FUNDAMENTAL INVESTIGATIONS OF PHOTOELECTROCHEMICAL WATER
SPLITTING WITH COPPER TUNGSTATE PHOTOANODES
By
Yuan Gao

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Chemistry – Doctor of Philosophy
2017

ABSTRACT
FUNDAMENTAL INVESTIGATIONS OF PHOTOELECTROCHEMICAL WATER
SPLITTING WITH COPPER TUNGSTATE PHOTOANODES
By
Yuan Gao
An increasing global energy demand along with detrimental consequences of extensive
usage of fossil fuel calls for the development of abundant, cost effective and carbon neutral
alternative energy sources. Solar energy is well distributed and is the most abundant energy source
available on Earth. This work focuses on using copper tungstate (CuWO4) photoelectrode to
harvest sunlight to oxidize water, which constitutes the rate limiting half reaction of
photoelectrochemical water splitting to generate clean and carbon neutral energy dense fuel,
hydrogen, and the byproduct, oxygen. CuWO4 is a promising photoanode material for water
splitting owing to its reasonably small bandgap, suitable band position, stability and abundance,
but has received limited attention due to its reported poor bulk properties and low hole collection
efficiency at the surface. A lack of fundamental insight into its limitations hinders rational design
of strategies to improve its performance. We developed an atomic layer stack deposition-annealing
(SDA) method derived from atomic layer deposition (ALD) to synthesize our CuWO4 thin films.
While keeping the benefit of precise thickness control of a uniform film as traditional ALD
procedures, this new approach of ALD of ternary metal oxides allows for a wider selection of
precursors since their ALD temperature windows do not have to overlap, and it allows for an easier
control of stoichiometry. The hole collection efficiency at the electrode surface was quantified by
comparing water oxidation and the oxidation of a suitable hole scavenger, as well as by intensity
modulated photocurrent spectroscopy (IMPS). Our results showed that water oxidation with
CuWO4 is limited by surface recombination at low applied bias, but is essentially quantitative at

higher potentials. A series of photoelectrochemical and electrochemical impedance spectroscopic
measurements were employed to elucidate the role of surface state of CuWO4 during water
oxidation. Our results suggested that the surface state of CuWO4 is related to a water oxidation
intermediate species rather than intrinsic, as is opposed to previous literature. Relatively thick
CuWO4 electrodes were made by spray pyrolysis and the pores were back-filled with Al2O3 to
produce compact films for bulk studies. Illumination direction and wavelength dependent
photoelectrochemical measurements were employed to investigate the bulk properties of CuWO4
for water oxidation, which suggested that the major problem of CuWO4 is a short electron
collection length.

ACKNOWLEDGEMENTS

When I look back on my five years’ graduate life, I found that there are so many people I
would like to express my gratitude. It was not only an experience of pursuing a PhD degree, but
also an unforgettable experience living in a different country. It is all of these people who have
helped and cared for me in my academic and personal life that got me through these years.
First of all, I would like to express my most sincere gratitude and appreciation to my
advisor, Professor Thomas W. Hamann, for his patience, guidance, encouragement and support.
Tom’s wisdom and critical scientific insight has always helped, enlightened and inspired me
throughout my research. He has always been a role model for his natural curiosity, passion and
dedication in science. My completion of my PhD degree would not be accomplished without the
help of my colleagues in Hamann Group. Many thanks to Ben, Kelley and Omid, who were the
past members of the Water Splitting project, for mentoring me and sharing their inspiring thoughts
and ideas about research. I would like to thank Yuling, who has graduated but has helped me a lot
in my research and also personal life. I would like to thank Hamed, as another member in Water
Splitting subgroup in the same year, for all the helpful discussions on our projects. I’d like to thank
Yujue, Qiong, Mandal, Dan, Josh, Faezeh, Geletu, Austin and Parisa for all the joy and fun in
Hamann Lab. I would also like to thank National Science Foundation for funding the research
presented herein, and also MSU Dissertation Completion Fellowship (Summer 2017) which
allowed me to complete this dissertation.
Many thanks to my committee members, Professor Gary Blanchard, Professor Rémi
Beaulac and Professor Benjamin Levine for their valuable suggestions and advice. I would also

iv

like to express my thanks to my previous committee member and second reader, Professor Merlin
Bruening, who has left MSU, for his help and advice during my PhD years.
I would like to thank my friends who cared for me. Many thanks to Chuanpeng Jiang, who
has helped me in both my research and personal life. Thanks to Xiaoran Zhang, Shuang Liang,
Ruiqiong Guo, Ruipeng Mu, Yiqing Yang and Zhihui Liu for their company in all these years.
Thanks to Nan Du, Ningliang Li, Yang Shangguan, Li Ke and Keying Wen, for the wonderful
time playing sports and games. Thanks to Yun Xian who is far away but gave me a lot of support.
I cannot express enough gratitude to my husband and best friend, Zhihao Jin, who is the
most important person that has helped me through these years. Any of my achievements would
not be possible without his patience, understanding, care and love. Many thanks to my relatives
for helping my family in China while I am absent. Finally, I would like to thank my parents for
their selfless love and support. No matter where I go, they are always the ones I care for.

v

TABLE OF CONTENTS

LIST OF TABLES.......................................................................................................................viii
LIST OF FIGURES ....................................................................................................................... ix
Chapter 1: Introduction ................................................................................................................... 1
1.1 Motivation ............................................................................................................................. 2
1.2 Approach ............................................................................................................................... 2
1.3 Performance Evaluation ........................................................................................................ 6
1.3.1 Current Density - Voltage (J-V) Measurements ............................................................. 6
1.3.2 Quantum Efficiencies ..................................................................................................... 8
1.4 Copper Tungstate (CuWO4) .................................................................................................. 8
1.5 Objectives ............................................................................................................................ 14
REFERENCES .......................................................................................................................... 15
Chapter 2: Atomic Layer Stack Deposition-Annealing Synthesis of CuWO4 .............................. 21
2.1 Abstract ............................................................................................................................... 22
2.2 Introduction ......................................................................................................................... 23
2.3 Experimental ....................................................................................................................... 25
2.3.1 Film Deposition ............................................................................................................ 25
2.3.2 Calculation of CuO Deposition Cycles ........................................................................ 26
2.3.3 Calculation of Film Thickness ...................................................................................... 26
2.3.4 Film Characterization ................................................................................................... 27
2.3.5 Photoelectrochemistry .................................................................................................. 28
2.4 Results and Discussion ........................................................................................................ 29
2.5 Conclusion........................................................................................................................... 52
APPENDIX ............................................................................................................................... 54
REFERENCES .......................................................................................................................... 56
Chapter 3: Quantifying Hole Collection Efficiency of CuWO4 for Water Oxidation .................. 62
3.1 Abstract ............................................................................................................................... 63
3.2 Introduction ......................................................................................................................... 64
3.3 Experimental ....................................................................................................................... 65
3.3.1 Film Preparation and Characterization ......................................................................... 65
3.3.2 Characterization ............................................................................................................ 65
3.3.3 Photoelectrochemical Measurements ........................................................................... 66
3.4 Results and Discussion .................................................................................................... 67
3.5 Conclusion........................................................................................................................... 78
APPENDIX ............................................................................................................................... 79
REFERENCES .......................................................................................................................... 85
Chapter 4: Elucidation of Role of CuWO4 Surface States During Photoelectrochemical Water
Oxidation....................................................................................................................................... 89
vi

