Anodic Etching of n-GaN Epilayer into Porous GaN and Its

Anodic Etching of n-GaN Epilayer into Porous GaN and Its Photoelectrochemical Properties ... Publication Date (Web): November 26, 2014. Copyright © 2...
0 downloads 0 Views 2MB Size
Subscriber access provided by University of Washington | Libraries

Article

Anodic Etching of n-GaN Epilayer into Porous GaN and its Photoelectrochemical Properties Wei-Jhih Tseng, Dennis Henri van Dorp, Ruben R Lieten, Philippe M Vereecken, and Gustaaf Borghs J. Phys. Chem. C, Just Accepted Manuscript • DOI: 10.1021/jp508314q • Publication Date (Web): 26 Nov 2014 Downloaded from http://pubs.acs.org on December 2, 2014

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

The Journal of Physical Chemistry C is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 17

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

Anodic Etching of n-GaN Epilayer into Porous GaN and its Photoelectrochemical Properties Wei-Jhih Tseng, *,†,‡ D. H. van Dorp,† R. R. Lieten,† P. M. Vereecken,†,§ and G. Borghs‡ †

imec, Kapeldreef 75, 3001 Leuven, Belgium



Department of Physics and Astronomy, K.U. Leuven, 3001 Leuven, Belgium

§

Centre of Surface Chemistry and Catalysis, K.U. Leuven, 3001 Leuven, Belgium

ABSTRACT: Porous n-GaN has been fabricated using electrochemical anodic etching in a 0.5 M H2SO4 solution in the dark for different biases (5.5 V- 18.0 V). The pore morphology of the porous GaN shows distinctive differences: from narrow branching pores to wide parallel pores for increasing applied bias. The pore formation process has been investigated using cyclic voltammetry and chronoamperometry. The photoelectrochemical properties of these porous n-GaN layers have been examined. For the porous GaN etched at 5.5-15.0 V, the plateau photocurrent increases over 4 times and the potential difference between the current onset and the plateau is reduced by 0.24 V with respect to unetched, planar n-GaN.

KEYWORDS: photoelectrochemistry; water splitting; porous GaN; photoelectrolysis INTRODUCTION Hydrogen is a potential candidate to replace fossil energy sources in the coming future as a clean, renewable energy carrier. Direct water electrolysis using solar energy is an attractive method for hydrogen production considering energy conversion efficiency and system simplicity. Since Fujishima and Honda demonstrated direct photoelectrolysis using a TiO2 photoanode in the 1970s,1 photoelectrochemical and photocatalytic properties of many semiconductors have been investigated extensively.2,3 GaN-based semiconductors have drawn much attention for the use in photoelectrolysis.4-8 Self-driven photoelectrolysis is possible using GaN for its favorable energetic alignment with the redox potentials of water.4 After alloying with indium, the bandgap of the semiconductor can cover the complete solar spectrum (0.7-3.4 eV).9 Surface nanostructuring of photocatalysts can improve charge separation efficiency. The migration distance for photocarriers to the semiconductor/solution interface is reduced in nanostructures. Carrier recombination in the bulk is therefore suppressed. Nanostructuring is particularly important for semiconductors with low absorption coefficient or short carrier diffusion length such as GaP,10 GaAs11 and most metal oxides.12 Although n-type GaN has a high absorption coefficient (105 cm-1) for supra bandgap light13 and superior minority carrier diffusion length (200-300 nm),14 recent studies have shown that the photoactivity of n-GaN can be significantly enhanced after surface roughening.5-7 Surface roughening using wet etching techniques is desirable due to process simplicity and low costs. However, the superior chemical stability of the GaN (0001) plane, which is the predominant plane for GaN epitaxy, prevents etching by most chemicals.15 To date, only the photoelectrochemical properties of n-GaN etched by hot concentrated H3PO4 have

