesis of 870°C, HCl flow in the range of 18-24 mls/min diluted in 500 mls/min of N2, NH3 flow of 1200 mls/min, and 3000 mls/min of N2 as a carrier gas were applied for runs of 8 to 15h. The growth rates observed for this geometry and the set of conditions varied from 100 to 200 μm /h. The different polarities of the two surfaces of the bulk sample were distinguished through the well-known reaction of the N-polar surface in KOH solutions . X-ray diffraction measurements confirm the polar face, showing the (002) and (004) peaks of the wurtzite structure (Fig. 1b).
Hall measurements indicate that free-standing GaN has a carrier density of about 2.4 x 1018 cm−3, whether measured from the Ga-polar or N-polar faces. The mobility was measured as 231 cm2V−1s−1 at the Ga-polar surface and 209 cm2V−1s−1 at the N-polar surface. Two 0.5 x 0.7 cm2 samples (Ga-polar and N-polar) were cut from the free standing GaN film for our study. In order to obtain the efficient current collection during the solar hydrogen gas production process, two strips of 0.1x1 cm of Ti(100 nm)/Au(20nm) were deposited at the edge of the samples. The electrodes were prepared by connecting Cu wires to the metal contacts with Ag glue. Afterwards, the metal area was covered with epoxy. The photoelectrochemical measurement was carried out in 1 M HCl under a 150-Watt Xe lamp light source at an intensity of 100 mW/cm2. A platinum counter electrode and a Ag/AgCl reference electrode under a Solartron 1470 E multichannel system were used for the measurements. The impedance spectra were measured from frequency range of 10 to 20 kHz in 1 M HCl solution with an applied potential ranging from 0.2 to 1.4 V.
3. Results and discussion
Figure 2(a) shows the photocurrent density versus potential curves of Ga-polar and N-polar of GaN photoelectrodes. Compared to the N-polar electrode, the Ga-polar electrode has a more negative onset potential, which is highly desirable for n-type semiconductors since the hydrogen gas generation is observed even at a potential that is much lower than the standard oxidation potential of water . Previous studies [17,18] show that under the experimental conditions, the anodic current is dominated by the oxidation of the Cl- in the working electrode, while the corresponding hydrogen gas production occurred at the cathodic platinum electrode. It is also shown that at higher applied voltages, about 0.4 V vs. Ag/AgCl, the photo-anodic current measured from the N-polar surface overtakes the photocurrent of the Ga-polar surface.
In order to assess the hydrogen gas generation efficiency, the photocurrents were measured in 1 M HCl using a two-electrode system, with a platinum sheet as the counter electrode as shown in the insert of Fig. 2(b). The photocurrent from the Ga-polar surface is higher than the N-polar surface at lower potential versus the counter electrode, VCE, but reverses at higher applied bias, similar to the profile we obtained from the three-electrode configuration. The saturated photocurrent is related to the maximum efficiency of charge-separation resulted from strong internal electric field via applied bias. At this time, the photocurrent cannot increase even more bias is further applied. This is why the decreasing efficiencies at high operating bias voltage. The surface morphology did not significantly change through SEM analysis. The projected efficiency values may be computed according to the formula:
where is the percent efficiency, is the current density, is the voltage for the electrode voltage difference of the hydrogen generation from the oxidation reaction, and is the intensity of light illuminating the sample. We used V for the standard Cl2/Cl- reduction potential since we used HCl solution . Apparently, the N-polar surface producesmuch higher conversion efficiency of 0.61% compared to that of 0.55% by using the Ga-polar surface.
In order to understand the observed phenomena, Mott-Schottky plot is applied to analyze the Ga- and N-polar systems. By using an equivalent circuit according to circuit B in reference 11, Mott-Schottky plot of capacitance versus bias voltage as shown in Fig. 3(a) can be obtained. The capacity for a semiconducting material to demonstrate photoactivity depends upon the band edge potentials and the resulting band bending profile of the material in solution. The Mott-Schottky equation relates the capacitance of the semiconductor to the carrier concentration (Nd) and the other constants such as the fundamental charge constant (e0), dielectric constant(), vacuum permittivity(), applied potential (Vapp), and the flatband potential():
The AC impedance measurements in the dark allow the estimation of the semiconductor capacitance in solution. Plotting 1/C2 versus Vapp allows the estimation of the flat band potential as the x-intercept. The Mott-Schottky results indicate that the flat band potential of the Ga-polar surface is −0.8 V vs. Ag/AgCl while that of the N-polar is −0.4 V. From the flat-band potentials, the band diagrams of the two polar surfaces can be constructed (Fig. 3(b)). The corresponding band edges are consistent with the literature report that the Ga-polar surface is about 0.3 ± 0.1 V more negative than the N-polar surface .
