Formation and Device Application of III-V …...Formation and Device Application of III-V Semiconductor-based Functional Microstructures Achieved by Electrochemical Process （電気化学的制御に基づくIII-V

Formation and Device Application of III-V

Semiconductor-based Functional Microstructures

Achieved by Electrochemical Process

（電気化学的制御に基づく III-V族化合物半導体の微細加工と機能素子応用に関する研究）

A dissertation submitted in partial fulfillment of the requirement for

the degree of Doctor of Philosophy (Engineering)

to Hokkaido University

February, 2017

by

Yusuke KUMAZAKI

Graduate School of information Science and Technology,

Hokkaido University

Dissertation Supervisor

Professor Taketomo SATO

Doctoral Dissertation

HOKKAIDO

© 2017 Yusuke Kumazaki

U N I V E R S I T Y

i

Acknowledgements

This dissertation describes a research work carried out at Research Center for

Integrated Quantum Electronics (RCIQE) under the supervision of Prof. Taketomo Sato,

while I was the graduate student of the Graduate School of Information Science

Technology, Hokkaido University, from 2012 to 2017.

I would like to express my deepest gratitude to Prof. Taketomo Sato for giving

me the opportunity to work in RCIQE and for spending a lot of time on supporting my

research, including guidance of experiment, advices for my research plans, discussions

on experimental results, and instructive corrections of my research papers and

presentations. Thanks to his enthusiasm supports and encouragements, I was able to

have a invaluable experience that will be of great help to my future work.

I would like to acknowledge the chief examiner and the co-examiners of my

dissertation, Prof. Seiya Kasai, Prof. Tamotsu Hashizume, Prof. Junichi Motohisa, and

Prof. Masamichi Akazawa, for spending their precious time on reviewing my

dissertation and providing valuable suggestions that helped me to improve the quality of

dissertation. I would also like to acknowledge other members of my dissertation

committee, Prof. Akihisa Tomita, Prof. Akihiro Murayama, Prof. Tetsuya Asai, Prof.

Tetsuya Uemura, Prof. Masato Motomura, Prof. Hiroshi Hirata, Prof. Kazuhisa Sueoka,

Prof. Eiichi Sano, and Prof. Yasuo Takahashi, for reviewing my dissertation and giving

me valuable comments at dissertation defense.

I would like to acknowledge Prof. Takashi Fukui, Prof. Kanji Yoh, Prof.

Shinjiro Hara, and Prof. Katsuhiro Tomioka, for their guidance and supports. I could do

a productive research because they create and maintain a wonderful research

environment, RCIQE.

I am very grateful to the members of EC group, Prof. Zenji Yatabe, Mr.

Tomohito Kudo, Mr. Yudai Imai, Mr. Ryohei Jinbo, Mr. Fumiyasu Makino, Mr. Akio

Watanabe, Mr. Soichiro Matsuda, Mr. Masaaki Edamoto, Ms. Sayaka Ohmi, Mr.

Hirofumi Kida, Mr. Xiaoyi Zhang, Mr. Keisuke Ito, Mr. Satoru Matsumoto, Mr.

Keisuke Uemura, and Mr. Masachika Toguchi, for their countless supports, including

ii Acknowledgements

guidance of experiments, constructive discussions at weekly meetings, and kind

personal consultations.

I am also grateful to the members of AD group, Prof. Joel T. Asubar, Dr.

Cheng-Yu Hu, Dr. Kota Ohi, Dr. Yujin Hori, Dr. Maciej Matys, Dr. Roman Stoklas, Dr.

Yin Xiang, Mr. Yuki Nakano, Mr. Toru Muramatsu, Mr. Bo Gao, Mr. Masaki Sato, Mr.

Takayuki Tanaka, Mr. Sungsik Kim, Mr. Naoki Azumaishi, Mr. Ryohei Kuroda, Mr.

Yuri Imai, Mr. Ma Wang-Cheng, Mr. Satoru Nakano, Mr. Takuma Nakano, Mr. Naoya

Inoue, Mr. Yushi Abe, Mr. Shinya Inoue, Mr. Kenya Nishiguchi, Mr. Yutaka Senzaki, Mr.

Masahito Chiba, Mr. Kento Shirata, Mr. Ryo Wakamiya, Mr. Takeshi Ishigooka, Mr. Joji

Ohira, Mr. Yoichiro Odanagi, Mr. Kentaro Sasaki, Mr. Shoma Okamoto, Mr. Yuki Inden,

Ms. Konomi Masuda, Mr. Atsushi Seino, Mr. Shota Kaneki, Mr. Shota Toiya, Mr. Kenta

Saito, Mr. Koki Abe, Mr. Naoshige Yokota, Mr. Taito Hasezaki, Mr. Yuji Ando, Mr.

Tatsuya Oyobiki, Mr. Katsuma Shimizu, Mr. Kazuki Inada, Mr. Shohei Kitajima, Mr.

Hajime Uetake, for constructive discussions at weekly meetings, which gave me the

opportunity to consider my research from various viewpoints.

Special thanks go to people who have (had) belonged to RCIQE. I would like

to thank researchers, Dr. Natsuo Nakamura, Dr. Takanori Maebashi, and Dr. Keita

Konishi, for giving me helpful comments on my research. I would like to thank my

seniors, Dr. Keitaro Ikejiri, Dr. Masatoshi Yoshimura, Dr. Yoshinori Kohashi, Dr. Tomo

Tanaka, Dr. Eiji Nakai, Dr. Tomotsugu Ishikura, Dr. Zhixin Cui, Dr. Shinya Sakita, Mr.

Kazuhiro Takahagi, Mr. Masatoshi Yatago, Mr. Takahito Endo, Mr. Yutaro Otsu, Mr.

Takehito Watanuki, Mr. Yoh Tanaka, and Ms. Ayana Yamamoto, for their guidance and

support which were of great help to my life in RCIQE. I would like to thank my

classmates, Mr. Fumiya Ishizaka, Mr. Takahiro Hiraki, Mr. Kohei Kamada, Mr. Toshiki

Wada, Mr. Hiromu Fujimagari, Mr. Shogo Yanase, and Ms. Aya Onodera, for sharing a

lot of time and having good relationships with me. I would like to thank my juniors, Mr.

Yungoo Ro, Mr. Hiroaki Kato, Mr. Takao Miyamoto, Mr. Dong Wang, Mr. Kazuki

Hiraishi, Mr. Toshihiro Wada, Mr. Muyi Chen, Mr. Yoshihiro Hiraya, Mr. Ryutaro

Kodaira, Mr. Akihito Sonoda, Mr. Seungwon Choi, Mr. Takehiro Kawauchi, Mr. Taro

Itatsu, Mr. Tomohide Yoshikawa, Mr. Naoto Tamaki, Mr. Kyohei Kabamoto, Mr.

Takuya Miyajima, Mr. Kosuke Wakita, Mr. Shun Takayashiki, Mr. Hiroki Kameda, Mr.

Ryoma Horiguchi, Mr. Dai Hasegawa, Mr. Shota Hiramatsu, Mr. Kohei Chiba, Mr.

Akinobu Yoshida, Mr. Tetsuro Kadowaki, Mr. Masaya Iida, Mr. Masahiro Sasaki, Mr.

Yusuke Minami, Mr. Shinya Yamamoto, for their personal friendships. I would like to

thank researches and students from Germany, Dr. Matthias T. Elm, Dr. Martin Fischer,

iii Acknowledgements

Dr. Martin Becker, Dr. Steve Petznick, Dr. Lennart-Kund Liefeith, Mr. Andy Rühl, and

Mr. Alexander Fabian, for their kindness and holding a number of cheerful parties. I

would like to thank technicians and secretaries, Mr. Kenji Takada, Mr. Kiyotake

Nagakura, Ms. Yuki Watanabe, Ms. Satoko Takeuchi, Ms. Yuka Kamoto, Ms. Mizuho

Tanaka, and Ms. Chieko Akiyama, for their countless supports. I could not accomplish

my research without the dedicated help of all the members of RCIQE.

I would like to express my appreciation to the Japan Society for the Promotion

of Science (JSPS) that supported me financially through the Research Fellowships for

Young Scientists from 2014 to 2017.

Finally, I wish to dedicate this work to my family, Tomohiko Kumazaki,

Yukiko Nishio, Keito Kumazaki, Mitsugi Kumazaki, Yaemi Kumazaki, Yosuke Nishio,

Kazuko Nishio, and Miho Kumazaki, whose encouragement helped me keep trying.

February 15, 2017

Yusuke Kumazaki

iv Acknowledgements

v

Contents

Acknowledgements ....................................................................................... i

Contents ........................................................................................................ v

Chapter 1: Introduction ................................................................................................. 1

1.1. Background .............................................................................................................. 1

1.1.1. Energy and environmental issues

1.1.2. Solutions and role of III-V semiconductors

1.1.3. Concerns about device processing

1.2. Objective .................................................................................................................. 5

1.3. Outline of this thesis ................................................................................................ 6

References ......................................................................................................................... 8

Chapter 2: Semiconductor electrochemistry .............................................................. 11

2.1. Introduction ............................................................................................................ 11

2.2. Basic principles of electrochemistry ...................................................................... 11

2.2.1. Redox reactions

2.2.2. Distribution of state density in the redox system

2.3. Electrochemistry at semiconductor electrode ........................................................ 15

2.3.1. Potential distribution of semiconductor/electrolyte interface

2.3.2. Redox reactions under polarization

2.3.3. Redox reactions under illumination

2.4. Practical use of semiconductor electrochemistry ................................................... 22

2.4.1. Anodic dissolution of III-V semiconductors

2.4.2. Semiconductor porous structures

2.4.3. Semiconductor photoelectrodes for water splitting

References ....................................................................................................................... 30

vi Contents

Chapter 3: Experimental technique ............................................................................ 33

3.1. Introduction ............................................................................................................ 33

3.2. Electrochemical process and characterization ....................................................... 33

3.2.1. Three-electrode electrochemical cell and potentiostat

3.2.2. Linear sweep and cyclic voltammetry

3.2.3. Photo-electrochemical measurement

3.3. Structural characterization ..................................................................................... 37

3.3.1. Scanning electron microscopy (SEM)

3.3.2. Atomic force microscopy (AFM)

3.4. Optical characterization ......................................................................................... 40

3.4.1. Ultraviolet-visible-infrared (UV-VIS-IR) spectroscopy

3.4.2. Photoluminescence (PL) spectroscopy

3.5. Electrical characterization ...................................................................................... 44

3.5.1. Current-voltage (I-V) measurement

3.5.2. Capacitance-voltage (C-V) measurement

References ....................................................................................................................... 45

Chapter 4: Formation and optical characterization of highly ordered pore arrays on InP for photo-electric conversion devices .......................................... 47

4.1. Introduction ............................................................................................................ 47

4.2. Fabrication of InP porous structures ...................................................................... 48

4.3. Optical absorption properties of InP porous structures .......................................... 51

4.3.1. Photo-electric conversion (PC) devices formed on p-n junction substrates

4.3.2. Basic operation properties of PC devices

4.3.3. Comparison between porous and non-porous devices

4.4. Platinum/porous InP Schottky junction PC devices ............................................... 58

4.4.1. Concept of PC devices utilizing large surface of porous layers

4.4.2. Formation of platinum on InP porous structures

4.4.3. I-V characteristics under illumination

4.5. Summary ................................................................................................................ 65

References ....................................................................................................................... 66

vii Contents

Chapter 5: Fabrication and size-modulation of GaN porous structures for EC energy-conversion systems ....................................................................... 71

5.1. Introduction ............................................................................................................ 71

5.2. Experimental details ............................................................................................... 72

5.3. GaN porous structures formed by photo-assisted EC etching ............................... 73

5.3.1. Formation and structural characterization

5.3.2. Pore formation mechanism in photo-assisted EC etching

5.3.3. Photoluminescence and photoreflectance properties

5.3.4. Photo-electrochemical response properties

5.4. GaN porous structures formed by anisotropic EC etching .................................... 81

5.4.1. EC current transition at high voltage


5.4.3. Pore formation mechanism in anisotropic EC etching

5.5. GaN porous structures formed by utilizing Franz-Keldysh effect ......................... 85

5.5.1. Observation of Franz-Keldysh effect (FKE) in EC etching with back-side

illumination mode

5.5.2. Structural characterization of GaN porous structures formed by

FKE-assisted EC etching

5.5.3. Pore formation mechanism in FKE-assisted EC etching

5.6. Precise structural control of GaN porous structures by post wet etching .............. 93

5.6.1. Post wet etching using TMAH

5.6.2. Correlation between structural and optical properties

5.7. Summary .............................................................................................................. 102

References ..................................................................................................................... 104

Chapter 6: Highly selective EC etching of p-GaN grown on AlGaN/GaN heterostructures ................................................................................................... 111

6.1. Introduction .......................................................................................................... 111

6.2. Experimental details ............................................................................................. 112

6.2.1. Basic concept of self-stopping etching using EC reactions

6.2.2. Experimental procedure

6.3. EC etching and structural characterizations ......................................................... 115

6.3.1. Reaction current transition during EC etching

viii Contents

6.3.2. Structural and chemical analyses of the etched surface

6.4. Electrical properties of Schottky diodes formed on the etched surface ............... 118

6.5. Summary .............................................................................................................. 119

References ..................................................................................................................... 120

Chapter 7: Self-terminating EC etching for recessed-gate AlGaN/GaN heterostructure field effect transistors ............................................................. 123

7.1. Introduction .......................................................................................................... 123

7.2. Experimental details ............................................................................................. 124

7.3. Basic photo-electrochemical behavior on AlGaN/GaN heterostructures............. 125

7.4. EC etching based on the regulation of photo-excited carriers ............................. 128

7.4.1. EC etching utilizing photo-carriers generated in GaN layer

7.4.2. EC etching utilizing photo-carriers generated in AlGaN layer

7.5. Electrical properties of recessed-gate AlGaN/GaN HEMTs ................................ 131

7.5.1. Capacitance-voltage characteristics of Schottky diode formed on etched

surface

7.5.2. Current-voltage characteristics of recessed-gate AlGaN/GaN HEMT

7.6. Summary .............................................................................................................. 134

References ..................................................................................................................... 135

Chapter 8: Conclusion ................................................................................................ 139

Appendix A: EC formation and optical characterization of Cu2O/GaN heterostructure for visible light responsive photoelectrode ........................... 143

A.1. Introduction .......................................................................................................... 143

A.2. Experimental details ............................................................................................. 144

A.3. Results and discussion ......................................................................................... 145

A.4. Summary .............................................................................................................. 150

References ..................................................................................................................... 151

ix Contents

Appendix B: Correlation between chapters and publications ................................... 153

Appendix C: List of figures and tables ......................................................................... 155

Research achievements ........................................................................... 167

x Contents

1

Chapter 1

Introduction

1.1. Background

1.1.1. Energy and environmental issues

Along with the rapid economic development and population growth in late

20th and 21st centuries, comprehensive issues related to depletion of energy resources

and environmental pollution (energy and environmental issues) grow into serious

problems. Figure 1-1 shows world primary energy consumption from 1990 to 2015

reported by British Petroleum [1]. It is obvious that overall energy consumption has

been growing and fossil fuels such as oil, coal, and nature gas have been dominant

energy resources. Fossil fuels, however, are unable to be used for eternity due to their

limited resources, indicated by the reserved-production ratios (R/P) of 51 for oil, 53 for

natural gas, and 114 for coals at 2015 [1]. If the fossil fuels are consumed at current rate,

these are at risk of depletion in the next several decades. The large consumption of

fossil fuels, in addition, cause various kinds of environmental problems. One of the

Figure 1-1. Stacked area chart of the world primary energy consumption from 1990 to 2015

reported by British Petroleum [1].

0

2000

4000

6000

8000

10000

12000

14000

1990 1995 2000 2005 2010 2015

Prim

ary

ener

gy

cons

ump

tion （m

illio

n to

e）

Year

Oil

Natural gas

Coal

Nuclear energy

Hydroelectricity

Other renewables

2 Chapter 1

serious environmental problems is global warming which have caused (is expected to

cause) extreme weather, sea level rise, and adverse effects on ecological systems. It is

said that primary origin of global warming is increasing concentration of greenhouse

gases represented by carbon dioxide (CO2). Since one with largest amount of CO2

emission is burning of fossil fuels, it is expected to increase CO2 emission with increase

of energy consumption. Since nuclear energy is known as low carbon power generation

methods of producing electricity, it accounted for 30 % of energy productions in Japan

in 2010. However, the Great East Japan Earthquake caused nuclear accidents in 2011,

and almost all of nuclear power plant in Japan has been shut down because they cannot

secure sufficient safety. As a result, energy deficiency had to be supplied by fossil fuels,

leading to increasing consumption of fossil fuels. Under these circumstances, innovative

technology for energy-creating and energy-saving are desired to solve energy and

environmental issues all over the world. In particular, there are high expectations on

semiconductor electronic fields as described in the next section.

1.1.2. Solutions and role of III-V semiconductors

As for the energy-creating technology to solve energy and environmental

issues, utilization of renewable energies are expected because they are sustainable on

human time scale, and can be utilized as energy without polluting environment. Among

the various renewable energies, solar light is an attractive candidate because of their

enormous energy and relatively low regional bias. The well-known application utilizing

solar light is photovoltaic cell which converts solar light directly to electricity using

semiconductors. Enormous researches on photovoltaic cell have been done, and

conversion energy has been improved step by step. Nevertheless, the current use of

solar light is only 0.4 % of overall energy consumption. Major challenge of photovoltaic

cell is high power-generating cost and low energy conversion efficiency. Although the

reduction of power generation cost and the improvement of conversion efficiency are

required to be comparable energy-creating technology to conventional large-scale power

generation methods, silicon (Si) series photovoltaic cells that are the mainstream of

photovoltaic cells will come closer to the limit of efficiency and cost in the near future.

As for the energy-saving technology, efficiency improvement of power

electronics is key factor. Power electronics are the technologies relevant to electricity

controlling and converting, which is applied to power source unit in every electronic

3 Introduction

devices and systems. In the actual power electronics, not all the electric power is

utilized because of power loss in the system. Such power electronic circuits consists of

various power semiconductor devices, which actually manage the efficiency of the

systems. Therefore, the realization of low-loss and high-performance power

semiconductor devices enables us to save energy. Currently, Si-based insulated gate

bipolar transistors (IGBTs) [2] and super junction metal-oxide-semiconductor

field-effect transistors (SJ-MOSFET) [3] dominate the market of power switching

devices. However, the characteristics of Si-based power switching devices are

approaching the performance limit as is the case with photovoltaic cells.

Thus, the replacement of Si in both energy-creating and energy-saving

technologies are required. III-V semiconductors have attracted much attention as device

materials going beyond Si. The physical parameters of typical III-V semiconductors are

shown in Table 1-I, including those of Si. Generally, III-V semiconductors shows

superior characteristics compared to Si as itemized below:

1) Most III-V semiconductors have direct bandgap, indicating higher absorption

coefficient and luminescence efficiency than those of Si.

2) The electron mass is smaller. Thus, high electron mobility and large saturation

velocity can be obtained for electronic devices.

3) III-V semiconductors have strong resistance against cosmic ray irradiation and

capability to work in high temperature environment due to their wide bandgap.

4) Wide bandgap materials show high breakdown field due to high impact

ionization energies, enabling low on resistance operation.

Table 1-I. Physical properties of Si and III-V semiconductors at RT.

Bandgap (eV)

Transition type

Electron mobility (cm2/V∙s)

Breakdown field

(MV/cm)

Saturation velocity

(×107 cm/s)

Si 1.12 indirect 1350 0.3 1.0

GaAs 1.43 direct 8500 0.4 2.0

InP 1.37 direct 5400 0.5 2.5

GaN 3.42 direct 1200 3.3 2.5

4 Chapter 1

5) Since wide varieties of mixed alloys can be obtained, physical properties can

be tuned flexibly as the situation demands.

In application to the photovoltaic cells, InP, GaAs and their alloys are attracted

much attentions. The Shockley-Queisser limit [4], which is calculated by examining the

amount of electrical energy that is extracted per photon of incoming solar light,

indicates the semiconductors with the bandgap of around 1.35 eV are favorable to

extract electric energy efficiently. In addition, multi-junction cells which can exceed the

Shockley-Queisser limit are often consist of III-V semiconductors owing to the bandgap

selectivity, and achieved high conversion efficiency under concentrated solar light

[5−7].

In application to the power switching devices, GaN has extensively studied as

one of the attractive candidate materials. Compared to Si, GaN has a threefold wilder

bandgap, leading to tenfold higher breakdown electric field. Although electron mobility

of GaN showed lower values than Si, that can be improved by formation of

heterostructures with other nitrides. Especially, AlGaN/GaN heterostructures where a

two-dimensional electron gas (2DEG) is formed by spontaneous and piezoelectric

polarizations shows higher electron mobility than Si. Thus, GaN-based devices are

highly advantageous for low-loss power switching devices [8,9].

1.1.3. Concerns about device processing

Despite the III-V semiconductors are significantly expected as device materials

going beyond Si, applications based on III-V semiconductors currently have not been

achieved to obtain superior performance as high as expected, which impedes

replacement of Si status. One of the key issues is the device processing technologies.

Device performance has been significantly improved by invention of device

structure as well as improvement of material quality. Development of fine processing

technologies, in other words, are crucial to improve device performance. In application

to the photovoltaic cells, for example, surface texturing techniques are promising to

reduce reflectance of devices, leading to improvement of conversion efficiency of

photovoltaic cells. As for the power switching devices, partial etching techniques are

frequently required to realize high-performance and reliable operation (detailed in

Chapter 6 and Chapter 7). Currently, major technique of processing of III-V

5 Introduction

semiconductors is plasma etching because of their superior controllability and

productivity. However, plasma-etched surface generally suffer from various types of

damage which could significantly degrade the device performance. Wet chemical

etching is also common technique as device processing because this induces few

damage on surface. However there are severe limitation in controllability, which make it

difficult to process precisely. In addition, chemically stable materials such as group-III

nitrides are barely reacted with solutions, indicating wet chemical etching is impossible

to use for such materials.

One alternative approach is a electrochemical (EC) etching which is the cyclic

process consists of an anodic oxidation and a subsequent dissolution of the resulting

oxide in an solution. Compared to plasma etching, EC etching is highly desirable in its

simplicity and the absence of plasma damage [10,11]. Besides, the electrochemical

process is applicable to various semiconductors even chemically stable materials such

as group-III nitrides [12]. Despite EC etching techniques of III-V semiconductors have

studied for a long time, etching controllability which is important aspect as device

processing has been insufficient yet. In addition, there are few reports which

demonstrate performance of application fabricated by EC etching technique. These are

why the EC etching cannot be the major technique in processing of III-V

semiconductors.

1.2. Objective

Although EC process is characterized as low damage and high productivity, EC

process has been suffer from etching controllability, which make it difficult to use as

device processing, as mentioned above. The objective of this work is to develop the EC

process technique and demonstrate applications of III-V semiconductors: InP, GaN and

AlGaN are chosen due to their desirable features as described in previous section.

Fundamental properties of EC reactions at semiconductor/solution interface are revealed,

and novel EC etching techniques are developed for application to the energy-creating

and energy-saving technologies, as follows:

1) EC-fabrication, optical characterization, and device application of InP porous

structures for photovoltaic cells.

2) EC-fabrication and size-modulation of GaN porous structures for EC

6 Chapter 1

energy-conversion systems.

3) Highly selective EC etching of p-GaN grown on AlGaN/GaN heterostructures

for power switching devices.

4) Self-terminating and well-controllable EC etching to fabricate recessed-gate

AlGaN/GaN hetero-structure field effect transistors for power switching

devices.

1.3. Outline of this thesis

This thesis consists of 8 chapters.

Chapter 2 describes some essential aspects of semiconductor electrochemistry,

especially related to practical etching of semiconductors. Electrochemical fundamentals

described in this chapter form the basis for the work in later chapters. In addition,

typical examples which shows unique and practical features of electrochemical process

on semiconductors were briefly presented.

Chapter 3 gives a brief overview of experimental techniques for formation and

characterization of III-V semiconductor microstructures is given. Experimental setup of

EC process is first described. Then, characterization techniques for structural, optical,

and electrical properties of III-V semiconductor microstructures are described in each

section.

Chapter 4 shows the results about formation, optical characterization, and

device application of InP porous structures. Optical absorption properties of InP porous

structures are investigated by using photo-electric conversion devices consist of top

porous layer and p-n junction where light converted into electric signals.

Photo-response properties of Schottky junction photo-electric conversion devices based

on platinum/porous InP are also investigated to clarify the impact of enhanced structural

and optical features on conversion efficiency.

Chapter 5 shows the results about fabrication and size-modulation of GaN

porous structures for EC energy-conversion systems. Structural properties of GaN

porous structures are modulated by EC conditions involving various charged-carrier

generation phenomena: band-edge absorption, Franz-Keldysh effect, and avalanche

7 Introduction

effect. Post treatment technique by conventional wet etching was also performed to

control structural properties precisely. Optical characterization revealed the importance

of precise structural controlling of porous structures on application to the EC energy

conversion systems.

Chapter 6 shows the results about highly selective EC etching of p-GaN

grown on AlGaN/GaN heterostructures. It is revealed that etching depth can be

precisely controlled and automatically stopped on the AlGaN surface by selecting the

optimal EC conditions. Electrical characterizations reveal that no significant damages

are introduced in the AlGaN/GaN heterostructures during the EC etching process.

Chapter 7 demonstrates the self-terminating EC etching for recessed-gate

AlGaN/GaN hetero-structure field effect transistors. Based on the regulation of carrier

transfer, we succeeded in developing self-terminating and depth-controllable EC etching

of AlGaN/GaN hetero-structure with very smooth surface. The recessed-gate

AlGaN/GaN HEMT fabricated by this technique shows positive threshold voltage,

reduction of static on-resistance, and improvement of transconductance compared to the

planar-gate AlGaN/GaN HEMT.

The research will be concluded in Chapter 8.

8 Chapter 1

Reference

[1] British Petroleum, "BP Statistical Review of World Energy June 2016", 2016.

[2] A. R. Hefner Jr, and D. L. Blackburn, "An analytical model for the steady-state

and transient characteristics of the power insulated-gate bipolar transistor",

Solid-State Electron., vol. 31, pp. 1513−1532, 1988.

[3] P. M. Shenoy, A. Bhalla, and G. M. Dolny, "Analysis of the effect of charge

imbalance on the static and dynamic characteristics of the super junction

MOSFET", Proc. ISPSD, pp. 99−102, 1999.

[4] W. Shockley, and H. J. Queisser, "Detailed Balance Limit of Efficiency of p-n

Junction Solar Cells", J. Appl. Phys., vol. 32, pp. 510−519, 1961.

[5] R. R. King, D. Bhusari, D. Larrabee, X. -Q. Liu, E. Rehder, K. Edmondson, H.

Cotal, R. K. Jones, J. H. Ermer, C. M. Fetzer, D. C. Law, and N. H. Karam,

"Solar cell generations over 40% efficiency", Prog. Photovoltaics Res. Appl.,

vol. 20, pp. 801−815, 2012.

[6] M. Yamaguchi, T. Takamoto and K. Araki, "Present and Future of Super High

Efficiency Multi-Junction Solar Cells", Proc. SPIE, vol. 6889, p. 688906, 2008.

[7] M. A. Green, K. Emery, Y. Hishikawa, W. Warta, and E. D. Dunlop, "Solar cell

efficiency tables (version 41)", Prog. Photovoltaics Res. Appl., vol. 21, pp.

1−11, 2013.

[8] U. K. Mishra, P. Parikh, and Y. -F. Wu, "AlGaN/GaN HEMTs – an overview of

device operation and applications", vol. 90, pp. 1022−1031, 2002.

[9] N. Ikeda, Y. Niiyama, H. Kambayashi, Y. Sato, T. Nomura, S. Kato, and S.

Yoshida, "GaN Power Transistors on Si Substrates for Switching Applications",

vol. 98, pp. 1151−1161, 2010.

[10] N. Shiozaki, T. Sato, and T. Hashizume, "Formation of Thin Native Oxide

Layer on n-GaN by Electrochemical Process in Mixed Solution with Glycol

and Water", Jpn. J. Appl. Phys., vol. 46, pp. 1471−1473, 2007.

[11] N. Shiozaki, and T. Hashizume, "Improvements of electronic and optical

characteristics of n-GaN-based structures by photoelectrochemical oxidation in

9 Introduction

glycol solution", J. Appl. Phys., vol. 105, p. 064912, 2009.

[12] T. Sato, Y. Kumazaki, M. Edamoto, M. Akazawa, and T. Hashizume, "Interface

control technologies for high-power GaN transistors: Self-stopping etching of

p-GaN layers utilizing electrochemical reactions", Proc. SPIE, vol. 9748, p.

97480Y, 2016.

10 Chapter 1

11

Chapter 2

Semiconductor electrochemistry

2.1. Introduction

Since Luigi Galvani found "animal electricity" and Alessandro Volt indicated

the relationship between electricity and chemistry in the late 18th century,

"Electrochemistry" has been intensively researched among the various research area for

a long time. Electrochemistry on semiconductor electrode is much useful because it is

applicable to various kinds of functionalization technique of semiconductors. In

Chapter 2, some essential aspects of semiconductor electrochemistry, especially related

to practical etching of semiconductors, will be briefly described. Electrochemical

fundamentals described in this chapter form the basis for the work in later chapters. In

addition, typical examples which shows unique and practical features of electrochemical

process on semiconductors were briefly presented.

2.2. Basic principles of electrochemistry [1]

2.2.1. Redox reactions

A redox reaction is a chemical phenomenon involving electron transfer

between two kinds of chemical particles. It can be classified into two groups: the

chemical reaction converting a reductant particle (RED) into an oxidant particle (OX)

with receiving electrons is "oxidation"; and the chemical reaction converting a OX into

a RED with consuming electrons is "reduction". Figure 2-1 schematically shows the

redox reactions of hydrates that could be presented as follows:

(1) REDred → OXred + estd : adiabatic electron release, ΔG = −ERED

(2) OXred → OXox : reorganization of hydrate structure, ΔG = −λOX

(3) OXox + estd → REDox : adiabatic electron absorption, ΔG = EOX

(4) REDox → OXred : reorganization of hydrate structure, ΔG = −λRED

where OXred (REDox) is the unstable oxidant (reductant) particle where hydrate structure

12 Chapter 2

is same with reductant (oxidant) particles, estd is the electron at vacuum level, ΔG is the

change in Gibbs free energy of each step, ERED (EOX) is the donor (acceptor) level of the

reductant (oxidant), and λ is the energy required to reorganize the hydrate structure. In

general, the electron transfer occurs adiabatically while the hydrate structure is frozen

which is followed by reorganization of the hydrate structure since the rate of electron

transfer is more than 100 times as high as that of molecular vibration. Figure 2-2(a)

shows schematic representation of the energy levels in the aforementioned redox

reaction. Since the energy should be conserved during redox reaction cycle, the

following equation can be obtained:

REDOXREDOX λλ EE . (2-1)

Eq. (2-1) indicates that the energy gap between the acceptor level of oxidant EOX and

the donor level of reductant ERED corresponds to the summation of the reorganization

energy of oxidant λOX and reductant λRED, as shown in Fig. 2-2(b). Thus, the electron

energy levels of oxidant and reductant are separated (EOX ≠ ERED), which is called

Frank-Condon splitting of electron states.

Figure 2-1. Redox reaction cycle in aqueous solution: λ is the reorganization energy, ERED is the

most probable donor level, and EOX is the most probable acceptor level.

OXred

unstable stable

REDred

OXox REDox

ERED

EOX

Ele

ctro

n en

erg

y E

0 estd

λRED

λOX

H2O

adiabatic electron transferhydrate structure reorganization

13 Semiconductor electrochemistry

2.2.2. Distribution of state density in the redox system

The localized electron energy level of hydrated particles in aqueous solutions

fluctuates in the range of reorganization energy λ which can be given by following

equation [2]:

sop0

2

ε

1

ε

1

πε8λ

a

q, (2-2)

where q is the elementary charge, a is the ion radius, ε0 is the permittivity of vacuum, εs

and εop are the static and high frequency (optical) dielectric constant of water (εs = 78.5

at room temperature and εop = 1.8), respectively. Since the energy fluctuation Δλ in the

hydrated structure follows the Boltzmann distribution, the resulting energy fluctuation

ΔE in the electron energy level is represented by a Gaussian normal distribution: namely,

the probability density of electron energy fluctuation W(E) can be represented by a

following equation:

kT

EEexp

kTEW

λ4λ4

1)(

20 , (2-3)

Figure 2-2. Schematic illustrations of (a) energy cycle and (b) electron energy levels for the

redox reaction of hydrated particles.

Ele

ctro

n en

erg

y E

OXred + estd

OXox + estd

REDox

REDred

−λRED

−λOX

EOX−ERED

Ele

ctro

n en

erg

y E

E = 0

EOX

ERED

λOX + λRED

electron transferreorganization

(a) (b)

14 Chapter 2

where T is the temperature, k is Boltzmann constant, and E0 is the most probable

electron energy level of hydrated particles. The electron state densities of hydrated

redox particles D(E) are given by the product of the probability densities W(E) and the

concentrations N of hydrated redox particles. Therefore the electron state densities of

hydrated redox particles in the donor bands DRED(E) and in the acceptor band DOX(E)

can be obtained by following equations, respectively:

kT

EEexp

kT

NNEWED

λ4λ4

2REDRED

REDREDRED , (2-4)

kT

EEexp

kT

NNEWED

λ4λ4

2OXOX

OXOXOX . (2-5)

The total state density DREDOX(E) is the summation of DRED(E) and DOX(E) as follow:

OXOXREDREDOXREDREDOX NEWNEWEDEDED . (2-6)

Figure 2-3 shows schematic representation of total state density of hydrated redox

particles when (a) oxidant concentration NOX equal to reductant concentration NRED, and

(b) NOX higher than NRED. As shown in Fig. 2-3, it can be assumed that the energy bands

are formed in the redox particles distribution: a donor (occupied) band is formed by the

REDs, and an acceptor (vacant) band is formed by the OXs. The energy level with

DRED(E) = DOX(E) is called the Fermi level of the electron in a redox system EF(REDOX)

that is similar to the Fermi level of electrons in a metal and a semiconductor EF [3].

Namely, the Fermi level of the electron in a redox system EF(REDOX) can be derived by

calculating using Eq. (2-4) and Eq. (2-5) with the assumption of DRED(E) = DOX(E) as

follow:

OX

RED0F(REDOX)

OX

REDREDOXF(REDOX) 2

1

N

NlnkTE

N

NlnkTEEE , (2-7)

where 0F(REDOX)E is called the standard Fermi level of the electron in a redox system.


2.3. Electrochemistry at semiconductor electrode

2.3.1. Potential distribution of semiconductor/electrolyte interface

Semiconductors are characterized by a valence band (VB) and conduction band

(CB) separated by a forbidden energy gap, the bandgap Eg. The position of Fermi level

EF, the chemical potential of electrons in semiconductor, depends strongly on the doping.

For an intrinsic semiconductor, the EF is located at the middle of bandgap. When

semiconductor is doped with atoms acting as electron donors, the n-type semiconductor

with EF close to the CB edge is obtained. In this case, EF is related to the CB electron

concentration in the bulk of semiconductor n0 as follow:

kT

EEexpNn FC

C0 , (2-8)

where NC is the effective density of states in the CB, EC is the energy level of the lower

edge of the CB. On the other hand, semiconductor doped with atoms acting as electron

acceptor is called a p-type semiconductor in which EF close to the VB edge. Such level

can accept electrons from VB, thereby giving rise to holes in VB. In this case, the hole

Figure 2-3. Distribution of the electron state density of hydrated redox particles: (a) oxidant

concentration OX equal to reductant concentration NRED, and (b) NOX higher than NRED.

(a) (b)

Ele

ctro

n en

erg

y E

Total state density DREDOX

NOX = NRED

Ele

ctro

n en

erg

y E

Total state density DREDOX

NOX > NRED

EOX

ERED

EF(REDOX)

EOX

ERED

EF(REDOX)

16 Chapter 2

concentration p0 is given by,

kT

EEexpNp VF

V0 , (2-9)

where NV is the effective density of states in the VB, EV is the energy level of the upper

edge of the CB.