been reported.8 Light-assisted electrochemical etching is a well-developed technique to roughen GaN. n-GaN is oxidized by photo-holes and subsequently dissolves in alkaline or acid solution. However, due to the higher carrier recombination rate, the density of photo-holes at the defect sites is less than that at the high quality GaN region. Consequently, in the photoelectrochemical etching, the etching rate of the high quality defect-free n-GaN is relatively faster, resulting in the deteriorated material quality after the etching.16 Recently, porosification of n-GaN using electrochemical anodic etching has been developed, and is mainly motivated by the need for a strain-free template.17,18 The surface morphologies of porous n-GaN etched in different solutions, such as HNO3,17 HF19 and H2C2O420, have been reported. Furthermore, the correlation between dopant density of n-GaN and pore morphology has been established.21 In this work, we use cyclic voltammetry, chronoamperometry and SEM measurements to discuss the mechanisms of anodic etching for n-GaN. Furthermore, the photoelectrochemical properties of porous n-GaN etched at different biases have been examined. The plateau photocurrents for all the porous GaN increase over a factor of 4 with respect to unetched n-GaN. The photoelectrochemical properties of the porous GaN is optimized when the anodic etching is a mixture of branching-pore etch and vertical-pore etch. EXPERIMENTAL SECTION A Ga-polar GaN (0001) layer was epitaxially grown on a Si (111) substrate by metal organic chemical vapor deposition (MOCVD).22 The top 1.7 μm of the n-GaN layer was highly Si doped, followed by the unintentionally-doped (UID) GaN and AlGaN buffer layers. The electron concentrations of the n-GaN and UID GaN layers are respectively

ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

6×1018 cm-3 and 5×1016 cm-3, assessed by Hall effect measurements. The full width at half maximum of the GaN X-ray diffraction measurement (Panalytical, Philips) are 350 and 550 arc second for the (0002) and (101̅ 2) reflection, respectively. Field-emission scanning electron microscopy (Nova, NanoSEM) was used to assess the surface morphology. The Ohmic contact to the n-GaN was formed by a metal stack deposition of (GaN/)Ti/Al/Mo/Au at the sample edge, alloyed at 700℃ for 1 minute in an N2 environment. Subsequently, the GaN sample was mounted on a copper stripecontaining glass slide using a conductive silver paste (Pelco 187, Ted Pella). PVC tape and epoxy (Torr Seal, Varian) were used to define the working area of the electrode and to shield other components from exposure to the solution. The anodic etching was performed in the dark at room temperature using a two electrode configuration. The working area of the n-GaN electrode and counter electrode (Pt mesh) are 0.25-0.5 cm2 and 4 cm2, respectively. A potentiostat (Autolab, Metrohm) was used for current-potential control. A 0.5 M H2SO4 (Sigma Aldrich) aqueous solution was used for the anodic etching. The solution was stirred using a magnetic stirring rod during the etching to remove the bubbles generated at the n-GaN surface. The photoelectrochemical measurements were performed in a UV-transparent quartz cell using a three-electrode configuration with an Ag/AgCl reference electrode (Metrohm). A 0.1 M HBr and 0.2 M Na2SO4 (Sigma Aldrich) solution was used, with Br- acting as a sacrificial reducing agent to avoid photo-corrosion of GaN during photoelectrolysis.23 The solution was stirred using a magnetic stirring rod during all photoelectrochemical measurements. A 150 W Xe lamp (LSB521, L.O.T.-oriel) was used as the light source for the photoelectrochemical measurements. An air mass 1.5 filter (LSZ289, L.O.T.-oriel) was applied to filter 97% of the UV light (λ < 400 nm) to reduce photo-corrosion of GaN during photoelectrolysis. The light intensity projected at the sample surface was 63 mW/cm2, measured by a thermopile detector (s302c, Thorlabs). During the anodic characterizations and etching, the external potential (Uetch) is biased between GaN and the counter electrode (Pt). For the photoelectrochemical studies, the potential (U) is biased between GaN and the reference electrode. RESULTS A. Cyclic voltammograms and chronoamperograms. Figure 1 shows cyclic voltammograms of n-GaN in the dark in a 0.5 M H2SO4 solution at the scan rates (v) of 5 V/s and 1 V/s. For Uetch < 4 V, current density is lower than 20 μA/cm2. A Schottky junction forms at the n-GaN/solution interface. No holes are available in the dark for the charge transfer. For Uetch > flat-band potential, electrons are depleted near the surface, leading to space charge region (SCR). For Uetch > 4 V, the anodic current arises, and porous etching of n-GaN is observed for Uetch ≥ 5.5 V, as shown later. For both scan rates, the current density increases monotonically with increasing potential in the forward. A kink is observed at Uetch ~ 10 V for v = 5 V/s and at ~8 V for v = 1 V/s. Once the bias is reversed, the current drops rapidly for

Page 2 of 17

Figure 1. Cyclic voltammograms for n-GaN in the dark in a 0.5 M H2SO4 solution, measured at the scan rates (v) of 1 V/s and 5V/s.