4.1 Abstract ............................................................................................................................... 90
4.2 Introduction ......................................................................................................................... 91
4.3 Experimental ....................................................................................................................... 91
4.3.1 Film Preparation ........................................................................................................... 91
4.3.2 Film Characterization ................................................................................................... 92
4.3.3 Electrochemical / Photoelectrochemical Measurements .............................................. 92
4.3.4 Measurements in Aqueous Electrolytes ....................................................................... 93
4.3.5 Measurements in Non-aqueous Electrolytes ................................................................ 93
4.4 Results and Discussion ........................................................................................................ 93
APPENDIX ............................................................................................................................. 107
REFERENCES ........................................................................................................................ 124
Chapter 5: Investigations of Charge Separation Limitations of CuWO4 .................................... 127
5.1 Abstract ............................................................................................................................. 128
5.2 Introduction ....................................................................................................................... 129
5.3 Approaches and Methods .................................................................................................. 132
5.4 Experimental ..................................................................................................................... 136
5.4.1 Film Preparation ......................................................................................................... 136
5.4.2 Film Characterization ................................................................................................. 137
5.4.3 Sandwich Cell Preparation ......................................................................................... 137
5.4.4 Photoelectrochemistry ................................................................................................ 138
5.5 Results and Discussion ...................................................................................................... 138
5.6 Conclusions ....................................................................................................................... 153
REFERENCES ........................................................................................................................ 155
Chapter 6: Conclusions and Future Directions ........................................................................... 158
6.1 Conclusions ....................................................................................................................... 159
6.2 Future Directions ............................................................................................................... 161
REFERENCES ........................................................................................................................ 164

vii

LIST OF TABLES

Table 4-1.

Flat band potential and dopant density calculated from Mott-Schottky plots in the
dark and under illumination ................................................................................ 101

Table A4-1.

Parameters extracted from fitting EIS data in 0.5 M Na2SO3 electrolyte in the
dark ..................................................................................................................... 110

Table A4-2.

Parameters extracted from fitting EIS data in 0.5 M Na2SO3 electrolyte under 1
sun illumination .................................................................................................. 111

Table A4-3.

Parameters extracted from fitting EIS data in aqueous electrolyte (pH9 KBi
buffer) in the dark ............................................................................................... 112

Table A4-4.

Parameters extracted from fitting EIS data in aqueous electrolyte (pH9 KBi
buffer) under 0.1 sun illumination ...................................................................... 113

Table A4-5.

Parameters extracted from fitting EIS data in aqueous electrolyte (pH9 KBi
buffer) under 1 sun illumination. ........................................................................ 114

viii

LIST OF FIGURES

Figure 1-1.

Diagram of PEC water splitting tandem cell. ......................................................... 5

Figure 1-2.

A three-electrode configuration for measuring the performance of a photoelectrode.
Reference, working, and counter electrodes are labeled as “RE”, “WE”, and “CE”,
respectively. The green “V” and red “I” indicate where the voltage and current are
measured, respectively. ........................................................................................... 6

Figure 1-3.

Simulated J-V curves of a tandem PEC water splitting cell. The intercept of the JV curves of the photocathode and photoanode determines the operating
photocurrent. ........................................................................................................... 7

Figure 1-4.

Band edge energy levels and bandgaps of candidates for PEC water oxidation
photoanodes. ......................................................................................................... 10

Figure 1-5.

Red curve: AM 1.5 solar spectrum; black vertical line: wavelength that corresponds
to the bandgap of CuWO4; yellow area: maximum photon flux that can be absorbed
by CuWO4 under 1 sun illumination..................................................................... 11

Figure 1-6.

Schematic diagram of PEC water oxidation at CuWO4 photoanode/electrolyte
interface. The red arrow indicates light absorption (charge generation), green
arrows represent charge separation, including diffusion in the quasi-neutral region
and drift due to the electric field in the depletion region, and blue arrows indicate
hole collection at CuWO4/electrolyte interface to oxidize water. Meanwhile, the
dark red arrows represent possible recombination pathways in the bulk or at the
interface................................................................................................................. 13

Figure 2-1.

Schematic illustration of a) ALD reactor and b) one cycle of ALD process. ....... 23

Figure 2-2.

Experimental X-ray photoelectron spectroscopy (XPS) spectra of W 4f and O 1s
peaks (black scatters) with fitting (red solid lines and dots) of as-deposited (top)
and annealed (bottom) WO3.................................................................................. 30

Figure 2-3.

X-ray diffraction (XRD) patterns of as-deposited (black) and annealed (red) WO3.
............................................................................................................................... 30

Figure 2-4.

Raman spectra of as-deposited (black) and annealed (red) WO3. ........................ 31

Figure 2-5.

Thickness of WO3 thin films vs number of deposition cycles using different metal
precursor pulse lengths. ........................................................................................ 31

Figure 2-6.

Experimental XPS spectra of Cu 2p and O 1s peaks (black scatters) with fitting (red
solid lines and dots) of as-deposited (top) and annealed (bottom) CuO. .............. 33
ix

Figure 2-7.

XRD patterns of as-deposited (black) and annealed (red) CuO. .......................... 33

Figure 2-8.

Raman spectra of as-deposited (black) and annealed (red) CuO. ......................... 34

Figure 2-9.

Thickness vs number of deposition cycles of CuO thin films with different metal
precursor pulse length. .......................................................................................... 34

Figure 2-10. Total thickness of a) WO3 on 100 cycles of CuO and b) CuO on 100 cycles of WO3
as a function of number of deposition cycles........................................................ 36
Figure 2-11. Sequential ALD of WO3/CuO and post-ALD annealing to make CuWO4 thin films
via stack deposition-annealing (SDA) approach................................................... 37
Figure 2-12. Cross-section SEM images of a) as-deposited WO3 and b) annealed CuWO4 thin
films on Si substrate. ............................................................................................. 39
Figure 2-13. AFM surface morphology image of a) as-deposited WO3, b) as-deposited CuO on
WO3, and c) annealed CuWO4 on Si substrate. .................................................... 40
Figure 2-14. XPS survey spectra of as-deposited (bottom) and annealed (top) film of CuO on top
of WO3. ................................................................................................................. 41
Figure 2-15. Experimental XPS spectra of a) Cu 2p and W 4f peaks (black scatters) with fitting
(red solid lines and dots) of CuO on top of WO3 as-deposited (top) and annealed
(CuWO4) (bottom) and O 1s peaks (black scatters) with fitting (red solid lines and
dots) of CuO on top of WO3 b) as-deposited and c) annealed (CuWO4). ............ 42
Figure 2-16. a) X-ray diffraction patterns and b) Raman spectra of CuO on WO3 as-deposited
(black) and annealed (CuWO4) (red). Peak positions of CuWO4 from literature are
plotted as vertical dashed gray lines. .................................................................... 44
Figure 2-17. Images of CuWO4 thin films deposited on FTO substrates with varying ALD cycles
of WO3 of 0, 500, 1000, 1500, 2000, and 2500. ................................................... 45
Figure 2-18. Thickness of CuWO4 films with different deposition cycles of WO3 measured with
AFM (black squares). Error bars represent the standard deviation of 10 different
measurements on each film. Red dash shows the linear fit of thickness vs. number
of cycles. ............................................................................................................... 46
Figure 2-19. Top: a) Absorbance of CuWO4 thin films with ALD cycles of WO3 of 500 (red
solid), 1000 (orange dash), 1500 (yellow short dash), 2000 (green dash dot), and
2500 (blue dot). Bottom: Absorbance of CuWO4 thin films at wavelength of b) 400
nm and c) 450 nm varying with different ALD cycles of WO3. Error bars represent
standard deviation of independent measurements on three different spots for each
thickness. ............................................................................................................... 47