When a semiconductor is immersed in an electrolyte, a redistribution of charge

occurs at the semiconductor/electrolyte interface. Figure 2-4 shows schematic

representations of energy diagram at n-type semiconductor/electrolyte interface when

they are at (a) non-equilibrium and (b) equilibrium. Obviously, similar considerations

also hold for p-type semiconductors. At non-equilibrium, electrons are transferred to

electrolyte because of the difference in Fermi level between semiconductor and

electrolyte (EF ≠ EF(REDOX)). Consequently, thermal equilibrium is realized in the

semiconductor/electrolyte interface (EF = EF(REDOX)), and space charge region (SCR)

consisting of uncompensated immobile ionized donors is formed within the

semiconductor as shown in Fig. 2-4(b). In SCR, there is a potential difference with

respect to the degree of band bending at semiconductor surface, and potential difference

at thermal equilibrium is called flat-band potential EFB.

Figure 2-4. Schematic representations of energy diagrams at n-type semiconductor/electrolyte

interface: (a) for the case in which the n-type semiconductor and electrolyte are not in equilibrium

(EF ≠ EF(REDOX)); (b) for the case in which the semiconductor is equilibrium with the electrolyte (EF

= EF(REDOX)).


2.3.2. Redox reactions under polarization

The redox reaction in which an electron transfers from the electrode to the

electrolyte is called the "cathodic reaction"; and the electrode reaction in which a hole

transfers from the electrode to the electrolyte is called the "anodic reaction". Further,

electrode at which the cathodic reaction takes place is called "cathode"; and the

electrode at which then anodic reaction takes place is called the "anode".

We consider redox reactions at n-type semiconductor electrodes which are

polarized (applied voltage) with an overpotential η relative to equilibrium redox

potential (the Fermi level EF(REDOX)). In general, redistribution of charge occurs to be at

thermal equilibrium as soon as immersing a semiconductor in an electrolyte as shown in

Fig. 2-4. Polarization of electrodes shifts the Fermi level of the electrode EF from the

Fermi level of the redox electron EF(REDOX) by an energy equivalent to the overpotential

η as described in following equation:

ηF(REDOX)F qEE . (2-10)

Figure 2-5 shows schematic representations of energy diagram at n-type

Figure 2-5. Schematic representations of energy diagram at n-type semiconductor/ electrolyte

interface under polarization with overpotential η: (a) under positive polarization (qη > EFB), (b)

under negative polarization (qη = EFB), and (c) under negative polarization (qη < EFB),

respectively.

(a) (b)

EF(REDOX)EC

EF

EV

positive polarization(qη > EFB)

n-typesemiconductor electrolyte

qη

negative polarization(qη = EFB)

qη = EFBqη

negative polarization(qη < EFB)

(c)

18 Chapter 2

semiconductor/electrolyte interface under (a) positive polarization, (b) negative

polarization with qη = EFB, and (c) negative polarization with qη < EFB, respectively.

Figure 2-6(a) shows typical current curves of n-type semiconductor being immersed in

an electrolyte as a function of overpotential qη. Under the positive polarization, there is

larger upward band bending at the semiconductor surface compared to that at

equilibrium as shown in Fig. 2-5(a). This upward band bending prevents electrons

(majority carriers) being transferred to electrolyte and allows holes (minority carriers)

being transferred to electrolyte, indicating that few redox reactions occur (few currents

flow) as shown in Fig. 2-6(a). When the negative potential is applied, potential drop

over the SCR decreases and becomes zero at flat-band potential EFB as shown in Fig.

2-5(b). A further increase in negative polarization causes accumulation of electrons

(majority carriers) because the band bend downwards. When the Fermi level may

intersect the CB-edge near the surface, the semiconductor becomes degenerate and

shows quasi-metallic behavior, causing increase of cathodic reactions (cathodic current

flow) as shown in Fig. 2-6(a).

So far, the redox reactions for n-type semiconductors has been discussed. Of

course, a similar argument also holds for p-type semiconductors. However, the

polarization dependence is reversed: positive polarization causes few redox reactions;

Figure 2-6. Typical current curves of (a) n-type and (b) p-type semiconductor plotted as a

function of overpotential qη. The carriers at n-type semiconductor/electrolyte interface are

depleted when qη > EFB and accumulated when qη < EFB. Polarization dependence of the

carriers at n-type semiconductor/electrolyte interface is reversed: they are depleted when qη <

EFB and accumulated when qη > EFB.

qηEFB

n-type semiconductor

depletion

accumulation

Cur

rent

qη

EFB

p-type semiconductor

depletion

accumulation

Cur

rent

(a) (b)


and the negative polarization with qη > EFB causes hole accumulation and anodic

reactions, as shown in Fig. 2-6(b).

2.3.3. Redox reactions under illumination

Under the illumination of light with energy above the bandgap Eg, the

semiconductor electrode absorbs the photon energy to produce photo-excited pairs of

electrons in the CB and holes in the VB. The photo-excited carriers in semiconductors

are relatively stable compared to those in metal electrodes. Therefore, the

illumination-effect on redox reactions can be observed more obviously with

semiconductors than with metals. For n-type semiconductors, the photo-excited

electrons hardly affect the concentration of electrons (majority carriers) in the CB, but

the photo-excited holes greatly increases the concentration of holes (minority carriers)

in the VB. In such situation, Since thermal equilibrium is not established between the

photo-excited electrons in the CB and the holes in the VB, the electrochemical

potentials of them can be defined individually, which are called the quasi-Fermi levels

[4,5].

Generally, redox reactions on semiconductor electrode proceed under the

condition of positive and negative polarization in which the Fermi level of the electron

in a semiconductor EF is deferent from the Fermi level of the electron in a redox system

EF(REDOX): cathodic reaction proceeds when EF > EF(REDOX); and anodic reaction

Figure 2-7. Schematic representations of energy diagrams for n-type semiconductor electrode

(a) in dark and (b) in the photo-excited state.

(a)

EF(H+/H2)EC

EF

EV

in dark

n-type semiconductor electrolyte

EF(O2/H2O)

EF(H+/H2)EC

EF

EV

in the photo-excited state


EF(O2/H2O)EFp

EFn

hν

(b)

electronelectron

hole

20 Chapter 2

proceeds when EF < EF(REDOX). Under light illumination, quasi-Fermi level of electrons

EFn and holes EFp determines the direction of redox electron transfer reactions: in other

words, cathodic reaction proceeds when EFn > EF(REDOX), and anodic reaction proceeds

when EFp < EF(REDOX). Figure 2-7 shows schematic representations of energy diagram

for n-type semiconductor electrode in aqueous electrolyte: (a) in dark, (b) in the

photo-excited state. Here, we assumed Fermi levels of redox reactions of hydrogen

EF(H+/H2) and oxygen EF(O2/H2O). When the n-type semiconductor electrode is in the dark

and the Fermi level of the semiconductor EF is located between EF(H+/H2) and EF(O2/H2O) as

shown in Fig. 2-7(a), it is thermodynamically impossible to cause both the electron

transfer reactions and the hole transfer reactions. When the light is irradiated on the

n-type semiconductor electrode, Fermi level of the semiconductor EF shifts by the

photo-potential ΔEph and splits into the two levels at semiconductor/electrolyte

interface: namely, quasi-Fermi levels of electrons EFn and holes EFp. Provided that the

energy conditions, EFn > EF(H+/H2) and EFp < EF(O2/H2O), are satisfied as shown in Fig.

2-7(b), both the electron transfer reactions and the hole transfer reactions is

thermodynamically possible.

Photo-excited carriers in the SCR migrate towards a electrolyte and cause

redox reactions there with providing the reaction current as shown in Fig. 2-8. Such a

current resulting from migration of photo-excited carriers is called the photocurrent. As

Figure 2-8. Schematic representations of redox reactions at (a) n-type and (b) p-type

semiconductor electrode under light illumination. Photo-excited holes cause anodic reactions

(RED → OX + e−) at n-type semiconductor, whereas photo-excited electrons cause cathodic

reactions (OX + e− → RED) at p-type semiconductor.

OX

RED

OX’

RED’

n-typesemiconductor electrolyte metal

electron

hole

OX’

RED’

OX

RED

p-typesemiconductor electrolyte metal

electron

hole

hνhν

EC

EF

EV

EF(M)

(a) (b)

EC

EF

EV

EF(M)


previously described, concentration of majority carriers are hardly influenced by photo

excitation, whereas concentration of minority carriers are greatly increase by photo

excitation. Consequently, the photocurrent can be observed only when the minority

carriers participate in redox reactions, that is, the anodic reaction at n-type

semiconductor and the cathodic reaction at p-type semiconductor as schematically

illustrated in Fig 2-9.

The generation rate of photo-excited electron-hole pairs G(x) is given as a

function of the light intensity at interface P0, the absorption coefficient of

semiconductor α, and the depth from the interface x, as follows [6]:

)xexp(PxG α)( 0 . (2-11)

Integration of G(x) yields the total photocurrent Iph [7]:

L

kTqexpqP

L

WexpqPI SCRD

α1

1φΔαλ21

α1

α1 0

SCR0ph , (2-12)

where WSCR is the thickness of the SCR, λD is the Debye length, ΔφSCR is the difference

of potential in the SCR, and L is the diffusion length of minority carriers. Eq. (2-12)

indicates that, if αWSCR is much larger than 1 (αWSCR >> 1), all the photo-excited

minority carriers would be consumed in the redox reactions (Iph = qP0). On the other

Figure 2-9. Typical current curves of (a) n-type and (b) p-type semiconductor electrode in dark

and under light illumination. Positive photocurrent Iph was observed at n-type semiconductor with

potential E > EFB, whereas negative Iph was observed at p-type semiconductor with E < EFB.

E

EFB

n-type semiconductorC

urre

nt E

p-type semiconductor

Cur

rent

(a) (b)

in dark

illumination

illumination

in darkIph

IphEFB

22 Chapter 2

hand, if αWSCR is much smaller than 1 (αWSCR << 1) and 1φΔαλ kTq SCRD is also

much smaller than one, Eq. (2-12) can deformed to following equation:

1φΔ

λ2αα 0SCR0ph kT

qLqPWLqPI SCR

D . (2-13)

Eq. (2-13) indicates that if the diffusion length L is much longer than the SCR thickness

WSCR (L << WSCR), the photocurrent would be constant and independent from the

potential.

2.4. Practical use of semiconductor electrochemistry

2.4.1. Anodic dissolution of III-V semiconductors

Most of semiconductors are not robust over anodic reactions caused by holes:

semiconductor electrodes themselves are dissolved anodically in electrolyte. The

presence of a hole (h+) in a localized surface bond means that one of the electrons in a

bonding orbital has been removed. In the case of gallium arsenide (GaAs), Gerischer

and Mindt [8] suggested that, this hole trapping at the semiconductor surface can be

represented by Fig. 2-10(a), in which a nucleophilic agent (X−) reacts with one of the

Ga As

X X

(a) h+ X−+ +Ga As Ga As

X

(b) h+ X−+ +Ga As

X

Ga As

X X

(c) 4h++ Ga3+ As3++

Figure 2-10. The flow of anodic dissolution of GaAs suggested by Gerischer and Mindt [8]: (a) A

nucleophilic agent (X−) reacts with one of the positively charged surface atoms from which an

electron is removed. (b) Surface state formed by unpaired electron traps hole and reacts with X−.

(c) One GaAs is dissolved into electrolyte by using four more holes.


positively charged surface atoms from which an electron is removed. The remaining

unpaired electron is no longer in a VB state but has an energy level in the bandgap

closer to the VB like an acceptor state. This acceptor-like state could capture a second

hole and react with X− as shown in Fig. 2-10(b). It has been determined that four more

holes are needed to dissolve one GaAs entirely according to Fig. 2-10(c). This is the

qualitative explanation of anodic dissolution of III-V semiconductors. It is obvious from

above discussion that anodic dissolution of III-V semiconductors requires holes at the

electrode surface. For p-type semiconductors, since holes are exist as majority carrier in

VB, anodic dissolution can be caused by application of potential (> EFB) on electrode as

shown in Fig. 2-6. The n-type semiconductors, on the other hand, intrinsically possess

few holes in VB, which makes it difficult to dissolve anodically in dark. Due to the lack

of intrinsic holes in VB of n-type semiconductor, light illumination is the major way to

cause anodic dissolution.

As discussed in Section 2.3.3, semiconductor electrodes absorb the energy of

photons to produce excited electrons and holes in the CB and VB under the light

illumination, which make it possible to cause anodic dissolution. Anodic dissolution

under light illumination is called the "photo-assisted electrochemical etching". In this

method, anodic reaction can be caused when holes can be transferred to electrolyte from

VB of semiconductor electrode: (1) illuminated light (photons) have larger energy than

bandgap, and (2) applied potential is larger than flat band potential. Because of these

features, photo-assisted electrochemical etching can realize superior etching selectivity

of bandgap [9], dopant-type [10], and so on [11].

Figure 2-11 shows typical example of material-selective photo-assisted

electrochemical etching reported by A. R. Stonas and co-workers [9]. In0.12Ga0.88N

Figure 2-11. (a) Schematic illustration of sample structure and (b) SEM image of sample after

material-selective etching by photo-assisted electrochemical process [9]: GaN cantilevers curved

upwards by the removal of InGaN layer and the relief of substrate-induced strain.

(a) (b)

24 Chapter 2

(lower bandgap energy with respect to GaN) layer over GaN could be selectively

removing using a GaN filter to restrict the illumination wavelength. After removing

InGaN selectively, GaN cantilevers curved upwards because of the relief of intrinsic,

as-grown strain gradient.

Dopant-selective photo-assisted electrochemical etching is also intensively

researched. As shown in Fig. 2-9, n-type semiconductors have lower flat band potential

than p-type semiconductors, which allows selective anodic etching of n-type

semiconductors over p-type semiconductors. Figure 2-12 shows typical example of

dopant-selective photo-assisted electrochemical etching reported by C. Youtsey and

co-workers [10]. The n-type GaN under layer was selectively etched and p-type GaN

top layer remained unchanged, resulting in mushroom-shaped structure.

Defect-selective etching is also one of famous practical application of

photo-assisted electrochemical etching. C. Youtsey and co-workers reported

whisker-shaped structure formed by photo-assisted electrochemical etching [11]. TEM

observation revealed that these whiskers correspond to edge and mixed dislocations as

shown in Fig. 2-13. Defects such as dislocations act as recombination center which

prevent holes from contributing to anodic reactions.

2.4.2. Semiconductor porous structures

In the late 20th century, the discovery of luminescent porous Si formed by

electrochemical process triggered considerable interest in porous semiconductors. At the

same time, porous structure have been formed on various III-V semiconductors to

emphasize their specific properties.

Figure 2-12. SEM images of p-on-n GaN columns (a) before and (b) after dopant-selective

etching by photo-assisted electrochemical process [10].

(a) (b)


Formation of porous structures is frequently reported on n-type III-V

semiconductors by anodic dissolution in dark. As mentioned in Section 2.3.2 and

Section 2.4.1, since the holes are minority carriers in n-type semiconductor, the amount

of anodic current caused by holes will be small in general. However, a steep anodic

current increase is usually observed when the electrode is highly polarized in anodic

direction. Such a anodic current increase is supposed to be related to holes generated by

avalanche effect in SCR. Under the high electric field, avalanche breakdown can be

initiated by a small number of electrons tunneling from the VB or RED to the CB. If the

field strength in SCR reaches the critical value, the tunneled electrons can be

accelerated to high enough to knock other bound electrons, leading to generating new

electron-hole pairs. Since avalanche breakdown is assumed to initiate preferentially at

dislocations, point defects, scratches, and so on, localized anodic dissolution could be

caused. This localized process induces new surface pits and the localized anodic

dissolution by avalanche breakdown can be caused at lower potentials as previous

process. Thus repeated cycle allow pore growth in n-type semiconductors. For the

p-type III-V semiconductors, uniformly distributed holes in VB often results in a

uniform dissolution which is called electropolishing.

Morphological characteristics are drastically changed by material, doping

density, crystal plane orientation, and electrochemical conditions. Table 2-I summarized

typical morphological characteristics of pores reported on III-V semiconductors.

Morphological characteristics of Si porous structures has been added for comparison.

Figure 2-13. (a) The SEM image of GaN whisker-shaped structure, and (b) the cross-sectional

TEM image of whisker, showing that both edge (e) and mixed (m) dislocations are associated

with whisker formation [11].

(a) (b)

26 Chapter 2

III-V semiconductor porous structures show some new and attractive properties

compared to bulk materials such as:

1) A dramatic increase in the band-edge light emission in the photoluminescence

(PL) properties on porous semiconductor with indirect bandgap as shown in

Fig. 2-14 [12].

2) Blue-shift of near-band-edge emission in PL due to quantum confinement

effect [13,14].

3) A dramatic increase in the optical second harmonic generation (SHG) from

porous GaP membranes [15].

4) Appearance of jet-black surface due to extremely low reflectance in visible

region as shown in Fig. 2-15 [16].

5) Porosity-induced modification of refractive index [17].

6) Reduction of substrate-induced strain which is typically observed on GaN

grown on foreign substrate [18,19].

Table 2-I. Typical morphological characteristics of pores reported on Si and III-V semiconductors.

Porous structure can be classified into three groups: macroporous with pore diameters of less

than 2 nm; mesoporous with pore diameters between 2 nm and 50 nm; and macroporous with

pore diameters of greater than 50 nm.

Typical dia. (µm) Classification Supply way

of holes Preferred growth

direction

Si 0.0011−10 Macroporous Mesoporous Microporous

Illumination Breakdown

<100> <113>

GaAs 0.5−2 Macroporous Breakdown <111B>

InP 0.05−1 Macroporous Breakdown <111B> <001B>

GaN 0.05−0.2 Macroporous Mesoporous Illumination No preferred

orientation


2.4.3. Semiconductor photoelectrodes for water splitting

Since Fujishima and Honda reported water photoelectrolysis on TiO2 using UV

light in early 1970 [20], semiconductor photoelectrodes are getting much attention as

the method to convert solar energy into chemicals. Although there are two kinds of the

semiconductor photoelectrode consist of either a couple of a metal and a semiconductor

or a couple of two kinds of semiconductors, we show a former type of photoelectrode

used in our study.

A typical photoelectrode splits water into gaseous oxygen and hydrogen

molecules under the light with energy hν. The overall reaction is expressed as follows:

Figure 2-14. The room temperature PL spectrum obtained on porous GaP and crystalline

(non-porous) GaP [12].

Figure 2-15. Photograph of planar InP reference, and porous sample with and without

photo-assisted electrochemical etching [16].

28 Chapter 2

22

ν

2 O2

1HOH

h, (2-14)

which consists of a anodic (oxygen generation) reaction and a cathodic (hydrogen

generation) reaction as follows:

22 O2

12h2OH H , anodic reaction, (2-15)

2He22H , cathodic reaction. (2-16)

Figure 2-16 shows schematic representations of energy diagrams of a

semiconductor photoelectrode for water splitting: this cell consists of a couple of a

metal electrode (cathode) and an n-type semiconductor electrode (anode); EF(H+/H2) and

EF(O2/H2O) are the Fermi levels of redox reactions of hydrogen and oxygen, respectively.

For water splitting, EF(H+/H2) and EF(O2/H2O) need to be located within bandgap of the

n-type semiconductor electrode. Here, it is assumed that the Fermi level of hydrogen

reaction EF(H+/H2) locates lower energy state than the Fermi level of semiconductor EF at

flat band potential EFB and it is also assumed that the EF(H+/H2) locates higher energy state

than the Fermi level of the metal EF(M). When the cell circuit is connected (Fig. 2-16(a)),

Fermi level is balanced between metal and semiconductor to be equilibrium, resulting in

(a)


EF

EV

EF(M)

EF(H+/H2)

EF(O2/H2O)

EC

in dark under illumination


EF

EV

EF(M)

EC

electron

hole

O2

evolution

H2

evolution

hν

(b)

Figure 2-16. Schematics representations of water splitting by semiconductor photoelectrode: (a)

in dark, redox reactions are thermodynamically impossible; (b) under illumination, water is split

into gaseous hydrogen and oxygen.


formation of SCR in semiconductor/electrolyte interface. In the dark, anodic reaction

cannot flow because the concentration of holes is small; and no cathodic reaction is

expected thermodynamically at the metallic electrode because EF(M) is lower than the

EF(H+/H2).

Under the light with energy above the bandgap of semiconductor (Fig. 2-16(b)),

the EF is raised by an energy equivalent to the photo-potential ΔEph as described in

Section 2.3.3; in addition, the EF(M) is also raised because anode and cathode are short

circuited. This illumination-assisted rise of the Fermi level allows cathodic reaction to

proceed thermodynamically at metal electrode. Further, photo-excitation increases the

concentration of holes and the raises the quasi-Fermi level of holes EFp to a energy level

that is lower than the Fermi level of oxygen reaction EF(O2/H2O), which allows anodic

reaction to proceed thermodynamically at n-type semiconductor electrode.

Consequently, the water can be split into hydrogen at a metal electrode and oxygen at an

n-type semiconductor electrode.

To satisfy above requirements, many reports relevant to water splitting are

oxide semiconductors such as TiO2 [20,21], SrTiO3 [22], ZnO [23], α-Fe2O3 [24], WO3

[25], and so on [26]. Recently, group-III nitrides such as GaN, InN, AlN, and its alloys

are getting much attention as photoelectrode because bandgap energy can be varied

from about 0.65 eV to 6.0 eV by alloying which enables us to design various functional

photoelectrode not only for spectral matching for solar light but also for the

electrochemical reduction of CO2 to carbohydrate [27].

30 Chapter 2

Reference

[1] N. Sato, "Electrochemistry at Metal and Semiconductor Electrode", Elsevier

Science, Amsterdam, 1998.

[2] R. A. Marcus, "On the Theory of Oxidation-Reduction Reactions Involving

Electron Transfer. I", J. Chem. Phys., vol. 24, pp. 966−978, 1956.

[3] P. Delehay, and C. W. Tobias, "Advance in Electrochemistry and

Electrochemical Engineering, volume 1: Electrochemistry", Interscience

Publishers, New York, 1961.

[4] W. Schokley, "Electrons and Holes in Semiconductors", Van Nostrand,

Newyork, 1950.

[5] H. Gerischer, "The impact of semiconductors on the concepts of

electrochemistry", Electrochim. Acta, vol. 35, pp. 1677−1699, 1990.

[6] M. A. Butler, "Photoelectrolysis and physical properties of the semiconducting

electrode WO2", J. Appl. Phys., vol. 48, pp. 1914−1920, 1977.

[7] R. Memming, "Charge transfer kinetics at semiconductor electrodes", Ber.

Bunsenges. Phys. Chem., vol. 91, pp. 353−361, 1987.

[8] H. Gerischer, W. Mindt, "The mechanisms of the decomposition of

semiconductors by electrochemical oxidation and reduction", Electrochim.

Acta, vol. 13, pp. 1328−1341, 1968.

[9] A. R. Stonas, N. C. MacDonald, K. L. Turner, S. P. DenBaars, and E. L. Hu,

"Photoelectrochemical undercut etching for fabrication of GaN

microelectromechanical systems", J. Vac. Sci. Technol. B, vol. 19, pp.

2838−2841, 2001.

[10] C. Youtsey, G. Bulman, and I. Adesida, "Dopant-selective photoenhanced wet

etching of GaN", J. Electron. Mater., vol. 27, pp. 282−287, 1998.

[11] C. Youtsey, L. T. Romano, and I. Adesida, "Gallium nitride whiskers formed by

selective photoenhanced wet etching of dislocations", Appl. Phys. Lett., vol. 73,

pp. 797−799, 1998.


[12] K. Kuriyama, K. Ushiyama, K. Ohbora, Y. Miyamoto, and S. Takeda,

"Characterization of porous GaP by photoacoustic spectroscopy: The relation

between band-gap widening and visible photoluminescence", Phys. Rev. B, vol.

58, pp. 1103−1105, 1998.

[13] L. T. Canham, "Silicon quantum wire array fabrication by electrochemical and

chemical dissolution of wafers", Appl. Phys. Lett., vol. 57, pp. 1046−1048,

1990.

[14] T. Fujino, T. Sato, and T. Hashizume, "Size-Controlled Porous Nanostructures

Formed on InP(001) Substrates by Two-Step Electrochemical Process", Jpn. J.

Appl. Phys., vol. 46, pp. 4375−4380, 2007.

[15] I. M. Tiginyanu, I. V. Kravetsky, G. Marowsky, and H. L. Hartnagel,

"Efficient Optical Second Harmonic Generation in Porous Membranes of GaP",

Phys. Status. Solidi A, vol. 175, pp. R5−R6, 1999.

[16] T. Sato, N. Yoshizawa, and T. Hashizume, "Realization of an extremely low

reflectance surface based on InP porous nanostructures for application to

photoelectrochemical solar cells", Thin Solid Films, vol. 518, pp. 4399-4402,

2010.

[17] E. Kikuno, M. Amiotti, T. Takizawa, and S. Arai, "Anisotropic Refractive

Index of Porous InP Fabricated by Anodization of (111)A Surface", Jpn. J. Appl.

Phys., vol. 34, pp. 177−178, 1995.

[18] A. P. Vajpeyi, S. J. Chua, S. Tripathy, and E. A. Fitzgerald, "Effect of carrier

density on the surface morphology and optical properties of nanoporous GaN

prepared by UV assisted electrochemical etching", Appl. Phys. Lett., vol. 91, p.

083110, 2007.

[19] K. Al-Heuseen, M. R. Hashim, and N. K. Ali, "Growth and Characterization of

Tree-Like Crystalline Structures during Electrochemical Formation of Porous

GaN", J. Electrochem. Soc., vol. 158, pp. D240−D243, 2011.

[20] A. Fujishima, and K. Honda, "Electrochemical Photolysis of Water at a

Semiconductor Electrode", Nature, vol. 238, pp. 37−38, 1972.

[21] G. K. Mor, K. Shankar, M. Paulose, O. K. Varghese, and C. A. Grimes,

32 Chapter 2

"Enhanced Photocleavage of Water Using Titania Nanotube Arrays", Nano

Lett., vol. 5, pp. 191−195, 2005.

[22] I. E. Paulauskas, J. E. Katz, G. E. Jellison Jr., N. S. Lewis, and L. A. Boatner,

"Photoelectrochemical studies of semiconducting photoanodes for hydrogen

production via water dissociation", Thin Solid Films, vol. 516, pp. 8175−8178,

2008.

[23] K. Keis, E. Magnusson, H. Lindstrӧm, S. -E. Lindquist, and A. Hagfeldt, "A

5% efficient photoelectrochemical solar cell based on nanostructured ZnO

electrodes", Sol. Energy Mater. Sol. Cells, vol. 73, pp. 51−58, 2002.

[24] K. Sivula, F. L. Formal, and M. Grätzel, "Solar Water Splitting: Progress Using

Hematite (α-Fe2O3) Photoelectrodes", ChemSusChem, vol. 4, pp. 432−449,

2011.

[25] Q. Mi, A. Zhanaidarova, B. S. Brunschwig, H. B. Gray, and N. S. Lewis, "A

quantitative assessment of the competition between water and anion oxidation

at WO3 photoanodes in acidic aqueous electrolytes", Energy Environ. Sci., vol.

5, pp. 5694−5700, 2012.

[26] I. E. Paulauskas, G. E. Jellison, and L. A. Boatner, "Photoelectrochemical

properties of n-type KTaO3 single crystals in alkaline electrolytes", J. Mater.

Res., vol. 25, pp. 52−62, 2010.

[27] S. Yotsuhashi, M. Deguchi, Y. Zenitani, R. Hinogami, H. Hashiba, Y. Yamada,

and K. Ohkawa, "Photo-induced CO2 Reduction with GaN Electrode in

Aqueous System", Appl. Phys. Express, vol. 4, p. 117101, 2011.

33

Chapter 3

Experimental technique

3.1. Introduction

Chapter 3 gives a brief overview of experimental techniques for formation and

characterization of III-V semiconductor microstructures. As previously described,

electrochemical (EC) process was used for the formation process of microstructures. EC

cell in which redox reactions take place, and a potentiostat which is needed to control

electrode potential precisely, are first described, followed by EC characterization

techniques in Section 3.2. III-V semiconductors after EC process are characterized by

various techniques: the structural characteristics were evaluated by scanning electron

microscopy (SEM) and atomic force microscopy (AFM) described in Section 3.3; the

optical characteristics were evaluated by ultraviolet-visible-infrared (UV-VIS-IR)

spectroscopy, photoluminescence (PL) spectroscopy described in Section 3.4; and the

electrical characteristics were evaluated by current-voltage (I-V) measurements, and

capacitance-voltage (C-V) measurements described in Section 3.5.

3.2. Electrochemical process and characterization

3.2.1. Three-electrode electrochemical cell and potentiostat

Most common EC cell setup used in electrochemistry is three-electrode EC cell

consists of a working electrode (WE), a reference electrode (RE), and a counter

electrode (CE) [1]. Figure 3-1(a) shows the schematic representation of three-electrode

EC cell used in this work. The roles of each electrode are described in following

paragraphs.

The WE is the electrode where the reaction of interest occurs. For etching

applications, the material set on a WE is actually etched with oxidation reactions. To

avoid unnecessary EC reactions, the low resistance materials except for objective

materials should be covered by insulated materials, as which Apiezon wax is used in

34 Chapter 3

this study, as schematically shown in Fig. 3-1(b).

The CE is the electrode used to close the circuit of the EC cell. An inert and

metallic material is usually chose as CE, as which platinum is used in this study, and it

does not usually dissolve into solution by EC reactions. Since the current flows between

a WE and a CE, it should not be a limiting factor in the kinetics of the EC process under

investigation.

The RE is the electrode used as a potential reference in the EC cell. To obtain

high stability of reference potential, a redox system with constant (buffered or saturated)

concentrations of each participants of the redox reaction is usually employed. In this

study, we adopted silver/silver chloride (Ag/AgCl) electrode immersed in saturated

potassium chloride (KCl). A typical Ag/AgCl electrode is consists of a Ag wire, the tip

of which is coated with a thin film of AgCl, immersed in saturated KCl. A porous plug

serves as the salt bridge. The Ag/AgCl electrode is based on the following overall redox

reaction:

(aq)ClAg(s)eAgCl(s) , (3-1)

in which the activity of Cl− determines the potential of the Ag/AgCl electrode. In the

case of Ag/AgCl electrode immersed in saturated KCl, the redox potential is +0.197 V

at 25°C.

Figure 3-1. Schematic representations of (a) three-electrode electrochemical cell used in this

work, and (b) top and cross-sectional schematics of working electrode (WE).

35 Experimental technique

In the three-electrode EC cell used in this study, a potentiostat with a Princeton

Applied Research VersaSTAT 4 controls the potential of a CE against a WE accurately

so that the potential difference between a WE and a RE corresponds to the setting value.

Figure 3-2 shows a schematic circuit diagram of a potentiostatic mode with a

VersaSTAT 4. In this mode, the VersaSTAT 4 controls the potential at the WE with

respect to the RE. The CE is driven to the potential required to establish the desired

WE's potential. The range over which the WE's potential can be controlled is ± 10 V.

3.2.2. Linear sweep and cyclic voltammetry

Linear sweep voltammetry and cyclic voltammetry is one of the most widely

used EC techniques. These measurements can be employed to study the electron

transfer kinetics and transport properties of electrolysis reactions.

In linear sweep voltammetry, the electrode potential is varied at a constant rate

throughout the scan and the resulting current is measured. Obtained current-voltage

curve is called linear sweep voltammogram in which the applied potential is plotted on

the x-axis and the resulting current on the y-axis. In cyclic voltammetry, the electrode

Figure 3-2. Schematic circuit diagram of a potentiostatic mode with a Princeton Applied

Research VersaSTAT 4.

CE

RE

WE

CELLSWITCH

POWERAMP

ELECTROMETER

POSITIVEFEEDBACK

CURRENTRANGE

PFIR

I/ECONVERTER

IREPORT

EREPORT

GROUND

PSREF

36 Chapter 3

potential is varied at a constant rate as is the case with linear sweep voltammetry, and

then reverses the scan, returning to the initial potential (repeatedly or not). Obtained

current-voltage curve is called cyclic voltammogram. The characteristics of the linear

sweep voltammogram recorded on a semiconductor depend on a number of factors

including: carrier type; carrier density; flat-band and band-edge potential; rate of carrier

transfer or ion transfer; and so on. In addition, the characteristics of the cyclic

voltammogram reflect morphological and electrical change of semiconductor electrodes

between repeated cycles.

3.2.3. Photo-electrochemical measurement

The semiconductor electrodes in photo-excited states show photo-effect on

redox reactions distinctly as described in Chapter 2. Since variation of the amount of

redox reactions leads to the variation of redox currents, voltammograms measured in

photo-excited states reflect photo-effect on redox reactions. In this study, measuring

voltammetry at semiconductor electrodes in photo-excited states is called

photo-electrochemical measurements. Figure 3-3 shows the simple description of

photo-effect in semiconductor, which described as follow: (a) incident of photons; (b)

Figure 3-3. The simple description of photo-effect in semiconductor: (a) incident of photons; (b)

absorption and excitation of minority charges (electron-hole pairs); (c) separation of electron-hole

pairs and transport toward electrolyte/semiconductor interface; (d) interfacial carrier transfer

across the semiconductor/electrolyte interface.

EC

EV


electron

hole

(a)

(b)

(c)

(c)

(d)

OX

RED


absorption and excitation of minority charges (electron-hole pairs); (c) separation of

electron-hole pairs and transport toward electrolyte/semiconductor interface; (d)

interfacial carrier transfer across the semiconductor/electrolyte interface. The

percentage of incident photons (Fig. 3-3(a)) converted to redox current (corresponding

to redox reactions as shown in Fig. 3-3(d)) is called incident photon to current

efficiency (IPCE) describing as following equation [2]:

100λ

IPCE(%)IN

darkph

Pq

JJhc, (3-2)

where h is Planck’s constant, c is speed of light, q is elementary charge, PIN is light

intensity, and Jph and Jdark are current density under irradiation and dark condition,

respectively. Thus we can calculate IPCE from the results measuring current of

electrodes under monochromatic light.

Since IPCE is affected by the efficiencies of each component in photo-effect, it

gives combined information about these components. In the case of III-V

semiconductors under the light with energy (= 1240/λ (nm)) below the bandgap, IPCE is

limited by efficiency of photo-absorption (Fig. 3-3(b)): it is almost zero. Under the light

with energy above the bandgap, efficiency of carrier separation usually limits the IPCE.

Since photo-carriers are separated in space charge region (SCR) at

electrolyte/semiconductor interface, the relationship between SCR thickness and light

penetration depth (inverse of absorption coefficient α) is important: if the light

penetration depth is much smaller than SCR thickness, most of photo-carriers can

contribute to redox reactions; If the light penetration depth is larger than SCR thickness,

a part of photo-carriers can contribute to redox reactions, others are recombined and

cannot contribute to redox reactions.

3.3. Structural characterization

3.3.1. Scanning electron microscopy (SEM)

Scanning electron microscopy (SEM) is a commonly used tool to characterize

structural properties. The spatial resolution of SEM is generally much higher as

compared to a state-of-the-art optical microscope because the wavelength of the

38 Chapter 3

electrons for typical acceleration voltages used in the SEM is significantly shorter than

that of visible light [3]. Figure 3-4 shows a simple schematic representation of an SEM,

which basically consists of an electron gun, condenser lenses, an aperture, scanning

coils, objective lens, and a specimen chamber. It is important to keep high vacuum in

the electron column to prevent electrical discharge, prolong the electron source lifetime,

and reduce un-wanted collisions between electrons and air molecules. The electron

beam trajectory and focus are further controlled by a set of magnetic condenser lenses,

which focal length is controlled by varying the current through coil. When the electron

beam hits the sample, it creates a certain excitation volume. Depending on the

acceleration voltage and the atomic density of the specimen, the electron beam will

penetrate the sample to different depths. For instance, a higher voltage (lower density)

will increase the excitation penetration depth. The electrons are also scattered from their

original trajectories leading to a three-dimensional excitation volume.

The interaction between the electron beam and the sample will result in

different kind of signals used for detection. The electrons will interact elastically and

inelastically with a sample, generating primary (such as backscattered) and secondary

electrons (such as secondary and auger electrons), respectively. The backscattered

Figure 3-4. Simple schematic representation of an SEM, which basically consists of an electron

gun, condenser lenses, an aperture, scanning coils, objective lens, and a specimen chamber.

e-sample

Electron gun

Condenser lens

evacuation

Scanning coil

Objective lens

Secondary electrondetector

ele

ctro

n b

ea

m


electrons are affected by the vicinity of a nucleus, hence altering its trajectory but

without losing much velocity. Some of these electrons are scattered back towards the

electron source (backscattered electrons), and can be detected by a detector above the

sample. The probability for scattering increases with atomic number, and lager fraction

is scattered by a heavier element. These electrons, consequently, contain the information

about the sample composition (atomic number contrast). The secondary electrons are

created after the collisions with a nucleus, which typically are sample response signals

that most commonly used for the detection in SEM. These electrons contain

topographical information, as they are typically emitted from a depth of a few tens of

nanometers.

