- Email: [email protected]
GaN nanocolumns fabricated by self-assembly Ni mask and its enhanced photocatalytic performance in water splitting
Accepted Manuscript Full Length Article GaN nanocolumns fabricated by self-assembly Ni mask and its enhanced photocatalytic performance in water splitting Xin Xi, Chao Yang, Haicheng Cao, Zhiguo Yu, Jin Li, Shan Lin, Zhanhong Ma, Lixia Zhao PII: DOI: Reference:
S0169-4332(18)32249-9 https://doi.org/10.1016/j.apsusc.2018.08.113 APSUSC 40150
To appear in:
Applied Surface Science
Received Date: Revised Date: Accepted Date:
26 March 2018 26 July 2018 13 August 2018
Please cite this article as: X. Xi, C. Yang, H. Cao, Z. Yu, J. Li, S. Lin, Z. Ma, L. Zhao, GaN nanocolumns fabricated by self-assembly Ni mask and its enhanced photocatalytic performance in water splitting, Applied Surface Science (2018), doi: https://doi.org/10.1016/j.apsusc.2018.08.113
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
GaN nanocolumns fabricated by self-assembly Ni mask and its enhanced photocatalytic performance in water splitting Xin Xi, Chao Yang, Haicheng Cao, Zhiguo Yu, Jin Li, Shan Lin, Zhanhong Ma, Lixia Zhao* †Semiconductor
Lighting
Research
and
Development
Center,
Institute
of
Semiconductors, Chinese Academy of Sciences, P. R. China ‡ College of Materials Science and Optoelectronic Technology, University of Chinese Academy of Sciences, Beijing 100049, People’s Republic of China *[email protected]
Highlights 1. GaN nanocolumns with different diameters were fabricated by Inductively coupled plasma using self-assembly Ni mask. 2. The absorption property of GaN nanocolumns was investigated by the PL spectrum and detailed mechanism was analyzed according to the data. 3. The GaN nanocolumns showed excellent photocatalytic activity. 4. A model was put forward to explain the relation between the photocatalytic property and the GaN diameters.
Keywords: GaN; nanocolumns; water splitting Abstract:
We investigate the influence of Gallium Nitride (GaN) nanocolumns on
the water splitting. The results show that with the GaN nanocolumns diameters decreasing from 250 to 170 nm, the photocurrents for the water splitting increase from 0.12 to 0.29 mA/cm2. The highest photocurrent is ~ 2.8 times higher than that of the planar GaN. The increase is mainly related to the surface states, which will lead to a band bending and help to separate the photo-generated carriers more effectively.
Furthermore, with the diameters decreasing, it will reduce the distance for the photo-generated carriers to the surface, which will suppress the recombination and increase the photocatalytic performance during the transport process. The work provides a cost-effective way for the applications of GaN nanocolumns in highly efficient water splitting.
Introduction Generation of hydrogen by photoelectrochemical (PEC) water splitting using solar energy has drawn extensive attention because of the requirements of clean and sustainable energy. Traditional semiconductors such as TiO2 and ZnO have received tremendous research because these powders can be obtained by chemical synthesis[1-4]. But these photocatalytic powders have low energy conversion efficiency, they can also be easily corrupted by electrolyte under light. III-V nitrides have recently attracted considerable attention in solar water splitting because of their advantages such as tunable band gap, high crystal quality, and material stability against photo corrosion[5-7]. The band gap of nitrides can be tuned from 3.4 to 0.7 eV by changing the Indium composition， which can help nitrides to absorb light over the entire solar spectrum[8]. Recent studies have also shown that GaN (energy bandgap 3.4 eV) has the thermodynamic and kinetic potentials for both water oxidation and proton reduction[9,10]. Furthermore, III-V nitrides can be grown epitaxially as single crystal, which is highly beneficial for efficient charge transfer. The strong ionic bonding characteristic of III-V nitrides can also decrease its chemical activity and prevent them from photo-corrosion[11-13]. The fabrications of GaN nanocolumns offer great potential to improve the water splitting efficiency over the planar counterparts. With the sidewalls on the nanocolumns, GaN has extremely large surface-to-volume ratio, which can significantly enhance light absorption[14]. Moreover, the distance for photo-generated carriers to semiconductor/solution interface can be effectively reduced by using the nanocolumns[15]. It will also decrease the recombination rate of the photo-generated
carriers to the surface. Furthermore, due to the band bending at the surface, the photo generated carriers can be separated efficiently and avoid the recombination during transport to the surface[16,17]. GaN nanocolumns can be fabricated by using either bottom-up or top-down technology. For the bottom-up methods, GaN nanostructures are synthesized by selective area epitaxial or liquid-solid growth[18-20]. During the growth, it is difficult to control the morphology and doping level of nanocolumns because of the unstable growth condition and crystal inherent properties. But for the top-down methods, nanocolumns can be fabricated by using Inductively Coupled Plasma (ICP) etching based on patterned GaN epilayers[21-23]. Because high quality GaN epilayers can be grown using stable and mature planar growth technology, during which the doping concentration can also be precisely controlled. Furthermore, the morphology of nanocolumns can be controlled by using well-established lithograph techniques with self-assembly metal mask[24,25]. Self-assemble Ni random mask is a cost-effective technique to produce the wafer-scale etch mask in which the diameters can be controlled by simply varying the thickness of the Ni deposition. Nanocolumns can improve the photocatalytic property due to its large surface-to-volume ratio. However, there are very few photoelectrochemical (PEC) studies on GaN nanocolumns fabricated using the top-down cost-effective approach. In this work, n-GaN nanocolumns were fabricated using top-down approach by ICP etching using self-assembled Ni mask. Compared to planar GaN, the nanocolumns exhibit an optimum enhancement of 3.8 times in energy conversion efficiency. Furthermore, the photocatalytic property increases with decreasing the diameters. A theoretic model has also been established. The surface state will lead to the band bending at the surface, which will accelerate the separation of photo-generated carriers and enhance the water splitting accordingly. The work provides a cost-effective way to fabricate the GaN nanocolumns for applications in highly efficient water splitting.