x

Figure 2-20. a) IPCE (red scatters) in pH9 KBi buffer at 1.23 V vs RHE and absorptance
spectrum (black solid line), b) Tauc plot, and c) absorption coefficient of CuWO4
thin film................................................................................................................. 48
Figure 2-21. a) J-V curves of an 80 nm thick CuWO4 electrode in the dark (black dash), under 1
sun illumination (blue solid) and chopped illumination (red solid). The vertical gray
dash shows the thermal dynamic potential of water oxidation. b) Stability test of
CuWO4 (red) and WO3 (black) at 1.23 V vs RHE under 1 sun illumination.
Experiments were made in 0.2 M KCl in pH9 KBi buffer. .................................. 51
Figure 2-22. J-V curve of a) CuO and b) WO3 electrode in the dark (black) and under 1 sun
illumination (red) measured in 0.2 M KCl in pH9 KBi buffer. ............................ 52
Figure A2-1. An example of an EDX spectrum of CuWO4 on FTO-coated glass. .................... 55
Figure 3-1.

J-V curves of a CuWO4 electrode immersed in a pH 9.0 KBi electrolyte in contact
with H2O (black dashed curve), Na2SO3 (red solid curve) and H2O2 (blue dash dot)
under 1 sun illumination. The vertical gray dash represents the thermodynamic
potential of water oxidation (1.23 V vs RHE). ..................................................... 68

Figure 3-2.

a) Photocurrent densities measured with a CuWO4 electrode in an aqueous
electrolyte containing Na2SO3 with various concentrations at 0.88 (red squares),
1.23 (green triangles) and 1.53 (blue circles) V vs RHE under 1 sun illumination.
b) Current transients measured under 1 sun illumination in 0.5 M Na2SO3
electrolyte at 0.88 (red solid), 1.23 (green dash) and 1.53 (blue dots) V vs RHE. 70

Figure 3-3.

Current transients measured with CuWO4 in an aqueous electrolyte (H2O) under 1
sun illumination at 0.88 (red solid), 1.23 (green dash) and 1.53 (blue dots) V vs
RHE....................................................................................................................... 71

Figure 3-4.

Nyquist plots of IMPS measured with CuWO4 in H2O under 470 nm
monochromatic light with intensity of 116 mW cm-2. Potential ranges from 0.73 to
1.73 V vs RHE with an interval of 0.1 V corresponding to the red, orange, yellow,
dark yellow, green, dark green, dark cyan, blue, navy, purple and wine colored
circles. ................................................................................................................... 73

Figure 3-5.

Nyquist plots of IMPS measurements under 116 mW cm-2 light at 470 nm measured
in H2O (black squares), Na2SO3 (red circles), and H2O2 (blue triangles) electrolyte
at a) 0.93 V and b) 1.23 vs RHE. .......................................................................... 75

Figure 3-6.

Current multiplication factor of CuWO4 in contact with H2O2 represented as
quantum efficiency (QE) extracted from IMPS data. ........................................... 76

Figure 3-7.

Calculated  HC of water oxidation with CuWO4 electrodes extracted from IMPS
data at 116 mW cm-2 with 470 nm LED (orange triangles) and from comparisons

xi

of J-V curves of water oxidation and oxidation of a) Na2SO3 (purple solid) and b)
H2O2 under 1 sun illumination (green solid)......................................................... 77
Figure A3-1. SEM images of a) as-deposited CuO on WO3, and CuWO4 annealed at b) 550 °C
and c) 600 °C……………………………………………………………………..80
Figure A3-2. Raman spectra of a) CuWO4 annealed at 550 °C and b) FTO. c) X-ray diffraction
pattern of CuWO4 annealed at 550 °C. Vertical gray dashed lines represent the peak
positions of CuWO4 recorded in literature32–34. .................................................... 81
Figure A3-3. Jdark-V (black dash) and Jtotal-V (red solid) curves of CuWO4 electrode measured in
a) H2O, b) 0.5 M Na2SO3 and c) 0.5 M H2O2 electrolytes in the dark and under 1
sun illumination. ................................................................................................... 83
Figure 4-1.

Jphoto-V curves of a CuWO4 electrode in contact with a pH 9.0 KBi electrolyte in
the absence (black solid curve) and presence (red dashed curve) of 0.5 M Na2SO3
under 1 sun illumination. The vertical gray dash represents the thermodynamic
potential of water oxidation (1.23 V vs. RHE). .................................................... 94

Figure 4-2.

Bode phase plots of CuWO4 electrode at 0.93 V vs. RHE a) in the dark and b) under
1 sun illumination and c) at 1.53 V vs. RHE under 1 sun illumination. Data
measured in H2O (black squares) and 0.5 M Na2SO3 (red circles) are both shown.
Fitting curves are shown as black dash for H2O and red dots for Na2SO3. The
resulted parameters from fitting are listed in Table A4-1 to Table A4-5. ............ 96

Figure 4-3.

a) Equivalent circuit for EIS data with two capacitive features. b) Randles circuit.
............................................................................................................................... 97

Figure 4-4.

Plots of equivalent circuit parameters obtained from fitting EIS data in the dark
(open squares), under 0.1 sun (blue triangles) and 1 sun (red circles) illumination
in H2O electrolyte. a) Surface state capacitance; b) Charge transfer resistance; c)
Bulk capacitance; d) Resistance of trapping of holes in surface states. ................ 98

Figure 4-5.

Mott-Schottky plots of CuWO4 immersed in a) aqueous electrolyte (H2O),
measured in the dark (open squares), under 0.1 sun (blue triangles) and 1 sun (red
circles) illumination; and in b) 0.5 M Na2SO3, measured in the dark (open squares)
and under 1 sun (red circles) illumination. ......................................................... 100

Figure 4-6.

Current transients of CuWO4 in H2O after turning the 1 sun illumination a) on and
b) off. The applied potential varies from 0.68, 0.73, 0.83, 1.03, 1.23 and 1.43 V vs.
RHE corresponding to red solid, orange dash, yellow dots, green dash dots, blue
short dash and purple short dots. c) Charge stored at CuWO4 surface calculated
from current transients after turning off the light (black squares) and Gaussian
fitting of the charge (red dash) at 0.68, 0.73, 0.83, 0.93, 1.03, 1.13, 1.23, 1.33 and
1.43 V vs. RHE. .................................................................................................. 102

xii

Figure 4-7.

a) CVs taken after holding potential at 2.3 V vs. RHE for 60 s under 1 sun
illumination at scan rates from 200, 400, 600, 800, 1000, 1200 and 1500 mV s-1
corresponding to red, orange, dark yellow, green, dark green, blue and purple solid.
b) CV after holding potential at ~1.8 V vs. Vfb for 60 s in the dark, measured in H2O
(black solid) and MeCN (red dash). Scan rate is 1500 mV s-1. .......................... 103

Figure 4-8.