In this study, a Hitachi SU-8010 system is used. The objective lens is

Semi-in-lens type and the electron gun is field emission type, enabling high resolution

of 1.0 nm.

3.3.2. Atomic force microscopy (AFM)

The atomic force microscopy (AFM), which is one of a scanning probe

microscopy (SPM), provides three-dimensional surface topography by detecting the

attractive and repulsive forces between a cantilever and an object surface [4]. Though

the lateral resolution of AFM is relatively low, the vertical resolution can be up to

sub-angstrom. The interaction between a cantilever and an object is monitored by an

optical probe to read a minute Z-direction displacement: changes in the cantilever

deflection or oscillation amplitude are determined by detecting the changes in the laser

beam reflected on the back of the cantilever. Z-axis drive of the cantilever can be

accurately operated by using a piezo device.

The operation modes of the AFM systems can be classified into two groups:

"contact mode" and "tapping mode". In the contact-mode AFM, cantilever deflection

affected by contacting a cantilever on an object are utilized to provide a topography. The

change in cantilever deflection is monitored by the optical probe while raster-scanning

the object. A cantilever deflection is kept constant by adjusting the vertical position of

scanner to maintain an optical signal constant. This feedback loop maintains a constant

force between a cantilever and an object during imaging. In the tapping-mode AFM,

cantilever oscillation during tapping the tip on the surface is utilized to provide a

topography. The optical probe monitors the root-mean-square (RMS) amplitude of

40 Chapter 3

cantilever oscillation. An oscillation amplitude is kept constant by adjusting the vertical

position of scanner. This feedback loop maintains a constant distance between a

cantilever and an object during imaging. Since tapping mode AFM can provide a

topography without contacting an object sample, no damage is induced during imaging.

Furthermore, the lateral, shear forces present in contact mode can be eliminated. These

features enable us to obtain the topography of soft, fragile, and adhesive surfaces

without damaging them.

In this study, we used OLYMPUS LEXT4500 which integrating optical

microscope and an AFM system, enabling to obtain both microscopic image and

topological image of the area of interest. AFM was operated at tapping mode to obtain

the surface topography without damaging on samples. Silicon tips (OLYMPUS

OMCL-AC240TS-C3) with radius of 7 nm and resonance frequency of 70 kHz were

used as cantilever.

3.4. Optical characterization

3.4.1. Ultraviolet-visible-infrared (UV-VIS-IR) spectroscopy

Ultraviolet-visible-infrared (UV-VIS-IR) spectroscopy is a non-contact,

nondestructive method of probing the interaction between objects and light. Each

semiconductor has the bandgap with energy in UV-VIS-IR region, resulting in specific

interaction with light in UV-VIS-IR region [5]. When the monochromatic light incidents

on the semiconductors, light is reflected, absorbed, and transmitted as shown in Fig.

3-5(a). In such situation, relationship between reflectance R, absorptance A, and

transmittance T is described as follow:

1 TAR . (3-3)

Following paragraphs explain the each parameter in detail.

Considering the case in which light incident from medium 1 to medium 2 as

shown in Fig. 3-5(b), reflectance R is calculated for p-polarized light and s-polarized

light by following equations, respectively:


2

1221

2112p θθ

θθ

cosncosn

cosncosnR , (3-4)

2

2211

2211s θθ

θθ

cosncosn

cosncosnR , (3-5)

where n1 and n2 are refractive index of medium 1 and medium 2, θ1 and θ2 are angle of

incidence and refraction. These are called Fresnel's equations. In the case of secular

reflection (θ1 and θ2 is zero degrees),

2

21

12Sp

nn

nnRRR . (3-6)

For the sake of simplicity, we consider the case of specular spectroscopic measurement

of semiconductor in air. In such situations, Eq. (3-6) can be deformed as follow:

2

1

1

n

nR , (3-7)

Figure 3-5. Schematic representations of (a) interaction between semiconductor and light

(reflection, absorption, and transmission), and (b) the case in which light incident from medium 1

to medium 2 with incident angle θ1 and reflection angle θ2.

medium 1 (n1)

θ1 θ2

ref lected light(Rp, Rs)

incident light

medium 2 (n2)

ref lectance

transmittance

incidentlight

semiconductor

absorptance

(a) (b)

42 Chapter 3

where n is refractive index of the semiconductor.

Next, we consider the light penetrating into semiconductor. Penetrating photon

flux at surface Φ0 is describes as follow:

hc

PR

E

PRR ININ

IN0λ

11Φ1Φ , (3-8)

where h is Planck’s constant, c is speed of light, and PIN is light intensity. Since the

photons are absorbed during progress in the semiconductor which has specific

absorption coefficient α, photon flux Φx at the distance from the surface x is:

xexp αΦΦ 0x , (3-9)

suggesting that photons penetrating in the semiconductor decrease exponentially with

distance.

Finally, we consider transmitted light which is the light emitted from the

semiconductor to air, in other words, transmitted light. Note that light is reflected at the

semiconductor/air interface. If the thickness of the semiconductor is assumed to be W,

the transmittance T is calculated as follow:

WexpE

PRWexpRRT α1αΦ1Φ1 IN2

0W , (3-10)

where ΦW is the photon flux at back side of the semiconductor. Thus, spectroscopic

light has information about physical properties of the semiconductor such as refractive

index, absorption coefficient, as well as geometric features.

In this study, we employed Shimadzu UV-1700 UV-VIS Spectrophotometer

with following specifications: spectral band width of 1 nm; wavelength range of 190 to

1100 nm. Light sources are halogen lamp and deuterium lamp, and photometric

system is double-beam method.

3.4.2. Photoluminescence (PL) spectroscopy

Photoluminescence (PL) spectroscopy is a non-contact, nondestructive method

of probing the electronic structure of materials [6]. When the light with energy above


the bandgap is irradiated onto a semiconductor, it is absorbed and electrons move into

permissible excited states, which is called the photo-excitation. When these electrons

return to their equilibrium states, which is called the recombination, the excess energy is

released either radiatively or non-radiatively. The light emitted through this process is

called photoluminescence (PL). The energy of the emitted light is relative to the

difference in energy levels between the excited state and the equilibrium state involved

in the radiative process. Most of the energy of the emitted light through the radiative

transition correspond to bandgap, others correspond to point defects such as vacancies,

impurities, and their complexes in the case of III-V semiconductors as shown in Fig.

3-6.

Figure 3-6. Schematic representations of various transition processes in semiconductors: (a)

direct band-to-band transition; (b) bound to free transition; (c) donor to acceptor pair transition;

(b) free to bound transition; (e) transition via localized state resulted from point defects.

CB

VB

donor state

acceptor state

trap(a)

(b)(c)

(d)

(e)

(e)

Figure 3-7. Schematic representations of PL measurement system. The He-Cd laser was

employed as light source with wavelength λ of 325 nm.

44 Chapter 3

The PL measurement system used in this study is illustrated schematically in

Figure 3-7. The He-Cd laser (KIMMON KOHA) emitting the light with wavelength of

325 nm was used as the excitation source. The laser light were focused to several

micro-meter in diameter by objective lens. The CCD camera combined with LED was

used to define the characterized area. The long pass filter was inserted to cut the light

with wavelength below 325 nm including reflected light of excitation laser.

3.5. Electrical characterization

3.5.1. Current-voltage (I-V) measurement

In this study, we measured current-voltage (I-V) characteristics of

Ni/AlGaN/GaN diodes and HEMTs after EC process. We employed Agilent B1500A

Semiconductor Device Analyzer in which built-in HDD and DVD drive, Windows 7 are

embedded. The instrument is totally operated with Agilent EasyEXPERT software

installed in B1500A itself. IV measurement module is Keysight B1517A High

Resolution Source/Measure Unit with following specification: source and measurement

range up to 100 V, 0.1 A; minimum measurement resolution of 1 fA, 0.5 µV; minimum

source resolution of 5 fA, 25 µV.

3.5.2. Capacitance-voltage (C-V) measurement

In this study, we measured capacitance-voltage (C-V) characteristics of

Ni/AlGaN/GaN diodes after EC process. We employed Yokogawa-Hewlett-Packard

4192A Impedance Analyzer. The instrument is connected to the external laptop

computer using GPIB interface. Measurements are automatically carried out using a

control-software which is developed with National Instrument LabVIEW. The

measurement duration per point and resolution are approximately 1.1 s and 0.1 pF,

respectively.


Reference

[1] J. Bard, and L. R. Faulkner, "Electrochemical Methods, Fundamentals and

Applications", John Wiley & Sons, New York, 2013.

[2] Z. Chen, H. N. Dinh, and E. Miller, "Photoelectrochemical Water Splitting:

Standards, Experimental Methods, and Protocols", Springer-Verlag New York,

New York, 2013.

[3] J. I. Goldstein, and H. Yakowitz, "Practical Scanning Electron Microscopy:

Electron and Ion Microprobe Analysis", Springer US, New York, 1975.

[4] G. Binnig, and C. F. Quate, "Atomic Force Microscope", Phys. Rev. Lett., vol.

56, pp. 930−933, 1986.

[5] B. L. Anderson, and R. L. Anderson, "Fundamentals of Semiconductor

Devices", McGraw-Hill Companies, New York, 2005.

[6] G. D. Gilliland, "Photoluminescence spectroscopy of crystalline

semiconductors", Mater. Sci. Eng., vol. R18, pp. 99−400, 1997.

46 Chapter 3

47

Chapter 4

Formation and optical characterization of highly

ordered pore arrays on InP for photo-electric

conversion devices

4.1. Introduction

The formation of semiconductor nanostructures targeting applications such as

quantum and optoelectronic devices has been intensely researched. Among the various

structures, porous structures, which are a high-density array of nanometer-sized pores

formed by using an electrochemical process as described in Chapter 2, are one of the

most promising nanostructures due to their unique features such as their large surface

and low optical reflectance [1]. Anodic porous etching on Si and Ge was first reported

by A. Uhlir at Bell Labs in 1956 [2]. After that, various compound semiconductors, such

as GaAs [3−5], InP [6−9], GaP [10,11], GaN [12−14] and SiC [15−17] were studied.

For porous InP in particular, it is known that straight pores can be uniformly formed in

the vertical direction under optimal conditions [18,19]. In addition, it has been reported

that an extremely low reflectance of below 0.4% was observed for porous InP in the UV,

visible, and near-infrared ranges [20]. These findings suggest that porous structures are

promising materials for use in photoelectric conversion devices such as solar cells

[21,22] and photo detectors [23]. Therefore, the optical absorption properties of porous

structures need to be determined so they can be used in such devices. It was previously

reported that the absorption efficiency is enhanced after the formation of porous

structures in indirect bandgap materials such as GaP [24−26] and SiC [27]. However,

there are few reports on direct bandgap materials. One report showed that the absorption

efficiency of InP decreased after the formation of porous structures due to the

high-density surface states and lower electric field of pore wall [28].

In Chapter 4, we aimed at clarifying the optical absorption properties of these

InP porous structures by conducting photocurrent measurements of photoelectric

conversion (PC) devices. First, we characterized optical absorption properties by using

48 Chapter 4

PC devices consist of top porous layer and p-n junction where light converted into

electric signals. We also fabricated Schottky junction PC devices based on

platinum/porous InP with large specific junction area.

4.2. Fabrication of InP porous structures

The experimental procedure and the applied waveform used in this study are

schematically shown in Fig. 4-1. The porous structures were first formed by an EC

reaction in the dark, and then, the photo-assisted EC etching of the porous surface was

performed in the same electrolyte under illumination. A series of EC processes was

conducted using a three-electrode EC cell as described in Chapter 3. The electrolyte

consisting of 1 M HCl (200 ml) with HNO3 (3 ml) was used throughout the experiment.

The constant bias was firstly applied to the semiconductor electrode to obtain high

density porous structures. After the pore formation, the photo-assisted EC etching was

carried out to remove the irregular top layer under illumination using a tungsten lamp

with an intensity of 10 mW/cm2. The anodic bias VEC was applied in ramped mode in

order to control the etching depth by cycle number of the ramped bias, as shown in Fig.

4-1(b). The VEC was swept at a rate of 50 mV/s starting from 0 V to 1 V in the positive

direction first, and then turning in the negative one.

Figure 4-2 shows plan-view scanning electron microscope (SEM) images of

(a) the template porous sample formed at VEC = 5 V for 60 s, the samples after the

ramped bias applied with (b) 3 cycles and (c) 6 cycles. The average pore diameter of

the present template was about 30 nm, as shown in Fig. 4-2(a). After the photo-assisted

Figure 4-1. Schematic illustrations of (a) experimental procedure and (b) applied waveform used

in this study.

(a) (b)

VEC (V)

Dp

Photo-assistedEC etching

Pore formation byEC etching

poreformation

illumination

Photo-assistedEC etching

illumination

49 Formation and optical characterization of highly ordered pore arrays on InP forphoto-electric conversion devices

EC etching, the surface morphology was changed and the pore diameter increased. As

shown in Fig. 4-2(c), the pore shape became a square defined by four equivalent 100

planes, whose feature is very similar to the case of the porous structures after the

cathodic decomposition process [29]. This is because the etching rate of the InP was

strongly dependent on crystal orientation, where the 100 planes preferentially

appeared on the wall surface due to their slow etching rate. After the ramped bias

applied with 6 cycles, the lateral thickness of the InP nanowalls was below 20 nm near

the surface, which was thinner than the initial value of the template porous sample.

These results indicate that the etching of the porous surface has been developed by the

photo-assisted EC etching. From the cross-sectional SEM image, it was found that the

straightness in vertical direction was improved after the photo-assisted EC etching.

Figure 4-3 compares the cross-sectional SEM images of the template porous sample

and the sample after the ramped bias applied with 6 cycles. Template porous sample has

Figure 4-2. Plan-view SEM images of porous samples (a) before and after photo-assisted EC

etching with (b) 3 cycles and (c) 6 cycles.

(a) (b) (c)

w/o photo-assisted ECetching

w/ photo-assisted ECetching (3 cycles)


Figure 4-3. Cross-sectional SEM images of porous samples before (left) and after (right)

photo-assisted EC etching.

w/o photo-assisted ECetching


50 Chapter 4

irregular layer partially on the top of the ordered porous structure. After the

photo-assisted EC etching, the irregular top layer was completely removed and the

regular array of straight pores appeared on the surface, as shown in Fig. 4-3. The initial

thickness of the porous structure was 27.6 µm including the irregular top region with a

thickness of 4.1 µm. After the ramped bias applied with 6 cycles, the thickness of

porous layer decreased to 22.9 µm due to the photo-assisted EC etching developed on

the surface.

The etching rate of the InP porous structures was investigated. Figure 4-4

shows the plot of the pore depth Dp measured as a function of the cycle number of the

ramped bias applied during photo-assisted EC etching. The measured anodic current

obtained on the InP porous electrode is also shown as a solid curve. It was found that

the surface of the InP porous structure started to be etched at the rate of 1.2 µm/cycle

and later in a gradual decline. This behavior is very consistent with the current transition,

as shown in Fig. 4-4. It appears certain that the etching rate became lower and lower

along with the anodic currents contributing the EC reaction of InP porous surface. In

this study, the irregular top layer with a thickness of 4.1 µm was completely removed

after the ramped bias applied with 5 cycles. The decline of the etching rate can be

explained by the fact that the amount of anodic photocurrents was strongly associated

Figure 4-4. Relationship between average pore depth Dp and cycle number of ramped bias

applied to sample. Anodic currents measured during photo-assisted EC etching is also shown as

a solid curve.

Dp

eye guide


with the surface morphology of the InP porous electrode [30]. At the beginning of the

photo-assisted EC etching, the surface etching started with a large photocurrent due to

the hole generation efficiently caused by the illumination in the irregular pore region. As

the photo-assisted EC etching proceeds, the pore diameter increased and the thickness

of InP nanowalls decreased, as shown in Fig. 4-2. In that case, the etching rate declined

because the effective area subjected to illumination became very small on the surface

with InP nanowalls. In addition to this, the etching rate of the surface further declined as

the 100 facets appeared on the wall surface during the photo-assisted EC etching.

This model qualitatively explains the observed behavior on the photo-assisted EC

etching of the InP porous structures. These results indicate that the irregular top layer

can be completely removed from the surface of InP porous structures by photo-assisted

EC etching controlling and monitoring the photocurrents.

4.3. Optical absorption properties of InP porous structures

4.3.1. Photo-electric conversion (PC) devices formed on p-n junction substrates

The device structure and the experimental setup for the photoelectrical

measurements are schematically shown in Fig. 4-5. This device consists of a top layer

such as a porous or non-porous layer, three ohmic electrodes, and a p-n junction with an

n-type InP layer (ND = 8 × 1017 cm−3) grown on a highly doped p-type substrate. SEM

images of the top and cross-section of the porous structures are shown in Fig. 4-6(a),

Figure 4-5. Schematic illustration of PC device structure and experimental setup for

photoelectric measurement.

I1

I2

Illumination(λ = 514.5 nm)

52 Chapter 4

and the photo image of the top of the PC device is shown in Fig. 4-6(b). The thickness

of the top layer dtop of porous and non-porous devices were 4.3 μm and 2 ~ 5 μm,

respectively. After the formation of the porous structures, the sample was partly etched

into a convex shape at a width of 1 mm using photolithography and a wet etching

process, in which the GeAu/Ni ohmic contacts were formed.

The optical absorption properties in this study were investigated by measuring

the photocurrents, I1 and I2, shown in Fig. 4-5. The measurements were carried out by

changing the irradiation-light power using an Ar+ ion laser at a wavelength of 514.5 nm.

Only the carriers excited near the p-n interface are separated by the electric field in the

depletion layer and collected by the electrodes. Since such carriers are excited by the

photons reaching the p-n interface through the top layer, the photocurrents provide us

with information on the optical absorption properties of the top layer. If high

photocurrents are observed, it indicates that the absorbance of the top layer is low.

However, the observation of low photocurrents indicates there is high absorbance of the

top layer. Thus, the optical absorption properties of the porous device can be evaluated

by comparing them with those of the non-porous device formed as a reference.

Figure 4-6. (a) Top and cross-sectional SEM images of InP porous structures, and (b) top photo

image of a PC device.

(a)

(b)

Illuminationarea


4.3.2. Basic operation properties of PC devices

The photocurrent measurements were carried out using the non-porous device,

which had no porous structures in the top layer, to clarify the basic operation properties

of our PC device. Figure 4-7 shows the current response properties of the non-porous

device measured using a light intensity PIN of 2500 μW, which was turned on at t = 500

s and off at t = 1500 s. As shown in Fig. 4-7, the currents, I1 and I2, respectively

measured by the back and top electrodes, responded to the light irradiation. Here, I1 was

positive due to the collection of holes by the back electrode. On the other hand, I2 was

negative due to the collection of electrons by the top electrode. There was no current

decay under the light irradiation, indicating that stable photocurrents with quick

responses were observed. When the ΔI1 and ΔI2 are defined as the photocurrents

measured by the back and top electrodes, respectively, we found that the ΔI1 was

approximately two times as large as that for the ΔI2.

The current ratio ΔI2/ΔI1 was measured on the non-porous device as a function

of the position of the light irradiation x defined in Fig. 4-6(b) in order to further clarify

the operation principle of the present PC device. As shown in Fig. 4-8, the ratio of

ΔI2/ΔI1 linearly increased as a function of x, and became nearly 50 % at x = 0. This

behavior can be explained by taking the collection rate of the photo-carriers in the top

and back electrodes into consideration. The excited electrons due to the photon

Figure 4-7. Current response properties of non-porous device with light intensity PIN of 2500

µW/cm2, which turns on at t = 500 s and off at t = 1500 s.

PIN = 2500 μW/cm2

ΔI1

ΔI2

54 Chapter 4

absorption were collected by the top two electrodes, A and B, whereas the holes were

collected by the one back electrode. In terms of the current continuity, the summation of

the currents observed in the top two electrodes should be equal to the current observed

in the back electrode. Since the irradiation point was close to electrode B, namely at a

small x, the excited electrons were preferentially collected in electrode B rather than in

electrode A. In such a situation, the proportion of the ΔI2 to the total current decreased,

resulting in a small ΔI2/ΔI1, as shown in Fig. 4-8. As the irradiation point moved

towards electrode A, the ΔI2/ΔI1 expectedly increased with the increase in the ΔI2

because the excited electrons were preferentially collected in electrode A. The linear

relationship shown in Fig. 4-8 indicates that the excited electrons near the p-n interface

were divided in rough proportion to the distance from the light irradiation points to each

top electrode. We found from this result that the light irradiation point can be adjusted

by monitoring the photocurrent ratio. In the following experiment, the ΔI2/ΔI1 value was

set at 50 % to irradiate the light at the center of the PC device.

Then, the effect of the top layer on the photocurrent response was investigated

using the non-porous PC device. Figure 4-9 shows the photocurrents ΔI1 measured by

changing the thickness of the top layer dtop defined in Fig. 4-5. We found that the ΔI1

exponentially increased with a decrease in dtop. The correlation coefficient R obtained

by the exponential fitting on the experimental data was 0.989. The number of

photo-carriers contributing photocurrents can be considered in proportion to the photon

flux reaching the p-n interface. The photon flux Φ is given by the following equation:

Figure 4-8. Correlation between current ratio ΔI2/ΔI1 and position of light irradiation x measured

on non-porous device.

PIN = 2500 μW/cm2

Pho

tocu

rren

t rat

io Δ

I 2/Δ

I 1(%

)

Position of light irradiation x (mm)


dαΦΦ 0 exp . (4-1)

In Eq. (4-1), α is the absorption coefficient, d is the optical-path length, and Φ0 is the

incident photon flux given by,

ININ

0λ

Φ PhcE

P , (4-2)

where h, λ, and c are the Planck's constant, the wavelength, and the velocity of the

incident light, respectively. The light intensity PIN and the product of the absorption

coefficient α and the optical-path length d are important for the discussion of optical

absorption properties. Since the photocurrents ΔI1 are assumed in proportion to the

photon flux Φ the following equation is derived by using Eqs. (4-1) and (4-2):

topIN

1 αλ

Δ dexphc

PI . (4-3)

The results shown in Fig. 4-9 are very consistent with that from using Eq. (4-3), where

the photocurrents, ΔI1, are described as an exponential function of the geometrical

thickness of the top layer dtop. We found from these results that the photocurrents

changed under the influence of the number of photo-carriers generated near the p-n

Figure 4-9. Photocurrent ΔI1 plots measured on non-porous device as function of thickness of

top layer dtop.

PIN = 2500 μW/cm2

Pho

tocu

rren

t ΔI 1

(µA

)

56 Chapter 4

interface, which gave us the information on the optical absorption properties of the top

layer.

4.3.3. Comparison between porous and non-porous devices

The photocurrent response of present PC devices was compared by changing

the top layer such as the non-porous and porous layers. Figure 4-10 shows the current

response of a porous PC device to the incident light at various power levels PIN. We

found that the photocurrents ΔI1 increased with the PIN in quick response to the light

switching. This is consistent with the theoretical description of the photocurrents

obtained using Eq. (4-3). For more discussion, the ΔI1 of non-porous and porous

devices are compared in Fig. 4-11, plotted as a function of the PIN. Both the non-porous

and porous devices showed a similar behavior, the ΔI1 linearly increased with PIN. The

correlation coefficients obtained by linear fitting on the experimental data were greater

than or equal to 0.998. However, the current value differed substantially, where the

photocurrent of the porous device was approximately 40 % that of the non-porous

device. As mentioned above for Eq. (4-3), a low photocurrent indicates a high

absorbance in the top layer since the photo-carriers generated near the p-n interface

decreased exponentially with an increase in the absorption coefficient α. These results

suggest that the absorption coefficient of the porous layer is higher than that of the

non-porous layer.

Figure 4-10. Current response properties of porous PC device based on incident light at various

light intensity PIN.

PIN = 2500 μW/cm2

2000 μW/cm2

1500 μW/cm2

500 μW/cm2

1000 μW/cm2


A previous study reported that a smaller absorption coefficient was obtained in

the porous InP, as compared with a bulk InP [28]. According to the literature, the

absorption coefficient was estimated from the photocurrents observed at the

electrolyte/InP interface inside the porous layer. The low photocurrents observed in the

porous layer were explained in terms of the surface states and lower electric field at the

pore walls. Since the surface states at the pore walls act as recombination centers, the

photocurrents observed at the electrolyte/InP interface decreased, resulting in the

underestimation of the absorption coefficient. In addition to this, the low built-in

potential at the electrolyte/InP interface reduces the charge separation and promotes the

recombination of excited carriers in the surface states. On the other hand, our PC

devices are designed to separate the porous layer from both the p-n interface and the

carrier-collecting electrode. Since the photocurrents flow in a channel far from the

porous layer in such a configuration, it can be measured to avoid the negative effects of

the surface states at the pore walls.

One of the possible reasons for the enhancement of the absorption property is

light scattering within the porous layer [25]. We believe that the number of absorbed

photons increases in the porous layer due to the significant increase in optical-path

length by the light scattering. Since the dtop is defined as a geometrical thickness of the

top layer, in this study, the effect of the increased optical-path length will be included in

the α, as shown in Eq. (4-3). Another possible reason is the sub-bandgap absorption in

Figure 4-11. Comparison of photocurrents between non-porous and porous devices plotted as

function of light intensity PIN.

Light intensity PIN (µW/cm2)

Pho

tocu

rren

t ΔI 1

(µA

)

58 Chapter 4

the porous layer [24,26]. Sub-bandgap is formed within the original bandgap of InP by

the localized levels such as the lattice defects, impurities in the bulk, and surface states.

The optical absorption process caused by the localized levels is schematically shown in

Fig. 4-12. Carriers can exist in the forbidden band due to the localized levels, which

allows carriers to excite (b) from the valence band (VB) to the localized levels, (c) from

the localized levels to the conduction band (CB), and (d) from the localized levels to

other localized levels in addition to (a), which is the general excitation from VB to CB.

Since the surface area of the porous layer is about fifty times as large as that of the

non-porous layer, the localized levels, especially those caused by the surface states,

might increase and lead to enhancement of the absorbance.

4.4. Platinum/porous InP Schottky junction PC devices

4.4.1. Concept of PC devices utilizing large surface of porous layers

The proposed device—based on a porous InP structure—is schematically

shown in Fig. 4-13. A potential barrier for separating the photo-carriers generated under

illumination is formed on the InP walls inside pores. In this study, a platinum film was

used for this purpose because the high Schottky barrier is provided to n-type

semiconductor due to a large work function of platinum (5.65 eV) [31]. The

photo-carriers generated under illumination are separated by the electric field in the

depletion region and collected on the top and back electrodes, as shown in Fig. 4-13.

Figure 4-12. Schematic illustration of carrier excitation with sub-bandgap absorption. (a) General

excitation from valence band (VB) to conduction band (CB), (b) from VB to localized levels, (c)

from localized levels to CB, and (d) from localized levels to other localized levels.

hν


The proposed device is expected to increase the efficiency of PC devices such as a

photo-detector and a solar cell due to its unique features, namely, large surface area

inside pores and low reflectance. As semiconductor materials for the device fabrication,

n-type InP (001) epitaxial wafers were used. Epitaxial layers with a thickness of 5 µm

were grown on n+-InP substrates by standard metal-organic vapor-phase epitaxy with

silicon doping of 1×1017 cm−3. To supply currents for forming the porous structures, a

Ge/Au/Ni ohmic contact layer was evaporated on the backside of the samples and

annealed in nitrogen for 5 min at 380˚C. Experimental setup of porous formation is

same with Section 4.2. The anodic bias and anodization time were set at 20 V and 5 s.

As previously mentioned, a disordered irregular layer formed during the initial stage of

pore formation when the pores formed and partly remained on top of the ordered porous

layer. To remove the irregular top layer, the porous surface was then

photo-electrochemically etched at an anodic bias of 1.5 V in the same electrolyte under

illumination. After the irregular top layer was removed by photo-assisted EC etching, a

cathodic bias was applied to a porous structure in H2PtCl6 electrolyte in order to form a

thin platinum film on the wall inside the pores. To improve the uniformity of the

platinum film, a pulsed bias mode was used. After the electrochemical process, the

sample was washed in deionized water and dried well for optical and electrical

characterizations performed in the air.

Structural properties and optical-reflectance properties were investigated by

using a Hitachi S-4100 scanning electron microscope (SEM) and a Shimadzu UV-1700

UV-VIS Spectrophotometer, respectively. I-V measurements were carried out in the air

Figure 4-13. Schematic illustration of the platinum/porous InP Schottky junction PC devices and

photo-carrier separation at platinum/pore interface.

Ohmic (Ge/Au/Ni)

60 Chapter 4

using a Keithley 2602A source meter with a contact probing system. A tungsten lamp

was used as a light source for the photoelectrical measurements. For purpose of

comparison, a platinum/planar InP without a porous structure was prepared by the

cathodic formation of platinum on the planar InP substrate.

4.4.2. Formation of platinum on InP porous structures

The effect of the irregular top layers on the formation of platinum films was

investigated first. SEM images of the reference surface just after the formation of the

porous structure and the etched surface after photo-assisted EC etching are shown in

Figs. 4-14(a) and (b), respectively. As shown in Fig. 4-14(a), small pores with a

diameter of about 80 nm were formed on the surface at the initial stage of the pore

formation. Furthermore, inner pores (with a large diameter) can be seen through the

irregular top layer as a lightly shaded circle around each pore. After photo-assisted EC

etching, as shown in Fig. 4-14(b), the pore diameter increased to 450 nm (on average).

This result indicates that the irregular top layer was completely removed by

photo-assisted EC etching and the inner pores appeared on the etched surface.

Figure 4-14. Top view SEM images of the InP porous structures and its schematic illustration. (a)

Sample just after the pore formation and (b) sample after the removal of the irregular top layer

formed on the surface.

(a)

(b)

3 μm

3 μm


Cross-sectional SEM images of the samples after the platinum formation on the

porous structure with and without the irregular top layer are shown in Figs. 4-15(a) and

(b), respectively. The coverage of the platinum films on the walls inside pores depends

strongly on the surface morphology of the porous structures. As shown in Fig. 4-15(a),

the platinum film formed only on the irregular top layer and not on the wall surface

inside the pores. This result indicates that each pore is completely closed by a platinum

film in the initial stage of platinum formation. On the contrary, the coverage of the

platinum film on the walls inside the pores is improved in the sample without the

irregular top layer, as shown in Fig. 4-15(b). A platinum film with an average thickness

of about 280 nm was formed on the wall surface inside the pores, indicating that

cathodic currents were supplied from the InP wall inside the pores. Similar structure has

been reported in the case using a conductive porous template such as a Si porous

structure [32]. Average thickness of the platinum film is plotted against processing time

of the cathodic formation in Fig. 4-15(c). This result indicates that the thickness of the

platinum film formed on the wall inside the pores can be controlled by the processing

time.

Figure 4-15. Cross-sectional SEM images of the sample after platinum formation: (a) on the

porous structure with the irregular top layer and (b) on the porous structure after the removal of

the irregular top layer. (c) Correlation between average thickness of platinum film formed on the

walls inside the pores and processing time for forming the cathodic platinum.

(a)

(b)

1 μm

1 μm

without irregular layer

200

50

100

200

150

250

300

400 600 800 10000

0

Thi

ckn

ess

of P

t film

(nm

)

Deposition time (s)

(c)

62 Chapter 4

The effect of the structural properties on the surface reflectance of the porous

structure was investigated by comparing three samples, as shown in Fig. 4-16. First, the

reflectance of the planar sample was higher than 30 % over the measurement range.

Typical peaks, attributed to the interband transitions, were observed around 250, 400,

and 850 nm [33]. The reflectance of the porous sample was 20 to 30 % lower than that

of the reference sample. Especially, the reflectance of the porous sample without the

irregular top layer was considerably decreased, namely, to lower than 3.2 % over the

measurement range. These results are very similar to those for a porous structure

without platinum films reported in a previous work [20]. It was found that the low-

reflectance properties of the InP porous structures remained after the formation of the

platinum film.

4.4.3. I-V characteristics under illumination

The basic photoelectrical properties of the porous sample were investigated by

comparing those of a planar sample formed as a reference. Figure 4-17 shows the

current-voltage (I-V) characteristics of (a) the platinum/planar InP and (b) the

Figure 4-16. Specular reflectance spectra obtained at planar InP, porous InP with irregular layer,

and porous InP without irregular layer as a function of wavelength λ.

100

80

60

40

20

0200 400 600 800 1000

planar (reference)

porous w/ irregular layer

Wavelength λ (nm)

Ref

lect

ance

（%

）

porous w/o irregular layer


platinum/porous InP sample without the irregular top layer, respectively. The average

thickness of the platinum film formed on the wall inside the pores was 72 nm. The

current density was calculated from the geometrical surface area of the top platinum

films where the forward bias was applied. As shown in Fig. 4-17, the I-V characteristics

obtained on both samples show a clear rectifying behavior, indicating the formation of a

Schottky barrier at the platinum/InP interface. However, the photocurrents measured

under illumination depend largely on the sample structure. The reverse current for the

platinum/planar InP sample is about 10 mA/cm2 under illumination, which is 3 mA

larger than that obtained in the dark, as shown in Fig. 4-17(a). On the contrary, the

reverse current of the platinum/porous InP sample is about 200 mA under illumination,

which is 175 mA larger than that obtained in the dark, as shown in Fig. 4-17(b).

To further clarify the photo current response properties of the porous sample,

the optical responsivity was investigated under monochromatic illumination by using a

band-path filter for the 514.5 nm wavelength. Responsivity R is defined as follow:

INL PJR= , (4-4)

where JL is the density of photocurrents, and PIN is incident light intensity. The

responsivity of the porous sample is compared with that of the reference planar sample

in Fig. 4-18. The photocurrents JL were obtained by applying a reverse bias from 0 to

Figure 4-17. Current-voltage (I-V) characteristics measured under illumination using a white light

for two samples. (a) Platinum/planar InP sample and (b) platinum/porous InP sample without

irregular top layer.

Voltage (V)1.0-1.0 -0.5 0 0.5

10-4

10-6

10-5

10-3

10-2

10-1

Cur

rent

den

sity

(A

/cm

2 )

3mA

in the dark

under illumination

Pt/planar InP10-7

100

Vp = 0.01 VJp = 0.26 mA/cm2

Vp = 0.02 VJp = 0.51 mA/cm2

(a) (b)

Voltage (V)1.0-1.0 -0.5 0 0.5

10-4

10-6

10-5

10-3

10-2

10-1

Cur

rent

den

sity

(A

/cm

2 ) 175mA

under illumination

in the dark

100

Pt/ porous InP10-7

64 Chapter 4

−1.0 V, and PIN was set at 3.7 mW/cm2. The responsivity of the reference planar sample

was below 0.15 A/W in the bias range used for the present measurements. On the

contrary, the responsivity of the porous sample depends largely on the applied reverse

bias, for example, it is 13 times larger than that of the reference planar sample at a bias

of −0.8 V. These results indicate that the conversion efficiency of the porous sample is

higher than that of the planar sample with the same geometrical area.

Large photocurrents on a porous InP electrode in an electrolyte during

electrochemical measurements have been reported [20,34]. Photocurrents (including

faradaic and non-faradaic currents) increased in proportion to pore depth or effective

area inside a pore. Similarly in the air, it seems that the major factor that enhances the

photocurrents observed in this study is the increase in the total area at the platinum/InP

interface. In consideration of the cylindrical pores, the effective surface area including

the wall surface was estimated to be about seven times larger than that of the planar

sample used in this study. It is concluded from this estimation that the increased

photocurrents shown in Fig. 4-18 cannot be explained only by the effect of the

increased surface area. Another possible reason for the increase in photocurrents by

formation of porous structure is the enhanced optical properties such as low reflectance

and high absorptance.

In contrast with the large photocurrents, the present device showed the small

photovoltages, as shown in Fig. 4-17(b). This is probably due to the poor potential

barrier formed at platinum/InP interface. For the photovoltaic application, further

Figure 4-18. Optical responsivity of Pt/planar InP and Pt/porous InP PC devices under

monochromatic light with the wavelength of 514.5 nm.