Experimental section Before the fabrication of the GaN nanocolumns, a 2 μm thick GaN epitaxial layer with 2.5×1018 carrier concentration was first grown on (0001) facet of sapphire wafer using metal organic chemical vapor? deposition (MOCVD). Figure 1 shows the schematic process to fabricate the GaN nanocolumns. First, 500 nm SiO2 mask was deposited on GaN by plasma enhanced chemical vapor deposition (PECVD). Then different thicknesses of Ni mask were grown above the SiO2 mask using e-beam evaporation (EB) by controlling the process time. After a Rapid Thermal Annealing (RTA) process under 800 ◦C for 3 minute, the Ni mask was self-assembled into nano-particles with random sizes. The patterns of the Ni nano-particles were transferred to SiO2 through ICP etching of SiO2. Afterwards, the GaN nanocolumns were fabricated by etching GaN under the SiO2 mask. The depth of nanorods prepared under each condition was ~1 μm. After the pattern transfer, the samples were dipped into fluorhydric acid and nitric acid subsequently to remove the remaining SiO2 and Ni nanoparticles. Finally, the GaN samples were prepared as photoanode with 0.5 1 cm2.
Figure 1. Illustration of the scheme for the fabrication of GaN nanocolumns using the top-down method.
The surface morphologies and crystal quality of the GaN nanocolumns was measured using scanning electron microscopy (SEM, Hitachi S-4800) and X-ray diffraction
(Empyrean) analysis. The chemical states of Ga and N ions in GaN were measured using X-ray photoelectron spectroscopy (XPS, Thermo ESCALAB 250 Xi). The optical properties of the nanocolumns were investigated by Photoluminescence (PL) measurements at room temperature using laser (325 nm) as the exiting source and optical reflectance spectra were recorded using a Cary 5000 UV-VIS-NIR spectrophotometer. PEC measurements of these photo electrodes were carried out under a Xe arc lamp (300W) with an intensity of 90 mW/cm2. The voltammogram was obtained using a CHI 660E Electrochemical workstation. The water splitting experiments were performed in a three-electrode quartz cell in 1 M NaOH electrolyte solution, consisting of the GaN nanocolumns working electrode, a Pt counter electrode and a Ag/AgCl reference electrode. H2 generation trial were performed using the same three electrodes set-up but with a different reaction solution and bias voltage. In the experiment, the bias voltage was 0.8 V and the solution was 0.1 mol Na2SO4.
Results and discussion Figure 2 shows the GaN nanocolumns with different diameters after removing the mask. The diameters of nanocolumns were controlled by deposition of the metal Ni using Electron Beam (EB) method. Different thicknesses of the metal Ni with 7, 11, 15 nm were deposited on the SiO2 to investigate the influence of metal thickness on the nanocolumns. Figures 2(a)-(c) show the top view SEM images of GaN nanocolumns with different Ni thicknesses and the distributions of the nanocolumns diameter. The statistic distribution of nanocolumns diameter is shown at the top right corner. The diameters of the nanocolumns increase from about 170 nm to 250 nm with the thickness of Ni mask increasing. The fluctuation of the diameters size is due to the instability during the RTA process. Figures 2(d)-(f) show the cross-sectional SEM images of GaN nanocolumns with 7, 11, 15 nm thickness of Ni mask. The depth is roughly about 1 μm. All the samples show a uniform length and relative smooth sidewalls irrespective of the ICP etching and different Ni mask thickness.
angle) (a)-(c) and cross-sectional SEM images (d)-(f) of GaN nanocolumns with increasing Ni mask thickness 7, 11, 15 nm.
Figure 3 (a) shows a typical XRD pattern of the planar GaN and GaN nanocolumns with diameters of 170, 200 and 250 nm. The two diffraction peaks correspond to the facets (0002) and (0004) of wurzite GaN structures. It suggests that the GaN grown on (0001) facet sapphire are oriented predominantly with the c-axis and formed with good single crystal quality. XPS was used to identify the elemental composition and chemical state of the GaN samples. Ga and N elements of the XPS spectra are shown in figure 3 (b), (c) and 3 (d). The shifts of peak position on the charge effect were calibrated using the binding energy of C 1s at 284.8 eV[26]. The XPS spectra of Ga were located at ~ 19.7 eV, which were referred as Ga 3d[27] and suggesting that the Gallium (Ga3+) was deposited in the form of Ga3+. However, the result of the peak fitting performed on N 1s for GaN samples shows three peaks, located at 391.17, 395.94 and 397.07 eV, respectively. The peak at 397.07eV can be assigned to Ga-N bond in GaN single crystal[26]. The peak at the 395.94 and 391.07eV are related to the surface Ga-N dangling bond and N-O bond for the O adsorption.
Figure 3. XRD pattern (a), XPS spectra for Ga 3d (b) and N 1s (c) of GaN nanocolumns with diameters of 170 nm, 200 nm, 250 nm and planar GaN, (b) XPS spectra fitting curve for N 1s.
Figure 4 (a) shows the room temperature PL spectrum ranging from 350 to 650 nm for the planar GaN and nanocolumns GaN. The planar GaN shows the strongest PL intensity among all samples. For the nanocolumns, the PL intensity decrease with the diameters decreasing. The PL intensity is normally related to the radical recombination of electrons and holes localized at the same sites. But because of the surface pinning effect, the holes are trapped at the surface which will result in a band bending. The band bending can separate the photo-generated electron-holes effectively and reduce the radical recombination severely. Although the radical recombination is reduced for GaN nanocolumns which result in low PL intensity, the prompt separations of photo-generated carriers will avoid the recombination during their transports to the surface. Figure 4(b) shows the detailed PL ranging from 350 to 395 nm, where the full-width-at-half-maximum (FWHM) becomes broader with the PL intensity increasing. It indicates that the holes at the surface have more opportunity to
recombine with the electrons closer to the core of nanocolumns. Because the nanocolumns with smaller diameters will have more severe band bending, so that the enhanced internal polarization electric field can separate the photo-generated carriers more obviously. The PL peaks also show a little blueshift for the nanocolumns because of the nonpolar surface appearing at sidewalls. During the water splitting, the surface band bending can separate the photo-generated carriers and transport the holes to the surface effectively. As a result, the electron-hole recombination will be prevented, which will increase the energy conversion efficiency accordingly.
Figure 4. (a) Room temperature photoluminesence (PL) of the as-fabricated GaN nanocolumns with diameter 170nm, 200nm, 250 nm and planr GaN, (b) The zoom-in PL of the planar GaN and GaN nanocolumns from 350 to 380 nm.