The difference in band-edge position greatly affects the onset potential of the two surfaces. At applied bias equal to the flat-band edge, no photocurrent results because of the absence of internal electric field that allows a more effective separation of the photogenerated charges. At applied biases more negative than the flat-band edge, the band bends up, electron accumulation occurs at the surface and substantial cathodic current is observed even in the absence of light. At applied biases more positive than the flat-band potential, the band bends downward, and in the presence of light, oxidative current can be observed. Hence it can be observed that the onset potential of Ga-polar surface is more negative than that of the N-polar. The 0.4 V-difference in the conduction band-edge potentials between the Ga-polar and the N-polar well agrees with the difference in their onset potentials. It should be noted, furthermore, that the onset potentials do not correspond to the exact positions of the band edges. For the Ga-polar electrode, the flat band is at −0.8V with the onset potential at −0.5 V. This is due to the overpotential of the system, which may be caused by the presence of surface states that is initially filled-up by the photogenerated charges. Certainly, more surface states could induce stronger surface band pinning effect. However, presently we cannot determine the amounts of surface states in both samples. The investigation on this issue is still in progress.
The relative photocurrents in photoelectrochemical behavior may be explained by the relative positions of the valence-band edges of the two surfaces and the barrier heights resulting from the different applied potentials. First, the barrier heights, and , dominate the rate of charge recombination (R). Calarco estimates that the recombination rate, is approximated by the exponential term ,
We therefore propose that these two factors play an important role in the photoelectrochemical behavior during the potential scan. At lower applied bias, the Ga-polar surface significantly experiences less recombination losses compared to the N-polar surface, and thus the photocurrent density of Ga-polar surface is much higher. While the applied potential is increased to more positive, the effect of valence band-edge becomes the dominating factor, and thus the N-polar surface would consequently yield much higher photocurrent density. This model assumes that other sensitive factors, such as carrier concentration and structural qualities, are quite similar for the two polar surfaces.
The photoelectrochemical behavior of the two different polar surfaces of n-GaN system was investigated. The Ga-polar surface was demonstrated to show a more negative onset potential compared to N-polar surface. At more positive applied voltages, however, the N-polar surface yielded much higher photocurrent with conversion efficiency of 0.61% compared to that of 0.55% by using the Ga-polar surface. A model was proposed to explain these observations through the variation of band structure at different polar facets. Based on this work, understanding the facet effect on photoelectrochemical activity can provide a blueprint for the design of materials in solar hydrogen applications. Moreover, as regards the decoration of sensitizer on GaN surface for improving visible light absorption, the facet effect can further play a significant role on the photo-conversion efficiency.
This work was supported by the National Natural Science Council, Ministry of Education, Taiwan, and AOARD under AFSOR, US. We gratefully thank NSC, IAMS, and NTU for financial support for this project.