-0.8 -1.00 -0.2 -0.4 -0.6

1.5

1.0

0.5

0

Pt/planar InP

Res

pon

sivi

ty R

(A/W

)

Pt/porous InP

×13

Voltage (V)


improvement on the device structure is necessary, such as employing the p-n junction

interface instead of the Schottky interface on the pore wall. We believe that the porous

structure—with unique features such as low optical reflectance and large wall

surface—is a good candidate for a building block of photoelectric conversion devices.

4.5. Summary

In Chapter 4, highly ordered pore arrays were formed on InP and

characterized by using two type of PC devices. Summary of this chapter is itemized

below.

1) Straight, homogeneous, and controllable porous structures could be formed on

InP by EC etching in dark and subsequent EC etching under light.

2) PC devices formed on p-n junction substrates showed that lower photocurrents

were observed on the porous device, as compared with that of the non-porous

device, indicating that the absorption properties of InP were enhanced after the

formation of porous structures.

3) The enhancement of absorption properties can be explained in terms of

absorption coefficient increased by the light scattering and the sub-bandgap

absorption in the porous layer.

4) Platinum/porous InP Schottky junction PC devices showed larger photocurrents

and higher responsivity than those of a reference planar sample.

From the above insights, we believe that InP porous structures are promising

materials for use in PC devices such as photovoltaic cells and photodetectors because of

their unique features such as their large surface area, low reflectance properties, and

high absorption properties.

66 Chapter 4

Reference

[1] H. Lv, H. Shen, Y. Jiang, C. Gao, H. Zhao, and J. Yuan, "Porous-pyramids

structured silicon surface with low reflectance over a broad band by

electrochemical etching", Appl. Surf. Sci., vol. 258, pp. 5451−5454, 2012.

[2] A. Uhlir, "Electrolytic shaping of germanium and silicon", Bell Syst. Tech. J.,

vol. 35, pp. 333−347, 1956.

[3] P. Schmuki, J. Fraser, C. M. Vitus, M. J. Graham, and H. S. Isaacs, "Initiation

and Formation of Porous GaAs", J. Electrochem. Soc., vol. 143, pp.

3316−3322, 1996.

[4] G. Oskam, A. Natarajan, P. C. Searson, and F. M. Ross, "The formation of

porous GaAs in HF solutions", Appl. Surf. Sci., vol. 119, pp. 160−168, 1997.

[5] I. M. Tiginyanu, V. V. Ursaki, E. Monaico, E. Foca, and H. Föll, "Pore Etching

in III-V and II-VI Semiconductor Compounds in Neutral Electrolyte",

Electrochem. Solid-State Lett., vol. 10, pp. D127−D129, 2007.

[6] T. Takizawa, S. Arai, and M. Nakahara, "Fabrication of Vertical and

Uniform-Size Porous InP Structure by Electrochemical Anodization", Jpn. J.

Appl. Phys., vol. 33, pp. L643−L645, 1994.

[7] A. Hamamatsu, C. Kaneshiro, H. Fujikura, and H. Hasegawa, "Formation of

<001>-aligned nano-scale pores on (001) n-InP surfaces by

photoelectrochemical anodization in HCl", J. Electroanal. Chem., vol. 473, pp.

223−229, 1999.

[8] H. Fujikura, A. Liu, A. Hamamatsu, T. Sato, and H. Hasegawa,

"Electrochemical Formation of Uniform and Straight Nano-Pore Arrays on

(001) InP Surfaces and Their Photoluminescence Characterizations", Jpn. J.

Appl. Phys., vol. 39, pp. 4616−4620, 2000.

[9] S. Langa, I. M. Tiginyanu, J. Carstensen, M. Christophersen, and H. Föll,

"Self-organized growth of single crystals of nanopores", Appl. Phys. Lett., vol.

82, pp. 278−280, 2003.

[10] A. Anedda, A. Serpi, V. A. Karavanskii, I. M. Tiginyanu, and V. M. Ichizli,


"Time resolved blue and ultraviolet photoluminescence in porous GaP", Appl.

Phys. Lett., vol. 67, pp. 3316−3318, 1995.

[11] K. Kuriyama, K. Ushiyama, K. Ohbora, Y. Miyamoto, and S. Takeda,

"Characterization of porous GaP by photoacoustic spectroscopy: The relation

between band-gap widening and visible photoluminescence", Phys. Rev. B, vol.

58, pp. 1103−1105, 1998.

[12] F. K. Yam, Z. Hassan, and S.S. Ng, "Porous GaN prepared by UV assisted

electrochemical etching", Thin solid films, vol. 515, pp. 3469−3474, 2007.

[13] A. Ramizy, Z. Hassan, and K. Omar, "Nanostructured GaN on silicon

fabricated by electrochemical and laser-induced etching", Mater. Lett., vol. 65,

pp. 61−63, 2011.

[14] K. Al-Heussen, M. R. Hashim, and N. K. Ali, "Effect of different electrolytes

on porous GaN using photo-electrochemical etching", Appl. Surf. Sci., vol. 257,

pp. 6197−6201, 2011.

[15] J. S. Shor, I. Grimberg, B. -Z. Weiss, and A. D. Kurtz, "Direct observation of

porous SiC formed by anodization in HF", Appl. Phys. Lett., vol. 62, pp.

2836−2838, 1993.

[16] A. Boukezzata, A. Keffous, A. Cheriet, Y. Belkacem, N. Gabouze, A. Manseri,

G. Nezzal, M. Kechouane, A. Bright, L. Guerbous, and H. Menari, "Structural

and optical properties of thin films porous amorphous silicon carbide formed

by Ag-assisted photochemical etching", Appl. Surf. Sci., vol. 256, pp.

5592−5595, 2010.

[17] A. Keffous, N. Gabouze, A. Cheriet, Y. Belkacem, and A. Boukezzata,

"Investigation of porous silicon carbide as a new material for environmental

and optoelectronic applications", Appl. Surf. Sci., vol. 256, pp. 5629−5639,

2010.

[18] H. Hasegawa, and T. Sato, "Electrochemical processes for formation,

processing and gate control of III-V semiconductor nanostructures",

Electrochim. Acta, vol. 50, pp. 3015−3027, 2005.

[19] T. Sato, T. Fujino, and H. Hasegawa, "Self-Assembled Formation of Uniform

68 Chapter 4

InP Nanopore Arrays by Eelectrochemical Anodization in HCl based

Electrolyte", Appl. Surf. Sci., vol. 252, pp. 5457−5461, 2006.



photoelectrochemical solar cells", Thin Solid Films, vol. 518, pp. 4399−4402,

2010.

[21] A. Ramizy, Z. Hassan, K. Omar, Y. Al-Douri, and M. A. Mahdi, "New optical

features to enhance solar cell performance based on porous silicon surfaces",

Appl. Surf. Sci., vol. 257, pp. 6112−6117, 2011.

[22] R. Jinbo, T. Kudo, Z. Yatabe, and T. Sato, "Large photocurrent-response

observed at Pt/InP Schottky interface formed on anodic porous structure", Thin

Solid Films, vol. 520, pp. 5710−5714, 2012.

[23] N. Naderi, and M. R. Hashim, "A combination of electroless and

electrochemical etching methods for enhancing the uniformity of porous

silicon substrate for light detection application", Appl. Surf. Sci., vol. 258, pp.

6436−6440, 2012.

[24] M. A. Butler, and D. S. Ginley, "Surface Treatment Induced Sub-Band Gap

Photoresponse of GaP Photoelectrode", J. Electrochem. Soc., vol. 128, pp.

712−714, 1981.

[25] B. H. Erné, D. Vanmaekelbergh, and J. J. Kelly, "Morphology and Strongly

Enhanced Photoresponse of GaP Electrodes Made Porous by Anodic Etching",

J. Electrochem. Soc., vol. 143, pp. 305−314, 1996.

[26] F. Iranzo Marín, M. A. Hamsira, and D. Vanmaekelbergh, "Greatly Enhanced

Sub-Bandgap Photocurrent in Porous GaP Photoanodes", J. Electrochem. Soc.,

vol. 143, pp. 1137−1142, 1996.

[27] J. van de Lagemaat, M. Plakman, D. Vanmaekelbergh, and J.J. Kelly,

"Enhancement of the light-to-current conversion efficiency in an

n-SiC/solution diode by porous etching", Appl. Phys. Lett., vol. 69, pp.

2246−2248, 1996.

[28] L. Santinacci, A. -M. Goncalves, N. Simon, and A. Etcheberry,


"Electrochemical and optical characterizations of anodic porous n-InP(100)

layers", Electrochim. Acta, vol. 56, pp. 878−888, 2010.

[29] T. Sato, T. Fujino, and T. Hashizume, "Electrochemical Formation of

Size-Controlled InP Nanostructures Using Anodic and Cathodic Reactions",

Electrochem. Solid State Lett., vol. 10, pp. H153−H155, 2007.

[30] T. Sato, and A. Mizohata, "Photoelectrochemical Etching and Removal of the

Irregular Top Layer Formed on InP Porous Nanostructures", Electrochem.

Solid State Lett., vol. 11, pp. H111−H113, 2008.

[31] H. B. Michaelson, "The work function of the elements and its periodicity", J.

Appl. Phys., vol. 48, pp. 4729−4733, 1977.

[32] K. Kobayashi, F. A. Harraz, S. Izuo, T. Sakka, and Y. H. Ogata, "Microrod and

Microtube Formation by Electrodeposition of Metal into Ordered Macropores

Prepared in p-Type Silicon", J. Electrochem. Soc., vol. 153, pp. C218−C222,

2006.

[33] P. Lautenschlager, M. Garriga, and M. Cardona, "Temperature dependence of

the interband critical-point parameters of InP", Phys. Rev. B, vol. 36, pp.

4813−4820, 1987.

[34] A. -M. Goncalves, L. Santinacci, A. Eb, I. Gerard, C. Mathieu, and A.

Etcheberry, "Pore Formation on n-InP(100) in Acidic Liquid Ammonia at 223

K", Electrochem. Solid-State Lett., vol. 10, pp. D35−D37, 2007.

70 Chapter 4

71

Chapter 5

Fabrication and size-modulation of GaN porous

structures for EC energy-conversion systems

5.1. Introduction

Electrochemical (EC) energy conversion systems based on semiconductor

photoelectrodes have recently attracted much attention due to their potential use in the

next generation of green technologies such as water splitting, artificial photosynthesis,

and so on [1−5]. Among the photoelectrode materials, GaN is one of the most attractive

because of its chemical stability and its potential to achieve direct photoelectrolysis by

solar power without the consumption of electric power [6−8]. In addition, the bandgap

energy of GaN-based materials can be varied from about 0.65 to 6.0 eV by alloying

them with InN and AlN, which enables us to design various functional photoelectrodes

not only for spectral matching of solar light but also for the EC reduction of CO2 to

carbohydrate [9]. One of the common approaches to improving conversion efficiency is

to form nanostructures on the photoelectrode surface in order to increase its surface area.

Most reported GaN nanostructures have been made using selective-area growth [10,11]

or a dry etching process such as reactive ion etching [12,13]. There are, however, severe

limitations on increasing the density of nanostructures because most approaches use

lithography for defining the size and position of the nanostructures. And when a dry

etching process is involved, the etching damage induced by ion bombardment is not

negligible [14−16] and could significantly degrade the conversion efficiency.

One alternative approach is an EC-fabrication process, which can form various

semiconductor nanostructures in a self-assembled fashion [17]. The most well-known

application of an EC process is the formation of porous structure by anodic etching in

which a high-density array of nanometer- or micrometer-sized pores is formed with high

productivity over a large area on the semiconductor surface as described in Chapter 4.

We have reported that InP porous structures showed low photo-reflectance and high

photo-absorption [18,19], which are very promising features for porous structures used

in the EC energy conversion systems. Besides, the EC process is applicable to various

semiconductors [20−24], even chemically stable materials such as GaN [25−29],

72 Chapter 5

without causing processing damage.

The EC conditions including applied bias and electrolyte solutions have been

investigated with regard to the formation of GaN porous structures, but most of the

previous studies targeted structural properties and only a few reported on the correlation

between the conditions and the resultant optical properties. This is partly because the

mechanism of the formation of GaN porous structures has not been clarified because of

sample-dependent inhomogeneity and insufficient material quality. It is well known that

GaN epitaxial layer grown on sapphire substrates always have a high density of

dislocations caused by strain at the lattice-mismatched interface between GaN and

sapphire substrates. In such a situation, the current supply for EC reactions would be

strongly affected by the dislocation density and dislocation distribution of the substrates

[30]. It is also difficult to judge whether the optical response obtained on sample

structures is due to the intrinsic properties of GaN porous structures or to photo-active

dislocations in the GaN epitaxial layer.

In Chapter 5, we first investigated the structural properties of GaN porous

structures formed on n-GaN homo-epitaxial layer grown on a free-standing GaN

substrate, which typically has a low dislocation density [31]. First of all, we tried

photo-assisted EC etching which is commonly used to fabricate GaN porous structures.

Then, we tried to improve structural controllability by focusing on two kinds of

charged-carrier generation phenomena: avalanche effect, and Franz-Keldysh effect. Post

treatment technique by conventional wet etching was also performed to control

structural properties precisely. Optical characterization such as photoluminescence (PL),

photoreflectance, and photoelectrochemical measurements revealed the importance of

precise structural controlling of porous structures on application to the EC energy

conversion systems.

5.2. Experimental details

An n-type GaN (0001) freestanding wafer with carrier concentration ND of 1.1

× 1018 cm−3 was employed in this study. The threading dislocation density (TDD) of

GaN substrates is approximately 2.2 × 106 cm−2, which is two or three orders of

magnitude less than that of hetero-epitaxial layers grown on sapphire substrates [32].

The EC current was supplied through the Au-ohmic contact on the backside of the

substrate, and the EC process was performed using a three-electrode EC cell, as

73 Fabrication and size-modulation of GaN porous structures for ECenergy-conversion systems

described in Chapter 3. We used a mixture of 1 mol/L H2SO4 and 1 mol/L H3PO4 (pH =

1.7) as an electrolyte to remove resulting oxides and expose the raw GaN surface for

further etching. Xe lamp was the light source in the photo-assisted EC etching forming

GaN porous structures.

5.3. GaN porous structures formed by photo-assisted EC etching


Typical cyclic voltammograms of an n-GaN epitaxial layer that were obtained

at light intensities PIN between 0 and 40 mW/cm2 by sweeping the potential VEC

between 0 and 4.0 V are shown in Fig. 5-1. Little or no current was measured in the

dark because of the absence of holes in the n-GaN layer. Under light irradiation, on the

other hand, current started to flow at about 1.0 V. Higher currents were measured at

higher light intensities, and these currents resulted in anodization of the n-GaN surface.

We therefore tried to form porous structures under irradiation; that is, by photo-assisted

EC etching. In view of low-energy processing as a first attempt, VEC was set at 1.0 V

and PIN was set at 5 mW/cm2.

Figures 5-2 show the top and cross-sectional SEM images of GaN porous

samples formed at VEC = 1.0 V, PIN = 5.0 mW/cm2, and various EC etching time tEC: (a)

tEC = 5 min (sample A), (b) tEC = 10 min (sample B), (c) tEC = 30 min (sample C), and

(d) tEC = 10 min with a chemical treatment (the process used to form sample B' is

Figure 5-1. Cyclic voltammograms measured on as-grown GaN electrode under light irradiation

with various PIN.

0.0

1.0

2.0

3.0

4.0

0.0 1.0 2.0 3.0 4.0

Cu

rre

nt d

ens

ity J

(m

A/c

m2 )

Potential vs. Ag/AgCl VEC

(V)

40 mW/cm2

20 mW/cm2

5 mW/cm2

dark

74 Chapter 5

described in Section 5.3.3). The pore density formed on the surface was estimated from

SEM images to be about 1010 cm−2, which is much higher than the dislocation density of

the substrates. This suggests that the pore formation in this study was substantially

unaffected by the dislocations in GaN substrates. In the cross-sectional views of

samples A and B we can clearly see pores with diameters over 30 nm, and in the top

views we can see pores with diameters under 10 nm. These results suggest that an

ultrathin porous layer having smaller-diameter pores formed on top of a thick porous

layer having larger-diameter pores. Figure 5-3 shows the relationship between pore

depth Dp and the charge density Q passing through the working electrode during the

anodization. The Dp increased with Q, and the slope of the experimental curve

decreased gradually. This gradual decrease indicates that the rate of pore formation in

the depth direction decreased with increasing Q. Considering the proportional

relationship between the amount of EC reaction and Q, as described by Faraday’s law,

the decreased slope shown in Fig. 5-3 indicates that lateral etching of pore walls

occurred in addition to the vertical etching in the depth direction.

As shown in Fig. 5-2(c), on the other hand, the structural features of sample C,

formed at tEC = 30 min, differ from those of samples A and B. In the top view, we can

clearly see pores over 30 nm in diameter, and in the cross-sectional view, we can see no

ultrathin porous layer with smaller pores. Furthermore, as shown in Fig. 5-3, the Dp in

sample C was smaller than that in samples formed at shorter tEC and smaller Q. These

results indicate that surface etching as well as vertical etching in the depth direction

occurred.

Figure 5-2. Top and cross-sectional SEM images of GaN porous nanostructures formed by

photo-assisted electrochemical etching at VEC = 1.0 V, PIN = 5 mW/cm2, and various tEC: (a) tEC =

5 min (sample A), (b) tEC = 10 min (sample B), (c) tEC = 30 min (sample C), and (d) tEC = 10 min

after H3PO4 treatment (sample B').


5.3.2. Pore formation mechanism in photo-assisted EC etching

Figure 5-4 shows the formation flow for GaN porous structures schematically.

At the initial stage, photo-generated holes transferred preferentially to the pore tips at

which electric field lines concentrated, leading to the formation of porous structures ((a)

→ (b)). Since the holes generated near the surface also transferred to the pore walls,

etching of pore walls and the top-surface occurred in addition to the vertical etching at

the pore tips ((b) → (c)). After reaching a critical depth, etching proceeded

preferentially at the top-surface removing the ultrathin porous layer with smaller size

pore ((c) → (d)), resulting in larger pores observed in sample C. There are two possible

reasons for this phenomenon. One is due to the reduced ions diffusion to pore tips in

deeper pores, and another is due to the decreased electric hole supply to pore tips. A

similar preferential top-surface etching is also observed on InP porous structures under

irradiation [33]. This formation model can compatibly explain both the SEM results in

Figs. 5-2 (a)−(c) and the correlation between pore depth and Q shown in Fig. 5-3.

The correlation with the crystal quality such as a dislocation density has been

frequently reported in the chemical and EC etching. A previous study found that

photo-assisted EC etching of GaN layers occurred only at dislocation-free areas because

the dislocations acted as recombination centers for the photo-generated carries [34].

Another study found that the etching using a HF solution proceeds thorough the

dislocations and grain boundaries, whereas the etching using a KOH solution takes

place at crystal grains [35]. These reports showed that the etching profile was strongly

Figure 5-3. Relationship between pore depth Dp and charge density Q passing through the

electrode during the photo-assisted EC etching.

PIN = 5 mW/cm2

VEC = 1.0 V

5 min(sample A)

20 min

15 min

10 min(sample B)

30 min(sample C)

0

100

200

300

400

500

600

700

0 200 400 600 800

Charge density Q (mC/cm2)

Po

re d

epth

Dp (

nm)

76 Chapter 5

affected by the dislocation density and distribution in GaN layers. However, that was

not the case with this study. The TDD of the GaN epitaxial layer used in this study is 2.2

× 106 cm−2, which means approximately 1 dislocation in 5 μm2. The pore density of the

present GaN porous structures, however, is estimated to be over 5000 pores per 5 μm2.

This fact indicates that the formation model proposed here is acceptable for the present

study free of influence from the dislocations in GaN layers.

5.3.3. Photoluminescence and photoreflectance properties

The correlation between the structural properties and optical properties was

investigated in the photoluminescence and photoreflectance measurements conducted

on the GaN porous samples shown in Figs. 5-2. As for sample B' shown in Fig. 5-2(d),

the ultrathin porous layer with smaller pores was removed by immersing the sample in a

heated H3PO4 solution for 1 min after the formation of porous structure with tEC = 600

sec. Figure 5-5(a) shows PL spectra obtained at room temperature from a non-porous

sample and porous sample C formed at tEC = 1800 s. Both samples exhibited strong

near-band-edge emission around 3.4 eV with a full width at half maximum (FWHM) of

~ 50 meV. In the present samples, two additional peaks related to longitudinal optical

(LO) phonon replicas were also observed. The position of main peak obtained from

sample C is at 3.404 eV which is about 5 meV higher than that of the main peak

obtained from the non-porous sample. We believe that the observed blue shift is most

probably due to the quantum confinement in pore walls, similarly to the case of porous

Figure 5-4. Schematic representations of formation flow of GaN porous structures by

photo-assisted EC etching. (a) Photo-holes are transferred preferentially to the pore tips at which

electric field lines concentrated. (b) Etching at initial stage proceeds in the vertical direction. (c)

Photo-holes are transferred to pore walls and the top-surface in addition to the pore tips. (d)

Etching of pore walls and the top-surface occurs.

n-GaN

light

electricfield

electrolyte

etching from top

photo-hole

(c) (d)(b)(a)


Si and InP [36,37]. The quantum confinement energy in GaN walls was calculated by

assuming a vacuum/GaN quantum well, whose thickness corresponds to the wall

thickness Wpw as follows.

*h

*e mmW

hE

11

8Δ

2pw

2

, (5-1)

where h is Planck’s constant, me*and mh

* are masses of electron and hall of GaN,

respectively. According to Eq. (5-1), a peak shift of 5 meV would be caused when the

thickness of pore walls was about 21 nm. This estimated thickness is very consistent

with the thickness measured in SEM observation. The positions of the main peaks

obtained from all porous samples are summarized in Fig. 5-5(b). The peak position of

sample A was almost same with that of the non-porous sample. The peak positions of

samples B and C, on the other hand, shifted higher position with increasing tEC. This

result suggests that the effect of the quantum confinement increased in the pore walls

being thinner with increase of tEC. It is to be also noted that the peak position of sample

B', whose ultrathin porous layer was removed by a H3PO4 treatment, is the same as that

of sample B formed with tEC = 600 s. As shown in Fig. 5-2 pores over 30 nm in

diameter were seen in top views of sample B', whereas there was almost no difference

between the features seen in cross-sectional views of samples B and B'. These results

Figure 5-5. (a) PL spectra obtained at room temperature from a non-porous sample and porous

sample C formed at tEC = 30 min, and (b) PL peak positions of porous samples and a non-porous

sample.

3.1 3.2 3.3 3.4 3.5 3.6

Sta

nda

rdiz

ed in

ten

sity

(a

.u.)

Photon energy E (eV)

3.398

3.400

3.402

3.404

3.406

Pos

itio

n o

f ma

in p

eak

(e

V)

non-porous

A B C B'

(a)

Sample

sample C

non-poroussample

2LO

1LO

main peak

(b)

78 Chapter 5

show that the H3PO4 treatment removed the ultrathin porous layer with smaller pores

without thinning the pore walls.

Figure 5-6 shows diffuse reflectance spectra of porous samples and a

non-porous sample measured as a reference. As expected, the photoreflectance R of all

porous samples was lower than that of the non-porous sample. The porous samples

could be separated into two groups according to reflectance value: a higher-R group

comprising samples like A and B, and a lower-R group comprising samples like C and

B'. SEM observation of the porous samples shown in Figs. 5-2 revealed that the

photoreflectance depends more on surface morphology than pore depth. Comparing

sample B' with sample B, one sees the reflectance surface etching using a chemical

treatment decreased reflectance even though the EC conditions forming porous structure

were the same for both samples. This phenomenon is very similar to the case of InP

porous structures [19]. In samples C and B', pores with enlarged openings appeared on

the surface, where air holes are closely aligned between the thin GaN walls. This kind

of air-dielectric composite has a refractive index n close to unity, leading to low

reflectance at the air interface.

5.3.4. Photo-electrochemical response properties

Photo-electrochemical characteristics of various GaN porous structures were

measured in 1 mol/L NaCl electrolyte under irradiation with monochromatic light (λ =

Figure 5-6. Diffuse reflectance spectra of non-porous and porous GaN: sample A formed at tEC =

5 min, sample B formed at tEC = 10 min, sample C formed at tEC = 30 min, and sample B' formed

at tEC = 10 min after H3PO4 treatment.

0

10

20

30

300 400 500

Ref

lect

anc

e R

(%

)

Wavelength (nm)

non-porous

AB

C

B'


350 nm). Figure 5-7 compares the photocurrent-voltage curves obtained on porous

samples and the non-porous sample. No current degradation or structural change was

observed in repeated experiments because oxidation of the GaN surface was suppressed

oxidation in the Cl−-containing electrolyte [6]. Here, the dominant EC reactions in this

system are thought here to be due to the photo-assisted electrolysis of water, the

reduction of H+, and the oxidation of Cl−. As shown in Fig. 5-7, it was found that

photocurrent of the non-porous sample was very low, i. e., only 0.02 µA/cm2 at 0 V and

9.4 µA/cm2 at 1.0 V applied bias. After the formation of porous structures, however,

photocurrents were drastically enhanced to as high as 80 µA/cm2 at 0 V. This might be

due to the unique features of porous structures, such as their large surface area and low

reflectance. These results enable us to conclude that the formation of porous structures

is effective for improving photo-electrochemical characteristics and is a promising

technique for use in photo-electrochemical water splitting.

It was also found that the photo-electrochemical characteristics were greatly

influenced by the structural variations of porous samples. Structural and optical

properties of porous samples are listed in Table 5-I. Photocurrents of sample B were

larger than those of sample A. Since the most apparent difference between sample A and

B was pore depth, it would appear that the improved photo-electrochemical

characteristics of sample B were due to its increased pore depth increasing the pores’

contribution to the sample’s surface area. On the other hand, the photocurrents of

sample C were lower than those of sample B in spite of the pore depths of the two

samples being almost the same. One possible reason for this is the thinning of the pore

Figure 5-7. Photoelectrochemical characteristics of non-porous and porous GaN: sample A

formed at tEC = 5 min, sample B formed at tEC = 10 min, sample C formed at tEC = 30 min, and

sample B' formed at tEC = 10 min after H3PO4 treatment.

-20

0

20

40

60

80

100

-0.5 0.0 0.5 1.0

Cur

ren

t de

nsity

J (A

/cm

2 )

Voltage (V)

dark

A

B

C

B'λ = 350 nm1M NaCl

non-porous

80 Chapter 5

walls in sample C. As described above, the thickness of pore wall decreased during the

prolonged process with tEC = 1800 s. The thinning of the pore walls probably decreased

thickness of photoabsorption region in the pore walls, and/or degraded carrier transport

properties by increasing resistivity.

The photocurrent of sample B which is the highest among porous samples A–C

was furthermore improved by H3PO4 treatment as observed at sample B'. SEM

observation and photoreflectance measurements revealed that its lower reflectance after

the removal of ultrathin porous layer was obtained without thinning the pore walls. The

typical relationship between photoreflectance R and photocurrent Jph is expressed by

)100(ΦINph RJ , (5-2)

where ΦIN is the irradiated photon flux. Assuming that the material properties of sample

B and B' are the same, we have the following equation:

100

100

(B)

)B(

(B)ph

)B(ph

R

R

J

J

. (5-3)

Substituting the values obtained in this study to Eq. (5-3), we obtained almost

equal values of 1.06 and 1.07, respectively, for the left hand side and right hand side of

the equation. Accordingly, the increase of photocurrent by H3PO4 treatment could be

explained by the effect of the decrease of photoreflectance in Eq. (5-3). In such a

situation, the number of photons absorbed at pore walls increased, resulting in the

improvement of photoelectrochemical conversion efficiency. These results indicate that

the control of porous structures features such as surface morphology, thickness of pore

Table 5-I. Structural and optical properties of porous samples.

Sample Pore depth Dp (nm)

PL peak shift(meV)

Photoreflectance R (%)

Photocurrent density Jph (µA/cm2)

A 200 0 9.6 55.3

B 340 3.0 9.4 77.7

C 350 5.0 4.1 71.2

B' 340 3.0 3.5 82.7


walls, and pore depth is crucial to the improvement of the photoelectrochemical

characteristics of GaN porous structures.

5.4. GaN porous structures formed by anisotropic EC etching

5.4.1. EC current transition at high voltage

Figure 5-8 shows current-time curves obtained during anodization at various

anode voltage VEC values. At the initial stage of electrochemical reactions, anodic

currents increased for one hundred seconds or more. The slope of the curve increased

with VEC: in other words, the reaction rate increased with VEC. In addition, we found that

anodic currents saturated for the data obtained at VEC = 8 and 10 V, but they kept

increasing at VEC = 15 V. These results indicate that the reaction rates at VEC = 8 and 10

V become constant after one hundred seconds, whereas the reaction rate at VEC = 15 V

increases with anodization time.


Figures 5-9 show SEM images of GaN porous structures formed with tEC = 20

min and various anode voltages: (a) VEC = 8 V, (b) VEC = 10 V, and (c) VEC = 15 V. The

Figure 5-8. Current-time curves obtained during EC etching in dark at various anode voltage VEC

values.

0

0.2

0.4

0.6

0.8

1

0 5 10 15 20

Cu

rre

nt d

en

sity

(m

A/c

m2)

EC etching time tEC

(min)

VEC = 8 V

VEC = 10 V

VEC = 15 V

in dark

82 Chapter 5

sample formed at VEC = 8 V (Fig. 5-9(a)) had straight pores oriented perpendicular to

the top surface, with pores branching from the straight pores. The sample formed at VEC

= 10 V (Fig. 5-9(b)) also had straight pores oriented perpendicular to the top surface,

but the pore branching observed at VEC = 8 V was significantly suppressed. The pore

diameter and depth were estimated as approximately 18 nm and 1360 nm, respectively,

throughout the entire porous region. On the other hand, the pores formed at VEC = 15 V

(Fig. 5-9(c)) had diminished linearity and size uniformity. This was probably because

the lateral etching of the pore walls was enhanced by the generation of holes in the walls.

From these results, we can conclude that pore morphologies are markedly affected by

VEC, and that electrochemical etching at VEC = 10 V can achieve the formation of linear,

uniform porous structures.

Figure 5-9(d) shows an SEM image of GaN porous structures formed with tEC

= 40 min and VEC = 10 V. As compared with those of the sample shown in Fig. 5-9(b),

both samples had two kinds of porous regions consisting of randomly oriented pores

near the top surface and vertically oriented pores in the [000−1] direction underneath.

Pore depth Dp increased from 1360 nm to 3000 nm with tEC increase, whereas pore

diameter Wp remained unchanged. This result clearly demonstrates that EC etching in

the dark proceeds anisotropically in the [000−1] direction and not preferentially in other

Figure 5-9. Cross-sectional SEM images of GaN porous structures formed in dark with various

EC conditions: (a) tEC = 20 min, VEC = 8 V; (b) tEC = 20 min, VEC = 10 V; (c) tEC = 20 min, VEC = 15

V; (d) tEC = 40 min, VEC = 10 V.


directions. Figure 5-10 summarizes tEC-dependence of (a) average pore depth Dp and

(b) average pore diameter Wp; black dots represent the results of photo-assisted EC

etching obtained in Section 5.3, red dots represent the results of anisotropic EC etching.

We found that anisotropic EC etching can form deeper pores that cannot be formed by

photo-assisted EC etching. We obtained deepest pores with Dp of 35.3 µm having the

incredible aspect ratio of 2000 or more, which is the highest value of a GaN

nanostructure to the best of our knowledge. In addition, Dp of pores formed by

anisotropic EC etching has linear relationship with tEC, indicating superior depth control

can be achieved. On the other hand, Wp could not be controlled with tEC, as shown in

Fig. 5-10(b). This is because the width of the pore wall kept a constant value throughout

the present EC etching due to its anisotropic nature.

5.4.3. Pore formation mechanism in anisotropic EC etching

We tried to describe the mechanism of anisotropic pore formation in EC

etching in the dark by referring to the “Beale model” which was first applied for porous

Si [38]. According to this model, the pores grow up along two notable phenomena: (1)

structural inhomogeneities induce localization of charge carrier flow by electric field

enhancement in the space charge region (SCR), and (2) fully depleted pore wall

Figure 5-10. (a) Average pore depth Dp and (b) diameter Wp of GaN porous structures formed by

photo-assisted EC etching (black symbols) and anisotropic EC etching (red symbols) plotted as a

function of EC etching time tEC.

(b)(a)

photo-assistedEC etching

anisotropicEC etchingetching rate:74.5 nm/min photo-assisted

EC etching

anisotropicEC etching

0

1000

2000

3000

4000

0 10 20 30 40 50

EC etching time tEC

(min)

Ave

rag

e po

re d

epth

Dp (

nm)

0

20

40

60

80

100

120

0 10 20 30 40 50

Ave

rage

po

re d

iam

eter

Wp (

nm

)EC etching time t

EC (min)

84 Chapter 5

(semiconductor between pores) is highly resistive, which prevents charge carrier flow

and further EC reactions.

On the basis of these pictures, anisotropic pore formation observed here is

schematically described in Fig. 5-11. At the initial stage, trenches are formed by EC

etching at where is chemically-unstable and/or occur concentration of electric field (Fig,

5-11(a)). Although it has remained unclear yet where initial etching occur, we believe

Ga-vacancies VGa (or related complexes) are possible initiation sites because

photoluminescence measurements revealed that yellow luminescence were diminished

by anisotropic EC etching. Once the trenches are formed, high electric field is induced

on trenches. If the enhanced electric field reaches breakdown voltage, free carriers are

excited by avalanche effect and consequently proceed etching from trenches, leading to

the formation of the porous region with randomly oriented pores (Fig. 5-11(b)). When

the neighbor pores approach so that their SCRs coalescent, free carriers can be excited

only underneath the pore tips because electric field at fully-depleted region becomes

small (Fig. 5-11(c)). In such situation, pores are prohibited growing except for in the

vertical direction, resulting in the formation of the porous region with vertically-aligned

pores (Fig. 5-11(d)).

To fully-deplete pore wall, the average width of pore wall Wpw must be less

than twice the width of SCR WSCR formed at electrolyte/GaN interface which is

calculated by


anisotropic EC etching. (a) Trenches are formed and induce electric field enhancement. (b)

Randomly oriented pores are formed by free carriers generated by avalanche effect. (c) The

SCRs of neighboring pores merge, restricting free carrier excitation except for underneath the

pore tips. (d) Pores are formed anisotropically in the vertical direction.

(c) (d)

SCR

trench formation

carriers

randomly orientedpore formation merging of SCR

vertically alignedpore formation

(b)(a)

electrolyte

n-GaN


21

FBECD

0GaNSCR

εε2/

q

kTVV

qNW

, (5-4)

where ε0 is permittivity of vacuum, εGaN is dielectric constant of GaN, k is Boltzmann

constant, T is absolute temperature, q is elementary charge, and VFB is flat band

potential which was −0.9 V experimentally obtained from a Mott-Schottky plot. We

obtained WSCR of 100.5 nm with VEC = 10 V from Eq. (5-4), and the Wpw is roughly

estimated at 80 nm from SEM observation on the vertically oriented porous region.

Here, one might think that the actual Wpw of 80 nm is too small if the lateral etching is

stopped by merging the neighbor SCRs, as described in Fig. 5-11(c). One reason for this

is because the SCR of the pore wall is thinner than that obtained by 1D calculation

using Eq. (5-4) because the electric field concentrates on pore tips. In such situation,

merging of SCRs occurs when the Wpw is sufficiently smaller than twice the WSCR.

5.5. GaN porous structures formed by utilizing Franz-Keldysh effect

5.5.1. Observation of Franz-Keldysh effect (FKE) in EC etching with back-side

illumination mode

We argued in Section 5.3 that the control of porous structure features such as

surface morphology, thickness of pore walls, and pore depth is crucial to the

improvement of the photoelectrochemical characteristics of GaN porous structures. As

Figure 5-12. Schematic illustration of the electrochemical setup used for both the formation of

porous structures and the spectro-electrochemical measurements.

86 Chapter 5

for the photo-assisted EC etching in front-side illumination (FSI) mode, however, there

are difficulty in improving structural controllability because carriers most of which were

generated near surface caused top and lateral etching along with pore growth. We

believe that the optimization of the supply of photo-carriers is one of the key issues for

controlling of the structural properties.

In this chapter, we formed GaN-porous structures using the photo-assisted EC

etching in the back-side illumination (BSI) mode and compared with it that in FSI mode.

The EC setup used in this study is schematically shown in Fig. 5-12. A custom-made

cell equipped with a crystal window and Indium Tin Oxide (ITO) plate was used for

both the photo-assisted EC etching and spectroscopic measurements.