To understand the effect of GaN nanostructure on light absorption, diffusive reflectance measurements were carried out, as shown in figure 5. All GaN nanocolumns show an extremely lower reflectance with a little redshift (about 33 nm) compared with the planar GaN when the incident light is below 406 nm , which is mainly due to the strain relaxation for the GaN nanocolumns[28,29]. Meanwhile, the reflectance spectra of GaN nanocolumns slightly increases with the diameter increasing in the light spectrum below 365 nm. The reflectance decreases for GaN nanocolumns are mainly result from the light absorption effect[30]. For GaN nanocolumns, the exposed surface areas are much larger than that for planar GaN, which will lead to more light absorption at the GaN surface. For the nanocolumns with different diameters, the surface to volume ratio increase with the diameters decreasing.
Figure 5. Reflectance spectra of the GaN nanocolumns with diameter 170nm, 200nm, 250 nm and planar GaN.
PEC performance of the planar GaN and nanocolumns were evaluated as working electrode in a three-electrode cell using an electrochemical workstation. Figure 6 (a) shows the linear sweep voltammetry of all the electrodes for dark and light currents. The dark currents for all samples are significantly low, which is represented by the dark line. As a reference, the photocurrents of planar GaN are saturated above -0.2 V at ~0.075 mA/cm2. With the diameters of GaN nanocolumns increasing from 170 to 250 nm, the onset potentials are saturated from the voltage -0.35, -0.08, and 0 V at ~0.12, 0.2 and 0.29 mA/cm2, respectively. For the nanocolumns, the photocurrents show a strong enhancement compared to the planar sample.
Figure 6. (a) Linear sweep voltammetry of GaN nanocolumns with different diameters and planar GaN. (b) Time dependent photocurrent density of nanocolumns with different diameters and plane GaN under zero bias voltage.
In general, there is a difference of Femi levels between the bulk and surface due to the surface trapping of holes，which will result in the imbalance of diffusion for electrons and holes and result in a charge depletion region near the semiconductor surface. The charge depletion will lead to a band bending at the surface. Compared to the planar GaN, photo-generated carriers within the depletion region can be effectively separated due to the band bending. Furthermore, the nanocolumns structure can increase the interface
between
the
semiconductor
and
electrolyte
due
to
the
large
surface-to-volume ratio. The increased reaction area can reduce the accumulation of carriers inside GaN, and enhance the water splitting efficiency. In addition, the small diameters of nanocolumns can also suppress the recombination of photo-generated carriers, because the lateral transport can reduce the distance for the carriers to the GaN/electrolyte interface. Therefore, the charge separation efficiency can be increased significantly for the nanocolumns. Because the nanocolumns with smaller diameter have more surface areas and shorter distance to the GaN/electrolyte interface, therefore a larger water splitting efficiency. With the diameters decreasing, there is an elevation on saturated potential for the nanocolumns. This positive shift is attributed to the surface damage caused by ICP etching which act as traps for the photo-generated charge carriers. When light
illuminates on the surface of nanocolumns, the photo-generated charge carriers can be trapped in the defect energy level caused by the surface damage. In this case, more energy will be needed to extract the trapped photo-generated charge carriers, and increase the saturation potential.
Figure 7. (a) The theoretic H2 generation amounts for the planar GaN and GaN nanocolumns. (b) The actual H2 generation amounts for the planar GaN and GaN nanocolumns.
To evaluate the self-driven performance of water-splitting, time-dependent photocurrents experiments were carried out under zero bias, as shown in figure 6 (b). The highest photocurrent about can reach 0.29 mA/cm2 for the nanocolumns with the smallest diameter 170 nm, which
is about 2.8 times higher than the planar sample.
With increasing the diameters, the photocurrent density of GaN nanocolumns decreases from 0.2 to 0.12 mA/cm2, but it still shows a significant increase for the planar GaN in water splitting efficiency. The dependence of water splitting performance on the diameters of nanocolumns can be deducted by the Faraday’s law of electrolysis, as described by the following equation [31]. Mole of
2
(1)
Where F is the Faraday constant (96,485 C/mol), I is the measured current and t is the time. According to the equation, the H2 generation rates for the GaN nanocolumns
with diameters 170, 200, 250 nm and planar GaN are 5.41, 3.63, 2.07 and 1.64 μmol/h, as shown in the figure 7 (a). This indicates that water splitting efficiency is strongly improved by the large surface area, the shorten electron transportation and the enlarged depletion region result from the nanocolumns with thinner diameter. Figure 7 (b) shows the actual generation amount of H2 . Considering the gas dissolution in the solution and slight anodic oxidation, the amount is a little less than the theoretic value.
Figure 8. a) Illumination of band bending of the Fermi level at the GaN nanocolumns surface, b) Bend energy diagram from the surface to inner neutral zone.
To understand the relation between the diameter and photocurrent, a model calculation based on the Poisson equation was proposed as equation 2:
d 2(r)
dr
2
eN (r ) s 0
(2)
where the ϕ(r) is the surface potential difference from the surface to position x, e is the elementary charge, N(r) is the charge density at position r, ε is the electric constant of the GaN, and ε0 is the permittivity of vacuum. Because of Fermi-level pinning at the surface, the electronic bands, conduction (Ec) and valence (Ev), are bent upward at the surface of the GaN nanocolumns, as shown in figure 8(a). Supposing the complete width of the depletion area is precise d0/2, the d0/2 can be calculated from the
following equation according to the boundary condition as shown in figure 8(b).
d(d 0 / 2) 0 dx
(3)
ϕ(0) = 0
(4)
ϕ(d0/2) =Vs
(5)
The critical radius at which the nanocolumns start to be depleted can be calculated as follows:
d0
8 s 0Vs
eN r
( )
where Vs is the band difference between the surface and the bulk. For Nr is 2.5×1018, according to the theoretical model, the optimum diameter d0 is about 140 nm. When the diameter is greater than d0, an increase of diameter leads to complete depletion region aroused from surface band bending. With the diameter increasing continously, the ratio of surface state will decrease due to the increase of flat band in the nanocolumns. Therefore, the nanocolumns exhibit more excellent water splitting property with the nanocolumn diameters decreasing, which is in good agreement with the experiments.