both scan rates. For v = 5 V/s, the current decreases continuously in the reverse scan. For v = 1 V/s, the current remains approximately the same for 8 V < Uetch < 14 V. Chronoamperometric experiments, as shown in Figure 2a, were performed as a complementary study of the cyclic voltammograms. Two potentiostatic etches, which are indicated as the 1st and 2nd etching, were sequentially performed at the same n-GaN sample at Uetch = 15 V. The integrated charge for the first and second etch are 1.23 C/cm 2 and 1.36 C/cm2, respectively. The porosification of the full n-GaN layer (1.7 μm) requires 4.59 C/cm2. A rapid current drop is observed at the first 15 s. The current then rises to a current peak (15-30 s), followed by a continuous current increase (t > 60 s). Three cyclic voltammograms performed before and after these two potentiostatic etches are shown in Figure 2b. The first, second and third cyclic voltammogram were measured when the pores are ~0.6 μm, ~1.0 μm

Figure 2. (a) Current density-time (j-t) plots for an n-GaN sample etched at an applied potential of 15 V for 160 seconds. (b) Cyclic voltammograms of n-GaN before and after the first and second potentiostatic etch. Solution: 0.5 M H2SO4; v = 5 V/s.

ACS Paragon Plus Environment

Page 3 of 17

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

Figure 3. (a) j-t plots of n-GaN etched in the 0.5 M H2SO4 solution at different Uetch in the dark. (b) The steady-state current density and (c) the integrated charge density of the potentiostatic etch as a function of Uetch.

and ~1.5 μm in depth, estimated from the integrated charges. The three current density-potential (j-U) characteristics are comparable, and are independent of the pore depth. Figure 3a shows the j-t plots of the potentiostatic etch at different Uetch. The sudden current drop at the end of each curve happens when the 1.7 μm n-GaN layer is completely porosified. Anodic etching of the succeeding UID GaN layer is not possible. The etching time for the full layer decreases when Uetch increases as a result of increasing etching rate. For Uetch ≥ 10.0 V, a rapid current drop is observed for the first 15 s, followed by a few oscillations until the current reaches a steady state. For the etching at Uetch ≤ 8.0 V, neither the current drop at the beginning nor the strong current oscillations are observed. Figure 3b shows the steady-state current density (jss) as a function of Uetch, obtained from Figure 3a. The determination of the jss is described in the supporting information (Figure S1-Figure S3). The etching experiments at Uetch = 18.0 V and 15.0 V were repeated for 3 and 5 times, respectively. The experiments using Uetch ≤ 13.0 V were repeated twice for each Uetch. For Uetch ≤ 13.0 V, the error is smaller than the size of the symbol in the figure. The jss-Uetch curve consists of three segments, which are separated at ~ 8 V and ~11.5 V. Figure 3c shows the integrated charge density (Q) for each potentiostatic etch. The error of each data point is all smaller than the symbol size in the figure. The increase of the integrated charge shows that the porosity of the GaN

film increases with increasing Uetch. For anodic etching, 3 holes are required to oxidize a nitride of GaN (oxidation state: -3 to 0). The oxidation of the 1.7 μm n-GaN layer therefore requires 3.6 C/cm2 of charge, indicated by a dashed line in Figure 3c. B. Morphology of porous GaN. The morphologies of the porous GaN were assessed by SEM measurements. Figure 4 shows the top-view SEM images for the porous GaN etched at 5.5 V, 8.0 V, 11.5 V and 15.0 V. The surface pore density and the pore size increase when Uetch increases. For Uetch = 5.5 V and 8.0 V, the density of the surface pores are 2.9×109 cm-2 and 2.6×1010 cm-2, respectively. For Uetch ≤ 8.0 V, the surface pores are isolated and distributed non-uniformly. For Uetch ≥ 11.5 V, the pores become so close that some of the pores are merged. At Uetch = 15.0 V, shallow pores (grey) exist next to the deep pores (black). Figure 5 shows 45˚ SEM images of the 1.7 μm porous nGaN layer etched at different Uetch. The pore growth stops at the UID GaN layer, which corresponds to the current drop at the end of each j-t plot (Figure 3a). Three different types of pores are observed. For Uetch = 5.5 V (Figure 5a), this pore morphology is characterized by the vertical pores with the side branching pores. The pores of n-GaN for 8.0V ≤ Uetch ≤ 11.5 V are branching-like (Figure 5b and 5c); the pores for Uetch ≤ 15.0 V are vertical (Figure 5d). The dimension of the GaN pores increases with increasing Uetch. The morphologies of pores for Uetch = 13.0 V and 18.0 V are shown in the supporting information (Figure S4). For Uetch ≥ 11.5 V, a transition layer is observed at the first ~100 nm from the surface of the GaN film (Figure 5c and 5d). Figure 6 shows a top-view SEM image of the porous GaN (Uetch = 15.0 V) with and without the transition layer. The surface transition layer is characterized by a higher

Figure 4. Top-view SEM images of the porous n-GaN etched at different potentials.

ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 17

Figure 7. Current density-potential (j-U) plots for unetched n-GaN and the 1.7 μm porous n-GaN samples etched at different potentials under illumination. The current density of unetched GaN is multiplied by three for easier comparison. Solution: 0.1 M HBr + 0.2 M Na2SO4; v: 100 mV/s. Figure 5. 45˚ SEM images of the porous n-GaN etched at different potentials.

pore density and smaller pore size with respect to the bulk pore. C. Photoelectrochemical properties of porous GaN. The photoelectrochemical properties of the porous n-GaN etched at different Uetch (5.5 V -15.0 V) are shown in Figure 7. The j-U plot of unetched GaN is also included in the figure and is multiplied by three for easier comparison. A 0.1 M HBr + 0.2 M Na2SO4 solution was used during the measurement. Table 1 lists the photoelectrochemical properties of these n-GaN photoanodes including the onset potential (Uonset), the potential difference between the current onset and the current plateau (∆Uonset-plateau) and the plateau current density (jplateau). The definitions of Uonset, Uplateau, ∆Uonset-plateau and jplateau are described in the footnotes of Table 1. Upon illumination, GaN absorbs supra-bandgap light and generates photocarriers. For U < Uonset, electrons and holes recombine directly or via defects. For Uonset < U < Uplateau, some of photocarriers are separated by the electric field of SCR and are used to oxidize Br-. For U > Uplateau, carrier recombi-

Figure 6. A top-view SEM image of the porous n-GaN (Uetch = 15.0 V) with (right) and without (left) a surface transition layer.

nation is suppressed; all photo-holes in the SCR are separated and contribute to the photocurrent. The Uonset of the porous GaN are all more positive than unetched GaN and shift toward negative potential with increasing Uetch. The ∆Uonset-plateau for the porous GaN are independent of the anodic etching bias, and are all smaller by ~0.24 V than that of unetched GaN. The plateau photocurrent of n-GaN increases over 400% after porosification. The highest plateau current is found at Uetch = 11.5 V, which is higher than unetched GaN by 470%. DISCUSSION A. Etching mechanism As commonly concluded in Si and III-V materials24, some important steps for anodic etching of n-GaN are (i) hole generation at the semiconductor/solution interface, (ii) oxidation of semiconductor surface, (iii) oxide dissolution and (v) the transport of dissolved ions away from the surface.21 Table 1. Photoelectrochemical properties of the planar n-GaN and the porous GaN. Uetch (V)

Uonseta (V)

∆Uonset-plateaub (V)

jplateaua (mA/cm2)

planar 5.5 8 11.5 15

-0.447 -0.248 -0.285 -0.298 -0.314

0.560 0.322 0.327 0.321 0.320

0.025 0.127 0.126 0.143 0.127

aIn the plateau region, the j-U relationship is a linear line. The plateau current density (jplateau) and the plateau potential (Uplateau) are reached when the current density deviates this linearity by 5 %. The onset potential (Uonset) is reached when the current density reach 5 % of the jplateau. b∆U

onset-plateau

= Uplateau - Uonset

ACS Paragon Plus Environment

Page 5 of 17

The Journal of Physical Chemistry 1

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

H2 O + 2h+ → O2 + 2H + 2

Figure 8. (a) A band diagram of a GaN/solution interface with a strong electric field in space charge region so that electron tunneling occurs and (b) a schematic of pore formation during anodic etching where r is the local curvature radius at the tip of pores.

Zener tunneling and avalanche breakdown are two main mechanisms for hole supply in anodic etching. Zener tunneling happens, as illustrated in Figure 8a, when the electric field of SCR is so strong that electrons can tunnel directly from valence band (VB) to conduction band (CB).25 Therefore, this effect preferentially happens at a highly-doped semiconductor due to the thin SCR. Avalanche breakdown occurs when charge carriers possess sufficient kinetic energy, by acceleration in an electric field, to free electronhole pairs via collisions with bound electrons.25 It preferentially happens at a lowly doped semiconductor since a longer SCR is needed for sufficient acceleration of charges. Considering the high doping concentration of our samples, we believe that Zener tunneling is the dominant mechanism for hole generation during anodic etching. At a planar GaN/solution interface, the electric field of SCR (ESC) is determined by the potential drop across the SCR (USC) and the width of the SCR (WSC). However, for the nonplanar interface where the local curvature radius ( r)