Figures 5-13 shows top and cross-sectional SEM images of the GaN-porous

samples formed after tEC = 10 and 30 min by comparing between the FSI and BSI

modes. The VEC and PIN were adjusted by monitoring the current to be about the same

between both modes. No anodic current and no porous structure was observed at the

same bias condition in the dark. For the FSI mode shown in Figs. 5-13(a) and (b), the

pore diameter measured from top images increased with anodization time, whereas the

pore depth Dp measured from the cross-section did not increase. On the other hand, Dp

Figure 5-13. Top and cross-sectional SEM images of GaN-porous samples formed in FSI mode

with VEC = 1.0 V, PIN = 5 mW/cm2 and different tEC: (a) tEC = 10 min and (b) tEC = 30 min, and

samples formed in BSI mode with VEC = 5.0 V, PIN = 65 mW/cm2 and different tEC: (c) tEC = 10 min

and (d) tEC = 30 min.


increased with anodization time in BSI mode, as shown in Figs. 5-13(c) and (d), where

the pore was the almost the same throughout all parts of the porous layer.

Figure 5-14 shows the relationship between Dp and the charge density Q

passing through the working electrode during the photo-assisted EC process. In FSI

mode, Dp increased with Q until around 400 mC/cm2 (tEC = 20 min), but it largely

decreased after that. In BSI mode, however, Dp increased almost linearly with Q,

showing no decrease in Q over 1000 mC/cm2 in this study. These results come from the

difference in the supply method of holes used for the anodization reaction. The FSI

mode generates holes near the top surface due to FSI. As Q increased to more than

about 400 mC/cm2 (tEC = 20 min), the pore wall thinned to breaking point and was

removed from the top-surface. In BSI mode, on the other hand, the holes were supplied

from the back surface or supplied only at the pore tips due to BSI. Since the anodic

reaction occurred only at the pore tips, the pore was etched in the vertical direction,

whose depth linearly increased with Q. From these results, we found that the BSI mode

is very powerful for controlling the structural properties of GaN-porous structures such

as pore diameter and Dp compared with the FSI mode. The BSI mode was first reported

on the formation of Si-porous structures [39]. Light with photon energy hν above 1.43

eV was irradiated, whose energy was larger than the bandgap of Si. The holes generated

at the back-surface diffused toward the pore tips to yield the anodic-reaction since the

diffusion length of minority carriers in Si was long enough. However, the situation in

BSI mode for GaN is quite different. Namely, the diffusion length of holes in n-type

Figure 5-14. Relationship between pore depth Dp and charge density Q passing through during

photo-assisted EC etching in FSI mode and BSI mode.

0

100

200

300

400

500

600

700

0 200 400 600 800 1000 1200

Po

re d

epth

Dp (

nm)

Charge density Q (mC/cm2)

FSI mode

BSI mode

88 Chapter 5

GaN is several hundred nm at most, which is much smaller than the sample thickness of

400 µm used in this study. In such a situation, the holes generated near the back-surface

cannot contribute to the anodic reaction because they recombine until reaching the pore

tips.

We conducted the photo-assisted EC process by changing monochromatic light

illumination using an optical filter. Table 5-II summarizes the anodic current and the

results of SEM observation on whether the porous structure was formed. The full width

at half maximum of the band-path filter was 10 nm. From a series of experiments, we

found that porous structures were formed on the GaN samples under monochromatic

light with wavelengths of 370 and 380 nm, whereas neither pore formation nor EC

etching was observed at wavelengths of 350, 390, and 400 nm. It should be noted that

the largest current was observed at a wavelength of 370 nm, corresponding to hν of 3.35

eV, which is smaller than the bandgap of GaN. These results clearly indicate that the

light with hν below the bandgap contributes to the formation of GaN porous structures

in BSI mode.

To clarify how the photo-absorption process affects EC reactions,

spectro-electrochemical measurements were conducted on planar GaN substrates.

Figure 5-15(a) shows transmittance spectra for T0V and T10V obtained by applying

anode voltages of VEC = 0 and 10 V, respectively, in the EC setup shown in Fig. 5-12.

We could observe a strong dependence of the transmittance on VEC at around 3.3 eV.

The transmittance difference T0V − T10V is plotted in Fig. 5-15(b) as a function of the

Table 5-II. Summary of the relationships between EC conditions and anodic reactions with

applied voltage VEC of 3.0 V and light intensity PIN of 3.0 mW/cm2.

Wavelength (nm) / Photon energy (eV) Current density (A/cm2) Etching (reaction) results

350 / 3.54 < 10−6

No etching

360 / 3.44 0.54 ~ 1.69×10−5

Shallow etching

370 / 3.35 2.76 ~ 8.39×10−4

Pore formation

380 / 3.26 0.44 ~ 2.37×10−4

Pore formation

390 / 3.18 < 10−6

No etching

400 / 3.10 < 10−6

No etching


photon energy hν. The difference was almost zero at energies higher than Eg because of

absorption by the bulk GaN. On the other hand, larger differences could be observed at

energies below Eg, reaching a maximum at around 3.3 eV and decreasing as hν

decreased. These results indicate that photoabsorption near the band edge can be

enhanced by the electric field mostly applied at the GaN/electrolyte interface.

One possible phenomenon to explain the above results is the Franz-Keldysh

effect (FKE), which causes the redshift of the absorption edge under a high electric field,

leading to the occurrence of absorption below the bulk bandgap Eg [40,41]. Generally, a

high electric field is applied by the formation of a potential barrier of about 1.0 eV at the

GaN/electrolyte interface [6,42]. When the doping density is as high as 1018 cm−3, for

example, the internal electric field in a thin depletion layer with a width of several

dozen nanometers reaches 5 × 105 V/cm. In such a situation, the electron and hole wave

functions become Airy functions, which extend into the bandgap, enabling band-to-band

transition at energies below Eg. In the depletion region, the Franz-Keldysh relative

absorption coefficient α can be approximated as [43,44],

23

g

g0 Δ

ναα

E

Ehexp~ . (5-5)

Figure 5-15. (a) Transmittance spectra of a GaN planar electrode, obtained by applying anode

voltages of VEC = 0 and 10 V. (b) Transmittance difference T0V − T10V plotted as a function of the

photon energy hν.

-0.01

0

0.01

0.02

0.03

0.04

0.05

3.1 3.2 3.3 3.4 3.5

Tra

nce

mitt

ance

diff

eren

ce T

0 V

- T

10

V

Photon energy (eV)

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

3.1 3.2 3.3 3.4 3.5

Tra

nsm

ittan

ce T

Photon energy (eV)

(a) (b)

T0V

T10V

EgEg − ΔEg

（i）（ii）（iii）

experimental

theoretical

90 Chapter 5

Here, α0 is the absorption coefficient at hν = Eg, and ΔEg is the redshift energy of the

absorption edge, expressed by

3

1

2

3

gξ

3

2Δ

*m

eE

, (5-6)

where q, ћ, ξ, and m* are the elementary charge, Dirac's constant, electric field, and

effective mass, respectively. The transmittance T is expressed as a function of the

reflectance R, absorption coefficient α, and optical path length x, as follows:

xexpRT α1 . (5-7)

Through calculations using Eqs. (5-5)−(5-7), we estimated the transmittance

difference T0V − T10V, plotted in Fig. 5-15(b) with a dotted line. The experimental data

are very consistent with the prediction of the FKE. In the case of illumination with hν >

Eg [case (i)], T0V − T10V was almost zero since all photons were absorbed near the

surface. For illumination with Eg − ΔEg < hν < Eg [case (ii)], the photons penetrated

through the bulk GaN but were absorbed at the GaN/electrolyte interface only under a

high electric field because of the FKE. In that situation, T0V − T10V increased up to an

observable level, as shown in Fig. 5-15(b). Finally, for illumination with hν < Eg − ΔEg

[case (iii)], since photons were not absorbed either in the bulk GaN or at the

GaN/electrolyte interface, T0V − T10V decreased again. These results strongly suggest

that the red-shift of the absorption edge observed for the formation of GaN-porous

structures arises from the FKE.

5.5.2. Structural characterization of GaN porous structures formed by

FKE-assisted EC etching

Figure 5-16(a) shows an SEM image of GaN porous structures formed with tEC

= 40 min and VEC = 10 V. As discussed in Section 5.4, vertically aligned pores can be

obtained by anisotropic EC etching with superior depth controllability. The pore

diameter was 18 nm, which could not be controlled throughout the EC etching in the

dark. Figures 5-16(b) and (c) show SEM images of GaN porous structures formed at

VEC = 10 V and tEC = 40 min, under monochromatic light with energy hν values of (b)

3.54 and (c) 3.26 eV. Under illumination with hν = 3.54 eV, which is larger than the


bandgap Eg of GaN, only pores near the top surface were affected by the light. In other

words, the pore diameter increased only in the top region, to a depth of about 200 nm,

as compared with that of the samples formed in the dark. From SEM observations, an

average pore diameter of 50 nm was obtained in the top region, which is approximately

30 nm larger than the pore diameter in the deeper region. On the other hand, under

illumination with hν = 3.26 eV (< Eg), the pore diameter increased to approximately 40

nm throughout the entire porous layer. Figure 5-17 shows the average pore diameter

plotted as a function of the light intensity PIN. We found that the pore diameter increased

linearly with PIN. These results clearly indicate that FKE involves in EC etching under

these conditions.

5.5.3. Pore formation mechanism in FKE-assisted EC etching

To investigate the possibility of applying the FKE, we calculated the potential

distribution of GaN porous structures in an electrolyte by solving a 3D Poisson equation.

Figure 5-18(a) shows the potential distribution of the conduction band minimum

measured from the Fermi level, EC − EF, with the solid lines indicating contours at 0.2,

Figure 5-16. Cross-sectional SEM images of GaN porous structures formed under various

electrochemical conditions: (a) dark, tEC = 40 min, VEC = 10 V; (b) illuminated, tEC = 40 min, VEC =

10 V, hν = 3.54 eV; (c) illuminated, tEC = 40 min, VEC = 10 V, hν = 3.26 eV.

92 Chapter 5

0.4, 0.6, 0.8, and 1.0 eV. The pore pitch, depth, and potential difference between the

GaN and electrolyte were specified as 25 nm, 50 nm, and 1.0 eV, respectively. Although

no external voltage is applied, the potential distribution is strongly modified near the

pore tips, where the contour intervals become narrow. Figure 5-18(b) shows

cross-sectional potential distributions obtained at the top surface (line i) and pore tips

(line ii). The potential variation at the pore tips is much larger than that at the top

surface, reflecting the high electric field applied to the tips.

On the basis of the above discussion, we propose a formation model for GaN

porous structures by FKE-assisted EC etching. For formation in the dark, the EC

etching is performed by the carriers generated by the avalanche effect as discussed in

Section 5.4. That is why the process of forming GaN porous structures in the dark can

achieve superior linearity and depth controllability but cannot vary the pore diameter.

Under illumination with hν > Eg, photons are preferentially absorbed near the top

surface because of band-edge absorption. Photo-generated carriers are then transferred

to the top surface and pore walls by the electric field, and contribute to the EC etching.

This explains the larger pore region with a thickness of 200 nm, a value consistent with

the light penetration depth in GaN [see Fig. 5-16(b)]. Under illumination with Eg − ΔEg

< hν < Eg, however, photons are transmitted in the top region but absorbed only at the

pore tips where a high electric field is applied. In this situation, the photo-carriers

generated by the FKE are consumed at the pore tips and assist the EC etching. As seen

Figure 5-17. Relationship between average pore diameter Wp and light intensity PIN with hν =

3.26 eV. Wp was estimated from SEM observation of GaN porous structure formed by

FKE-assisted EC etching.

0

10

20

30

40

50

0 2 4 6 8 10

Ave

rage

po

re d

iam

eter

Wp (

nm)

Light intensity PIN

(mW/cm2)

formed under illuminationhν = 3.26 eVVEC = 10 V


from Fig. 5-17, in fact, the pore diameter increased with the light intensity, as the

number of carriers contributing to the EC etching increased. This formation model can

consistently explain both the experimental results and the theoretical model, enabling us

to conclude that the application of the FKE is useful in controlling the structural

properties of GaN porous structures.

5.6. Precise structural control of GaN porous structures by post wet etching

5.6.1. Post wet etching using TMAH

To explore the possibility of diameter control, we conducted wet chemical

etching utilizing TMAH as subsequent treatment after the EC etching. Figure 5-19

shows top and cross-sectional SEM images of GaN porous structures formed by

anisotropic EC etching (a) without and (b) with TMAH etching for 60 min. In order to

evaluate structural properties accurately, randomly oriented pore regions were removed

from the sample surface by reactive ion beam etching. By applying TMAH etching,

pore shape changed from a round to a hexagonal shape surrounded with 6 facets of

1−100. Average pore diameter Wp increased from 18 nm to 34 nm, but average pore

depth Dp remained unchanged after the TMAH treatment. These results indicate that

TMAH etching proceeds anisotropically in the horizontal direction, but not in the

Figure 5-18. (a) Potential distribution in GaN porous structures drawn with contour levels at 0.2,

0.4, 0.6, 0.8, and 1.0 eV, and (b) the cross-sectional potential distribution at the top surface (line

i) and pore tips (line ii) obtained by solving 3D Poisson equation.

-0.2

0

0.2

0.4

0.6

0.8

1

1.2

0 10 20 30 40 50

Pot

ent

ial E

C -

EF (

eV)

Depth from surface (nm)

line ii (pore tip)

line i (top)

i ii

(a) (b)

EC − EF

x

y

94 Chapter 5

vertical direction of the substrate.

Generally, anisotropy in wet chemical etching is attributed to difference of

etching rate; in other words, a crystallographic face with the minimum etching rate

would appear. From this point of view, etching rate of the (0001) plane was almost zero,

and etching rate of the 1−100 plane was smaller than those of the other planes except

for the (0001) plane. Similar results are reported by a number of previous studies on wet

chemical etching of GaN using alkaline solution [45−48]. Etching-rate difference on

crystallographic face is qualitatively explained by the difference of dangling bond

density (DBD), which is a function of surface energy [45,46]. The DBDs of the (0001)

and 1−100 planes are 11.4 and 12.1 nm−2, respectively, which is smaller than those of

the other planes (e.g., 11−20 plane (14.0 nm−2) and 1−101 plane (16.0 nm−2)),

indicating that these planes are relatively stable toward wet chemical etching [49]. In

addition, surface polarity is also associated with etching selectivity [47,48]. On the

Ga-polar (0001) plane, used in this study, three occupied dangling bonds of nitrogen

appear after the 1st gallium layer is removed by the attack of hydroxide ions in the

alkaline solution. Hydroxide ions cannot attack the 2nd nitrogen layer after that because

of the large coulomb repulsion between hydroxide ions and negatively charged dangling

bonds. Such self-stopping behavior would not appear on N-polar and non-polar surfaces.

We believe both etching mechanisms related to DBD and surface polarity are involved,

and hexagonal pores surrounded with facets of 1−100 can be obtained by TMAH

etching.

Figure 5-19. Top and cross-sectional SEM images of GaN porous structures formed by

anisotropic EC etching (a) without TMAH etching, (b) with TMAH etching for 60 min.


Figure 5-20 shows the average pore diameter Wp plotted as a function of

TMAH etching time tTMAH, and insets show top SEM images of pores with different

tTMAH. There is a linear relationship between Wp and tTMAH with the slow etching rate of

0.27 nm/min. Hence, post wet etching using TMAH enables us to control pore diameter

precisely as well as independently from pore depth because of its anisotropic nature.

5.6.2. Correlation between structural and optical properties

In order to gain deeper clarity about the superiority of the present GaN porous

structures, optical and photoelectrochemical measurements were conducted. Figure

5-21(a) shows normalized PL spectra in the UV region obtained at room temperature

from non-porous GaN and porous GaN without and with TMAH etching for 60 min.

Although all samples exhibited strong near-band-edge (NBE) emission around 3.4 eV,

emission peaks of two porous samples shifted toward higher energies—so-called “blue

shift”. The amounts of blue shift were 5.9 meV and 13.9 meV for the sample just after

the anisotropic EC etching and for the sample with subsequent TMAH etching,

respectively. Observation of blue shift by formation of porous structures is completely

different from a number of studies on porous GaN, where the emission peaks shifted

toward lower energies—so-called “red shift” [25,28]. In the literature, it has been

reported that red shift of NBE originates from a strain-relaxation effect, which could be

Figure 5-20. Relationship between average pore diameter Wp and TMAH etching time tTMAH.

Insets shows top SEM images of pores with different tTMAH.

10

20

30

40

50

0 10 20 30 40 50 60

TMAH etching time tTMAH

(min)

Ave

rag

e po

re d

iam

ete

r W

p (nm

)

etching rate:0.27 nm/min

0 min 15 min 30 min 45 min 60 min

96 Chapter 5

caused by forming nanostructures on hetero-epitaxial GaN layers having high residual

stress. In contrast, the freestanding GaN substrates used in this study were prepared by a

substrate separation process that reduced substrate-induced strain [50], where the

strain-relaxation effect after the porous formation could be negligible. We believe that

the observed blue shift is probably due to the quantum confinement in pore walls,

similar to the case of porous Si, InP and our previous reports on porous GaN. After the

formation of porous structures, carriers were confined in two dimensions in thinned

pore-walls, forming higher quantized-energy levels than the bandgap energy level.

Since the quantum confinement energy is inversely proportional to the square of pore

wall width Wpw, larger blue shift was observed on the porous structures with thinner

walls obtained after TMAH etching.

Figure 5-21(b) shows the PL spectra obtained in the region from 1.8 eV to 2.6

eV, which are smaller than the bandgap of GaN. Yellow luminescence (YL) was clearly

observed around 2.2 eV on the non-porous GaN. The origin of YL has been intensively

discussed for a long time from both the experimental and theoretical points of view, and

the accepted stance is that it is likely caused by Ga vacancy VGa and its complex with O

or C [51−53]. Application of dry etching often leads to an increase of YL indicating the

formation of high-density point defects. As shown in Fig. 5-21(b), YL intensity was

decreased by the formation of porous structures, suggesting that VGa-related defects

were preferentially removed by EC etching. In addition, TMAH etching barely affected

Figure 5-21. Normalized photoluminescence spectra of (a) UV region, and (b) visible region

obtained at room temperature: Black line represents planar GaN, blue line represents porous

GaN without TMAH etching, and red line represents porous GaN with TMAH etching for 60 min.


intensity of the YL peak. These results indicate that anisotropic EC and TMAH etching

can process GaN without any damage induced.

Figure 5-22(a) shows specular reflectance spectra of a non-porous and porous

GaN with different tTMAH. Reflectance spectra of all porous samples were lower than

those of the non-porous sample, as expected. In addition, reflectance oscillations were

observed only on porous samples in the visible and near infrared regions. Local peak

positions and period were changed with tTMAH, indicating that oscillations were

originated from optical interference between reflected light of the air/porous layer

interface and that of the porous layer/substrate interface.

Referring to the effective medium approximation (EMA) model, we assumed

the porous layer as an air-composite material that had an effective refractive index np

calculated by the deformed Bruggeman equation as follows:

012p

2air

2p

2air

GaN2p

2GaN

2p

2GaN

GaN

nn

nnf

nn

nnf , (5-8)

where nair and nGaN are refractive index of air and GaN and fGaN is volume fraction of

Figure 5-22. (a) Specular reflectance spectra of planar (black line) and porous GaN (colored

line) with different TMAH etching time tTMAH. (b) The tTMAH dependency of effective refractive

index of porous GaN calculated in accordance with effective medium approximation model (red

dots) and thin-film interference model (black line).

98 Chapter 5

GaN in the porous layer. In the case of the porous structures formed in this study, the

fGaN can easily be estimated from SEM images because the pores with a constant

diameter were formed straightly in the vertical direction. In addition to the Bruggeman

approach, we tried to estimate np with a thin-film interference model from reflectance

properties. In this configuration (nair < np < nGaN), constructive interference occurred

when the difference of optical path length corresponded with integral multiple of

wavelength as follows:

lmaxpp λ2 mDn (5-9)

where m is an integer and λlmax is specific wavelength of local maximum.

The np values calculated from the structural properties using Eq. (5-8) (red

circles) and reflectance properties using Eq. (5-9) (black solid line) are plotted in Fig.

5-22(b) as a function of TMAH etching time. Application of these models seems to be

valid and feasible because the red circles and the black solid line are consistent. Both

models indicated np decreased approximately linearly with tTMAH whose change rate of

5.2 × 10−3 min−1 was very slow. From these results, superior structural controllability of

anisotropic EC and TMAH etching enable us to obtain a surface with a desirable

refractive index.

Finally, we evaluated the capability of our GaN porous structures as EC energy

conversion systems. Figure 5-23(a) shows the current-voltage (I-V) characteristics of

various GaN electrodes measured in 0.1 mol/L PBS under irradiation of monochromatic

light (λ = 350 nm, PIN = 0.1 mW/cm2). For the I-V curves measured in the dark, all

samples showed rectifying properties similar to the case of Schottky barrier diodes.

Here, negative currents were observed at negative bias because cathodic reactions were

caused by electrons, which are majority carriers in n-GaN. On the other hand, few

positive currents were observed at the entire bias range because anodic reactions were

barely caused due to lack of holes, which are minority carriers in n-GaN.

Under irradiation of light, positive currents were observed as photocurrents on

all samples, indicating photo-generated carriers were separated by potential gradient in

SCR and contributed to EC reactions. Rest potentials, at which positive current started

to flow, of all samples were almost the same value of −0.58 V, but the bias-dependent

behavior of photocurrents is changed by the surface geometry of GaN. The photocurrent

of non-porous GaN started to flow at around −0.5 V and kept increasing slightly until


1.0 V. This is a typical behavior observed on planar photoelectrodes, where the amount

of charge carriers contributing to the photocurrents increased with the SCR width WSCR

as a function of the applied voltage, as described in Eq. (5-4) [6]. On the other hand, the

photocurrents of porous GaNs saturated at specific voltage, showing less

dependent-change with the applied voltage. In porous samples, charge carriers would be

generated in the porous region, not in the bulk region underneath, because light

penetration depth is much shorter than pore depth. Therefore, the charge carriers can

easily reach the SCR formed in the pore walls even though the pore wall is not

completely depleted and WSCR is small at low bias such as 0 V. In such situation, the

charge carriers generated in the pore walls can be immediately separated, leading to the

steep increase of photocurrents up to the saturation value. This phenomenon of

"efficient carrier separation" is one of the attractive features and big advantages of

porous structures for EC energy conversion systems because the highest photocurrent

can be obtained even at a lower bias such as 0 V. In addition, we found that

photocurrent-saturated bias shifted from 0.6 V to −0.1 V by subsequent TMAH etching.

This was because the increase of pore diameter Wp enhanced "efficient carrier

separation" due to the increase of the ratio of WSCR to the pore wall width Wpw.

Figure 5-23(b) shows the incident-photon-to-current conversion efficiency

(IPCE) measured at 0 V as a function of light wavelength λ. IPCE of 54 % at 350 nm

Figure 5-23. (a) Current-voltage characteristics under irradiation of monochromatic light (λ = 350

nm, PIN = 0.1 mW/cm2), and (b) IPCE at 0V plotted as a function of wavelength: Black line

represents planar GaN, blue line represents porous GaN without TMAH etching, and red line

represents porous GaN with TMAH etching for 45 min.

-0.01

0

0.01

0.02

0.03

0.04

-1 -0.5 0 0.5 1

Cu

rren

t de

nsity

(m

A c

m-2

)

Voltage (V)

dark current

planar GaN

porous GaN w/ TMAH

porous GaN w/o TMAH

0

20

40

60

80

100

120

350 360 370 380 390 400

Wavelength (nm)

IPC

E (

%)

planar GaN

porous GaN w/o TMAH

porous GaN w/ TMAH

(a) (b)

100 Chapter 5

was obtained on non-porous GaN, which is almost the same value as a previous report

[54]. Compared to that, IPCE was enhanced by the formation of porous structures to as

high as 70 %, and enhanced furthermore by subsequent TMAH etching to as high as

91 %. These results clearly demonstrate that the formation of porous structures is

effective for yielding high IPCE in EC energy conversion systems due to their attractive

features mentioned above. It is also noted that photocurrent under the light with photon

energy below the GaN bandgap was dramatically improved by the formation of porous

structures. We believe this is due to enhancement of Franz-Keldysh effect: namely, the

high electric field induced at pore tips causes redshift of the photoabsorption edge [55].

We further investigated the effect of TMAH etching time tTMAH on IPCE. As

shown in Fig. 5-24(a), IPCE increased from 70 % to 91 % with tTMAH up until 45 min,

but it decreased with tTMAH = 60 min, although the structural change was only several

nanometers. In order to explain the IPCE behavior toward tTMAH, we assumed that the

separation- and correction-efficiency of the photo-carriers generated in the pore walls is

strongly affected by the width of pore wall Wpw. Figs. 5-24(b) and (c) shows the width

of quasi-neutral region WQNR and flat band potential VFB, respectively, whose definition

and relation are schematically described in Fig. 5-24(d). If the Wpw is more than twice

the size of the WSCR, the pore wall consists of SCRs and the quasi-neutral region

Figure 5-24. The tTMAH dependency of IPCE at 0 V,

(b) Width of quasi-neutral region WQNR, and (c)

modulus of flat band potential |VFB|. (d) Schematic

representations of energy band diagrams in pore

wall if Wpw > 2WSCR (left) or Wpw < 2WSCR (right). 0.3

0.4

0.5

0.6

0.7

0.8

0.9

0 10 20 30 40 50 60

Fla

t ban

d po

tent

ial |

VF

B| (

eV)

TMAH etching time tTMAH

(min)

0

4

8

12

16

20

24

Wid

th o

f qua

si-n

eutr

alre

gion

WQ

NR (

nm)

60

70

80

90

100

IPC

E (

%)

(a)

(b)

(c)

(d)

EC

EV

quasi-neutralregion

SCR SCR SCR

hole

electron

WSCRWQNR

Wpw Wpw

hole

electron

Wpw > 2WSCR Wpw < 2WSCR

VFBVFB


illustrated in the left diagram of Fig. 5-24(d). Here, WQNR and VFB can be represented as

SCRpwQNR 2WWW . (5-10)

2SCR

0GaN

DFB εε2

WqN

V . (5-11)

If the Wpw is less than twice the size of the WSCR, the pore wall is fully depleted

and the quasi-neutral region disappears, as illustrated in the right diagram of Fig.

5-24(d). In such situation, WQNR will be zero and VFB can be obtained by the following

approximation:

2

pw

0GaN

DFB 2εε2

WqNV . (5-22)

As shown in Fig. 5-24(b), WQNR decreased with tTMAH in a reflection to the Wpw

changed by the TMAH treatment, where WQNR becomes zero with tTMAH ≥ 45 min. The

VFB was changed closely in the relation between WQNR and Wpw as shown in Fig. 5-24(c).

Namely, the VFB started to decrease at around tTMAH = 45 min when the pore walls were

fully depleted and the WQNR disappeared. The VFB further decreased to 0.4 eV with

tTMAH = 60 min, whose value is almost half of the initial value obtained with tTMAH = 0

min.

The semi-qualitative analysis above indicates that the highest IPCE value was

obtained when the WQNR is just on zero at around tTMAH = 45 min. The contribution of

the photo-carriers for the photocurrents becomes larger as the WQNR decreases, because

"efficient carrier separation" was enhanced. On the other hand, the ability of the

carrier-separation becomes smaller as the VFB decreases, because the electric field in the

SCR decreases. Lower VFB also led to photocurrent reduction by recombination in SCR

region, as predicted by Sah et al [56]. From the above discussion, one possible reason

the highest IPCE could be obtained with tTMAH = 45 min is that recombination is

suppressed because Wpw is almost the same with twice the WSCR. We conclude that

precise structural controlling is crucially important to obtain superior capability for EC

energy conversion systems, and the two-step process utilizing anisotropic EC etching

and TMAH etching is very promising as a nanostructure fabrication technique.

102 Chapter 5

5.7. Summary

In Chapter 5, we have developed various EC etching techniques to fabricate

GaN porous structures for the application to the EC energy conversion systems.

Summary of this chapter is itemized below.

1) By adopting free-standing GaN substrate as starting substrate, various kinds of

porous structures could be formed independent of dislocations which made it

difficult to control structural properties precisely.

2) In photo-assisted EC etching, pore depth was nonlinear against etching time

and there were difficulty in formation of deeper pores than micro-meter

because carriers most of which were generated near surface caused top and

lateral etching along with pore growth.

3) EC etching related to avalanche effect proceeded in the direction of [000−1]

anisotropically, enabling superior controllability in depth with incredible aspect

ratio.

4) Franz-Keldysh effect (FKE) gave us further possibility of improving structural

controllability: FKE occurred only at pore tips where a high electric field was

induced, enabling the tuning of the diameter without diminishing the linearity

or depth controllability which was achieved by avalanche effect.

5) Post wet etching utilizing TMAH was also a good technique to control pore

diameter because of its anisotropic nature.

6) From PL measurements, porous samples showed blue shift of NBE emission

due to quantum confinement in the pore wall. Reduction of YL intensity were

also observed, indicating VGa-related defects were preferentially removed and

additional damage was not induced by this method.

7) Photoreflectance measurement revealed that porous samples had effective

refractive index that could be controlled by TMAH etching time.

8) In photo-electrochemical measurement, IPCE was dramatically enhanced to as

high as 91 % by formation of porous structures with adequate structural

properties.


These results clearly denote that precisely-controlled GaN porous structure by

EC etching process are very promising as photoelectrode in EC energy conversion

systems.

104 Chapter 5

Reference

[1] A. Fujishima, and K. Honda, "Electrochemical Photolysis of Water at a

Semiconductor Electrode", Nature, vol. 238, pp. 37−38, 1972.

[2] K. Maeda, K. Teramura, D. L. Lu, T. Takata, N. Saito, Y. Inoue, and K. Domen,

"Photocatalyst releasing hydrogen from water", Nature, vol. 440, p. 295, 2006.

[3] E. E. Barton, D. M. Rampulla, and A. B. Bocarsly, "Selective Solar-Driven

Reduction of CO2 to Methanol Using a Catalyzed p-GaP Based

Photoelectrochemical Cell", J. Am. Chem. Soc., vol. 130, pp. 6342−6344,

2008.

[4] S. Sato, T. Arai, T. Morikawa, K. Uemura, T. M. Suzuki, H. Tanaka, and T.

Kajino, "Selective CO2 Conversion to Formate Conjugated with H2O

Oxidation Utilizing Semiconductor/Complex Hybrid Photocatalysts", J. Am.

Chem. Soc., vol. 133, pp. 15240−15243, 2011.

[5] B. Oregan, and M. Gratzel, "A low-cost, high-efficiency solar cell based on

dye-sensitized colloidal TiO2 films", Nature, vol. 353, pp. 737−740, 1991.

[6] I. M. Huygens, K. Strubbe, and W. P. Gomes, "Electrochemistry and

Photoetching of n-GaN", J. Electrochem. Soc., vol. 147, pp. 1797−1802, 2000.

[7] J. D. Beach, R. T. Collins, and J. A. Turner, "Band-Edge Potentials of n-Type

and p-Type GaN", J. Electrochem. Soc., vol. 150, pp. A899−A904, 2003.

[8] K. Fujii, T. K. Karasawa, and K. Ohkawa, "Hydrogen Gas Generation by

Splitting Aqueous Water Using n-Type GaN Photoelectrode with Anodic

Oxidation", Jpn. J. Appl. Phys., vol. 44, pp. L543−L545, 2005.

[9] S. Yotsuhashi, M. Deguchi, Y. Zenitani, R. Hinogami, H. Hashiba, Y. Yamada,

and K. Ohkawa, "Photo-induced CO2 Reduction with GaN Electrode in

Aqueous System", Appl. Phys. Express, vol. 4, p. 117101, 2011.

[10] Y. D. Wang, K. Y. Zang, and S. J. Chua, "Nonlithographic nanopatterning

through anodic aluminum oxide template and selective growth of highly

ordered GaN nanostructures", J. Appl. Phys., vol. 100, p. 054306, 2006.

[11] I. Waki, D. Cohen, R. Lal, U. Mishra, S. P. DenBaars, and S. Nakamura,


"Direct water photoelectrolysis with patterned n-GaN", Appl. Phys. Lett., vol.

91, p. 093519, 2007.

[12] H. Ono, Y. Ono, K. Kasahara, J. Mizuno, and S. Shoji, "Fabrication of

High-Intensity Light-Emitting Diodes Using Nanostructures by Ultraviolet

Nanoimprint Lithography and Electrodeposition", Jpn. J. Appl. Phys., vol. 47,

pp. 933−935, 2008.

[13] W. M. Zhou, G. Q. Min, Z. T. Song, J. Zhang, Y. B. Liu, and J. P. Zhang,

"Enhanced efficiency of light emitting diodes with a nano-patterned gallium

nitride surface realized by soft UV nanoimprint lithography", Nanotechnology,

vol. 21, p. 405304, 2010.

[14] E. D. Haberer, C. H. Chen, A. Abare, M. Hansen, S. Denbaars, L. Coldren, U.

Mishra, and E. L. Hu, "Channeling as a mechanism for dry etch damage in

GaN", Appl. Phys. Lett., vol. 76, pp. 3941−3943, 2000.

[15] R. Dimitrov, V. Tilak, W. Yeo, B. Green, H. Kim, J. Smart, E. Chumbes, J. R.

Shealy, W. Schaff, L. F. Eastman, C. Miskys, O. Ambacher, and M. Stutzmann,

"Influence of oxygen and methane plasma on the electrical properties of

undoped AlGaN/GaN heterostructures for high power transistors", Solid-State

Electron., vol. 44, pp.1361−1365, 2000.

[16] F. A. Khan, L. Zhou, V. Kumar, and I. Adesida, "Plasma-induced damage study

for n-GaN using inductively coupled plasma reactive ion etching", J. Vac. Sci.

Technol. B, vol. 19, pp. 2926−2929, 2001.

[17] A. Uhlir, "Electrolytic shaping of germanium and silicon", Bell Syst. Tech. J.,

vol. 35, pp. 333−347, 1956.

[18] Y. Kumazaki, T. Kudo, Z. Yatabe, and T. Sato, "Investigation on optical

absorption properties of electrochemically formed porous InP using

photoelectric conversion devices", Appl. Surf. Sci., vol. 279, pp. 116−120,

2013.



photoelectrochemical solar cells", Thin Solid Films, vol. 518, pp. 4399−4402,

2010.

106 Chapter 5

[20] V. Lehmann, "The Physics of Macropore Formation in Low Doped n-Type

Silicon", J. Electrochem. Soc., vol. 140, pp. 2836−2843, 1993.

[21] X. G. Zhang, "Morphology and Formation Mechanisms of Porous Silicon", J.

Electrochem. Soc., vol. 151, pp. C69−C80, 2004.

[22] B. H. Erne, D. Vanmaekelbergh, and J. J. Kelly, "Morphology and Strongly

Enhanced Photoresponse of GaP Electrodes Made Porous by Anodic Etching",

J. Electrochem. Soc., vol. 143, pp. 305−314, 1996.

[23] W. Shin, T. Hikosaka, W. S. Seo, H. S. Ahn, N. Sawaki, and K. Koumoto,

"Fibrous and Porous Microstructure Formation in 6H-SiC by Anodization in

HF Solution", J. Electrochem. Soc., vol. 145, pp. 2456−2460, 1998.

[24] T. Sato, T. Fujino, and T. Hashizume, "Electrochemical Formation of

Size-Controlled InP Nanostructures Using Anodic and Cathodic Reactions",

Electrochem. Solid State Lett., vol. 10, pp. H153−H155, 2007.

[25] A. P. Vajpeyi, S. J. Chua, S. Tripathy, and E. A. Fitzgerald, "Effect of carrier

density on the surface morphology and optical properties of nanoporous GaN

prepared by UV assisted electrochemical etching", Appl. Phys. Lett., vol. 91, p.

083110, 2007.

[26] H. Hartono, C. B. Soh, S. J. Chua, and E. A. Fitzgerald, "High Quality GaN

Grown from a Nanoporous GaN Template", J. Electrochem. Soc., vol. 154, pp.

H1004−H1007, 2007.

[27] D. T. Chen, H. D. Xiao, and J. Han, "Nanopores in GaN by electrochemical

anodization in hydrofluoric acid: Formation and mechanism", J. Appl. Phys.,

vol. 112, p. 064303, 2012.

[28] K. Al-Heuseen, M. R. Hashim, and N. K. Ali, "Growth and Characterization of

Tree-Like Crystalline Structures during Electrochemical Formation of Porous

GaN", J. Electrochem. Soc., vol. 158, pp. D240−D243, 2011.