Conclusion In summary, we have fabricated GaN nanocolumns by using top-down technology with self-assembly Ni as etch mask. Compared to planar GaN, the GaN nanocolumns show an enhancement of ~ 3.8 times on water splitting efficiency. The reason is mainly due to the more surface states in the nanocolumns GaN than planar GaN. The surface states can result in the band bending which separate the photo-generated carriers promptly. In addition, the nanocolumns can reduce the distance for carriers to the GaN surface. A theoretical model has been established to explain the correlation
between the surface state and diameters and find that the surface states increase with the reduction of the diameter which result in the high water splitting efficiency. ACKNOWLEDGMENTS： This work was supported by the National High Technology Program of China (2017YFB0403602 and 2017YFB0403601) and the National Natural Science Foundation of China (11574306 and 61774148).
Reference [1] F. Wang, W. Ge, T. Shen, B. Ye, Z. Fu, Y. Lu, Applied Surface Science, 410 (2017) 513-518. [2] J. Fang, H. Fan, Y. Ma, Z. Wang, Q. Chang, Applied Surface Science, 332 (2015) 47-54. [3] M. Wang, X. Xi, C. Gong, X.L. Zhang, G. Fan, Materials Research Bulletin, 74 (2016) 258-264. [4] S. Liu, H. Tang, H. Zhou, G. Dai, W. Wang, Applied Surface Science, 391 (2016) 542-547. [5] B. AlOtaibi, S. Fan, S. Vanka, M. Kibria, Z. Mi, Nano Letters, 15 (2015) 6821-6828. [6] S. Fan, B. AlOtaibi, S.Y. Woo, Y. Wang, G.A. Botton, Z. Mi, Nano letters, 15 (2015) 2721-2726. [7] M. Kibria, F. Chowdhury, S. Zhao, B. AlOtaibi, M. Trudeau, H. Guo, Z. Mi, Nature communications, 6 (2015) . [8] P.G. Moses, C.G.V.D. Walle, Applied Physics Letters, 96 (2010) 3675. [9] L. Caccamo, J. Hartmann, C. Fàbrega, S. Estradé, G. Lilienkamp, J.D. Prades, M.W.G. Hoffmann, J. Ledig, A. Wagner, X. Wang, Acs Applied Materials & Interfaces, 6 (2014) 2235. [10] M. Ebaid, J.H. Kang, S.W. Ryu, Journal of the Electrochemical Society, 162 (2015) 1-7. [11] M. Ebaid, J.H. Kang, S.H. Lim, J.S. Ha, J.K. Lee, Y.H. Cho, S.W. Ryu, Nano Energy, 12 (2015) 215-223. [12] J. Wallys, S. Hoffmann, F. Furtmayr, J. Teubert, M. Eickhoff, Nanotechnology, 23 (2012) 165701. [13] K. Aryal, B.N. Pantha, J. Li, J.Y. Lin, H.X. Jiang, Applied Physics Letters, 96 (2010) 37. [14] C. Yang, L. Liu, S. Zhu, Z. Yu, X. Xi, S. Wu, H. Cao, J. Li, L. Zhao, Journal of Physical Chemistry C, 121 (2017). [15] C. Yang, X. Xi, Z.G. Yu, H. Cao, J. Li, S. Lin, Z. Ma, L. Zhao, Acs Appl Mater Interfaces, (2018). [16] R. Calarco, M. Marso, T. Richter, A.I. Aykanat, R. Meijers, A. vd Hart, T. Stoica, H. Lüth, Nano letters, 5 (2005) 981-984. [17] T. Richter, H.L.R. Meijers, R. Calarco, M. Marso, Nano Letters, 8 (2008) 3056-3059. [18] W. Bergbauer, M. Strassburg, C. Koelper, N. Linder, C. Roder, J. Laehnemann, A. Trampert, S. Fuendling, S.F. Li, H.H. Wehmann, A. Waag, Nanotechnology, 21 (2010). [19] Y.T. Lin, T.W. Yeh, P.D. Dapkus, Nanotechnology, 23 (2012) 6. [20] S.D. Hersee, X. Sun, X. Wang, Nano Letters, 6 (2006) 1808-1811. [21] D. Paramanik, A. Motayed, G.S. Aluri, J.-Y. Ha, S. Krylyuk, A.V. Davydov, M. King, S. McLaughlin, S. Gupta, H. Cramer, Journal of Vacuum Science & Technology B, 30 (2012). [22] Y.D. Wang, S.J. Chua, S. Tripathy, M.S. Sander, P. Chen, C.G. Fonstad, Applied Physics Letters, 86 (2005) 3.
[23] Z. Zhuang, X. Guo, G. Zhang, B. Liu, R. Zhang, T. Zhi, T. Tao, H. Ge, F. Ren, Z. Xie, Y. Zheng, Nanotechnology, 24 (2013). [24] S.C. Zhu, Z.G. Yu, L.X. Zhao, J.X. Wang, J.M. Li, Opt. Express, 23 (2015) 13752-13760.
[25] Z.G. Yu, P. Chen, G.F. Yang, B. Liu, Z.L. Xie, X.Q. Xiu, Z.L. Wu, F. Xu, Z. Xu, X.M. Hua, P. Han, Y. Shi, R. Zhang, Y.D. Zheng, Chin. Phys. Lett., 29 (2012) 4. [26] X. Cao, L. Cao, W. Yao, X. Ye, Surf. Interf. Anal., 24 (1996) 662. [27] D.S. Li, M. Sumiya, S. Fuke, D.R. Yang, D.L. Que, Y. Suzuki, Y. Fukuda, Journal of Applied Physics, 90 (2001) 4219-4223. [28] H.J. Kim, J. Park, B.U. Ye, C.J. Yoo, J.L. Lee, S.W. Ryu, H. Lee, K.J. Choi, J.M. Baik, Acs Applied Materials & Interfaces, [Vajpeyi, 2005 #233]8 (2016) 18201. [29] A.P. Vajpeyi, S.J. Chua, S. Tripathy, E.A. Fitzgerald, Electrochem. Solid-State Lett., 8 (2005) G85-G88. [30] J.F. Muth, J.H. Lee, I.K. Shmagin, R.M. Kolbas, H.C.C. Jr, B.P. Keller, U.K. Mishra, S.P. Denbaars, Applied Physics Letters, 71 (1997) 2572-2574. [31] S.H. Kim, M. Ebaid, J.H. Kang, S.W. Ryu, Applied Surface Science, 305 (2014) 638-641.