CV after holding potential at ~1.8 V vs. Vfb for 60 s in the dark, measured in 0.2 M
TBAPF6 in dichloromethane in the glovebox. Scan rates varies from 200, 800, 1500
and 2000 mV / s corresponding to pink, yellow, cyan and purple solid lines. Initial
scan is in the cathodic direction. ......................................................................... 105

Figure A4-1. Jdark-V (black dash) and Jtotal-V (red solid) curves of CuWO4 electrode measured in
a) H2O and b) 0.5 M Na2SO3 electrolytes in the dark and under 1 sun illumination.
............................................................................................................................. 108
Figure A4-2. Nyquist plots of CuWO4 electrode at 0.93 V vs. RHE a) in the dark and b) under 1
sun illumination and c) at 1.53 V vs. RHE under 1 sun illumination. Data measured
in H2O (black squares) and 0.5 M Na2SO3 (red circles) are both shown. Fitting
curves are shown as black dash for H2O and red dots for Na2SO3. .................... 109
Figure A4-3. CV taken after holding potential at 2.3 V vs. RHE for 60 s in the dark. Scan rates
varies from 200, 400, 600, 800, 1000, 1200 and 1500 mV s-1 corresponding to red,
orange, dark yellow, green, dark green, blue and purple solid. The peak intensity
after the dark pretreatment is slightly smaller compared to the light pretreatment.
............................................................................................................................. 115
Figure A4-4. Mott-Schottky plots (black scattered points) and linear fit (red dash) of CuWO4 in
a) MeCN electrolyte, b) MeCN electrolyte containing 2% H2O, and c)
dichloromethane electrolyte. ............................................................................... 116
Figure A4-5. CV after holding potential at ~1.8 V vs. Vfb for 60 s in the dark, measured in H2O
(black solid), MeCN (red dash) and MeCN with 2% H2O (blue dot). Scan rate is
1500 mV s-1. ........................................................................................................ 117
Figure A4-6. a) Raman spectrum and b) XRD pattern of CuWO4 synthesized by spray pyrolysis
and annealed at 550 °C for 1 hr (SP-CuWO4). Experimental data of CuWO4 are
shown as black curves and peak positions from literature are shown as vertical dark
yellow dash. The XRD pattern of FTO substrate is shown as blue curve as control
experiment, and the diffraction peaks from FTO are labeled as *. ..................... 118
Figure A4-7. J-V curves of SP-CuWO4 under 1 sun (red solid), 0.33 sun (green dash), 0.1 sun
(blue dots) illumination and in the dark (black dash dot) measured in pH 9 aqueous
(H2O) electrolyte. ................................................................................................ 119
Figure A4-8. Bode phase plots of SP-CuWO4 at 0.93V vs. RHE in the dark (black squares) and
under 1 sun illumination (red circles) measured in pH 9 aqueous (H2O) electrolyte.
xiii

Fitting curves are shown as black dash and red dots for dark and 1 sun
measurements, respectively. ............................................................................... 120
Figure A4-9. Css obtained from fitting EIS measured in pH 9 aqueous (H2O) electrolyte under
0.1 sun (blue triangles), 0.33 sun (green diamonds) and 1 sun (red circles)
illumination with SP-CuWO4. ............................................................................ 120
Figure A4-10. Mott-Schottky plots of SP-CuWO4 measured in the dark in pH 9 aqueous (H2O)
electrolyte. ........................................................................................................... 121
Figure A4-11. CVs taken after holding potential at 2.3 V vs. RHE for 60 s under 1 sun illumination
at scan rates from 200, 400, 600, 800, 1000, 1200 and 1500 mV s-1 corresponding
to red, orange, dark yellow, green, dark green, blue and purple. Experiment was
conducted in pH 9 aqueous electrolyte. .............................................................. 122
Figure 5-1.

Scheme of charge separation in a) planar CuWO4 electrode and b) optimized
nanostructured CuWO4 photoanode. .................................................................. 131

Figure 5-2.

Schematic illustration of hole generation and charge separation in CuWO4
photoanode illuminated from a) electrolyte-electrode (front) side and b) substrateelectrode (back) side. The red arrows represent hole collection and green arrows
represent electron transport. ................................................................................ 133

Figure 5-3.

Scheme of a) a nanostructured CuWO4 film, b) a nanostructured film with pores
filled with excess Al2O3 and c) after etching the top Al2O3 layer with 1 M KOH to
expose CuWO4 electrode to the electrolyte. ....................................................... 134

Figure 5-4.

Schematic illustration of light absorption in CuWO4 electrode by a) front and b)
back illumination. Ln is the electron collection length. The light red, orange, green,
blue and purple arrows represent illumination from relatively long wavelengths to
short wavelengths................................................................................................ 136

Figure 5-5.

a) Absorption coefficient and b) Tauc plot of CuWO4 indicating its indirect
bandgap. The red dashed line represents the fitting result of the linear region in
Tauc plot. The blue dash represents the corrected baseline according to the plateau
region in Tauc plot. ............................................................................................. 139

Figure 5-6.

a) Calculated raw data and b) corrected absorptance of CuWO4........................ 140

Figure 5-7.

Jphoto-V response of CuWO4 in a pH 9 KBi electrolyte with 0.5 M Na2SO3 (blue
dash) and 200 mM KI / 5 mM I2 (red solid) under 1 sun illumination. .............. 141

Figure 5-8.

Dark J-V curves of CuWO4 measured in a) Na2SO3 and b) I-/I3- electrolyte. ..... 141

Figure 5-9.

Charge separation efficiency of CuWO4 electrode. ............................................ 142

xiv

Figure 5-10. SEM image of a) top-down and b) cross-section view of bare CuWO4 by spray
pyrolysis, c) top-down and d) cross-section view of CuWO4 deposited with 1500
cycles (~150 nm) of Al2O3 and etched in 1 M KOH for 40 min. ....................... 144
Figure 5-11. J-V curves of CuWO4 coated with ~150 nm of Al2O3 (black dash) and etched in 1
M KOH for 30 min (red), 32 min (orange), 34 min (yellow), 36 min (green), 38 min
(dark cyan), 40 min (blue), 42 min (purple) and 44 min (wine). These J-V curves
are also compared with that of bare CuWO4 (pink dots). All measurements are done
in pH 9 KBi electrolyte under 1 sun illumination. .............................................. 145
Figure 5-12. Schematic illustration of CuWO4 sandwich cell and light illumination. ............ 146
Figure 5-13. Absorptance of FTO-coated glass (black solid) and platinized FTO-coated glass
(blue dash). Both were measured by illuminating from the glass side. .............. 147
Figure 5-14. Absorbance of different concentrations of I2 in pH9 KBi buffer with 200 mM KI at
a) 350 nm, b) 405 nm and c) 470nm. .................................................................. 148
Figure 5-15. a) Absorptance of the integrated sandwich solar cell with (red solid) and without
(black dash) electrolyte by illuminating from the front side. b) Absorptance of the
electrolyte inside the sandwich cell. ................................................................... 149
Figure 5-16. a) J-V response of CuWO4 sandwich by illuminating from front (red dash) and back
(blue solid) side under AM 1.5 1 sun illumination and in the dark (black dots). b)
Jphoto-V of CuWO4 sandwich cell illuminated from front (red dash) and back (blue
solid) side. ........................................................................................................... 150
Figure 5-17. a) J-V response, b) Jphoto-V response and c) quantum efficiency of CuWO4 sandwich
cell with front (red dash) and back (blue solid) illumination and in the dark (black
dots). Light source was 470 nm LED with 218 mW cm-2 power density. .......... 151
Figure 5-18. a) Original data and b) corrected IPCE of CuWO4 sandwich cell under front (red
circles) and back (blue triangles) illumination.................................................... 152
Figure 5-19. Absorbed photon-to-current efficiency of CuWO4 sandwich cell under front (red
circles) and back (blue triangles) illumination.................................................... 153