References and links
1. Y. G. Lin, Y. K. Hsu, Y. C. Chen, L. C. Chen, S. Y. Chen, and K. H. Chen, “Visible-light-driven photocatalytic carbon-doped porous ZnO nanoarchitectures for solar water-splitting,” Nanoscale 4(20), 6515–6519 (2012). [CrossRef] [PubMed]
2. Y. G. Lin, Y. K. Hsu, Y. C. Chen, S. B. Wang, J. T. Miller, L. C. Chen, and K. H. Chen, “Plasmonic Ag@Ag3(PO4)1−x nanoparticle photosensitized ZnO nanorod-array photoanodes for water oxidation,” Energy Environ. Sci. 5(10), 8917–8922 (2012). [CrossRef]
3. H. Kato, K. Asakura, and A. Kudo, “Highly efficient water splitting into H2 and O2 over Lanthanum-doped NaTaO3 photocatalysts with high crystallinity and surface nanostructure,” J. Am. Chem. Soc. 125(10), 3082–3089 (2003). [CrossRef] [PubMed]
4. K. Domen, J. N. Kondo, M. Hara, and T. Takata, “Photo- and mechano-catalytic overall water splitting reactions to form hydrogen and oxygen on heterogeneous catalysts,” Bull. Chem. Soc. Jpn. 73(6), 1307–1331 (2000). [CrossRef]
5. H. Kadowaki, J. Sato, H. Kobayashi, N. Saito, H. Nishiyama, Y. Simodaira, and Y. Inoue, “Photocatalytic activity of the RuO2-dispersed composite p-block metal oxide LiInGeO4 with d10-d10 configuration for water decomposition,” J. Phys. Chem. B 109(48), 22995–23000 (2005). [CrossRef] [PubMed]
7. D. Wang, A. Pierre, M. G. Kibria, K. Cui, X. Han, K. H. Bevan, H. Guo, S. Paradis, A. R. Hakima, and Z. Mi, “Wafer-level photocatalytic water splitting on GaN nanowire arrays grown by molecular beam epitaxy,” Nano Lett. 11(6), 2353–2357 (2011). [CrossRef] [PubMed]
8. S. Y. Liu, J. K. Sheu, Y. C. Lin, S. J. Tu, F. W. Huang, M. L. Lee, and W. C. Lai, “Mn-doped GaN as photoelectrodes for the photoelectrolysis of water under visible light,” Opt. Express 20(S5Suppl 5), A678–A683 (2012). [CrossRef] [PubMed]
10. D. Zhuang and J. H. Edgar, “Wet etching of GaN, AlN, and SiC: a review,” Mater. Sci. Eng. Rep. 48(1), 1–46 (2005). [CrossRef]
11. J. D. Beach, R. T. Collins, and J. A. Turner, “Band-edge potentials of n-type and p-type GaN,” J. Electrochem. Soc. 150(7), A899–A904 (2003). [CrossRef]
12. X. Shen, Y. A. Small, J. Wang, P. B. Allen, M. V. Fernandez-Serra, M. S. Hybertsen, and J. T. Muckerman, “Photocatalytic water oxidation at the GaN (101̅0)−water interface,” J. Phys. Chem. C 114(32), 13695–13704 (2010). [CrossRef]
13. Y. K. Hsu, Y. G. Lin, and Y. C. Chen, “Polarity-dependent photoelectrochemical activity in ZnO nanostructures for solar water splitting,” Electrochem. Commun. 13(12), 1383–1386 (2011). [CrossRef]
14. K. Fujii, Y. Iwaki, H. Masui, T. J. Baker, M. Iza, H. Sato, J. Kaeding, T. Yao, J. S. Speck, S. P. Denbaars, S. Nakamura, and K. Ohkawa, “Photoelectrochemical Properties of Nonpolar and Semipolar GaN,” Jpn. J. Appl. Phys. 46(10A), 6573–6578 (2007). [CrossRef]
15. M. Stutzmann, O. Ambache, M. Eickhoff, U. Karrer, A. L. Pimenta, R. Neuberger, J. Schalwig, R. Dimitrov, P. J. Schuck, and R. D. Grober, “Playing with polarity,” Phys. Status Solidi 228(2), 505–512 (2001). [CrossRef]
17. M. Ono, K. Fujii, T. Ito, Y. Iwaki, A. Hirako, T. Yao, and K. Ohkawa, “Photoelectrochemical reaction and H2 generation at zero bias optimized by carrier concentration of n-type GaN,” J. Chem. Phys. 126(5), 054708 (2007). [CrossRef] [PubMed]
18. I. M. Huygens, A. Theuwis, W. P. Gomes, and K. Strubbe, “Photoelectrochemical reactions at the n-GaN electrode in 1 M H2SO4 and in acidic solutions containing Cl− ions,” Phys. Chem. Chem. Phys. 4(11), 2301–2306 (2002). [CrossRef]
19. B. J. Rodriquez, W. C. Yang, R. J. Nemanich, and A. Gruverman, “Scanning probe investigation of surface charge and surface potential of GaN-based heterostructures,” Appl. Phys. Lett. 86(11), 112115 (2005). [CrossRef]
20. R. Calarco, M. Marso, T. Richter, A. I. Aykanat, R. Meijers, A. V D Hart, T. Stoica, and H. Lüth, “Size-dependent Photoconductivity in MBE-Grown GaN-Nanowires,” Nano Lett. 5(5), 981–984 (2005). [CrossRef] [PubMed]