S0169-4332(18)32249-9 https://doi.org/10.1016/j.apsusc.2018.08.113 APSUSC 40150
To appear in:
Applied Surface Science
Received Date: Revised Date: Accepted Date:
26 March 2018 26 July 2018 13 August 2018
Please cite this article as: X. Xi, C. Yang, H. Cao, Z. Yu, J. Li, S. Lin, Z. Ma, L. Zhao, GaN nanocolumns fabricated by self-assembly Ni mask and its enhanced photocatalytic performance in water splitting, Applied Surface Science (2018), doi: https://doi.org/10.1016/j.apsusc.2018.08.113
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
GaN nanocolumns fabricated by self-assembly Ni mask and its enhanced photocatalytic performance in water splitting Xin Xi, Chao Yang, Haicheng Cao, Zhiguo Yu, Jin Li, Shan Lin, Zhanhong Ma, Lixia Zhao* †Semiconductor
Lighting
Research
and
Development
Center,
Institute
of
Semiconductors, Chinese Academy of Sciences, P. R. China ‡ College of Materials Science and Optoelectronic Technology, University of Chinese Academy of Sciences, Beijing 100049, People’s Republic of China *[email protected]
Highlights 1. GaN nanocolumns with different diameters were fabricated by Inductively coupled plasma using self-assembly Ni mask. 2. The absorption property of GaN nanocolumns was investigated by the PL spectrum and detailed mechanism was analyzed according to the data. 3. The GaN nanocolumns showed excellent photocatalytic activity. 4. A model was put forward to explain the relation between the photocatalytic property and the GaN diameters.
Keywords: GaN; nanocolumns; water splitting Abstract:
We investigate the influence of Gallium Nitride (GaN) nanocolumns on
the water splitting. The results show that with the GaN nanocolumns diameters decreasing from 250 to 170 nm, the photocurrents for the water splitting increase from 0.12 to 0.29 mA/cm2. The highest photocurrent is ~ 2.8 times higher than that of the planar GaN. The increase is mainly related to the surface states, which will lead to a band bending and help to separate the photo-generated carriers more effectively.
Furthermore, with the diameters decreasing, it will reduce the distance for the photo-generated carriers to the surface, which will suppress the recombination and increase the photocatalytic performance during the transport process. The work provides a cost-effective way for the applications of GaN nanocolumns in highly efficient water splitting.
Introduction Generation of hydrogen by photoelectrochemical (PEC) water splitting using solar energy has drawn extensive attention because of the requirements of clean and sustainable energy. Traditional semiconductors such as TiO2 and ZnO have received tremendous research because these powders can be obtained by chemical synthesis[1-4]. But these photocatalytic powders have low energy conversion efficiency, they can also be easily corrupted by electrolyte under light. III-V nitrides have recently attracted considerable attention in solar water splitting because of their advantages such as tunable band gap, high crystal quality, and material stability against photo corrosion[5-7]. The band gap of nitrides can be tuned from 3.4 to 0.7 eV by changing the Indium composition， which can help nitrides to absorb light over the entire solar spectrum[8]. Recent studies have also shown that GaN (energy bandgap 3.4 eV) has the thermodynamic and kinetic potentials for both water oxidation and proton reduction[9,10]. Furthermore, III-V nitrides can be grown epitaxially as single crystal, which is highly beneficial for efficient charge transfer. The strong ionic bonding characteristic of III-V nitrides can also decrease its chemical activity and prevent them from photo-corrosion[11-13]. The fabrications of GaN nanocolumns offer great potential to improve the water splitting efficiency over the planar counterparts. With the sidewalls on the nanocolumns, GaN has extremely large surface-to-volume ratio, which can significantly enhance light absorption[14]. Moreover, the distance for photo-generated carriers to semiconductor/solution interface can be effectively reduced by using the nanocolumns[15]. It will also decrease the recombination rate of the photo-generated
carriers to the surface. Furthermore, due to the band bending at the surface, the photo generated carriers can be separated efficiently and avoid the recombination during transport to the surface[16,17]. GaN nanocolumns can be fabricated by using either bottom-up or top-down technology. For the bottom-up methods, GaN nanostructures are synthesized by selective area epitaxial or liquid-solid growth[18-20]. During the growth, it is difficult to control the morphology and doping level of nanocolumns because of the unstable growth condition and crystal inherent properties. But for the top-down methods, nanocolumns can be fabricated by using Inductively Coupled Plasma (ICP) etching based on patterned GaN epilayers[21-23]. Because high quality GaN epilayers can be grown using stable and mature planar growth technology, during which the doping concentration can also be precisely controlled. Furthermore, the morphology of nanocolumns can be controlled by using well-established lithograph techniques with self-assembly metal mask[24,25]. Self-assemble Ni random mask is a cost-effective technique to produce the wafer-scale etch mask in which the diameters can be controlled by simply varying the thickness of the Ni deposition. Nanocolumns can improve the photocatalytic property due to its large surface-to-volume ratio. However, there are very few photoelectrochemical (PEC) studies on GaN nanocolumns fabricated using the top-down cost-effective approach. In this work, n-GaN nanocolumns were fabricated using top-down approach by ICP etching using self-assembled Ni mask. Compared to planar GaN, the nanocolumns exhibit an optimum enhancement of 3.8 times in energy conversion efficiency. Furthermore, the photocatalytic property increases with decreasing the diameters. A theoretic model has also been established. The surface state will lead to the band bending at the surface, which will accelerate the separation of photo-generated carriers and enhance the water splitting accordingly. The work provides a cost-effective way to fabricate the GaN nanocolumns for applications in highly efficient water splitting.