xv

Chapter 1:
Introduction

1

1.1 Motivation
The world’s total primary energy consumption rate in 2013 was estimated to be 18.0 TW
(i.e. 5.67 × 1020 J/yr).1 Due to the worldwide population increase, economic growth and
industrialization, the energy demand is projected to increase to 27 TW by the year 2050.2 So far,
fossil fuels have been the major primary energy source, and are expected to remain the main source
in the next several decades.1,2 The combustion of fossil fuels, however, releases excessive amount
of carbon dioxide and has caused global warming, which is now widely accepted.3 Besides, the
release of other harmful byproducts such as various forms of sulfur oxide and nitrogen oxide, toxic
mineral particles, etc., has caused environmental issues especially in developing countries.4 In
addition to these environmental concerns, fossil fuels are not always distributed where they are
most needed, therefore causing geopolitical conflicts and economic issues. All these factors
motivate development and utilization of carbon-neutral and easily distributable energy sources as
an alternative to fossil fuels.
1.2 Approach
Sunlight strikes the earth with a power of ~120,000 TW, making it the largest energy source
available. Utilizing only ~0.01% of solar energy would have the potential to meet the world’s
energy demand.2 Being a carbon-neutral and clean energy source, it is also well-distributed over
the most populous areas, which minimizes the transportation concerns of this energy source. These
all make solar energy the most promising means of energy supply in the future.
One possible way to utilize solar energy is to harvest photons and convert them to
electricity using photovoltaic (PV) systems, such as Si-based solar cells which have been
investigated for decades. The purity requirements of these PV cells result in their relatively high
cost, averagely $0.35 / kW-hr for typical Si solar cells, comparing with traditional electricity

2

derived from combustion of fossil fuels ($0.02-0.05 / kW-hr).2 Another issue associated with PV
cells is the storage of electricity. Current development in batteries is not sufficient to store
electricity cost-effectively on a large scale, and scientists and engineers are working on improving
the energy capacity and solving some stability and safety issues associated with battery
technologies.
An alternative to PV cells is to store solar energy in chemical bonds (i.e. solar fuels), which
is analogue to the photosynthesis process found in nature. Two of the most studied approaches are
photoelectrochemical (PEC) hydrogen production and carbon dioxide reduction.2,5–7 While the
conversion of CO2 to CO or organic fuels provides an attractive way of solar energy utilization,
capture of CO2 from the atmosphere remains challenging. Besides, the combustion of these carboncontaining fuels generates CO2, which raises the problem of capturing CO2 efficiently.
Alternatively, utilizing the solar energy to split water to get hydrogen appears an attractive and
promising approach. Hydrogen can be used as a clean secondary energy source in fuel cells and
other systems. The only combustion product of hydrogen is water, which is clean and carbonneutral, and hydrogen has a high mass energy density.8 Of course, water splitting can be
accomplished by coupling a PV cell with an external electrolyzer,9,10 but the lack of long-term
durability, toxicity and high cost of most high-performance excellent PV materials remain a
problem. Instead, using earth abundant and robust semiconductors to harness solar energy and split
water in situ would be more realistic.
The overall PEC water splitting reaction requires a Gibbs free energy change of +237
kJ/mol H2 at standard temperature (298 K) and unit activities, which corresponds to 1.229 V
potential difference between the reduction and oxidation half reactions. The overall reaction is
described by the following equation:

3

h
2H2O 
 2H2 + O2

G=237 kJ/mol

(1.1)

For an alkaline environment, the reduction and oxidation half reactions can be written as:
4H 2O + 4e4OH- + 4h +

2H 2 + 4OH-

E 0red = +0.000 V vs. RHE

(1.2)

0
E ox
= -1.229 V vs. RHE

(1.3)

2H 2O + O2

If a single material is to be used for photoelectrochemical water splitting, it has to meet the
following requirements:
1) The conduction band (CB) edge and valence band (VB) edge should straddle the
reduction and oxidation potentials of water to drive each half reaction effectively;
2) The bandgap should be reasonably small to harvest a sufficient portion of solar spectrum;
3) The semiconductor has to be composed of earth-abundant elements to scale up
production;
4) The photoelectrode should be stable under harsh PEC water splitting conditions.
In reality, in order to provide fast kinetics to drive the two half reactions efficiently, an
overpotential of 600 - 900 mV will likely be required in addition to the Nernstian potential
difference of 1.23 V. Therefore, a bandgap energy of 1.8 - 2.2 eV would be ideal for a
semiconductor to be used in a single photoelectrode configuration.11,12 Unfortunately, mainly due
to the restriction in bandgap energy and band edge positions, as well as the instability of many IIIV semiconductors in aqueous solutions under PEC conditions, no single material can fulfill all
these requirements for efficient water splitting.13
An attractive alternative is to couple a n-type photoanode with a p-type photocathode to
form a tandem cell and separate the two half reactions,13 as shown in Figure 1-1. In this
configuration, one of the photoelectrode would absorb short wavelength light while the other one
would absorb longer wavelengths. Apart from efficient utilization of solar spectrum, this
4

configuration allows us to investigate and optimize the two half reactions separately, and the usage
of an ion-selective membrane would allow for the separation of H2 and O2 products to avoid
potential explosion.

Figure 1-1. Diagram of PEC water splitting tandem cell.

According to equations 1.2 and 1.3, the generation of 1 mol of H2 involves the participation
of 2 mol of electrons (e-), while the evolution of 1 mol of O2 involves 4 mol of holes (h+). The
four-h+ process of water oxidation makes it a more challenging reaction than water reduction,
suffering from sluggish kinetics and unfavorable recombination at photoanode/electrolyte
interface.14–18 In addition to this, water oxidation is a universal oxidation half reaction for the
production of many other solar fuels, e.g. CO2 reduction. Therefore, the investigation of this half
reaction would generally benefit the development of solar fuels. Thus, our research interest lies in
searching for potentially promising photoanodes for water oxidation.

5

1.3 Performance Evaluation
1.3.1 Current Density - Voltage (J-V) Measurements
Current density - voltage (J-V) measurements are the most straightforward and important
technique to evaluate the performance of photoelectrodes. The J-V measurements are usually
measured in a three-electrode configuration as shown in Figure 1-2 in order to individually
characterize and optimize single photoelectrodes (e.g. a photoanode) as polarization losses at the
counter electrode can be neglected. The onset potential of photocurrent is related to the extra
energy needed to initiate the corresponding reaction, and the photocurrent evaluates the rate of the
reaction if the Faradaic efficiency of water oxidation reaction is 100%. Therefore, for water
oxidation half reaction, low onset potentials and high photocurrents are desired.

Figure 1-2. A three-electrode configuration for measuring the performance of a photoelectrode.
Reference, working, and counter electrodes are labeled as “RE”, “WE”, and “CE”, respectively.
The green “V” and red “I” indicate where the voltage and current are measured, respectively.