[29] K. P. Beh, F. K. Yam, L. K. Tan, S. W. Ng, C. W. Chin, and Z. Hassan,

"Photoelectrochemical Fabrication of Porous GaN and Their Applications in

Ultraviolet and Ammonia Sensing", Jpn. J. Appl. Phys., vol. 52, p. 08JK03,

2013.


[30] D. Zhuang, and J. H. Edgar, "Wet etching of GaN, AlN, and SiC: a review",

Mater. Sci. Eng. R, vol. 48, pp. 1−46, 2005.

[31] Y. Oshima, T. Eri, M. Shibata, H. Sunakawa, K. Kobayashi, T. Ichihashi, and A.

Usui, "Preparation of Freestanding GaN Wafers by Hydride Vapor Phase

Epitaxy with Void-Assisted Separation", Jpn. J. Appl. Phys., vol. 42, pp. L1−L3,

2003.

[32] P. Visconti, K. M. Jones, M. A. Reshchikov, R. Cingolani, H. Morkoc, and R. J.

Molnar, "Dislocation density in GaN determined by photoelectrochemical and

hot-wet etching", Appl. Phys. Lett., vol. 77, pp. 3532−3534, 2000.

[33] T. Sato, and A. Mizohata, "Photoelectrochemical Etching and Removal of the

Irregular Top Layer Formed on InP Porous Nanostructures", Electrochem.

Solid State Lett., vol. 11, pp. H111−H113, 2008.

[34] C. Youtsey, L. T. Romano, and I. Adesida, "Gallium nitride whiskers formed by

selective photoenhanced wet etching of dislocations", Appl. Phys. Lett., vol. 73,

pp. 797−799, 1998.

[35] C. B. Soh, C. B. Tay, R. J. N. Tan, A. P. Vajpeyi, I. P. Seetoh, K. K.

Ansah-Antwi, and S. J. Chua, "Nanopore morphology in porous GaN template

and its effect on the LEDs emission", J. Phys. D: Appl. Phys., vol. 46, p.

365102, 2013.

[36] L. T. Canham, "Silicon quantum wire array fabrication by electrochemical and

chemical dissolution of wafers", Appl. Phys. Lett., vol. 57, pp. 1046−1048,

1990.

[37] T. Fujino, T. Sato, and T. Hashizume, "Size-Controlled Porous Nanostructures

Formed on InP(001) Substrates by Two-Step Electrochemical Process", Jpn. J.

Appl. Phys., vol. 46, pp. 4375−4380, 2007.

[38] M. I. J. Beale, N. G. Chew, M. J. Uren, A. G. Cullis, and J. D. Benjamin,

"Microstructure and formation mechanism of porous silicon", Appl. Phys. Lett.,

vol. 46, pp. 86−88, 1985.

[39] V. Lehmann, and H. Fӧll, "Formation Mechanism and Properties of

Electrochemically Etched Trenches in n-Type Silicon", J. Electrochem. Soc.,

108 Chapter 5

vol. 137, pp. 653−659, 1990.

[40] W. Franz, "Einfluß eines elektrischen Feldes auf eine optische

Absorptionskante", Z. Naturforsch. vol. 13a, pp. 484−489, 1958. [in German]

[41] L. V. Keldysh, "Behavior of Non-metallic Crystals in Strong Electric Fields",

Sov. Phys. JETP , vol. 6, pp. 763−770, 1958.

[42] S. S. Kocha, M. W. Peterson, D. J. Arent, J. M. Redwing, M. A. Tischler, and J.

A. Turner, "Electrochemical Investigation of the Gallium Nitride-Aqueous

Electrolyte Interface", J. Electrochem. Soc., vol. 142, pp. L238−L240, 1995.

[43] H. Y. Peng, M. D. McCluskey, Y. M. Gupta, M. Kneissl, and N. M. Johnson,

"The Franz-Keldysh effect in shocked GaN:Mg", Appl. Phys. Lett., vol. 82, pp.

2085−2087, 2003.

[44] A. Cavallini, L. Polenta, and M. Rossi, "Franz-Keldysh Effect in GaN

Nanowires", Nano Lett., vol. 7, pp. 2166−2170, 2007.

[45] J. Liu, J. Huang, X. Gong, J. Wang, K. Xu, Y. Qiu, D. Cai, T. Zhou, G. Ren,

and H. Yang, "A practical route towards fabricating GaN nanowire arrays",

CrystEngComm, vol. 13, pp. 5929−5935, 2011.

[46] W. Chen, J. Lin, G. Hu, X. Han, M. Liu, Y. Yang, Z. Wu, Y. Liu, and B. Zhang,

"GaN nanowire fabricated by selective wet-etching of GaN

microtruncated-pyramid", J. Cryst. Growth, vol. 426, pp. 168−172, 2015.

[47] D. Li, M. Sumiya, S. Fuke, D. Yang, D. Que, Y. Suzuki, and Y. Fukuda,

"Selective etching of GaN polar surface in potassium hydroxide solution

studied by x-ray photoelectron spectroscopy", J. Appl. Phys., vol. 90, pp.

4219−4223, 2001.

[48] M. Itoh, T. Kinoshita, C. Koike, M. Takeuchi, K. Kawasaki, and Y. Aoyagi,

"Straight and smooth etching of GaN (1-100) plane by combination of reactive

ion etching and KOH wet etching techniques", Jpn. J. Appl. Phys., vol. 45, pp.

3988−3991, 2006.

[49] K. Hiramatsu, K. Nishiyama, A. Motogaito, H. Miyake, Y. Iyechika, and T.

Maeda, "Recent Progress in Selective Area Growth and Epitaxial Lateral


Overgrowth of III-Nitrides: Effects of Reactor Pressure in MOVPE Growth",

Phys. Status Solidi A, vol. 176, pp. 535−543, 1999.

[50] V. Darakchieva, T. Paskova, P. P. Paskov, B. Monemar, N. Ashkenov, and M.

Schubert, "Residual strain in HVPE GaN free-standing and re-grown

homoepitaxial layers", Phys. Status Solidi A, vol. 195, pp. 516−522, 2003.

[51] T. Ogino, and M. Aoki, "Mechanism of Yellow Luminescence in GaN", Jpn. J.

Appl. Phys., vol. 19, pp. 2395−2405, 1980.

[52] J. Neugebauer, and C. G. Van de Walle, "Gallium vacancies and the yellow

luminescence in GaN", Appl. Phys. Lett., vol. 69, pp. 503−505, 1996.

[53] M. A. Reshchikov, and H. Morkoç, "Luminescence properties of defects in

GaN", J. Appl. Phys., vol. 97, p. 061301, 2005.

[54] T. Hayashi, M. Deura, and K. Ohkawa, "High Stability and Efficiency of GaN

Photocatalyst for Hydrogen Generation from Water", Jpn. J. Appl. Phys., vol.

51, p. 112601, 2012.

[55] T. Sato, Y. Kumazaki, H. Kida, A. Watanabe, Z. Yatabe, and S. Matsuda,

"Large photocurrents in GaN porous structures with a redshift of the

photoabsorption edge", Semicond. Sci. Technol., vol. 31, p. 014012, 2016.

[56] C. -T. Sah, R. N. Noyce, and W. Shockley, "Carrier Generation and

Recombination in P-N Junctions and P-N Junction Characteristics", Proc. IRE,

vol. 45, pp. 1228−1243, 1957.

110 Chapter 5

111

Chapter 6

Highly selective EC etching of p-GaN grown on

AlGaN/GaN heterostructures

6.1. Introduction

AlGaN/GaN-based high electron mobility transistors (HEMTs) and vertical

field effect transistors (FETs) are promising for high-power electronics applications

because high blocking voltage and yet low on-resistance can be achieved. In these

devices, the selective etching of p-type GaN or AlGaN layer is inevitable process. For

example, one of the promising approaches to realize a normally-off operation of

AlGaN/GaN HEMTs is by using a p-type (Al)GaN gate structure [1−3], in which the

p-type layer in the access regions should be selectively removed, as shown in Fig. 6-1.

For another example, the AlGaN/GaN vertical FETs equipped with a current aperture

structure [4−6] required the selective etching of p-GaN layer and regrowth of

AlGaN/GaN heterostructures on the etched surface.

Usually, the etching processes for nitrides are carried out in high-energy

environments because nitrides typically have high bond energies. These high energies

are usually supplied by accelerated ions or plasma-assisted species. However, use of

high-energy reactions tend to generate surface damage and defects [7,8], leading to

high-density surface states [9,10]. As for the AlGaN/GaN HEMTs, the various types of

damage will be induced in the AlGaN layer and/or the two-dimensional electron gas

(2DEG) channel region [11,12], causing to unintentional surface charge up and

degradation of electron mobility. Moreover, the precise control of etching depth is

strongly required because the excess etching of the AlGaN layer in the access region

may lead to the degradation of device performance due to increased access resistance.

Alternative approach for the low-damage etching is wet-chemical process in

which the etching reactions proceed by surface oxidation and dissolution of oxidants in

a chemical solution. Since the nitride materials show high-chemical stability, the

photo-assisted oxidation and etching have been commonly performed under

112 Chapter 6

illumination for the supply of electric holes [13,14]. Especially for n-type GaN, the

illumination of ultra-violet (UV) light with a photon energy above the bandgap is

absolute required because the electric holes are minority carriers in n-type material.

Recently, we have shown that the n-GaN and AlGaN/GaN structures applied by photo

electrochemical oxidation in a glycol solution improve the device performance by

suppressing the surface states due to the nature of low-energy process [15,16].

In Chapter 6, we demonstrated the EC etching of p-GaN layer grown on

AlGaN/GaN structures in the dark condition. Etching depth was precisely controlled by

the EC conditions. A series of experimental results showed that the self-stopping etching

was achieved with highly material-selectivity.


6.2.1. Basic concept of self-stopping etching using EC reactions

The EC etching process has been widely investigated on III-V semiconductor

materials. In one example for InP, the damage-free and flat-etched surface has been

obtained due to the nature of the low-energy process utilizing the EC reactions [17,18].

The EC etching of semiconductor surface is achieved by a cyclic process consists of an

anodic oxidation and a subsequent dissolution of the resulting oxide in an electrolyte. A

typical anodic reaction of GaN is described as follows:

2232 NO3HOGa4h6OHGaN (6-1)

Figure 6-1. Schematic illustrations of a typical AlGaN/GaN HEMT structure using p-GaN gate for

normally-off operation.

113 Highly selective EC etching of p-GaN grown on AlGaN/GaN heterostructures

Here, the Ga2O3 is easily dissolved into the pH-controlled acid or alkali electrolyte,

resulting the etching of GaN surface. It is noted that the morphology of the etched

surface strongly depends on the anodic conditions. For example, a porous nanostructure

is obtained by controlling an anodic voltage, anodic currents, illumination condition and

so on [19−21]. In this study, we optimized the anodic voltage to achieve the flat and

smooth surface for AlGaN/GaN transistors. The basic concept of self-stopping etching

process is schematically shown in Fig. 6-2. Holes are absolutely necessary for the EC

oxidation reaction, as shown in Eq. (6-1). It follows that a self-stopping etching of

p-GaN is possible due to lack of holes in i-AlGaN layer.

6.2.2. Experimental procedure

A p-GaN/AlGaN/GaN heterostructure schematically shown in Fig. 6-3(a) were

used as starting wafer. A Ni/Au ohmic contact was first fabricated on the p-GaN surface.

A SiO2 film was then deposited as an etching mask. EC etching experiments were

performed using a three-electrode EC cell as described in Chapter 3. The ohmic contact

formed on the p-GaN surface was electrically connected to an outer circuit, in which the

potential of the sample electrode was controlled with respect to the reference electrode

by a potentiostat. The electrolyte was a tartaric-acid-based solution. In this study, the

triangular waveforms of voltage VEC were repeatedly applied, as shown in Fig. 6-3(b).

The p-GaN surface was first oxidized at higher voltage, and then dissolved into the

pH-controlled electrolyte at lower voltage. In such a manner, the step-by-step etching of

p-GaN surface was controlled with the number of process cycle, as described in Fig. 6-2.

Figure 6-2. Schematic illustrations of electrochemical etching cycle: oxidation is caused by ions

and holes, followed by dissolution of resulting oxide.

Oxide (Ga2O3)

Dissolution

p-GaN p-GaN p-GaN

electrolyte electrolyte electrolyte

i-AlGaN (no holes) i-AlGaN (no holes) i-AlGaN (no holes)

ionion ion ion ion ion

Oxidation

hh

h

114 Chapter 6

The reaction current density, JEC between the sample and the counter electrodes was

monitored during the entire etching process.

The sample structure and surface morphology after the EC etching were first

characterized by atomic force microscopy (AFM) and transmission electron microscopy

(TEM) measurements. Elemental analysis on the etched surface was conducted using

micro-Auger electron spectroscopy (µ-AES) measurement. To evaluate electrical

properties of the etched surface, current-voltage (I-V) and capacitance-voltage (C-V)

Figure 6-4. (a) JEC-VEC characteristics of the p-GaN and i-AlGaN electrodes, and (b) JEC

observed on the sample electrode when a train of triangular-wave voltage pulses was applied.

The decrease of JEC with cycle was resulted from the decrease of thickness of p-GaN layer.

-2

0

2

4

6

8

10

12

0 1 2 3 4

Cur

rent

den

sity

(m

A/c

m2 )

Voltage (V)

p-GaN

i-AlGaN

-2

0

2

4

6

8

10

12

-1

0

1

2

3

4

5

6

0 200 400 600 800 1000

Voltag

e (V)

Cu

rre

nt d

ens

ity (

mA

/cm

2)

Time (s)

Sweep rate : 50 mV/sdark

Red

Ox

⇒

EC

EF

EV

p-GaNelectrolyte

ERedox

oxidationeV

(a) (b)

Figure 6-3. Schematic illustrations of (a) a sample structure and (c) triangular waveform of

applied voltage. An ohmic contact and SiO2 mask were fabricated on the p-GaN surface.

SiO2 mask

i-GaN

buffer

i-AlGaN : 22 nm

p-GaN : 100 nm

p-Si substrate

Ohmic(Ti/Al/Ti/Au)

(a)

Vol

tage

VE

C(V

)

Time (s)

・・・

4

0

① ② n

(b)


measurements were performed on Schottky contacts fabricated on the AlGaN/GaN

structures after removing the p-GaN layer.

6.3. EC etching and structural characterizations

6.3.1. Reaction current transition during EC etching

In advance of the etching experiments, the typical JEC-VEC characteristics for a

p-GaN electrode and an i-AlGaN electrode were compared. As shown in Fig. 6-4(a), the

reaction currents due to the surface oxidation were clearly observed on the p-GaN

electrode. This is due to the fact that the p-GaN can easily supply the required holes to

the electrolyte under the anodic bias as schematically shown in the inset of Fig. 6-4(a).

In contrast, an i-AlGaN electrode showed no reaction current, indicating no oxidation

reaction on i-AlGaN surface due to lack of holes. Figure 6-4(b) shows JEC observed on

a p-GaN electrode when a train of triangular-wave voltage was applied, where the VEC

was swept from −0.5 V to 4.0 V. Reaction current density JEC decreased with the

increasing number of cycles, finally being 0.1 % of or less than the initial peak current.

This behavior of JEC indicates that the surface oxidation and the dissolution repeatedly

occurred on p-GaN surface, resulting in the decrease of JEC with decreasing thickness of

p-GaN layer, as described in Fig. 6-2.

Figure 6-5. The typical (a) SEM and (b) AFM images of the sample after the electrochemical

etching. Vertical steps can be observed between the masked and unmasked regions, indicating

that the p-GaN layer was etched along the mask pattern.

(b)

(a)

unetched

200

0

(nm)

0

0

1515(µm)

(µm)

SiO2

p-GaN

116 Chapter 6

6.3.2. Structural and chemical analyses of the etched surface

Typical SEM and AFM images of the sample after the electrochemical etching

are shown in Figs. 6-5(a) and (b), respectively. Vertical steps can be observed between

the masked and unmasked regions, indicating that the p-GaN layer was etched along the

mask pattern. Figure 6-6 shows AFM cross-sectional profiles of samples (a) before

etching, after etching with (b) n = 3 cycles and (c) n = 5 cycles. Vertical steps can be

observed between the masked and unmasked regions, indicating that the p-GaN layer

was etched along the mask pattern. The difference in height before the process, 114 nm,

is corresponding to the thickness of SiO2 mask deposited on the p-GaN surface. After

the 3 cycles and 5 cycles of waveforms were applied, the difference in height gradually

increased, indicating that the p-GaN surface was etched by the EC reaction. The etching

depth D was plotted in Fig. 6-6(d) as a function of the number of process cycle n.

During the initial stage, the D can be linearly controlled by n at a rate of 25 nm/cycle.

However, the D value saturated and the etching stopped after 5 cycles applied, whose D

values were consistent with the initial thickness of p-GaN layer. The RMS roughness

slightly increased during the process, however, it decreased again and reached to about

1 nm or less. The smallest RMS value obtained after the 5-cycled etching was

comparable to the initial surface.

Figure 6-7 shows the cross-sectional TEM images obtained on the boundary

between the etched and unetched region after the 5 cycles of waveforms were applied.

Figure 6-6. AFM cross-sectional profiles and histograms of samples (a) before etching, (b) after

etching with n = 3 cycles and (c) n = 5 cycles. (d) Etching depth as a function of the number of

cycles, n. During the initial stage of the electrochemical etching, the etching depth can be linearly

controlled by n at a rate of 25 nm/cycle.

initial p-GaN thickness

25 nm/cycle

Etc

hin

g d

ep

th(n

m)

Cycle number

120

100

80

60

40

20

00 1 2 3 4 5

124 nm

163 nm

220

220

(a) initial

(b) after3 cycles

SiO2

SiO2

184 nm

100 (µm)0

(c) after5 cycles SiO2

220

(nm

)(n

m)

(nm

)

(d)

25 nm/cycle

Cycle number


The surface at the etched region was overall smooth without any texture and p-GaN

layer was completely removed. From the high-magnification image, it was found that

the thickness of AlGaN layer remained at the etched region was just 20 nm, which was

comparable to the designed value of the epitaxial layer. This result indicates that the EC

etching stopped just on AlGaN surface. Figure 6-8 shows the differential intensity of

AES spectra obtained at the etched-region that is marked in Fig. 6-7. The specific peak

was observed at 1383 eV but not observed at the unetched region. This peak was

assigned as an aluminum (Al) peak originated from the AlGaN layer. These results

indicate that the EC process is useful for the selective etching of p-GaN layer on

i-AlGaN in a self-stopping fashion.

Figure 6-7. Cross-sectional TEM images of the sample after the electrochemical etching. Top

p-GaN layer was completely removed and AlGaN layer was not etched at all.

etchedunetched

1 µm

protective film

20 nm

AlGaN

GaN

p-GaN

20 nm

etched

Figure 6-8. The differential intensity of AES spectra obtained on unetched region and etched

region after the 5 cycles applied. The p-GaN layer with thickness of 100 nm was just etched and

the i-AlGaN layer appeared on the surface.

0 200 400 600 800 1000 1200 1400 1600

dN/d

E (

arb.

uni

ts)

Kinetic energy (ev)

unetched

etched

Ga LMMN KLL O KLL

Al KLL

etched

118 Chapter 6

6.4. Electrical properties of Schottky diodes formed on the etched surface

To evaluate electrical properties of the etched surface, we fabricated Schottky

diodes on the AlGaN/GaN structures after removing the p-GaN layer. A ring-shaped

ohmic contact was formed on AlGaN surface by a standard evaporation and annealing

process. Then, the circular Schottky electrode was formed at the center of ohmic ring by

evaporation process. Figure 6-9(a) compares the I-V characteristics of two kinds of

diodes formed on ICP-etched and electrochemically etched samples. Both diodes

showed typical rectifying characteristics of Schottky contacts on the HEMT structures,

including the 2DEG depletion behavior observed at around −3 V. However, the

electrochemically etched sample showed lower leakage currents, indicating less

tunneling leakage at the Schottky interface. Figure 6-9(b) shows the C-V curve

measured at 1 MHz for the Schottky diode formed on the electrochemically etched

sample. The data showed a clear pinch-off behavior at around −3 V being consistent

with the I-V characteristics. In addition, the experimental data was well reproduced by

the theoretical curve calculated using the AlGaN thickness of 20 nm. The carrier density

profile obtained from the C-V result showed the 2DEG peak at the AlGaN/GaN

interface as designed. These results indicate formation of good Schottky interface,

suggesting no significant damages were introduced in the AlGaN/GaN heterostructures

during the EC etching process.

Figure 6-9. (a) I-V characteristics of the two kinds of diodes formed on ICP-etched and

electrochemically-etched samples. The electrochemically-etched sample showed lower leakage

currents, indicating less tunneling leakage at the Schottky interface. (b) C-V curve of the Schottky

Schottky diode formed on the electrochemically-etched sample. The experimental data are well

reproduced by the calculation (black solid line).

(a) (b)

Ni/AlGaN/GaN

EC etching

experiment(EC etching)

Ni/AlGaN/GaN


6.5. Summary

In Chapter 6, we developed the selective and well-controllable etching process

for a p-GaN layers utilizing low-energy EC reactions. Summary of this chapter is

itemized below.

1) Etching depth was linearly controlled by cycle number of triangular waveform

at a rate of 25 nm/cycle.

2) The AFM, SEM, TEM and µ-AES analyses showed that the top p-GaN layer

was completely removed after 5 cycles applied, and the etching reaction was

automatically stopped on the AlGaN surface.

3) From the I-V and C-V measurements, it was found that no significant damages

were induced in the AlGaN/GaN heterostructures.

The present data showed that the EC process is promising for selective and

self-stopping etching of p-GaN on AlGaN/GaN structures.

120 Chapter 6

Reference

[1] M. Ishida, T. Ueda, T. Tanaka, and D. Ueda, “GaN on Si Technologies for

Power Switching Devices”, IEEE Trans. Electron Devices, vol. 60, pp.

3053−3059, 2013.

[2] I. Hwang, H. Choi, J. W. Lee, H. S. Choi, J. Kim, J. Ha, C. -Y. Um, S. -K.

Hwang, J. Oh, J. -Y. Kim, J. K. Shin, Y. Park, U. -I. Chung, I. -K. Yoo, and K.

Kim, "1.6kV, 2.9 mΩ cm2 normally-off p-GaN HEMT device", Proc. ISPSD,

pp. 41−44, 2012.

[3] L. -Y. Su, F. Lee, and J. J. Huang, "Enhancement-Mode GaN-Based

High-Electron Mobility Transistors on the Si Substrate With a P-Type GaN

Cap Layer", IEEE Trans. Electron Devices, vol. 61, pp. 460−465, 2014.

[4] I. Ben-Yaacov, Y. -K. Seck, U. K. Mishra, and S. P. DenBaars, "AlGaN/GaN

current aperture vertical electron transistors with regrown channels", J. Appl.

Phys., vol. 95, pp. 2073−2078, 2004.

[5] M. Kanechika, M. Sugimoto, N. Soejima, H. Ueda, O. Ishiguro, M. Kodama, E.

Hayashi, K. Itoh, T. Uesugi, and T. Kachi, "A Vertical Insulated Gate

AlGaN/GaN Heterojunction Field-Effect Transistor", Jpn. J. Appl. Phys., vol.

46, pp. L503−L505, 2007.

[6] R. Yeluri, J. Lu, C A. Hurni, D. A. Browne, S. Chowdhury, S. Keller, J. S.

Speck, and U. K. Mishra, "Design, fabrication, and performance analysis of

GaN vertical electron transistors with a buried p/n junction", Appl. Phys. Lett.,

vol. 106, p. 183502, 2015.

[7] T. Hashizume, and R. Nakasaki, "Discrete surface state related to

nitrogen-vacancy defect on plasma-treated GaN surfaces", Appl. Phys. Lett.,

vol. 80, pp. 4564−4566, 2002.

[8] T. Hashizume, J. Kotani, A. Basile, and M. Kaneko, "Surface Control Process

of AlGaN for Suppression of Gate Leakage Currents in AlGaN/GaN

Heterostructure Field Effect Transistors", Jpn. J. Appl. Phys., vol. 45, pp.

L111−L113, 2006.

[9] J. Neugebauer, and C. G. Van de Walle, "Atomic geometry and electronic

structure of native defects in GaN", Phys. Rev. B, vol. 50, pp. 8067−8070,

1994.


[10] P. Bogusławski, E. L. Briggs, and J. Bernholc, "Native defects in gallium

nitride", Phys. Rev. B, vol. 51, pp. 17255−17258, 1995.

[11] Z. Mouffak, A. Bensaoula, and L. Trombetta, "The effects of nitrogen plasma

on reactive-ion etching induced damage in GaN", J. Appl. Phys., vol. 95, pp.

727−730, 2004.

[12] K. Tang, W. Huang, and T. P. Chow, "GaN MOS Capacitors and FETs on

Plasma-Etched GaN Surfaces", J. Electron. Mater., vol. 38, pp. 523−528, 2009.

[13] C. Youtsey, I. Adesida, L. T. Romano, and G. Bulman, "Smooth n-type GaN

surfaces by photoenhanced wet etching", Appl. Phys. Lett., vol. 72, pp.

560−562, 1998.

[14] T. Rotter, J. Aderhold, D. Mistele, O. Semchinova, J. Stemmer, D. Uffmann,

and J. Graul, "Smooth GaN surfaces by photoinduced electro-chemical

etching", Mater. Sci. Eng. B, vol. 59, pp. 350−354, 1999.







[17] T. Sato, S. Uno, T. Hashizume, and H. Hasegawa, "Large Schottky Barrier

Heights on Indium Phosphide-Based Materials Realized by In-Situ

Electrochemical Process", Jpn. J. Appl. Phys., vol. 36, pp. 1811−1817, 1997.

[18] C. Kaneshiro, T. Sato, and H. Hasegawa, "Electrochemical Etching of Indium

Phosphide Surfaces Studied by Voltammetry and Scanned Probe Microscopes",

Jpn. J. Appl. Phys., vol. 38, pp. 1147−1152, 1999.

[19] Y. Kumazaki, A. Watanabe, Z. Yatabe, and T. Sato, "Correlation between

Structural and Photoelectrochemical Properties of GaN Porous Nanostructures

Formed by Photo-Assisted Electrochemical Etching", J. Electrochem. Soc., vol.

161, pp. H705−H709, 2014.

[20] A. Watanabe, Y. Kumazaki, Z. Yatabe, and T. Sato, “Formation of

GaN-porous Structures using Photo-assisted Electrochemical Process in

Back-side Illumination Mode”, ECS Electrochem. Lett., vol. 4, pp. H11−H13,

2015.

122 Chapter 6



photoabsorption edge", Semicond. Sci. Technol., vol. 31, p. 014012, 2016.

123

Chapter 7

Self-terminating EC etching for recessed-gate

AlGaN/GaN heterostructure field effect tran-

sistors

7.1. Introduction

AlGaN/GaN high-electron-mobility transistors (HEMTs) are promising

candidate for high-power electronics applications because high blocking voltage and yet

low on-resistance RON can be achieved [1−4]. From the fail-safe viewpoint of

power-switching devices, it is indispensable to attain the normally-off operation. One of

the promising approaches to realize a normally-off operation is by adopting

recessed-gate structure [5−7], which can be realized by thinning the AlGaN layer

beneath the gate electrode. Dry etching process is commonly used for the thinning of

AlGaN because wet etching process is not applicable due to the chemical stability of

group-III nitrides [8−10]. However, dry-etched surface generally suffer from various

types of damage which may lead to the degradation of device performance [11,12].

Moreover, unintentional variations in recess depth make it difficult to control threshold

voltage (Vth) precisely.

One alternative approach is a photo-assisted EC etching which is the cyclic

process consists of an anodic oxidation and a subsequent dissolution of the resulting

oxide in an electrolyte. Compared to dry etching, photo-assisted EC etching is

highly-desirable in its simplicity and the absence of plasma damage [13,14]. Besides,

the EC process is applicable to various semiconductors even chemically stable materials

such as group-III nitrides [15].

Recently, some papers reported recessed gate AlGaN/GaN HEMTs using

photo-assisted EC etching. Y. L. Chiou and co-workers have improved transconductance

Gm by photo-assisted EC etching and subsequent oxidation techniques [16]. N. Harada

and co-workers realized normally-off operation by parallel process of photo-assisted EC

etching and oxidation [17]. Z. Zhang and co-workers showed availability of

124 Chapter 7

photo-assisted EC etching in ionic liquid etchant by providing high performance

normally-off HEMT [18]. Despite the device performances of recessed gate

AlGaN/GaN HEMTs fabricated by photo-assisted EC etching have been demonstrated

so far, the important aspects for recess etching such as depth-controllability and surface

roughness have been insufficient yet. Although photo-assisted EC etching is caused by

photo-excited carriers, carrier transfer in AlGaN/GaN heterostructure has not been fully

understood, which make it difficult to optimize EC conditions to obtain desirable

etching features.

In Chapter 7, we investigated basic photo-electrochemical behavior of

AlGaN/GaN heterostructure under monochromatic light to clarify carrier transfer

process in photo-assisted EC etching. Based on the regulation of carrier transfer, we

succeeded in developing self-terminating and depth-controllable etching of AlGaN/GaN

heterostructure with very smooth surface. We also examined electrical properties of

recessed-gate AlGaN/GaN HEMT fabricated by photo-assisted EC etching with carrier

regulation.


The i-Al0.25Ga0.75N/i-GaN heterostructure grown on Si substrate was used as

starting wafer as shown in Fig. 7-1(a). The thickness of AlGaN barrier layer was 25 nm.

A Ti/Al/Ti/Au (20/50/20/50 nm) multilayer ohmic electrode was formed by

electron-beam evaporation method, followed by annealing at 830°C for 1 min in N2

ambient. Then, SiO2 film (100 nm) was formed by sputtering and lithography to define

Figure 7-1. Schematic representations of (a) sample structure, and (b) the experimental setup of

photo-assisted electrochemical etching.

125 Self-terminating EC etching for recessed-gate AlGaN/GaN heterostructure fieldeffect transistors

etching region. Photo-assisted EC etching were performed using three-electrode EC cell

(Fig. 7-1(b)) immersed in electrolyte; mixture of 1 mol/L H2SO4 and 1 mol/L H3PO4

(pH = 1.7). The ohmic contact was electrically connected to an outer circuit, in which

the potential of sample electrode was controlled with respect to the RE by a potentiostat

with a Princeton Applied Research VersaSTAT 4. Xenon (Xe) lamp was used as a light

source and monochromatic light passing through band-path filters were irradiated from

top side of samples.

In advance of etching experiments, basic photo-electrochemical behavior of

AlGaN/GaN heterostructure were investigated by measuring current under

monochromatic light. Considering the photo-electrochemical behavior with potential

distribution, relationships between EC reactions and carrier transfer process were

discussed. Based on the regulation of carrier transfer, two kinds of EC conditions were

compared to investigate the influence of carrier transfer process on etching behavior. We

then fabricated Schottky diode and Schottky HEMT to evaluate electrical properties of

sample fabricated by photo-assisted EC etching.

7.3. Basic photo-electrochemical behavior on AlGaN/GaN heterostructures

Figure 7-2 shows current-voltage (JEC-VEC) characteristics of AlGaN/GaN

heterostructure immersed in electrolyte in dark (black line) and under light with

wavelength λ of 300 nm (red line), 360 nm (blue line), and 400 nm (green line),

respectively. Figure 7-3 shows potential distribution of electrolyte/AlGaN/GaN

structure at voltage VEC of (a) −1.0 V, (b) 1.0 V, and (c) 3.0 V, respectively, calculated by

one-dimensional Poisson equation.

Negative currents (reduction currents) were observed at potentials negative to

around −0.8 V under all light conditions. Since AlGaN barrier become small by

application of negative voltage as shown in Fig. 7-3(a), electrons in 2DEG flow toward

electrolyte/AlGaN interface and cause reduction reaction. The electron concentration in

2DEG is generally as high as 1019 cm−3, this is why the reduction currents are barely

influenced by light illumination. At the positive voltage, few currents were observed on

AlGaN/GaN heterostructure in dark and under light with λ = 400 nm since electrons

were restricted to flow by AlGaN barrier and few holes exist in both AlGaN and GaN

layers. Although light with λ = 400 nm is absorbed in Si substrate, photo-excited

126 Chapter 7

carriers cannot flow due to the existence of high-resistive buffer layer. This rectifying

behavior observed on AlGaN/GaN heterostructure was similar to the n-type

semiconductor electrodes in dark, whereas positive current behaviors by photo-excited

carriers were unlike those. Under the light with λ = 360 nm, which penetrates AlGaN

layer and is absorbed by GaN layer, positive current (oxidation current) was observed

with voltage larger than around 1.0 V. Considering the potential distribution at 1.0 V

(Fig. 7-3(b)), electrons exist in 2DEG with high concentration which prevent

photo-excited holes from flowing toward electrolyte/AlGaN interface; slight oxidation

current was observed at VEC = 1.0 V. Concentration of electrons, however, decreases

with VEC and would be fully depleted at VEC = 3.0 V as shown in Fig. 7-3(c). In such

situation, photo-excited holes can flow toward electrolyte/AlGaN interface and cause

oxidation reactions; oxidation current increased rapidly. Obviously, oxidation reactions

in this conditions were caused by photo-holes generated in GaN layer.

On the other hand, under the light with λ = 300 nm, which is absorbed by both

AlGaN and GaN layer, oxidation current was observed at VEC as high as −0.5 V, and

subsequently saturated at around VEC = 0 V. Since strong polarization exists in AlGaN

layer even at low VEC as shown in Fig. 7-3(b), photo-excited electrons and holes are

transferred to 2DEG and electrolyte/AlGaN interface, respectively; oxidation current

Figure 7-2. Current-voltage characteristics of AlGaN/GaN hetero-structure immersed in

electrolyte in dark (black line) and under light with wavelength λ of 300 nm (red line), 360 nm

(blue line), and 400 nm (green line), respectively. Sweep direction was positive, and sweep rate

was set in 50 mV/s.

-0.01

0

0.01

0.02

-1 0 1 2 3 4

Cu

rre

nt d

en

sity

JE

C (

mA

/cm

2)

Voltage v.s. Ag/AgCl VEC

(V)

λ = 300 nm

360 nm

400 nm

dark

PIN = 0.1 mW/cm2


flows. As previously discussed, photo-generated carries in GaN layer cannot flow at VEC

below 1.0 V due to the existence of high concentration of electrons in 2DEG; it is

obvious that oxidation reactions in this conditions are caused by holes generated in

AlGaN layer.

Thus, it is found that oxidation reactions on AlGaN/GaN heterostructure are

caused by two types of photo-holes: those generated in GaN layer; and those generated

in AlGaN layer. In addition, we can regulate supply way of photo-holes by selecting

proper light wavelength λ and electrode voltage VEC. Next section shows the importance

of the photo-hole regulation on achieving homogeneous surface in photo-assisted EC

etching.

Figure 7-3. Potential distribution of electrolyte/AlGaN/GaN structure at voltage VEC of (a) −1.0 V,

(b) 1.0 V, and (c) 3.0 V, respectively, calculated by one-dimensional Poisson equation. Schottky

barrier height was assumed to be 1.0 eV.

0 10 20 30 40 50 60

Pot

entia

l (eV

/div

)

Distance from surface (nm)

(b)VEC = 1.0 V

(c)VEC = 3.0 V

AlGaN GaN

λ < 365 nmλ < 315 nm

hole

electron

EC

EV

EC

EV

EC

EV

(a)VEC = −1.0 Velectron

hole

electron

λ < 365 nmλ < 315 nm

2DEG

2DEG

128 Chapter 7

7.4. EC etching based on the regulation of photo-excited carriers

7.4.1. EC etching utilizing photo-carriers generated in GaN layer

We first performed photo-assisted EC etching by utilizing photo-holes

generated in GaN layer. Figure 7-4(a) shows the AFM image of sample after

photo-assisted EC etching with VEC = 5.0 V, λ = 360 nm, PIN = 1.0 mW/cm2, and tEC =

16 min. There was no difference in height between etched and un-etched region,

indicating no homogeneous etching occurred. From the top SEM image shown in Fig.