Experimental section Before the fabrication of the GaN nanocolumns, a 2 μm thick GaN epitaxial layer with 2.5×1018 carrier concentration was first grown on (0001) facet of sapphire wafer using metal organic chemical vapor? deposition (MOCVD). Figure 1 shows the schematic process to fabricate the GaN nanocolumns. First, 500 nm SiO2 mask was deposited on GaN by plasma enhanced chemical vapor deposition (PECVD). Then different thicknesses of Ni mask were grown above the SiO2 mask using e-beam evaporation (EB) by controlling the process time. After a Rapid Thermal Annealing (RTA) process under 800 ◦C for 3 minute, the Ni mask was self-assembled into nano-particles with random sizes. The patterns of the Ni nano-particles were transferred to SiO2 through ICP etching of SiO2. Afterwards, the GaN nanocolumns were fabricated by etching GaN under the SiO2 mask. The depth of nanorods prepared under each condition was ~1 μm. After the pattern transfer, the samples were dipped into fluorhydric acid and nitric acid subsequently to remove the remaining SiO2 and Ni nanoparticles. Finally, the GaN samples were prepared as photoanode with 0.5 1 cm2.
Figure 1. Illustration of the scheme for the fabrication of GaN nanocolumns using the top-down method.
The surface morphologies and crystal quality of the GaN nanocolumns was measured using scanning electron microscopy (SEM, Hitachi S-4800) and X-ray diffraction
(Empyrean) analysis. The chemical states of Ga and N ions in GaN were measured using X-ray photoelectron spectroscopy (XPS, Thermo ESCALAB 250 Xi). The optical properties of the nanocolumns were investigated by Photoluminescence (PL) measurements at room temperature using laser (325 nm) as the exiting source and optical reflectance spectra were recorded using a Cary 5000 UV-VIS-NIR spectrophotometer. PEC measurements of these photo electrodes were carried out under a Xe arc lamp (300W) with an intensity of 90 mW/cm2. The voltammogram was obtained using a CHI 660E Electrochemical workstation. The water splitting experiments were performed in a three-electrode quartz cell in 1 M NaOH electrolyte solution, consisting of the GaN nanocolumns working electrode, a Pt counter electrode and a Ag/AgCl reference electrode. H2 generation trial were performed using the same three electrodes set-up but with a different reaction solution and bias voltage. In the experiment, the bias voltage was 0.8 V and the solution was 0.1 mol Na2SO4.
Results and discussion Figure 2 shows the GaN nanocolumns with different diameters after removing the mask. The diameters of nanocolumns were controlled by deposition of the metal Ni using Electron Beam (EB) method. Different thicknesses of the metal Ni with 7, 11, 15 nm were deposited on the SiO2 to investigate the influence of metal thickness on the nanocolumns. Figures 2(a)-(c) show the top view SEM images of GaN nanocolumns with different Ni thicknesses and the distributions of the nanocolumns diameter. The statistic distribution of nanocolumns diameter is shown at the top right corner. The diameters of the nanocolumns increase from about 170 nm to 250 nm with the thickness of Ni mask increasing. The fluctuation of the diameters size is due to the instability during the RTA process. Figures 2(d)-(f) show the cross-sectional SEM images of GaN nanocolumns with 7, 11, 15 nm thickness of Ni mask. The depth is roughly about 1 μm. All the samples show a uniform length and relative smooth sidewalls irrespective of the ICP etching and different Ni mask thickness.
angle) (a)-(c) and cross-sectional SEM images (d)-(f) of GaN nanocolumns with increasing Ni mask thickness 7, 11, 15 nm.
Figure 3 (a) shows a typical XRD pattern of the planar GaN and GaN nanocolumns with diameters of 170, 200 and 250 nm. The two diffraction peaks correspond to the facets (0002) and (0004) of wurzite GaN structures. It suggests that the GaN grown on (0001) facet sapphire are oriented predominantly with the c-axis and formed with good single crystal quality. XPS was used to identify the elemental composition and chemical state of the GaN samples. Ga and N elements of the XPS spectra are shown in figure 3 (b), (c) and 3 (d). The shifts of peak position on the charge effect were calibrated using the binding energy of C 1s at 284.8 eV[26]. The XPS spectra of Ga were located at ~ 19.7 eV, which were referred as Ga 3d[27] and suggesting that the Gallium (Ga3+) was deposited in the form of Ga3+. However, the result of the peak fitting performed on N 1s for GaN samples shows three peaks, located at 391.17, 395.94 and 397.07 eV, respectively. The peak at 397.07eV can be assigned to Ga-N bond in GaN single crystal[26]. The peak at the 395.94 and 391.07eV are related to the surface Ga-N dangling bond and N-O bond for the O adsorption.
Figure 3. XRD pattern (a), XPS spectra for Ga 3d (b) and N 1s (c) of GaN nanocolumns with diameters of 170 nm, 200 nm, 250 nm and planar GaN, (b) XPS spectra fitting curve for N 1s.
Figure 4 (a) shows the room temperature PL spectrum ranging from 350 to 650 nm for the planar GaN and nanocolumns GaN. The planar GaN shows the strongest PL intensity among all samples. For the nanocolumns, the PL intensity decrease with the diameters decreasing. The PL intensity is normally related to the radical recombination of electrons and holes localized at the same sites. But because of the surface pinning effect, the holes are trapped at the surface which will result in a band bending. The band bending can separate the photo-generated electron-holes effectively and reduce the radical recombination severely. Although the radical recombination is reduced for GaN nanocolumns which result in low PL intensity, the prompt separations of photo-generated carriers will avoid the recombination during their transports to the surface. Figure 4(b) shows the detailed PL ranging from 350 to 395 nm, where the full-width-at-half-maximum (FWHM) becomes broader with the PL intensity increasing. It indicates that the holes at the surface have more opportunity to
recombine with the electrons closer to the core of nanocolumns. Because the nanocolumns with smaller diameters will have more severe band bending, so that the enhanced internal polarization electric field can separate the photo-generated carriers more obviously. The PL peaks also show a little blueshift for the nanocolumns because of the nonpolar surface appearing at sidewalls. During the water splitting, the surface band bending can separate the photo-generated carriers and transport the holes to the surface effectively. As a result, the electron-hole recombination will be prevented, which will increase the energy conversion efficiency accordingly.
Figure 4. (a) Room temperature photoluminesence (PL) of the as-fabricated GaN nanocolumns with diameter 170nm, 200nm, 250 nm and planr GaN, (b) The zoom-in PL of the planar GaN and GaN nanocolumns from 350 to 380 nm.