6

In a tandem PEC water splitting cell, assuming 100% Faradaic efficiency for each half
reaction, if the J-V curves of the photocathode and photoanode have an intercept as shown in
Figure 1-3, the operating photocurrent, Jop, would be the current at this matching point. If no
external bias is needed to drive water splitting at this operating condition, the solar-to-hydrogen
efficiency, hSTH , can be calculated using this equation:11

 STH 

J op  1.23 V
Pin

(1.4)

Here, J op is the operating current density of the cell, and Pin is the incident light power density
(100 mW cm-2 for 1 sun). According to calculations by Lewis and coworkers, a tandem PEC water
splitting system exhibits a maximum hSTH of 29.7% with optimized band gap combination of 1.60
eV and 0.95 eV.19

Figure 1-3. Simulated J-V curves of a tandem PEC water splitting cell. The intercept of the J-V
curves of the photocathode and photoanode determines the operating photocurrent.

7

1.3.2 Quantum Efficiencies
Apart from measuring the water oxidation J-V curves under simulated 1 sun illumination,
the quantum efficiency under monochromatic illumination can be measured to assess the electrode
performance. The incident photon-to-current efficiency, IPCE, also called the external quantum
efficiency, is calculated by dividing the measured photocurrent by the number of incident photons
at a certain wavelength using the following equation:
IPCE 

J ph  hc

 Pin  q

(1.5)

Where Jph is the measured photocurrent density, h is the Planck’s constant, c is the speed of light,
λ is the wavelength of incident light, Pin is the power density of the incident illumination, q is the
elementary charge. IPCE considers all the incident photons, thus not correcting for the photons
that are transmitted or reflected by the electrode. The absorbed photon-to-current efficiency,
APCE, also called the internal quantum efficiency, can be calculated by dividing IPCE by the
absorptance as in the following equation:

APCE 

IPCE
1 T  R

(1.6)

Where T and R are the transmittance and reflectance of the photoanode. APCE corrects for the
photons that are transmitted and reflected by the electrode. Therefore, it is a more useful parameter
to characterize and optimize a photoelectrode since APCE values lower than unity indicate
recombination losses in the bulk or at the surface of the material.
1.4 Copper Tungstate (CuWO4)
For an n-type semiconductor to be used as the photoanode material for water oxidation, it
should meet several requirements:
1) It should possess a reasonably small bandgap to absorb a substantial fraction of sunlight;
8

2) The valence band edge potential should be positive enough to provide sufficient
overpotential to drive water oxidation reaction;
3) The photoanode should facilitate efficient separation of photogenerated holes and
electrons in the bulk of the material, and should oxidize water to oxygen at the
photoanode/electrolyte interface selectively and efficiently;
4) The material has to be composed of earth-abundant elements for large scale production;
5) It should have long term stability under PEC water oxidation conditions.
Among various semiconductor materials, transitional metal oxides are on top of the
candidate list because of their generally excellent stability. The band energy levels of promising
water oxidation photoanodes20–25 are depicted in Figure 1-4. As the first photoelectrode used for
water splitting, TiO2 has a large bandgap of 3.2 eV.20 Therefore, unfortunately, it is only able to
absorb the UV portion of solar spectrum. WO3 has been investigated as a promising water splitting
photoanode due to its earth-abundant composition and stability in acidic solutions.26–29 However,
it also suffers from a relatively large bandgap of 2.7 eV, so it only absorbs up to 460 nm. Besides,
it is unstable in neutral and basic electrolytes, and it forms peroxide species during water oxidation
which leads to photocorrosion. Even in acidic solutions, it tends to oxidize anions which leads to
less-than-unity Faradaic efficiency, making it difficult to study its water oxidation
performance.21,30,31 α-Fe2O3, known as “rust”, has attracted extensive interest due to its small
bandgap of 2.1 eV, its earth abundance, and stability in alkaline environment.32,33 In fact, α-Fe2O3
has been investigated extensively and systematically to understand its limiting processes in water
oxidation, and various strategies have been designed to improve its performance.25,33–41 The
lessons and investigation methods learnt from α-Fe2O3 can be applied to other systems to discover
new promising materials for water oxidation and to better understand the general rules of tuning

9

the optoelectronic and catalytic properties of them. While there is a limited library of binary metal
oxides as candidates of water oxidation photoanode, ternary and quaternary oxides have attracted
emerging interest recently due to their diversity and tunability.23,42,43

Figure 1-4. Band edge energy levels and bandgaps of candidates for PEC water oxidation
photoanodes.

CuWO4 has a distorted wolframite structure, which consists of corner-linked CuO6 and
WO6 octahedra.18 It has an indirect bandgap of ~2.3 eV,19 which allows it to absorb photons up to
539 nm. The reduction of the bandgap compared with WO3 is due to the valence band edge of
CuWO4 being more negative than that of WO3, while the conduction band edge stays
comparable.17 This is favorable because the holes in the valence band of WO3 have more than
sufficient energy for water oxidation. Integration of AM 1.5 solar spectrum with this bandgap gives
a maximum photocurrent density of ~9 mA cm-2 (Fig 1-5). Assuming that we could pair CuWO4
with a suitable photocathode to operate the cell at the maximum photocurrent density of CuWO4
10

and no additional overpotential is required, the maximum theoretical solar-to-hydrogen efficiency
is ~11% based on calculation via equation 1.1. Besides, as mentioned above, CuWO4 shows great
stability over long periods of illumination in neutral and slightly basic electrolytes, 16 and is made
of earth abundant elements.

Figure 1-5. Red curve: AM 1.5 solar spectrum; black vertical line: wavelength that corresponds
to the bandgap of CuWO4; yellow area: maximum photon flux that can be absorbed by CuWO4
under 1 sun illumination.

Despite all these favorable characteristics, the overall water oxidation efficiency with
CuWO4 has been low. The best photocurrent density achieved with CuWO4 to date is only 0.3 mA
cm-2.44 In addition, the photocurrent onset potential is several hundred millivolts positive of the
flat band potential.45–47 The solar energy conversion efficiency of water oxidation with CuWO4 is
controlled by three factors: the light harvesting efficiency (  LH ), charge separation efficiency

11

( CS ), and hole collection efficiency (  HC ). Under the operating conditions of PEC water
oxidation with CuWO4 (Fig 1-6), light is absorbed by the photoanode, generating holes in the
valence band (VB) and electrons in the conduction band (CB). The fraction of photons absorbed
by the semiconductor determines the light harvesting efficiency,  LH . This term is determined by
the absorption coefficient (α) of CuWO4 and the thickness of the electrode. Due to the weak
absorption of this indirect semiconductor, a thickness of several hundred nm to several µm is
needed to absorb 95% of incident light, depending on the wavelength and synthetic methods. After
charge generation, electrons transport to the back contact, while holes go through diffusion in the
quasi-neutral region in the bulk of the material and drift in the space charge region to reach the
surface. Meanwhile, the charge carriers can undergo unfavorable recombination in the bulk. The
portion of charge carriers that reach the interface is defined as CS . For n-type semiconductors,
since electrons are in large excess, usually the transport of holes limit their bulk properties. After
the holes reach the CuWO4/electrolyte interface, they either get collected by the electrolyte species
to generate oxygen, or recombine with electrons from CB. The fraction of holes at the surface that
contribute to water oxidation is described by  HC .