7-4(b), however, small pores with diameter of 20 nm or less could be observed. Since

pore depth was estimated to be 240 nm from the cross-sectional SEM image shown in

Fig. 7-4(c), pores seemed to pierce AlGaN layer. Pore diameter in AlGaN layer was

estimated to be 20 nm or less as was the case with the value estimated from the top

SEM image, whereas pore diameter in GaN layer was estimated to be about 50 to 100

nm. Increase of etching time lead to increase of pore diameter in GaN layer (pore

diameter in AlGaN layer remained unchanged), indicating carrier transfer at

electrolyte/GaN interface occurred rather than that at electrolyte/AlGaN interface.

From above results, we can conclude that photo-holes generated in GaN layer

cause inhomogeneous etching. Generally, inhomogeneous etching of semiconductor is

Figure 7-4. (a) 3D-AFM image, (b) top SEM image, and (c) cross-sectional SEM image of

sample after photo-assisted EC etching with photo-carriers generated in GaN layer: VEC = 5.0 V,

λ = 360 nm, PIN = 1.0 mW/cm2.


resulted from localized carrier transfer at electrolyte/semiconductor interface [19−21].

Similar to the case with the carrier transfer in Schottky barrier diode [22−25], localized

carrier transfer may be caused by inhomogeneous potential distribution due to the

concentration of electric field on crystallographic disorders such as dislocations and

vacancies. Since inhomogeneous etching observed with this condition will be the origin

of rough surface, utilization of photo-holes generated in GaN layer is unfavorable for

the fabrication of recessed-gate structure.

7.4.2. EC etching utilizing photo-carriers generated in AlGaN layer

Next, we performed photo-assisted EC etching by utilizing photo-holes

generated in AlGaN layer. Figure 7-5(a) shows the AFM image of sample after

photo-assisted EC etching with VEC = −0.2 V, λ = 300 nm, PIN = 1.0 mW/cm2, and tEC =

80 min. The definite step could be observed between etched and un-etched region unlike

the sample etched by photo-holes generated in GaN layer. The root mean square (RMS)

of etched region was only 0.41 nm. Figures 7-5(b) and (c) show the top and

cross-sectional SEM images of the sample etched with this condition. There seemed to

be no pores and irregularity on etched region, indicating homogeneous etching can be

Figure 7-5. (a) 3D-AFM image, (b) top SEM image, and (c) cross-sectional SEM image of

sample after photo-assisted EC etching with photo-carriers generated in AlGaN layer: VEC = −0.2

V, λ = 300 nm, PIN = 1.0 mW/cm2.

130 Chapter 7

achieved by utilizing photo-holes generated in AlGaN layer.

Figure 7-6(a) shows relationship between etching depth and etching time tEC

obtained on the sample etched with PIN = 0.5 mW/cm2. Etching depth increased lineally

with tEC at the initial stage. However, we found that etching was terminated at the

specific etching depth, in other words, self-termination phenomenon occurred in this

condition. Since initial AlGaN thickness was 25 nm, self-termination occurred in the

process of AlGaN layer etching. Figure 7-6(b) shows relationship between

self-termination depth and light intensity PIN. We found that self-termination depth

increased lineally with PIN, indicating self-termination phenomenon is related to the

amount of photo-holes generated in AlGaN layer.

From above results, we considered mechanism of photo-assisted EC etching

utilizing photo-holes generated in AlGaN layer. As previously mentioned, photo-holes

generated in AlGaN layer are transferred to electrolyte/AlGaN interface by internal

polarization. They cause oxidation reactions of AlGaN and resulting oxides such as

Ga2O3 and Al2O3 dissolve in electrolyte, leading to etching of AlGaN layer. Since

photo-holes generated in AlGaN layer decrease with thinning of AlGaN, etching

reactions are suppressed at specific AlGaN thickness which cannot generate sufficient

photo-holes; self-termination phenomenon occurred. Since the amount of photo-holes

Figure 7-6. (a) Relationship between etching depth and etching time, and (b) relationship

between self-termination depth and light intensity obtained on the sample etched with

photo-carriers generated in AlGaN layer.

VEC = −0.2 V vs. Ag/AgClλ = 300 nm

0.12 nm/min

VEC = −0.2 Vλ = 300 nmPIN = 0.5 mW/cm2

Self-terminating

(a) (b)

initial AlGaN thickness

14

16

18

20

22

24

26

28

0 0.5 1 1.5 2 2.5

Light intensity PIN

(mW/cm2)

Sel

f-te

rmin

atin

g de

pth

(nm

)

0

5

10

15

20

0 50 100 150 200 250

Etc

hing

dep

th (

nm)

Etching time (min)


are varied with light intensity PIN, self-terminating depth is a function of PIN as shown

in Fig. 7-6(b).

Thus, the regulation of photo-holes affect critically on etching behavior.

Homogeneous etching can be achieved by preventing photo-holes generated in GaN

layer from participating in oxidation reactions. In addition, photo-assisted EC etching

utilizing photo-holes generated in AlGaN layer is terminated spontaneously, and

self-terminating depth can be controlled by light intensity PIN. Self-termination

phenomenon would help us to suppress unintentional variations of recess depth.

Photo-assisted EC etching based on the regulation of photo-holes, therefore, has

attractive features for use in fabricating recessed-gate structure.

7.5. Electrical properties of recessed-gate AlGaN/GaN HEMTs

7.5.1. Capacitance-voltage characteristics of Schottky diode formed on etched

surface

We evaluated electrical properties of AlGaN/GaN heterostructures after

Figure 7-7. Capacitance-voltage characteristics measured at 100 kHz for Schottky diode

fabricated on planar (black) and etched (red) samples: symbols represent experimental results,

and solid lines represent theoretical curve assuming AlGaN thickness of 25 nm for planar sample

and 8 nm for etched sample.

Recessed-gate(AlGaN 8 nm)

Planar-gate(AlGaN 25 nm)

：experimental：calculation

0

200

400

600

800

1000

-5 -4 -3 -2 -1 0 1

Gate voltage (V)

Cap

acita

nce

(nF

/cm

2)

132 Chapter 7

photo-assisted EC etching by fabricating Schottky diode. The AlGaN thickness of

unetched and etched region is 25 nm and 8 nm, respectively. Figure 7-7 shows C-V

characteristics measured at 100 kHz for Schottky diode fabricated on planar (black) and

etched (red) samples: symbols represent experimental results; and solid lines represent

theoretical curve assuming AlGaN thickness of 25 nm for planar sample and 8 nm for

etched sample. Threshold voltage and the saturated capacitance value were increased by

thinning of AlGaN thickness as expected. The experimental data was well reproduced

by the theoretical curve, and no hysteresis was observed between positive and negative

sweep directions, suggesting ideal Schottky interface was formed on both samples.

7.5.2. Current-voltage characteristics of recessed-gate AlGaN/GaN HEMT

We evaluated electrical properties of recessed-gate AlGaN/GaN HEMT by

fabricating Schottky gate on the same sample with Section 7.5.1: the AlGaN thickness

of unetched and etched region is 25 nm and 8 nm, respectively.

Figures 7-8(a) and (b) show drain current-voltage (IDS-VDS) characteristics of

planar-gate and recessed-gate AlGaN/GaN Schottky HEMTs with a gate length LG of 10

µm and a source-drain spacing of 30 µm. Both sample shows good I-V curves with

Figure 7-8. Drain current-voltage (IDS-VDS) characteristics of (a) planar-gate and (b)

recessed-gate AlGaN HEMTs with gate length of 10 µm and source-drain spacing of 30 µm. The

static on-resistance RON were estimated from the inverse of slope in linear region.

0

10

20

30

40

50

60

0 2 4 6 8 10

Drain voltage VDS

(V)

Dra

in c

urre

nt

I DS (

mA

/mm

)

0

20

40

60

80

100

120

0 2 4 6 8 10

Dra

in c

urre

nt

I DS (

mA

/mm

)

Drain voltage VDS

(V)

VGS = 0 V

−0.5 V

−1.0 V

−1.5 V

−2.0 V

−2.5 V

< −3.0 V

VGS = 1.0 V

0.8 V

0.6 V

0.4 V

0.2 V

< 0 V

(a) (b)

Planar-gate Recessed-gate


constant saturation currents and a pinch-off behavior. The static RON of recessed-gate

HEMT extrapolated from linear region were 22.5 Ω∙mm which is better than that of

planar-gate HEMT (25.9 Ω∙mm). The transfer characteristics of planar-gate and

recessed-gate AlGaN/GaN Schottky HEMTs in the saturated region (VDS = 10 V) are

compared in Fig. 7-9. The threshold voltage Vth determined by the linear extrapolation

method are −2.42 and +0.30 V for the planar-gate and recessed-gate HEMTs,

respectively. Positive shift of Vth could be confirmed with recessed-gate structure, and

no hysteresis was observed between positive and negative sweep directions. In addition,

transconductance Gm of the recessed-gate HEMT showed better value than that of the

planar-gate HEMT due to the thinning of AlGaN barrier layer.

These electrical characterization suggests that no significant damages are

induced in the AlGaN/GaN heterostructures during photo-assisted EC etching. We could

conclude that photo-assisted EC etching with carrier regulation is desirable for

fabricating recessed-gate structure.

Figure 7-9. The transfer characteristics of planar-gate (black) and recessed-gate (red) Schottky

HEMTs in the saturated region (VDS = 10 V): symbols represent drain current IDS, and solid lines

represent transconductance Gm.

0

20

40

60

80

100

120

0

10

20

30

40

50

60

70

-4 -3 -2 -1 0 1

Transconductance G

m (m

S/m

m)

Dra

in c

urr

ent

ID

S (

mA

/mm

)

Gate voltage VGS

(V)

VDS = 10 V

Recessed(AlGaN 8 nm)

Planar-gate(AlGaN 25 nm)

Gm

Gm

IDSIDS

134 Chapter 7

7.6. Summary

In Chapter 7, photo-assisted EC etching process is demonstrated to fabricate

recessed-gate AlGaN/GaN HEMTs without inducing damage. Summary of this chapter

is itemized below.

1) Basic photo-electrochemical characteristics of AlGaN/GaN heterostructures

revealed that two kinds of photo-holes generated in either AlGaN layer or GaN

layer could be supplied to solid/liquid interface separately by selecting proper

light wavelength and voltage.

2) The hole-regulation by these conditions significantly affected on etching

behavior: photo-holes generated in GaN layer caused inhomogeneous etching;

those generated in AlGaN layer caused homogeneous etching.

3) In the photo-assisted EC etching with hole regulation, self-termination

phenomenon was observed, and self-terminating depth could be controlled by

light intensity, enabling us to obtain desired AlGaN thickness without

unintentional variation.

4) The recessed-gate AlGaN/GaN HEMT showed positive threshold voltage,

reduction of static on-resistance, and improvement of transconductance

compared to the planar-gate AlGaN/GaN HEMT.

Photo-assisted EC etching based on the regulation of photo-holes, therefore, is

very attractive for the use in fabricating recessed-gate structure.


Reference

[1] Y. Uemoto, M. Hikita, H. Ueno, H. Matsuo, H. Ishida, M. Yanagihara, T. Ueda,

T. Tanaka, and D. Ueda, " Gate Injection Transistor (GIT)—A Normally-Off

AlGaN/GaN Power Transistor Using Conductivity Modulation" IEEE Trans.

Electron Devices, vol. 54, pp. 3393−3399, 2007.

[2] T. Kikkawa, K. Makiyama, T. Ohki, M. Kanamura, K. Imanishi, N. Hara, and

K. Joshin, "High performance and high reliability AlGaN/GaN HEMTs", Phys.

Status Solidi A, vol. 206, pp. 1135−1144, 2009.

[3] R. Chu, A. Corrion, M. Chen, R. Li, D. Wong, D. Zehnder, B. Hughes, and K.

Boutros, "1200-V Normally Off GaN-on-Si Field-Effect Transistors With Low

Dynamic ON-Resistance", IEEE Electron Device Lett., vol. 32, pp. 632−634,

2011.

[4] M. Kuzuhara, J. T. Asubar, and H. Tokuda, "AlGaN/GaN high-electron-

mobility transistor technology for high-voltage and low-on-resistance

operation", Jpn. J. Appl. Phys., vol. 55, p. 070101, 2016.

[5] T. Palacios, C. S. Suh, A. Chakraborty, S. Keller, S. P. DenBaars, and U. K.

Mishra, "High-Performance E-Mode AlGaN/GaN HEMTs", IEEE Electron

Device Lett., vol. 27, pp. 428−430, 2006.

[6] S. Maroldt, C. Haupt, W. Pletschen, S. Müller, R. Quay, O. Ambacher, C.

Schippel, and F. Schwierz, "Gate-Recessed AlGaN/GaN Based Enhancement-

Mode High Electron Mobility Transistors for High Frequency Operation", Jpn.

J. Appl. Phys., vol. 48, p. 04C083, 2009.

[7] S. Huang, X. Liu, J. Zhang, K. Wei, G. Liu, X. Wang, Y. Zheng, H. Liu, Z. Jin,

C. Zhao, C. Liu, S. Liu, S. Yang, J. Zhang, Y. Hao, and K. J. Chen, "High RF

Performance Enhancement-Mode Al2O3/AlGaN/GaN MIS-HEMTs Fabricated

With High-Temperature Gate-Recess Technique", IEEE Electron Device Lett.,

vol. 36, pp. 754−756, 2015.

[8] R. S. Qhalid Fareed, X. Hu, A. Tarakji, J. Deng, R. Gaska, M. Shur, and M. A.

Khan, "High-power AlGaN/InGaN/AlGaN/GaN recessed gate heterostructure

field-effect transistors", Appl. Phys. Lett., vol. 86, p. 143512, 2005.

136 Chapter 7

[9] T. Oka, and T. Nozawa, "AlGaN/GaN Recessed MIS-Gate HFET With

High-Threshold-Voltage Normally-Off Operation for Power Electronics

Applications", IEEE Electron Device Lett., vol. 29, pp. 668−670, 2008.

[10] N. Maeda, M. Hiroki, S. Sasaki, and Y. Harada, "High-Temperature

Characteristics in Normally Off AlGaN/GaN Heterostructure Field-Effect

Transistors with Recessed-Gate Enhanced-Barrier Structures", Appl. Phys.

Express, vol. 5, p. 084201, 2012.

[11] Z. Mouffak, A. Bensaoula, and L. Trombetta, "The effects of nitrogen plasma

on reactive-ion etching induced damage in GaN", J. Appl. Phys., vol. 95, pp.

727−730, 2004.

[12] K. Tang, W. Huang, and T. P. Chow, "GaN MOS Capacitors and FETs on

Plasma-Etched GaN Surfaces", J. Electron. Mater., vol. 38, pp. 523−528, 2009.









p-GaN layers utilizing electrochemical reactions", Proc. SPIE, vol. 9748, p.

97480Y, 2016.

[16] Y. L. Chiou, L. H. Huang, and C. T. Lee, "Photoelectrochemical Function in

Gate-Recessed AlGaN/GaN Metal–Oxide–Semiconductor High-Electron-

Mobility Transistors", IEEE Electron Device Lett., vol. 31, pp. 183−185, 2010.

[17] N. Harada, Y. Hori, N. Azumaishi, K. Ohi, and T. Hashizume, "Formation of

Recessed-Oxide Gate for Normally-Off AlGaN/GaN High Electron Mobility

Transistors Using Selective Electrochemical Oxidation", Appl. Phys. Express,

vol. 4, p. 021002, 2011.

[18] Z. Zhang, S. Qin, K. Fu, G. Yu, W. Li, X. Zhang, S. Sun, L. Song, S. Li, and R.


Hao, "Fabrication of normally-off AlGaN/GaN metal-insulator-semiconductor

high-electron-mobility transistors by photo-electrochemical gate recess etching

in ionic liquid", Appl. Phys. Express, vol. 9, p. 084102, 2016.



Formed by Photo-Assisted Electrochemical Etching", J. Electrochem. Soc., vol.

161, pp. H705−H709, 2014.

[20] A. Watanabe, Y. Kumazaki, Z. Yatabe, and T. Sato, “Formation of GaN-porous

Structures using Photo-assisted Electrochemical Process in Back-side

Illumination Mode”, ECS Electrochem. Lett., vol. 4, pp. H11−H13, 2015.

[21] Y. Kumazaki, Z. Yatabe, and T. Sato, "Formation of GaN porous structures

with improved structural controllability by photoassisted electrochemical

etching", Jpn. J. Appl. Phys., vol. 55, p. 04EJ12, 2016.

[22] O. Mitrofanov, and M. Manfra, "Poole-Frenkel electron emission from the

traps in AlGaN/GaN transistors", J. Appl. Phys., vol. 95, pp. 6414−6419, 2004.

[23] H. Zhang, E. J. Miller, and E. T. Yu, "Analysis of leakage current mechanisms

in Schottky contacts to GaN and Al0.25Ga0.75N∕GaN grown by molecular-beam

epitaxy", J. Appl. Phys., vol. 99, p. 023703, 2006.

[24] H. Hasegawa, and S. Oyama, "Mechanism of anomalous current transport in

n-type GaN Schottky contacts", J. Vac. Sci. Technol. B, vol. 20, pp. 1647−1655,

2002.

[25] J. Kotani, T. Tamotsu, and H. Hasegawa, "Analysis and control of excess

leakage currents in nitride-based Schottky diodes based on thin surface barrier

model", J. Vac. Sci. Technol. B, vol. 22, pp. 2179−2189, 2004.

138 Chapter 7

139

Chapter 8

Conclusion

This work aimed at developing various EC etching techniques on III-V

semiconductors for the application to the energy-creating and energy-saving

technologies. As for the energy-creating technology, surface porosification techniques

are developed for InP and GaN, and it was revealed that their optical properties were

dramatically improved. As for the energy-saving technology, selective etching

techniques are developed for two types of AlGaN/GaN HEMTs, and it was revealed that

self-termination phenomena were occurred and etching depth could be controlled

precisely. Summaries of each chapter are described in the following paragraphs.

In Chapter 4, highly ordered pore arrays were formed on InP and

characterized by using two type of photoelectric conversion devices. The photoelectric

conversion devices formed on p-n junction substrates showed that lower photocurrents

were observed on the porous device, as compared with that of the non-porous device,

indicating that the absorption properties of InP were enhanced after the formation of

porous structures. The enhancement of absorption properties can be explained in terms

of absorption coefficient increased by the light scattering and the sub-bandgap

absorption in the porous layer. Platinum/porous InP Schottky junction photoelectric

conversion devices showed larger photocurrents and higher responsivity than those of a

reference planar sample. We believe that porous structures are promising materials for

use in photoelectric conversion devices such as solar cells and photodetectors because

of their unique features such as their large specific surface area, low reflectance

properties, and high absorption properties.

In Chapter 5, we have developed various EC etching technique to fabricate

GaN porous structures. By adopting free-standing GaN substrate as starting substrate,

various kinds of porous structures could be formed independent of dislocations which

made it difficult to control structural properties precisely. In photo-assisted EC etching,

pore depth was nonlinear against etching time and there were difficulty in formation of

deeper pores than micro-meter because carriers most of which were generated near

surface caused top and lateral etching along with pore growth. On the other hand, EC

140 Chapter 8

etching related to avalanche effect proceeded in the direction of [000−1] anisotropically,

enabling superior controllability in depth with incredible aspect ratio. Franz-Keldysh

effect gave us further possibility of improving structural controllability: Franz-Keldysh

effect occurred only at pore tips where a high electric field was induced, enabling the

tuning of the diameter without diminishing the linearity or depth controllability which

was achieved by avalanche effect. Post wet etching utilizing TMAH was also good

technique to control pore diameter because of its anisotropic nature. From PL

measurements, porous samples showed blue shift of NBE emission due to quantum

confinement in the pore wall. Reduction of YL intensity were also observed, indicating

VGa-related defects were preferentially removed and additional damage was not induced

by this method. Photoreflectance measurement revealed that porous samples had

effective refractive index which could be controlled by TMAH etching time. In

photoelectrochemical measurement, IPCE was drastically enhanced to as high as 91 %

by formation of porous structures with adequate structural properties. These results

clearly denote that precisely controlled GaN porous structure by EC etching process are

very promising as photoelectrode in EC energy conversion systems.

In Chapter 6, we developed the selective and well-controllable etching process

for a p-GaN layers utilizing low-energy EC reactions. Etching depth was linearly

controlled by cycle number of triangular waveform at a rate of 25 nm/cycle. The AFM,

TEM and µ-AES results showed that the top p-GaN layer was completely removed after

5 cycles applied, and the etching reaction was automatically stopped on the AlGaN

surface. From the I-V and C-V measurements, it was found that no significant damages

were induced in the AlGaN/GaN heterostructures. The present data showed that the EC

process is promising for selective and self-stopping etching of p-GaN on AlGaN/GaN

structures.

In Chapter 7, EC etching process is demonstrated to fabricate recessed-gate

AlGaN/GaN HEMTs without inducing damage. Basic photo-electrochemical

characteristics of AlGaN/GaN hetero-structures revealed that two kinds of photo-holes

generated in either AlGaN layer or GaN layer could be supplied to solid/liquid interface

separately by selecting proper light wavelength and voltage. The hole-regulation by

these conditions significantly affected on etching behavior: photo-holes generated in

GaN layer caused inhomogeneous etching; those generated in AlGaN layer caused

homogeneous etching. Moreover in hole-regulated photo-assisted EC etching,

self-termination phenomenon was observed, and self-terminating depth could be

controlled by light intensity, enabling us to obtain desired AlGaN thickness without

141 Conclusion

unintentional variation. The recessed-gate AlGaN/GaN HEMT showed positive

threshold voltage, reduction of static on-resistance, and improvement of

transconductance compared to the planar-gate AlGaN/GaN HEMT. Photo-assisted EC

etching based on the regulation of photo-holes, therefore, is very attractive for the use in

fabricating recessed-gate structure.

Thus, EC etching techniques of III-V semiconductors have attractive features:

self-aligned nanostructures can be formed and their structural and optical properties are

controlled precisely; homogeneous, well-controllable, and self-terminating etching can

be achieved by selecting proper conditions. These are specific features obtained in EC

etching techniques: in other words, it is difficult to obtain these features in others such

as plasma etching technique. Based on the insights obtained in this work, it can be

concluded that EC etching techniques of III-V semiconductors are very promising as

device processing for both energy-creating and energy-saving technologies.

142 Chapter 8

143

Appendix A

EC formation and optical characterization of

Cu2O/GaN heterostructure for visible light

responsive photoelectrode

A.1. Introduction

As described in Chapter 5, GaN is one of the most attractive photoelectrode

materials because of its chemical stability and its potential to achieve direct

photoelectrolysis by solar power without the consumption of electric power. However,

GaN can absorb light only in UV region accounting for only 3 % of entire solar light

energy, which make it difficult to improve conversion efficiency dramatically.

Capability for utilizing visible light is necessary to overcome this issue. Cuprous oxide

(Cu2O) is known as visible light responsive photoelectrode because it has direct

bandgap with energy of about 2.1 eV corresponding to theoretical light to hydrogen

conversion efficiency of 18 % [1−3]. In addition, band alignment of Cu2O/GaN

heterostructure classified into type II with relatively low conduction band offset as

shown in Fig. A-1, which allows interactive and low loss carrier transportation [4,5].

Although Cu2O/GaN heterostructure is assumed to work as visible light responsive

Figure A-1. Schematic representation of energy band diagram of a GaN and a Cu2O reported by

B. Kramm and co-workers [4]. The band alignment of Cu2O/GaN heterostructure classified into

type II with conduction band offset of 0.24 eV.

GaN Cu2O

1.47 eV

2.1 eV

0.24 eV CB

VB

CB

VB

144 Appendix A

photoelectrode without the consumption of electric power, it has never been

demonstrated.

In this work, Cu2O/GaN heterostructure photoelectrodes were fabricated by EC

deposition process, and their capability as a photoelectrode material were demonstrated.

EC deposition process was first optimized to form a homogeneous Cu2O film on GaN

substrate. Then Spectroscopic measurements and photo-electrochemical measurements

are conducted.

A.2. Experimental details

An n-type GaN epitaxial layers (ND = 2 × 1016 cm−3) grown on free-standing

GaN substrates (ND > 1 × 1018 cm−3) were used as initial wafer. EC deposition process

was first performed using a standard three-electrode EC cell, as described in Chapter 3.

We used a mixture of 0.2 mol/L CuSO4 and 1.5 mol/L C3H6O3 as an electrolyte. PH of

electrolyte was adjusted by saturated potassium hydroxide. Cathodic voltage are applied

for deposition of Cu2O. Bias application mode (constant or pulse), bath temperature TEC,

and deposition time tEC are varied to obtain homogeneous Cu2O film.

Figure A-2. Top SEM images of Cu2O deposited on GaN by EC deposition process with various

conditions: (a) constant mode, TEC = 25, tEC = 10 min; (b) constant mode, TEC = 25, tEC = 20

min; (c) constant mode, TEC = 25, tEC = 80 min; (d) constant mode, TEC = 75, tEC = 20 min; (e)

pulsed mode, TEC = 75, tEC = 20 min.

145 EC formation and optical characterization of Cu2O/GaN heterostructure forvisible light responsive photoelectrode

A.3. Results and discussion

Figure A-2 shows top SEM images of Cu2O deposited on GaN by EC process

with various conditions: (a) constant bias mode, TEC = 25, tEC = 10 min; (b) constant

bias mode, TEC = 25, tEC = 20 min; (c) constant bias mode, TEC = 25, tEC = 80 min;

(d) constant bias mode, TEC = 75, tEC = 20 min; (e) pulsed bias mode, TEC = 75, tEC

= 20 min. At bath temperature TEC of 25, secondary particles are observed due to

aggregation, and their size increased with deposition time tEC. The surface coverage also

increased with deposition time tEC, but it had not been 100 % even with the prolonged

time as shown in Fig. A-2(c). On the other hand, GaN surface was fully covered with

Cu2O at bath temperature TEC of 75 as shown in Fig. A-2(d). In general, Deposition

process is supposed to consist of electron transfer reactions and chemical reactions

(such as ligand dissociation). Since the rate of chemical reactions would be temperature

dependent, electrochemically active species existed sufficiently at GaN surface, which

may allow this process to enhance coverage of Cu2O. However, the size variation

between primary and secondary particles was enhanced, which should be suppressed to

obtain homogeneous Cu2O film. Figure A-3 schematically illustrates the formation flow

of Cu2O with constant bias mode. Chemical and electron transfer reactions were

Figure A-3. Schematic representations of formation flow of Cu2O with constant bias mode. (a)

Chemical and electron transfer reactions are occurred at GaN/electrolyte interface. (b) Cu2O

particles are formed. (c) Concentration of interfacial ions become low (diffusion-limited reaction),

resulting in (d) Cu2O particles are aggregated.

Cu2+

H2O

H2OCu2+

n-GaN

e

H2O

H2O

electrolyte

e

ee

e

e

ee

Cu2+

Cu2+

Cu2+

n-GaN

Cu2+

H2O

electrolyte

H2O

Cu2+

e

e

ee

e

e

ee

Cu2+

n-GaN

Cu2+

H2O

electrolyte

H2O

Cu2+

e

e

ee

e

e

ee

Cu2+

n-GaN

Cu2+

Cu2+

electrolyte

H2O

H2O

H2O

e

e

ee

e

e

ee

Cu2O

(a) (b)

(c) (d)

146 Appendix A

occurred at GaN/electrolyte interface (Fig. A-3(a)), followed by the formation of Cu2O

particles (Fig. A-3(b)). Since the rate of electron transfer was faster than that of ion

transfer, the concentration of chemical species decreased and deposition reactions

proceeded with diffusion-limited mode (Fig. A-3(c)). The diffusion-limited deposition

reactions occurred preferentially at secondary particles where diffusion length required

for adsorption is relatively short, resulting in excess aggregation (Fig. A-3(d)). From the

above discussion, diffusion-limited reactions need to be avoided to form a homogeneous

Cu2O film. Hence, pulsed bias were applied and aggregation of Cu2O are dramatically

suppressed as shown in Fig. A-2(e). These results clearly denote that the pulsed bias

deposition is effective to avoid aggregation and obtain homogeneous Cu2O film.

Figure A-4 shows AES differential spectra obtained at a bare GaN (a black

line) and a Cu2O film deposited on GaN (a red line). After deposition, Ga and N peaks

were disappeared and Cu peak was appeared. Relative concentration of Cu and O,

calculated by using relative sensitivity coefficient, was approximately 2 : 1, indicating

that Cu2O was deposited stoichiometry. Figure A-5 shows X-ray diffraction pattern of a

bare GaN (a black line) and a Cu2O film deposited on GaN (a red line). A Cu2O film

shows diffraction peaks at 2θ of 52.45° corresponding to (112) and 61.36°

corresponding to (220). Hence Cu2O deposited with this condition was a polycrystalline

film with preferred orientation of 112 and 110.

Figure A-4. AES differential spectra obtained at a bare GaN (a black line) and a Cu2O film

deposited on GaN (a red line).

200 400 600 800 1000 1200 1400

Diff

ere

ntia

l in

ten

sity

(a

rb.

uni

ts)

Kinetic energy (eV)

ONCGa

OC

Cu

bare GaN

Cu2O/GaN


Figure A-6 shows specular reflectance R (blue lines) and transmittance T

(green lines) spectra of (a) a bare GaN and (b) a Cu2O/GaN heterostructure. As for a

bare GaN, transmittance T at photon energy hν above its bandgap of 3.4 eV was nearly

zero due to the band-edge absorption, and transmittance T at photon energy hν below its

bandgap reached as high as 70 %. On the other hand, the threshold point of

transmittance T of a Cu2O/GaN heterostructure shifted towards lower photon energies

than that of a bare GaN. The absorption coefficient α can be roughly calculated from

Figure A-5. X-ray diffraction pattern of a bare GaN (a black line) and a Cu2O film deposited on

GaN (a red line).

20 30 40 50 60 70 80 90 100Two theta (degrees)

Inte

nsity

(cp

s) bare GaN

Cu2O/GaN

GaN(002)

GaN(004)

Cu2O(220)

Cu2O(112)

Figure A-6. Transmittance, reflectance spectra, and calculated (αhν)2-hν plots (Tauc plots)

obtained at a bare GaN and a Cu2O/GaN heterostructure.

Transmittance

Reflectance

Tauc plot Transmittance

Reflectance

Tauc plot

2.28 eV3.3 eV

(a) (b)

0

20

40

60

80

100

0.0

0.5

1.0

1.5

2.0

1 2 3 4Tra

nsm

itta

nce

T,

Re

flect

ance

R (

%)

Photon energy h (eV)

(h 2 ( x10

10 cm

-2 eV2)

0

20

40

60

80

100

0.0

0.5

1.0

1.5

2.0

1 2 3 4

(h 2 ( x10

5 cm-2 eV

2)

Photon energy h (eV)

Tra

nsm

itta

nce

T,

Re

flect

ance

R (

%)

148 Appendix A

spectroscopic properties by using following simplified equation [6,7]:

21

αR

TWexp

, (A-1)

where W is the thickness of a film. The transition type can be determined by plotting

either α1/2 or (αhν)2 as a function of photon energy hν: namely, linear relation between

(αhν)2 and hν indicates a direct transition type, and linear relation between α1/2 and hν

indicates an indirect transition type. Then, (αhν)2-hν plots, which is called the Tauc plot,

are shown in Fig. A-6, indicating the electro-deposited Cu2O film was direct transition

type. In addition, the x-intercept of the extrapolated line from Tauc plot gave the

bandgap Eg of 2.28 eV, which is consistent value compared to previous reports.

Therefore, a Cu2O/GaN heterostructure are capable of absorbing visible light with

photon energy above 2.28 eV.

Finally, we evaluated the capability of a Cu2O/GaN heterostructure as

photoelectrodes. Figure A-7 shows current photoresponse properties of (a) a bare GaN

electrode and (b) a Cu2O/GaN heterostructure electrode in 1 mol/L NaCl obtained at

voltage V of 0 V vs. Ag/AgCl under monochromatic light that was tuned on at 10 s and

tuned off at 30 s. We found that a Cu2O/GaN heterostructure electrode acted as

photoanode: photo-excited holes caused anodic reactions at Cu2O/electrolyte interface,

and photo-excited electrons caused cathodic reactions at metal/electrolyte interface

Figure A-7. Photoresponse properties of a bare GaN and a Cu2O/GaN heterostructure obtained

at voltage V of 0 V under monochromatic light. The light started to be irradiated at 10 s and was

turned off at 30 s.

-5

0

5

10

15

20

25

30

0 5 10 15 20 25 30 35 40

Elapsed Time (s)

Cur

ren

t de

nsi

ty J

(A

/cm

2 )

-0.5

0.0

0.5

1.0

1.5

2.0

0 5 10 15 20 25 30 35 40

Cur

ren

t de

nsi

ty J

(A

/cm

2 )

Elapsed Time (s)

(a) (b)

λ = 350 nm

λ > 400 nm

light on off

350 nm

400 nm

λ = 450 nm

500 nm

550 nm

600 nm 650 nm700 nm

light on off


(counter electrode). A bare GaN electrode shows photocurrent under the light with

wavelength shorter than the bandgap of GaN as expected, whereas a Cu2O/GaN

heterostructure electrode shows photocurrent under the light with wavelength longer

than the bandgap of GaN, as well as that with wavelength shorter than the bandgap of

GaN. Figure A-8 shows the incident-photon-to-current conversion efficiency (IPCE)

measured at voltage V of 0 V as a function of light wavelength λ. It was found that a

Cu2O/GaN heterostructure electrode exhibited photocurrent under visible light with

wavelength shorter than 600 nm that almost corresponds to absorption-edge of Cu2O.

From these results, it can be concluded that photo-carriers excited in Cu2O can cause

redox reaction and Cu2O/GaN heterostructure acts as visible light responsive

photoelectrode. The highest IPCE, however, was 3.6 % at λ = 450 nm, which must be

improved for the practical use in photoelectrode. It is assumed that carrier transport

properties of Cu2O is one of the most serious issues. Since the XRD pattern shown in

Fig. A-5 indicates crystal quality of Cu2O is poor, carriers are frequently trapped at

defects and vanished by recombination process. Further optimization of deposition

conditions such as bath temperature, pH, and voltage is necessary. Surface texturing and

modification may also be effective to improve the interfacial carrier transport.

Figure A-8. The incident-photon-to-current conversion efficiency (IPCE) calculated at voltage V

of 0V plotted as a function of wavelength: black symbols represent a bare GaN electrode, and

red symbols represent a Cu2O/GaN heterostructure electrode.

0

1

2

3

4

5

6

7

350 400 450 500 550 600 650 700

IPC

E (

%)

Wavelength (nm)

bare GaN

Cu2O/GaN

PIN = 0.1 mW/cm2

V = 0 V 1M NaCl

150 Appendix A

A.4. Summary

In this work, we performed the EC deposition of a Cu2O film on GaN and

evaluated the capability of a Cu2O/GaN heterostructure as a photoelectrode material.

1) Pulsed bias deposition is effective to avoid aggregation and obtain

homogeneous Cu2O film.

2) Deposited Cu2O was polycrystalline film with preferred orientation of 112

and 110, and almost corresponded to stoichiometric composition.

3) Estimated bandgap of Cu2O was 2.28 eV, indicating electro-deposited Cu2O is

capable of absorbing visible light.

4) Photo-electrochemical measurements revealed that a Cu2O/GaN

heterostructure could work as visible light responsive photoelectrode.

5) The highest IPCE was 3.6 % at λ = 450 nm, which must be improved for the

practical use in photoelectrode.


Reference

[1] A. Paracchino, V. Laporte, K. Sivula, M. Grätzel, and E. Thimsen, "Highly

active oxide photocathode for photoelectrochemical water reduction", Nature

Mater., vol. 10, pp. 456−461, 2011.

[2] M. Hara, T. Kondo, M. Komoda, S. Ikeda, K. Shinohara, A. Tanaka, J. N.

Kondo, and K. Domen, "Cu2O as a photocatalyst for overall water splitting

under visible light irradiation ", Chem. Commun., vol. 3, pp. 357−358, 1998.

[3] W. Siripala, A. Ivanovskaya, T. F. Jaramillo, S. -H. Baeck, and E. W.

McFarland, "A Cu2O/TiO2 heterojunction thin film cathode for

photoelectrocatalysis", Sol. Energy Mater. Sol. Cells, vol. 77, pp. 229−237,

2003.

[4] B. Kramm, A. Laufer, D. Reppin, A. Kronenberger, P. Hering, A. Polity, and B.

K. Meyer, "The band alignment of Cu2O/ZnO and Cu2O/GaN heterostructures",

Appl. Phys. Lett., vol. 100, p. 094102, 2012.