To understand the effect of GaN nanostructure on light absorption, diffusive reflectance measurements were carried out, as shown in figure 5. All GaN nanocolumns show an extremely lower reflectance with a little redshift (about 33 nm) compared with the planar GaN when the incident light is below 406 nm , which is mainly due to the strain relaxation for the GaN nanocolumns[28,29]. Meanwhile, the reflectance spectra of GaN nanocolumns slightly increases with the diameter increasing in the light spectrum below 365 nm. The reflectance decreases for GaN nanocolumns are mainly result from the light absorption effect[30]. For GaN nanocolumns, the exposed surface areas are much larger than that for planar GaN, which will lead to more light absorption at the GaN surface. For the nanocolumns with different diameters, the surface to volume ratio increase with the diameters decreasing.
Figure 5. Reflectance spectra of the GaN nanocolumns with diameter 170nm, 200nm, 250 nm and planar GaN.
PEC performance of the planar GaN and nanocolumns were evaluated as working electrode in a three-electrode cell using an electrochemical workstation. Figure 6 (a) shows the linear sweep voltammetry of all the electrodes for dark and light currents. The dark currents for all samples are significantly low, which is represented by the dark line. As a reference, the photocurrents of planar GaN are saturated above -0.2 V at ~0.075 mA/cm2. With the diameters of GaN nanocolumns increasing from 170 to 250 nm, the onset potentials are saturated from the voltage -0.35, -0.08, and 0 V at ~0.12, 0.2 and 0.29 mA/cm2, respectively. For the nanocolumns, the photocurrents show a strong enhancement compared to the planar sample.
Figure 6. (a) Linear sweep voltammetry of GaN nanocolumns with different diameters and planar GaN. (b) Time dependent photocurrent density of nanocolumns with different diameters and plane GaN under zero bias voltage.
In general, there is a difference of Femi levels between the bulk and surface due to the surface trapping of holes，which will result in the imbalance of diffusion for electrons and holes and result in a charge depletion region near the semiconductor surface. The charge depletion will lead to a band bending at the surface. Compared to the planar GaN, photo-generated carriers within the depletion region can be effectively separated due to the band bending. Furthermore, the nanocolumns structure can increase the interface
between
the
semiconductor
and
electrolyte
due
to
the
large
surface-to-volume ratio. The increased reaction area can reduce the accumulation of carriers inside GaN, and enhance the water splitting efficiency. In addition, the small diameters of nanocolumns can also suppress the recombination of photo-generated carriers, because the lateral transport can reduce the distance for the carriers to the GaN/electrolyte interface. Therefore, the charge separation efficiency can be increased significantly for the nanocolumns. Because the nanocolumns with smaller diameter have more surface areas and shorter distance to the GaN/electrolyte interface, therefore a larger water splitting efficiency. With the diameters decreasing, there is an elevation on saturated potential for the nanocolumns. This positive shift is attributed to the surface damage caused by ICP etching which act as traps for the photo-generated charge carriers. When light
illuminates on the surface of nanocolumns, the photo-generated charge carriers can be trapped in the defect energy level caused by the surface damage. In this case, more energy will be needed to extract the trapped photo-generated charge carriers, and increase the saturation potential.
Figure 7. (a) The theoretic H2 generation amounts for the planar GaN and GaN nanocolumns. (b) The actual H2 generation amounts for the planar GaN and GaN nanocolumns.
To evaluate the self-driven performance of water-splitting, time-dependent photocurrents experiments were carried out under zero bias, as shown in figure 6 (b). The highest photocurrent about can reach 0.29 mA/cm2 for the nanocolumns with the smallest diameter 170 nm, which
is about 2.8 times higher than the planar sample.
With increasing the diameters, the photocurrent density of GaN nanocolumns decreases from 0.2 to 0.12 mA/cm2, but it still shows a significant increase for the planar GaN in water splitting efficiency. The dependence of water splitting performance on the diameters of nanocolumns can be deducted by the Faraday’s law of electrolysis, as described by the following equation [31]. Mole of
2
(1)
Where F is the Faraday constant (96,485 C/mol), I is the measured current and t is the time. According to the equation, the H2 generation rates for the GaN nanocolumns
with diameters 170, 200, 250 nm and planar GaN are 5.41, 3.63, 2.07 and 1.64 μmol/h, as shown in the figure 7 (a). This indicates that water splitting efficiency is strongly improved by the large surface area, the shorten electron transportation and the enlarged depletion region result from the nanocolumns with thinner diameter. Figure 7 (b) shows the actual generation amount of H2 . Considering the gas dissolution in the solution and slight anodic oxidation, the amount is a little less than the theoretic value.
Figure 8. a) Illumination of band bending of the Fermi level at the GaN nanocolumns surface, b) Bend energy diagram from the surface to inner neutral zone.
To understand the relation between the diameter and photocurrent, a model calculation based on the Poisson equation was proposed as equation 2:
d 2(r)
dr
2
eN (r ) s 0
(2)
where the ϕ(r) is the surface potential difference from the surface to position x, e is the elementary charge, N(r) is the charge density at position r, ε is the electric constant of the GaN, and ε0 is the permittivity of vacuum. Because of Fermi-level pinning at the surface, the electronic bands, conduction (Ec) and valence (Ev), are bent upward at the surface of the GaN nanocolumns, as shown in figure 8(a). Supposing the complete width of the depletion area is precise d0/2, the d0/2 can be calculated from the
following equation according to the boundary condition as shown in figure 8(b).
d(d 0 / 2) 0 dx
(3)
ϕ(0) = 0
(4)
ϕ(d0/2) =Vs
(5)
The critical radius at which the nanocolumns start to be depleted can be calculated as follows:
d0
8 s 0Vs
eN r
( )
where Vs is the band difference between the surface and the bulk. For Nr is 2.5×1018, according to the theoretical model, the optimum diameter d0 is about 140 nm. When the diameter is greater than d0, an increase of diameter leads to complete depletion region aroused from surface band bending. With the diameter increasing continously, the ratio of surface state will decrease due to the increase of flat band in the nanocolumns. Therefore, the nanocolumns exhibit more excellent water splitting property with the nanocolumn diameters decreasing, which is in good agreement with the experiments.