12

Figure 1-6. Schematic diagram of PEC water oxidation at CuWO4 photoanode/electrolyte
interface. The red arrow indicates light absorption (charge generation), green arrows represent
charge separation, including diffusion in the quasi-neutral region and drift due to the electric field
in the depletion region, and blue arrows indicate hole collection at CuWO4/electrolyte interface to
oxidize water. Meanwhile, the dark red arrows represent possible recombination pathways in the
bulk or at the interface.

The poor PEC performance of CuWO4 has been attributed to both limited charge separation
and hole collection efficiencies.45,47–49 Bohra et al50 and Gaillard et al45 compared the PEC
performance by front (electrolyte) and back (substrate) illumination, and concluded that the
electron transport is limiting the bulk property of CuWO4. To overcome the opposing properties
of a long light absorption depth48,51 and poor charge separation efficiency, strategies such as
nanostructuring52,53 and doping45 have been pursued. However, there has only been limited
improvement through these modifications, mainly due to a lack of knowledge of its fundamental
properties such as charge carrier collection length. Prior reports have also suggested that the hole
13

collection efficiency is low, due to slow water oxidation kinetics and the presence of a mid-gap
surface state which mediates the oxygen evolution reaction (OER) on CuWO4.45,47,49,54 However,
to date only one catalyst (MnPO)47 has moderately improved the performance of CuWO4, while
other catalysts either make no change in its performance or make it worse. A limited number of
previous work on the investigations of CuWO4 indicates a lack of knowledge to direct scientists
and engineers to effectively optimize its performance. Therefore, this motived us to conduct
fundamental studies with this material to better understand its limits and properties.
1.5 Objectives
It is the purpose of this dissertation to deeply understand the limitations of CuWO4 in terms
of charge separation in the bulk of the material and charge transfer (hole collection) at
semiconductor/electrolyte interface. Specifically, we mainly utilized atomic layer deposition
(ALD) to deposit CuWO4 electrodes48 to conduct these fundamental studies. The self-limiting
behavior of ALD allows for deposition of conformal flat films with precise thickness control.55
This eliminates the complexity that nanostructured or porous films can cause in separating bulk
and surface properties. A new ALD-based stack deposition-annealing method was used to deposit
CuWO4 as discussed in Chapter 2. In Chapter 3, we introduce using a suitable hole scavenger to
quantify the hole collection efficiency which would benefit understanding the underlying reasons
for the effect of water oxidation catalysts on CuWO4 electrode.46 In Chapter 4, complementary
electrochemical and photoelectrochemical measurements were used to elucidate the nature of
surface states of CuWO4 which would enlighten understanding of the general role of surface states
on other semiconductor photoanodes. In Chapter 5, spray pyrolysis was used to deposit thick
CuWO4 electrodes in order to determine the limiting charge carriers, which are the electrons. The
future promise of CuWO4 as water splitting photoanode was discussed in Chapter 6.

14

REFERENCES

15

REFERENCES

(1)

US. Energy Information Administration: eia.gov.

(2)

Lewis, N. S.; Nocera, D. G. Powering the Planet: Chemical Challenges in Solar Energy
Utilization. Proc. Natl. Acad. Sci. 2007, 104, 15729–15835.

(3)

Kerr, R. A. How Urgent Is Climate Change? Science 2007, 318, 1230–1231.

(4)

United States Environmental Protection Agency: www.epa.gov.

(5)

Zhou, X.; Liu, R.; Sun, K.; Chen, Y.; Verlage, E.; Francis, S. A.; Lewis, N. S.; Xiang, C.
Solar-Driven Reduction of 1 Atm of CO2 to Formate at 10% Energy-Conversion Efficiency
by Use of a TiO2 -Protected III–V Tandem Photoanode in Conjunction with a Bipolar
Membrane and a Pd/C Cathode. ACS Energy Lett. 2016, 764–770.

(6)

Sun, K.; Liu, R.; Chen, Y.; Verlage, E.; Lewis, N. S.; Xiang, C. Solar-Driven Water
Splitting: A Stabilized, Intrinsically Safe, 10% Efficient, Solar-Driven Water-Splitting Cell
Incorporating Earth-Abundant Electrocatalysts with Steady-State pH Gradients and Product
Separation Enabled by a Bipolar Membrane. Adv. Energy Mater. 2016, 6 (13), 1–7.

(7)

Shaner, M. R.; McDowell, M. T.; Pien, A.; Atwater, H. A.; Lewis, N. S. Si/TiO2 TandemJunction Microwire Arrays for Unassisted Solar-Driven Water Splitting. J. Electrochem.
Soc. 2016, 163 (5), H261–H264.

(8)

Thomas, G. Overview of Storage Development DOE Hydrogen Program eere.energy.gov.

(9)

Jingshan, L.; Jeong-Hyeok, I.; Mayer, M. T.; Schreier, M.; Nazeeruddfn, M. K.; Nam-Gyu,
P.; Tilley, S. D.; Hong Jin, F.; Gratzel, M. Water Photolysis at 12.3% Efficiency via
Perovskite Photovoltaics and Earth-Abundant Catalysts. Science 2014, 345, 1593–1596.

(10)

Khaselev, O.; Turner, J. A. A Monolithic Photovoltaic-Photoelectrochemical Device for
Hydrogen Production via Water Splitting. Science 1998, 280 (1998), 425–427.

(11) Walter, M. G.; Warren, E. L.; McKone, J. R.; Boettcher, S. W.; Mi, Q.; Santori, E. A.; Lewis,
N. S. Solar Water Splitting Cells. Chem. Rev. 2010, 110, 6446–6473.
(12)

Bolton, J. R.; Strickler, S. J.; Connolly, J. S. Limiting and Realizable Efficiencies of Solar
Photolysis of Water. Nature 1985, 316, 495–500.

(13)

Hamann, T. W. Splitting Water with Rust: Hematite Photoelectrochemistry. Tetrahedron
2011, 67 (45), 8639–8647.

(14)

Dare-Edwards, M. P.; Goodenough, J. B.; Hamnett, A.; Trevellick, P. R. Electrochemistry
16

and Photoelectrochemistry of iron(III) Oxide. J. Chem. Soc. Faraday Trans. 1983, 79 (9),
2027.
(15)

Dotan, H.; Sivula, K.; Gr, M.; Warren, S. C. Probing the Photoelectrochemical Properties
of Hematite (α-Fe2O3) Electrodes Using Hydrogen Peroxide as a Hole Scavenger. 2011, 4,
958–964.

(16) Cowan, A. J.; Barnett, C. J.; Pendlebury, S. R.; Barroso, M.; Sivula, K.; Grätzel, M.; Durrant,
J. R.; Klug, D. R. Activation Energies for the Rate-Limiting Step in Water Photooxidation
by Nanostructured α-Fe2O3 Nd TiO2. J. Am. Chem. Soc. 2011, 133 (26), 10134–10140.
(17)

Itoh, K. Thin Film Photoelectrochemistry: Iron Oxide. J. Electrochem. Soc. 1984, 131 (6),
1266–1271.

(18)

Peter, L. M.; Upul, K. G.; Tahir, A. A. Kinetics of Light-Driven Oxygen Evolution at αFe2O3 Electrodes. Faraday Discuss. 2012, 155, 309–322.