[5] B. K. Meyer, A. Polity, D. Reppin, M. Becker, P. Hering, P. J. Klar, Th. Sander,

C. Reindl, J. Benz, M. Eickhoff, C. Heiliger, M. Heinemann, J. Bläsing, A.

Krost, S. Shokovets, C. Müller, and C. Ronning, "Binary copper oxide

semiconductors: From materials towards devices", Phys. Status Solidi B, vol.

249, pp. 1487−1509, 2012.

[6] J. I. Pankove, "Optical Processes in Semiconductors", Dover Publications, New

York, 1971.

[7] T. D. Golden, M. G. Shumsky, Y. Zhou, R. A. VanderWerf, R. A. Van Leeuwen,

and J. A. Switzer, "Electrochemical Deposition of Copper(I) Oxide Films",

Chem. Mater., vol. 8, pp. 2499−2504, 1996.

152 Appendix A

153

Appendix B

Correlation between chapters and publications

The experimental results and discussions described in each chapter have been

partly published in the corresponding articles as listed below.

Chapter 4:

[1] Y. Kumazaki, T. Kudo, Z. Yatabe, and T. Sato, "Investigation on Optical

Absorption Properties of Electrochemically Formed Porous InP using

Photoelectric Conversion Devices", Applied Surface Science, vol. 279, pp.

116−120, 2013.

[2] R. Jinbo, Y. Kumazaki, Z. Yatabe, and T. Sato, "Formation and Photoelectrical

Measurements of Pt Schottky Interfaces on InP Porous Structures", ECS

Transactions, vol. 50, pp. 247−252, 2013.

[3] 熊崎祐介, 神保亮平, 谷田部然治, 佐藤威友, "電気化学的手法によ

る InP多孔質構造の光吸収特性と光電変換", 電子情報通信学会信学技報,

vol. 113, ED2013-27, pp. 61−64, 2013 (in Japanese).

Chapter 5:

[1] 熊崎祐介, 渡部晃生, 谷田部然治, 佐藤威友, "電気化学的手法によ

る GaN 多孔質構造の形成と光電極特性", 電子情報通信学会信学技報,

vol. 113, ED2013-88, pp. 113−116, 2013 (in Japanese).



Formed by Photo-Assisted Electrochemical Etching", Journal of The

Electrochemical Society, vol. 161, pp. H705−H709, 2014.

[3] A. Watanabe, Y. Kumazaki, Z. Yatabe, and T. Sato, "Formation of GaN-porous


Illumination Mode", ECS Electrochemistry Letters, vol. 4, pp. H11−H13, 2015.

154

Appendix B

[4] 熊崎祐介, 渡部晃生, 谷田部然治, 佐藤威友, "分光電気化学法によ

る GaN/電解液界面の評価とナノ構造形成への応用", 電子情報通信学会

信学技報, vol. 115, ED2015-28, pp. 63−66, 2015 (in Japanese).



etching", Japanese Journal of Applied Physics, vol. 55, p. 04EJ12, 2016.

[6] Y. Kumazaki, S. Matsumoto, and T. Sato, "Precise Structural Controlling of

GaN Porous Nanostructures Using Anisotropic Electrochemical and Wet

Etching Technique", Journal of The Electrochemical Society (under review).

Chapter 6:



p-GaN layers utilizing electrochemical reactions", Proceedings of SPIE, vol.

9748, p. 97480Y, 2016.

Chapter 7:

[1] 熊崎祐介 , 植村圭佑 , 佐藤威友 , "電気化学加工技術を利用した

AlGaN/GaN ヘテロ構造の低損傷リセスエッチング", 電子情報通信学会

信学技報, vol. 116, ED2016-66, pp. 45−50, 2016 (in Japanese).

[2] Y. Kumazaki, K. Uemura, T. Sato, and T. Hashizume, "Precise thickness

control in recess etching of AlGaN/GaN heterostructures by

photocarrier-regulated electrochemical process", Japanese Journal of Applied

Physics (under review).

155

Appendix C

List of figures and tables

Figures:

Chapter 1

Figure 1-1. Stacked area chart of the world primary energy consumption from 1990 to

2015 reported by British Petroleum [1]. ..................................................... 1

Chapter 2

Figure 2-1. Redox reaction cycle in aqueous solution: λ is the reorganization energy,

ERED is the most probable donor level, and EOX is the most probable

acceptor level. ........................................................................................... 12

Figure 2-2. Schematic illustrations of (a) energy cycle and (b) electron energy levels

for the redox reaction of hydrated particles. ............................................ 13

Figure 2-3. Distribution of the electron state density of hydrated redox particles: (a)

oxidant concentration NOX equal to reductant concentration NRED, and (b)

NOX higher than NRED. .............................................................................. 15

Figure 2-4. Schematic representations of energy diagrams at n-type

semiconductor/electrolyte interface: (a) for the case in which the n-type

semiconductor and electrolyte are not in equilibrium (EF ≠ EF(REDOX)); (b)

for the case in which the semiconductor is equilibrium with the electrolyte

(EF = EF(REDOX)). ....................................................................................... 16

Figure 2-5. Schematic representations of energy diagram at n-type semiconductor/

electrolyte interface under polarization with overpotential η: (a) under

positive polarization (qη > EFB), (b) under negative polarization (qη = EFB),

and (c) under negative polarization (qη < EFB), respectively. .................. 17

156 Appendix C

Figure 2-6. Typical current curves of (a) n-type and (b) p-type semiconductor plotted

as a function of overpotential qη. The carriers at n-type

semiconductor/electrolyte interface are depleted when qη > EFB and

accumulated when qη < EFB. Polarization dependence of the carriers at

n-type semiconductor/electrolyte interface is reversed: they are depleted

when qη < EFB and accumulated when qη > EFB. .................................... 18

Figure 2-7. Schematic representations of energy diagrams for n-type semiconductor

electrode (a) in dark and (b) in the photo-excited state. ........................... 19

Figure 2-8. Schematic representations of redox reactions at (a) n-type and (b) p-type

semiconductor electrode under light illumination. Photo-excited holes

cause anodic reactions (RED → OX + e−) at n-type semiconductor,

whereas photo-excited electrons cause cathodic reactions (OX + e− →

RED) at p-type semiconductor. ................................................................ 20

Figure 2-9. Typical current curves of (a) n-type and (b) p-type semiconductor

electrode in dark and under light illumination. Positive photocurrent Iph

was observed at n-type semiconductor with potential E > EFB, whereas

negative Iph was observed at p-type semiconductor with E < EFB. ........... 21

Figure 2-10. The flow of anodic dissolution of GaAs suggested by Gerischer and Mindt

[8]: (a) A nucleophilic agent (X−) reacts with one of the positively charged

surface atoms from which an electron is removed. (b) Surface state formed

by unpaired electron traps hole and reacts with X−. (c) One GaAs is

dissolved into electrolyte by using four more holes. ................................ 22

Figure 2-11. (a) Schematic illustration of sample structure and (b) SEM image of

sample after material-selective etching by photo-assisted electrochemical

process [9]: GaN cantilevers curved upwards by the removal of InGaN

layer and the relief of substrate-induced strain. ....................................... 23

Figure 2-12. SEM images of p-on-n GaN columns (a) before and (b) after dopant-

selective etching by photo-assisted electrochemical process [10]. ............ 24

Figure 2-13. (a) The SEM image of GaN whisker-shaped structure, and (b) the

cross-sectional TEM image of whisker, showing that both edge (e) and

mixed (m) dislocations are associated with whisker formation [11]. ....... 25

Figure 2-14. The room temperature PL spectrum obtained on porous GaP and

crystalline (non-porous) GaP [12]. ........................................................... 27

157 List of figures and tables

Figure 2-15. Photograph of planar InP reference, and porous sample with and without

photo-assisted electrochemical etching [16]. ........................................... 27

Figure 2-16. Schematics representations of water splitting by semiconductor

photoelectrode: (a) in dark, redox reactions are thermodynamically

impossible; (b) under illumination, water is split into gaseous hydrogen

and oxygen. .............................................................................................. 28

Chapter 3

Figure 3-1. Schematic representations of (a) three-electrode electrochemical cell used

in this work, and (b) top and cross-sectional schematics of working

electrode (WE). ........................................................................................ 34

Figure 3-2. Schematic circuit diagram of a potentiostatic mode with a Princeton

Applied Research VersaSTAT 4. .............................................................. 35

Figure 3-3. The simple description of photo-effect in semiconductor: (a) incident of

photons; (b) absorption and excitation of minority charges (electron-hole

pairs); (c) separation of electron-hole pairs and transport toward

electrolyte/semiconductor interface; (d) interfacial carrier transfer across

the semiconductor/electrolyte interface. .................................................. 36

Figure 3-4. Simple schematic representation of an SEM, which basically consists of an

electron gun, condenser lenses, an aperture, scanning coils, objective lens,

and a specimen chamber. .......................................................................... 38

Figure 3-5. Schematic representations of (a) interaction between semiconductor and

light (reflection, absorption, and transmission), and (b) the case in which

light incident from medium 1 to medium 2 with incident angle θ1 and

reflection angle θ2. .................................................................................... 41

Figure 3-6. Schematic representations of various transition processes in

semiconductors: (a) direct band-to-band transition; (b) bound to free

transition; (c) donor to acceptor pair transition; (b) free to bound transition;

(e) transition via localized state resulted from point defects. ................... 43

Figure 3-7. Schematic representations of PL measurement system. The He-Cd laser

was employed as light source with wavelength λ of 325 nm. .................. 43

158 Appendix C

Chapter 4

Figure 4-1. Schematic illustrations of (a) experimental procedure and (b) applied

waveform used in this study. .................................................................... 48

Figure 4-2. Plan-view SEM images of porous samples (a) before and after

photo-assisted EC etching with (b) 3 cycles and (c) 6 cycles. ................. 49

Figure 4-3. Cross-sectional SEM images of porous samples before (left) and after

(right) photo-assisted EC etching. ............................................................ 49

Figure 4-4. Relationship between average pore depth Dp and cycle number of ramped

bias applied to sample. Anodic currents measured during photo-assisted

EC etching is also shown as a solid curve. ............................................... 50

Figure 4-5. Schematic illustration of PC device structure and experimental setup for

photoelectric measurement. ...................................................................... 51

Figure 4-6. (a) Top and cross-sectional SEM images of InP porous structures, and (b)

top photo image of a PC device. .............................................................. 52

Figure 4-7. Current response properties of non-porous device with light intensity PIN

of 2500 µW/cm2, which turns on at t = 500 s and off at t = 1500 s. .......... 53

Figure 4-8. Correlation between current ratio ΔI2/ΔI1 and position of light irradiation x

measured on non-porous device. .............................................................. 54

Figure 4-9. Photocurrent ΔI1 plots measured on non-porous device as function of

thickness of top layer dtop. ........................................................................ 55

Figure 4-10. Current response properties of porous PC device based on incident light at

various light intensity PIN. ........................................................................ 56

Figure 4-11. Comparison of photocurrents between non-porous and porous devices

plotted as function of light intensity PIN. ................................................. 57

Figure 4-12. Schematic illustration of carrier excitation with sub-bandgap absorption.

(a) General excitation from valence band (VB) to conduction band (CB),

(b) from VB to localized levels, (c) from localized levels to CB, and (d)

from localized levels to other localized levels. ........................................ 58

Figure 4-13. Schematic illustration of the platinum/porous InP Schottky junction PC

devices and photo-carrier separation at platinum/pore interface. ............. 59


Figure 4-14. Top view SEM images of the InP porous structures and its schematic

illustration. (a) Sample just after the pore formation and (b) sample after

the removal of the irregular top layer formed on the surface. .................. 60

Figure 4-15. Cross-sectional SEM images of the sample after platinum formation: (a)

on the porous structure with the irregular top layer and (b) on the porous

structure after the removal of the irregular top layer. (c) Correlation

between average thickness of platinum film formed on the walls inside the

pores and processing time for forming the cathodic platinum. ................ 61

Figure 4-16. Specular reflectance spectra obtained at planar InP, porous InP with

irregular layer, and porous InP without irregular layer as a function of

wavelength λ. ............................................................................................ 62

Figure 4-17. Current-voltage (I-V) characteristics measured under illumination using a

white light for two samples. (a) Platinum/planar InP sample and (b)

platinum/porous InP sample without irregular top layer. ......................... 63

Figure 4-18. Optical responsivity of Pt/planar InP and Pt/porous InP PC devices under

monochromatic light with the wavelength of 514.5 nm. .......................... 64

Chapter 5

Figure 5-1. Cyclic voltammograms measured on as-grown GaN electrode under light

irradiation with various PIN. ..................................................................... 73

Figure 5-2. Top and cross-sectional SEM images of GaN porous nanostructures

formed by photo-assisted electrochemical etching at VEC = 1.0 V, PIN = 5

mW/cm2, and various tEC: (a) tEC = 5 min (sample A), (b) tEC = 10 min

(sample B), (c) tEC = 30 min (sample C), and (d) tEC = 10 min after H3PO4

treatment (sample B'). .............................................................................. 74

Figure 5-3. Relationship between pore depth Dp and charge density Q passing through

the electrode during the photo-assisted EC etching. ................................ 75


photo-assisted EC etching. (a) Photo-holes are transferred preferentially to

the pore tips at which electric field lines concentrated. (b) Etching at initial

stage proceeds in the vertical direction. (c) Photo-holes are transferred to

pore walls and the top-surface in addition to the pore tips. (d) Etching of

pore walls and the top-surface occurs. ..................................................... 76

160 Appendix C

Figure 5-5. (a) PL spectra obtained at room temperature from a non-porous sample and

porous sample C formed at tEC = 30 min, and (b) PL peak positions of

porous samples and a non-porous sample. ............................................... 77

Figure 5-6. Diffuse reflectance spectra of non-porous and porous GaN: sample A

formed at tEC = 5 min, sample B formed at tEC = 10 min, sample C formed

at tEC = 30 min, and sample B' formed at tEC = 10 min after H3PO4

treatment. .................................................................................................. 78

Figure 5-7. Photoelectrochemical characteristics of non-porous and porous GaN:

sample A formed at tEC = 5 min, sample B formed at tEC = 10 min, sample

C formed at tEC = 30 min, and sample B' formed at tEC = 10 min after

H3PO4 treatment. ...................................................................................... 79

Figure 5-8. Current-time curves obtained during EC etching in dark at various anode

voltage VEC values. ................................................................................... 81

Figure 5-9. Cross-sectional SEM images of GaN porous structures formed in dark with

various EC conditions: (a) tEC = 20 min, VEC = 8 V; (b) tEC = 20 min, VEC =

10 V; (c) tEC = 20 min, VEC = 15 V; (d) tEC = 40 min, VEC = 10 V. .......... 82

Figure 5-10. (a) Average pore depth Dp and (b) diameter Wp of GaN porous structures

formed by photo-assisted EC etching (black symbols) and anisotropic EC

etching (red symbols) plotted as a function of EC etching time tEC. ........ 83


anisotropic EC etching. (a) Trenches are formed and induce electric field

enhancement. (b) Randomly oriented pores are formed by free carriers

generated by avalanche effect. (c) The SCRs of neighboring pores merge,

restricting free carrier excitation except for underneath the pore tips. (d)

Pores are formed anisotropically in the vertical direction. ....................... 84

Figure 5-12. Schematic illustration of the electrochemical setup used for both the

formation of porous structures and the spectro-electrochemical

measurements. .......................................................................................... 85

Figure 5-13. Top and cross-sectional SEM images of GaN-porous samples formed in

FSI mode with VEC = 1.0 V, PIN = 5 mW/cm2 and different tEC: (a) tEC = 10

min and (b) tEC = 30 min, and samples formed in BSI mode with VEC = 5.0

V, PIN = 65 mW/cm2 and different tEC: (c) tEC = 10 min and (d) tEC = 30

min. ........................................................................................................... 86


Figure 5-14. Relationship between pore depth Dp and charge density Q passing through

during photo-assisted EC etching in FSI mode and BSI mode. ............... 87

Figure 5-15. (a) Transmittance spectra of a GaN planar electrode, obtained by applying

anode voltages of VEC = 0 and 10 V. (b) Transmittance difference T0V −

T10V plotted as a function of the photon energy hν. .................................. 89

Figure 5-16. Cross-sectional SEM images of GaN porous structures formed under

various electrochemical conditions: (a) dark, tEC = 40 min, VEC = 10 V; (b)

illuminated, tEC = 40 min, VEC = 10 V, hν = 3.54 eV; (c) illuminated, tEC =

40 min, VEC = 10 V, hν = 3.26 eV. ............................................................ 91

Figure 5-17. Relationship between average pore diameter Wp and light intensity PIN

with hν = 3.26 eV. Wp was estimated from SEM observation of GaN

porous structure formed by FKE-assisted EC etching. ............................ 92

Figure 5-18. (a) Potential distribution in GaN porous structures drawn with contour

levels at 0.2, 0.4, 0.6, 0.8, and 1.0 eV, and (b) the cross-sectional potential

distribution at the top surface (line i) and pore tips (line ii) obtained by

solving 3D Poisson equation. ................................................................... 93

Figure 5-19. Top and cross-sectional SEM images of GaN porous structures formed by

anisotropic EC etching (a) without TMAH etching, (b) with TMAH

etching for 60 min. ................................................................................... 94

Figure 5-20. Relationship between average pore diameter Wp and TMAH etching time

tTMAH. Insets shows top SEM images of pores with different tTMAH. ....... 95

Figure 5-21. Normalized photoluminescence spectra of (a) UV region, and (b) visible

region obtained at room temperature: Black line represents planar GaN,

blue line represents porous GaN without TMAH etching, and red line

represents porous GaN with TMAH etching for 60 min. ......................... 96

Figure 5-22. (a) Specular reflectance spectra of planar (black line) and porous GaN

(colored line) with different TMAH etching time tTMAH. (b) The tTMAH

dependency of effective refractive index of porous GaN calculated in

accordance with effective medium approximation model (red dots) and

thin-film interference model (black line). ................................................ 97

162 Appendix C

Figure 5-23. (a) Current-voltage characteristics under irradiation of monochromatic

light (λ = 350 nm, PIN = 0.1 mW/cm2), and (b) IPCE at 0V plotted as a

function of wavelength: Black line represents planar GaN, blue line

represents porous GaN without TMAH etching, and red line represents

porous GaN with TMAH etching for 45 min. .......................................... 99

Figure 5-24. The tTMAH dependency of IPCE at 0 V, (b) Width of quasi-neutral region

WQNR, and (c) modulus of flat band potential |VFB|. (d) Schematic

representations of energy band diagrams in pore wall if Wpw > 2WSCR (left)

or Wpw < 2WSCR (right). .......................................................................... 100

Chapter 6

Figure 6-1. Schematic illustrations of a typical AlGaN/GaN HEMT structure using

p-GaN gate for normally-off operation. ................................................. 112

Figure 6-2. Schematic illustrations of electrochemical etching cycle: oxidation is

caused by ions and holes, followed by dissolution of resulting oxide. .. 113

Figure 6-3. Schematic illustrations of (a) a sample structure and (c) triangular

waveform of applied voltage. An ohmic contact and SiO2 mask were

fabricated on the p-GaN surface. ............................................................ 114

Figure 6-4. (a) JEC-VEC characteristics of the p-GaN and i-AlGaN electrodes, and (b)

JEC observed on the sample electrode when a train of triangular-wave

voltage pulses was applied. The decrease of JEC with cycle was resulted

from the decrease of thickness of p-GaN layer. ..................................... 114

Figure 6-5. The typical (a) SEM and (b) AFM images of the sample after the

electrochemical etching. Vertical steps can be observed between the

masked and unmasked regions, indicating that the p-GaN layer was etched

along the mask pattern. ........................................................................... 115

Figure 6-6. AFM cross-sectional profiles and histograms of samples (a) before etching,

(b) after etching with n = 3 cycles and (c) n = 5 cycles. (d) Etching depth

as a function of the number of cycles, n. During the initial stage of the

electrochemical etching, the etching depth can be linearly controlled by n

at a rate of 25 nm/cycle. ......................................................................... 116


Figure 6-7. Cross-sectional TEM images of the sample after the electrochemical

etching. Top p-GaN layer was completely removed and AlGaN layer was

not etched at all. ..................................................................................... 117

Figure 6-8. The differential intensity of AES spectra obtained on unetched region

and etched region after the 5 cycles applied. The p-GaN layer with

thickness of 100 nm was just etched and the i-AlGaN layer appeared on

the surface. ............................................................................................. 117

Figure 6-9. (a) I-V characteristics of the two kinds of diodes formed on ICP-etched and

electrochemically-etched samples. The electrochemically-etched sample

showed lower leakage currents, indicating less tunneling leakage at the

Schottky interface. (b) C-V curve of the Schottky diode formed on the

electrochemically-etched sample. The experimental data are well

reproduced by the calculation (black solid line). ................................... 118

Chapter 7

Figure 7-1. Schematic representations of (a) sample structure, and (b) the experimental

setup of photo-assisted electrochemical etching. ................................... 124

Figure 7-2. Current-voltage characteristics of AlGaN/GaN hetero-structure immersed

in electrolyte in dark (black line) and under light with wavelength λ of 300

nm (red line), 360 nm (blue line), and 400 nm (green line), respectively.

Sweep direction was positive, and sweep rate was set in 50 mV/s. ....... 126

Figure 7-3. Potential distribution of electrolyte/AlGaN/GaN structure at voltage VEC of

(a) −1.0 V, (b) 1.0 V, and (c) 3.0 V, respectively, calculated by

one-dimensional Poisson equation. Schottky barrier height was assumed to

be 1.0 eV. ................................................................................................ 127

Figure 7-4. (a) 3D-AFM image, (b) top SEM image, and (c) cross-sectional SEM

image of sample after photo-assisted EC etching with photo-

carriers generated in GaN layer: VEC = 5.0 V, λ = 360 nm, PIN = 1.0

mW/cm2. ................................................................................................. 128

Figure 7-5. (a) 3D-AFM image, (b) top SEM image, and (c) cross-sectional SEM

image of sample after photo-assisted EC etching with photo-

carriers generated in AlGaN layer: VEC = −0.2 V, λ = 300 nm, PIN = 1.0

mW/cm2. ................................................................................................. 129

164 Appendix C

Figure 7-6. (a) Relationship between etching depth and etching time, and (b)

relationship between self-termination depth and light intensity obtained on

the sample etched with photo-carriers generated in AlGaN layer. ......... 130

Figure 7-7. Capacitance-voltage characteristics measured at 100 kHz for Schottky

diode fabricated on planar (black) and etched (red) samples: symbols

represent experimental results, and solid lines represent theoretical curve

assuming AlGaN thickness of 25 nm for planar sample and 8 nm for

etched sample. ........................................................................................ 131

Figure 7-8. Drain current-voltage (IDS-VDS) characteristics of (a) planar-gate and (b)

recessed-gate AlGaN HEMTs with gate length of 10 µm and source-drain

spacing of 30 µm. The static on-resistance RON were estimated from the

inverse of slope in linear region. ............................................................ 132

Figure 7-9. The transfer characteristics of planar-gate (black) and recessed-gate (red)

Schottky HEMTs in the saturated region (VDS = 10 V): symbols represent

drain current IDS, and solid lines represent transconductance Gm. ......... 133

Appendix A

Figure A-1. Schematic representation of energy band diagram of a GaN and a Cu2O

reported by B. Kramm and co-workers [4]. The band alignment of

Cu2O/GaN heterostructure classified into type II with conduction band

offset of 0.24 eV. .................................................................................... 143

Figure A-2. Top SEM images of Cu2O deposited on GaN by EC deposition process

with various conditions: (a) constant mode, TEC = 25, tEC = 10 min; (b)

constant mode, TEC = 25, tEC = 20 min; (c) constant mode, TEC = 25,

tEC = 80 min; (d) constant mode, TEC = 75, tEC = 20 min; (e) pulsed

mode, TEC = 75, tEC = 20 min. ............................................................ 144

Figure A-3. Schematic representations of formation flow of Cu2O with constant bias

mode. (a) Chemical and electron transfer reactions are occurred at

GaN/electrolyte interface. (b) Cu2O particles are formed. (c) Concentration

of interfacial ions become low (diffusion-limited reaction), resulting in (d)

Cu2O particles are aggregated. ............................................................... 145

Figure A-4. AES differential spectra obtained at a bare GaN (a black line) and a Cu2O

film deposited on GaN (a red line). ........................................................ 146


Figure A-5. X-ray diffraction pattern of a bare GaN (a black line) and a Cu2O film

deposited on GaN (a red line). ................................................................ 147

Figure A-6. Transmittance, reflectance spectra, and calculated (αhν)2-hν plots (Tauc

plots) obtained at a bare GaN and a Cu2O/GaN heterostructure. ........... 147

Figure A-7. Photoresponse properties of a bare GaN and a Cu2O/GaN heterostructure

obtained at voltage V of 0 V under monochromatic light. The light started

to be irradiated at 10 s and was turned off at 30 s. ................................. 148

Figure A-8. The incident-photon-to-current conversion efficiency (IPCE) calculated at

voltage V of 0V plotted as a function of wavelength: black symbols

represent a bare GaN electrode, and red symbols represent a Cu2O/GaN

heterostructure electrode. ....................................................................... 149

Tables:

Chapter 1

Table 1-I. Physical properties of Si and III-V semiconductors at RT. ........................ 3

Chapter 2

Table 2-I. Typical morphological characteristics of pores reported on Si and III-V

semiconductors. Porous structure can be classified into three groups:

macroporous with pore diameters of less than 2 nm; mesoporous with pore

diameters between 2 nm and 50 nm; and macroporous with pore diameters

of greater than 50 nm. .............................................................................. 26

Chapter 5

Table 5-I. Structural and optical properties of porous samples. ............................... 80

Table 5-II. Summary of the relationships between EC conditions and anodic

reactions with applied voltage VEC of 3.0 V and light intensity PIN of 3.0

mW/cm2. ................................................................................................... 88

166 Appendix C

167

Research achievements

Publications:

Journal papers

[1] Y. Kumazaki, T. Kudo, Z. Yatabe, and T. Sato, "Investigation on Optical

Absorption Properties of Electrochemically Formed Porous InP using

Photoelectric Conversion Devices", Applied Surface Science, vol. 279, pp.

116−120, 2013.

[2] R. Jinbo, Y. Kumazaki, Z. Yatabe, and T. Sato, "Formation and Photoelectrical

Measurements of Pt Schottky Interfaces on InP Porous Structures", ECS

Transactions, vol. 50, pp. 247−252, 2013.



Formed by Photo-Assisted Electrochemical Etching", Journal of The

Electrochemical Society, vol. 161, pp. H705−H709, 2014.

[4] A. Watanabe, Y. Kumazaki, Z. Yatabe, and T. Sato, "Formation of GaN-porous


Illumination Mode", ECS Electrochemistry Letters, vol. 4, pp. H11−H13, 2015.

[5] T. Sato, H. Kida, Y. Kumazaki, and Z. Yatabe, "Bias-dependent

Photoabsorption Properties of GaN Porous Structures under Back-side

Illumination", ECS Transactions, vol. 69, pp. 161−166, 2015.



photoabsorption edge", Semiconductor Science and Technology, vol. 31, p.

14012, 2016.



etching", Japanese Journal of Applied Physics, vol. 55, p. 04EJ12, 2016.

Kumazaki

タイプライターテキスト

-link-

Kumazaki


-link-

Kumazaki


-link-

Kumazaki


-link-

Kumazaki


-link-

Kumazaki


-link-

Kumazaki


-link-

Kumazaki


Kumazaki


http://www.sciencedirect.com/science/article/pii/S016943321300737X

http://ecst.ecsdl.org/content/50/37/247.abstract

http://jes.ecsdl.org/content/161/10/H705.short

http://eel.ecsdl.org/content/4/5/H11.abstract

http://ecst.ecsdl.org/content/69/2/161.abstract

http://iopscience.iop.org/article/10.1088/0268-1242/31/1/014012/meta

http://iopscience.iop.org/article/10.7567/JJAP.55.04EJ12/meta

168 Research achievements



p-GaN layers utilizing electrochemical reactions", Proceedings of SPIE, vol.

9748, p. 97480Y, 2016.

[9] Y. Kumazaki, S. Matsumoto, and T. Sato, "Precise Structural Controlling of

GaN Porous Nanostructures Using Anisotropic Electrochemical and Wet

Etching Technique", Journal of The Electrochemical Society (under review).

[10] Y. Kumazaki, K. Uemura, T. Sato, and T. Hashizume, "Precise thickness

control in recess etching of AlGaN/GaN heterostructures by

photocarrier-regulated electrochemical process", Japanese Journal of Applied

Physics (under review).

[11] T. Sato, X. Zhang, K. Ito, S. Matsumoto, and Y. Kumazaki, "Electrochemical

Formation of N-type GaN and N-type InP Porous Structures for Chemical

Sensor Applications", IEEE Sensors Journal (to be submitted).

Technical reports (in Japanese)


る InP多孔質構造の光吸収特性と光電変換", 電子情報通信学会信学技報,

vol. 113, ED2013-27, pp. 61−64, 2013.


る GaN 多孔質構造の形成と光電極特性", 電子情報通信学会信学技報,

vol. 113, ED2013-88, pp. 113−116, 2013.



信学技報, vol. 115, ED2015-28, pp. 63−66, 2015.

[4] 喜田弘文, 熊崎祐介, 谷田部然治, 佐藤威友, "電気化学的手法によ

る GaN 多孔質構造の形成と紫外光応答特性", 電子情報通信学会信学技

報, vol. 115, ED2015-53, pp. 51−54, 2015.



信学技報, vol. 116, ED2016-66, pp. 45−50, 2016.

Kumazaki


-link-

http://spie.org/Publications/Proceedings/Paper/10.1117/12.2211964


Conferences:

International conferences

[1] Y. Kumazaki, T. Kudo, Z. Yatabe, and T. Sato, "Optical Absorption Properties

of InP Porous Structures Formed by Electrochemical Process", 2012

International Conference on Solid State Devices and Materials (SSDM2012),

PS-6-5, Kyoto, Japan, September 25−27, 2012.

[2] Y. Kumazaki, A. Watanabe, R. Jinbo, Z. Yatabe, and T. Sato, "Electrochemical

Formation and Optical Characterization of GaN Porous Structures", The 40th

International Symposium on Compound Semiconductors (ISCS2013),

MoPC-05-22, Kobe, Japan, May 19−23, 2013.

[3] Y. Kumazaki, A. Watanabe, R. Jinbo, Z. Yatabe, and T. Sato, "Electrochemical

Formation and Optical Characterization of GaN Porous Structures", 32nd

Electronic Materials Symposium (EMS32), We2-15, Shiga, Japan, July 10−12,

2013.

[4] Y. Kumazaki, N. Azumaishi, H. Ueda, M. Kanechika, H. Tomita, T. Sato, and T.

Hashizume, "Selective etching of p-GaN layers for normally-off AlGaN/GaN

HEMTs by electrochemical process", 10th International Conference on Nitride

Semiconductors (ICNS-10), DP2.20, Washington, DC, USA, August 25−30,

2013.

[5] Y. Kumazaki, A. Watanabe, Z. Yatabe, and T. Sato, "Structural and Optical

Characterization of GaN Porous Structures Formed by Photo-assisted

Electrochemical Process", 224th ECS meeting, A1-51, San Francisco, USA,

October 27−November 1, 2013.


Structural and Optical Properties of GaN Porous Structures Formed by

Photo-assisted Electrochemical Etching", 8th International Workshop on

Nitride Semiconductors (IWN2014), WeBP17, Wrocław, Poland, August

24−29, 2014.

[7] Y. Kumazaki, T. Sato, and Z. Yatabe, "Formation of Self-aligned Pore Arrays

on n-GaN Substrates by Photo-assisted Electrochemical Etching Process",

2015 International Conference on Solid State Devices and Materials

(SSDM2015), D-7-5, Sapporo, Japan, September 27−30, 2015.


[8] Y. Kumazaki, Z. Yatabe, and T. Sato, "Size-Controlled Formation of

High-Aspect Ratio Porous Nanostructures on GaN Substrates Utilizing

Photo-Assisted Electrochemical Etching for Photovoltaic Applications", 2015

MRS Fall Meeting & Exhibit, NN20.20, Boston, USA, November

29−December 4, 2015.

[9] Y. Kumazaki, S. Matsumoto, and T. Sato, "Precise Structural Tuning of Porous

GaN Using Two-Step Anisotropic Etching for Optical and Photo-Electric

Applications", 2016 Pacific Rim Meeting on Electrochemistry and Solid State

Science (PRiME 2016), C04-1253, Honolulu, USA, October 2−7, 2016.

Domestic conferences (in Japanese)

[1] 熊崎祐介, 工藤智人, 谷田部然治, 佐藤威友, "pn 接合基板上に形成

した InP 多孔質構造の光学応答特性の評価", 第 73 回応用物理学会学術

講演会, 13a-F1-2, 愛媛大学, 2012 年 9 月 11 日−14 日.

[2] 熊崎祐介, 渡部晃生, 神保亮平, 谷田部然治, 佐藤威友, "電気化学

的手法による GaN 多孔質構造の形成と光学特性の評価", 第 60 回応用物

理学会春季学術講演会, 30a-G20-14, 神奈川工科大学, 2013 年 3 月 27 日

−30 日.


る InP多孔質構造の光吸収特性と光電変換", 電子情報通信学会電子デバ

イス研究会, 13, 豊橋技術科学大学, 2013 年 5 月 16 日−17 日.


る GaN 多孔質構造の形成と光学特性の評価", 第 29 回ライラックセミ

ナー・第 19 回若手研究者交流会, 06, 小樽, 2013 年 6 月 15 日−16 日.

[5] 熊崎祐介, 渡部晃生, 谷田部然治, 佐藤威友, "GaN 多孔質ナノ構造

の表面形状と光電気化学特性", 第 74 回応用物理学会秋季学術講演会,

16p-B5-2, 同志社大学, 2013 年 9 月 16 日−20 日.


る GaN 多孔質構造の形成と光電極特性", 電子情報通信学会電子デバイ

ス研究会, 26, 大阪大学, 2013 年 11 月 28 日−29 日.

[7] 熊崎祐介, 佐藤威友, 橋詰保, "AlGaN/GaN ヘテロ構造上に形成した

p-GaN 層の選択的電気化学エッチング", 第 61 回応用物理学会春季学術

講演会, 19a-D8-8, 青山学院大学, 2014 年 3 月 17 日−20 日.


[8] 熊崎祐介, 渡部晃生, 谷田部然治, 佐藤威友, "電気化学エッチング

による GaN 多孔質構造の形成と形状制御の向上", 第 75 回応用物理学会

秋期学術講演会, 18p-A27-1, 北海道大学, 2014 年 9 月 17 日−20 日.

[9] 熊崎祐介, 近江沙也夏, 谷田部然治, 佐藤威友, "電気化学堆積法に

よる Cu2O/GaN ヘテロ構造の形成と特性評価", 第 62 回応用物理学会春

季学術講演会, 11a-D10-6, 東海大学, 2015 年 3 月 11 日−14 日.



電子デバイス研究会, 14, 豊橋技科大学, 2015 年 5 月 28 日−29 日.

[11] 熊崎祐介, 近江沙也夏, 谷田部然治, 佐藤威友, "Cu2O/GaN ヘテロ構

造の電気化学形成と光学的特性評価", 第 76 回応用物理学会秋期学術講

演会, 14p-PA13-19, 名古屋国際会議場, 2015 年 9 月 13 日−16 日.

[12] 熊崎祐介, 松本悟, 佐藤威友, "異方性ウェットエッチングによる

GaN 多孔質構造の作製と光学特性評価", 第 63 回応用物理学会春季学術

講演会, 19p-H121-14, 東京工科大学大岡山キャンパス, 2015 年 3 月 19

日−22 日.

[13] 熊崎祐介, 植村圭佑, 佐藤威友, "AlGaN/GaN ヘテロ構造の光電気化

学特性と低損傷加工技術への応用", 第 77 回応用物理学会秋季学術講演

会, 16p-B1-9, 朱鷺メッセ, 2016 年 9 月 13 日−16 日.



電子デバイス研究会, 10, 京都大学桂キャンパス, 2016 年 12 月 12 日−13

日.

Award (in Japanese):

[1] 2013 年度電子情報通信学会電子デバイス研究会論文発表奨励賞

講演題目 "電気化学的手法による GaN 多孔質構造の形成と光電極特性"

Patent (in Japanese):

[1] 佐藤威友, 熊崎祐介, 橋詰保, "半導体微細加工法"（出願予定）.


Documents

Formation and Device Application of III-V …...Formation and Device Application of III-V Semiconductor-based Functional Microstructures Achieved by Electrochemical Process （電気化学的制御に基づくIII-V