Conclusion In summary, we have fabricated GaN nanocolumns by using top-down technology with self-assembly Ni as etch mask. Compared to planar GaN, the GaN nanocolumns show an enhancement of ~ 3.8 times on water splitting efficiency. The reason is mainly due to the more surface states in the nanocolumns GaN than planar GaN. The surface states can result in the band bending which separate the photo-generated carriers promptly. In addition, the nanocolumns can reduce the distance for carriers to the GaN surface. A theoretical model has been established to explain the correlation
between the surface state and diameters and find that the surface states increase with the reduction of the diameter which result in the high water splitting efficiency. ACKNOWLEDGMENTS： This work was supported by the National High Technology Program of China (2017YFB0403602 and 2017YFB0403601) and the National Natural Science Foundation of China (11574306 and 61774148).
Reference [1] F. Wang, W. Ge, T. Shen, B. Ye, Z. Fu, Y. Lu, Applied Surface Science, 410 (2017) 513-518. [2] J. Fang, H. Fan, Y. Ma, Z. Wang, Q. Chang, Applied Surface Science, 332 (2015) 47-54. [3] M. Wang, X. Xi, C. Gong, X.L. Zhang, G. Fan, Materials Research Bulletin, 74 (2016) 258-264. [4] S. Liu, H. Tang, H. Zhou, G. Dai, W. Wang, Applied Surface Science, 391 (2016) 542-547. [5] B. AlOtaibi, S. Fan, S. Vanka, M. Kibria, Z. Mi, Nano Letters, 15 (2015) 6821-6828. [6] S. Fan, B. AlOtaibi, S.Y. Woo, Y. Wang, G.A. Botton, Z. Mi, Nano letters, 15 (2015) 2721-2726. [7] M. Kibria, F. Chowdhury, S. Zhao, B. AlOtaibi, M. Trudeau, H. Guo, Z. Mi, Nature communications, 6 (2015) . [8] P.G. Moses, C.G.V.D. Walle, Applied Physics Letters, 96 (2010) 3675. [9] L. Caccamo, J. Hartmann, C. Fàbrega, S. Estradé, G. Lilienkamp, J.D. Prades, M.W.G. Hoffmann, J. Ledig, A. Wagner, X. Wang, Acs Applied Materials & Interfaces, 6 (2014) 2235. [10] M. Ebaid, J.H. Kang, S.W. Ryu, Journal of the Electrochemical Society, 162 (2015) 1-7. [11] M. Ebaid, J.H. Kang, S.H. Lim, J.S. Ha, J.K. Lee, Y.H. Cho, S.W. Ryu, Nano Energy, 12 (2015) 215-223. [12] J. Wallys, S. Hoffmann, F. Furtmayr, J. Teubert, M. Eickhoff, Nanotechnology, 23 (2012) 165701. [13] K. Aryal, B.N. Pantha, J. Li, J.Y. Lin, H.X. Jiang, Applied Physics Letters, 96 (2010) 37. [14] C. Yang, L. Liu, S. Zhu, Z. Yu, X. Xi, S. Wu, H. Cao, J. Li, L. Zhao, Journal of Physical Chemistry C, 121 (2017). [15] C. Yang, X. Xi, Z.G. Yu, H. Cao, J. Li, S. Lin, Z. Ma, L. Zhao, Acs Appl Mater Interfaces, (2018). [16] R. Calarco, M. Marso, T. Richter, A.I. Aykanat, R. Meijers, A. vd Hart, T. Stoica, H. Lüth, Nano letters, 5 (2005) 981-984. [17] T. Richter, H.L.R. Meijers, R. Calarco, M. Marso, Nano Letters, 8 (2008) 3056-3059. [18] W. Bergbauer, M. Strassburg, C. Koelper, N. Linder, C. Roder, J. Laehnemann, A. Trampert, S. Fuendling, S.F. Li, H.H. Wehmann, A. Waag, Nanotechnology, 21 (2010). [19] Y.T. Lin, T.W. Yeh, P.D. Dapkus, Nanotechnology, 23 (2012) 6. [20] S.D. Hersee, X. Sun, X. Wang, Nano Letters, 6 (2006) 1808-1811. [21] D. Paramanik, A. Motayed, G.S. Aluri, J.-Y. Ha, S. Krylyuk, A.V. Davydov, M. King, S. McLaughlin, S. Gupta, H. Cramer, Journal of Vacuum Science & Technology B, 30 (2012). [22] Y.D. Wang, S.J. Chua, S. Tripathy, M.S. Sander, P. Chen, C.G. Fonstad, Applied Physics Letters, 86 (2005) 3.
[23] Z. Zhuang, X. Guo, G. Zhang, B. Liu, R. Zhang, T. Zhi, T. Tao, H. Ge, F. Ren, Z. Xie, Y. Zheng, Nanotechnology, 24 (2013). [24] S.C. Zhu, Z.G. Yu, L.X. Zhao, J.X. Wang, J.M. Li, Opt. Express, 23 (2015) 13752-13760.
[25] Z.G. Yu, P. Chen, G.F. Yang, B. Liu, Z.L. Xie, X.Q. Xiu, Z.L. Wu, F. Xu, Z. Xu, X.M. Hua, P. Han, Y. Shi, R. Zhang, Y.D. Zheng, Chin. Phys. Lett., 29 (2012) 4. [26] X. Cao, L. Cao, W. Yao, X. Ye, Surf. Interf. Anal., 24 (1996) 662. [27] D.S. Li, M. Sumiya, S. Fuke, D.R. Yang, D.L. Que, Y. Suzuki, Y. Fukuda, Journal of Applied Physics, 90 (2001) 4219-4223. [28] H.J. Kim, J. Park, B.U. Ye, C.J. Yoo, J.L. Lee, S.W. Ryu, H. Lee, K.J. Choi, J.M. Baik, Acs Applied Materials & Interfaces, [Vajpeyi, 2005 #233]8 (2016) 18201. [29] A.P. Vajpeyi, S.J. Chua, S. Tripathy, E.A. Fitzgerald, Electrochem. Solid-State Lett., 8 (2005) G85-G88. [30] J.F. Muth, J.H. Lee, I.K. Shmagin, R.M. Kolbas, H.C.C. Jr, B.P. Keller, U.K. Mishra, S.P. Denbaars, Applied Physics Letters, 71 (1997) 2572-2574. [31] S.H. Kim, M. Ebaid, J.H. Kang, S.W. Ryu, Applied Surface Science, 305 (2014) 638-641.