(19)

Hu, S.; Xiang, C.; Haussener, S.; Berger, D.; Lewis, N. S. An Analysis of the Optimal Band
Gaps of Light Absorbers in Integrated Tandem Photoelectrochemical Water-Splitting
Systems. Energy Environ. Sci. 2013, 6, 2984–2993.

(20)

Fujishima, A.; Honda, K. Electrochemical Photolysis of Water at a Semiconductor
Electrode. Nature 1972, 238, 37–38.

(21)

Mi, Q.; Zhanaidarova, A.; Brunschwig, B. S.; Gray, H. B.; Lewis, N. S. A Quantitative
Assessment of the Competition between Water and Anion Oxidation at WO3 Photoanodes
in Acidic Aqueous Electrolytes. Energy Environ. Sci. 2012, 5, 5694.

(22)

Seabold, J. a.; Choi, K. S. Efficient and Stable Photo-Oxidation of Water by a Bismuth
Vanadate Photoanode Coupled with an Iron Oxyhydroxide Oxygen Evolution Catalyst. J.
Am. Chem. Soc. 2012, 134, 2186–2192.

(23)

Park, Y.; McDonald, K. J.; Choi, K.-S. Progress in Bismuth Vanadate Photoanodes for Use
in Solar Water Oxidation. Chem. Soc. Rev. 2013, 42, 2321–2337.

(24)

Lhermitte, C. R.; Bartlett, B. M. Advancing the Chemistry of CuWO4 for
Photoelectrochemical Water Oxidation. Acc. Chem. Res. 2016, 49, 1121–1129.

(25)

Riha, S. C.; Klahr, B. M.; Tyo, E. C.; Seifert, S.; Vajda, S.; Pellin, M. J.; Hamann, T. W.;
Martinson, A. B. F. Atomic Layer Deposition of a Submonolayer Catalyst for the Enhanced
Photoelectrochemical Performance of Water Oxidation with Hematite. ACS Nano 2013, 7
(3), 2396–2405.

(26)

Butler, M. A.; Nasby, R. D.; Quinn, R. K. Tungsten Trioxide as an Electrode for
Photoelectrolysis of Water. Solid State Commun. 1976, 19 (10), 1011–1014.

17

(27)

Liu, R.; Lin, Y.; Chou, L. Y.; Sheehan, S. W.; He, W.; Zhang, F.; Hou, H. J. M.; Wang, D.
Water Splitting by Tungsten Oxide Prepared by Atomic Layer Deposition and Decorated
with an Oxygen-Evolving Catalyst. Angew. Chemie Int. Ed. 2011, 50, 499–502.

(28)

Rodríguez-Pérez, M.; Chacón, C.; Palacios-González, E.; Rodríguez-Gattorno, G.; Oskam,
G. Photoelectrochemical Water Oxidation at Electrophoretically Deposited WO3 Films as a
Function of Crystal Structure and Morphology. Electrochim. Acta 2014, 140, 320–331.

(29)

Weinhardt, L.; Blum, M.; Bär, M.; Heske, C.; Cole, B.; Marsen, B.; Miller, E. L. Electronic
Surface Level Positions of WO3 Thin Films for Photoelectrochemical Hydrogen Production.
J. Phys. Chem. C 2008, 112 (8), 3078–3082.

(30)

Klepser, B. M.; Bartlett, B. M. Anchoring a Molecular Iron Catalyst to Solar-Responsive
WO3 Improves the Rate and Selectivity of Photoelectrochemical Water Oxidation. J. Am.
Chem. Soc. 2014, 136, 1694–1697.

(31)

Hill, J.; Choi, K. Effect of Electrolytes on the Selectivity and Stability of N-Type WO3
Photoelectrodes for Use in Solar Water Oxidation Cited by Me. J. Phys. Chem. C 2012, 116,
7612–7620.

(32)

Klahr, B.; Gimenez, S.; Fabregat-santiago, F.; Hamann, T.; Bisquert, J. Water Oxidation at
Hematite Photoelectrodes: The Role of Surface States. 2012, 134, 4294-4302.

(33)

Klahr, B.; Gimenez, S.; Fabregat-Santiago, F.; Bisquert, J.; Hamann, T. W. Electrochemical
and Photoelectrochemical Investigation of Water Oxidation with Hematite Electrodes.
Energy Environ. Sci. 2012, 5, 7626–7636.

(34)

Klahr, B. M.; Hamann, T. W.; Klahr, B. M.; Hamann, T. W. Voltage Dependent
Photocurrent of Thin Film Hematite Electrodes Voltage Dependent Photocurrent of Thin
Film Hematite Electrodes. 2011, 99, 63508.

(35)

Klahr, B. M.; Hamann, T. W. Current and Voltage Limiting Processes in Thin Film
Hematite Electrodes. 2011, 115, 8393–8399.

(36)

Klahr, B.; Gimenez, S.; Fabregat-santiago, F.; Hamann, T.; Bisquert, J. Water Oxidation at
Hematite Photoelectrodes: The Role of Surface States. J. Am. Chem. Soc. 2012, 134, 4294–
4302.

(37)

Klahr, B.; Gimenez, S.; Fabregat-Santiago, F.; Bisquert, J.; Hamann, T. W.
Photoelectrochemical and Impedance Spectroscopic Investigation of Water Oxidation With
“Co-Pi”-Coated Hematite Electrodes. J. Am. Chem. Soc. 2012, 134 (40), 16693–16700.

(38)

Zandi, O.; Beardslee, J. A; Hamann, T. W. Substrate Dependent Water Splitting with
Ultrathin α-Fe2O3 Electrodes. 2014, 118, 16494-16503.

(39)

Zandi, O.; Hamann, T. W. Enhanced Water Splitting Efficiency Through Selective Surface
18

State Removal. J. Phys. Chem. Lett. 2014, 5 (9), 1522–1526.
(40)

Zandi, O.; Klahr, B. M.; Hamann, T. W. Highly Photoactive Ti-Doped α-Fe2O3 Thin Film
Electrodes: Resurrection of the Dead Layer. Energy Environ. Sci. 2013, 6 (2), 634–642.

(41)

Zandi, O.; Hamann, T. W. Determination of Photoelectrochemical Water Oxidation
Intermediates on Haematite Electrode Surfaces Using Operando Infrared Spectroscopy. Nat.
Chem. 2016, 8, 778–783.

(42)

Benko, F. A.; MacLaurin, C. L.; Koffyberg, F. P. CuWO4 and Cu3WO6 as Anodes for the
Photoelectrolysis of Water. Mat. Res. Bull. 1982, 17, 133–136.
Guo, W.; Chemelewski, W. D.; Mabayoje, O.; Xiao, P.; Zhang, Y.; Mullins, C. B. Synthesis
and Characterization of CuV2O6 and Cu2V2O7 : Two Photoanode Candidates for
Photoelectrochemical Water Oxidation. J. Phys. Chem. C 2015, 119, 27220–27227.

(43)

(44)

Zhang, H.; Yilmaz, P.; Ansari, J. O.; Khan, F. F.; Binions, R.; Krause, S.; Dunn, S.
Incorporation of Ag Nanowires in CuWO4 for Improved Visible Light-Induced Photoanode
Performance. J. Mater. Chem. A 2015, 3, 9638–9644.

(45)

Gaillard, N.; Chang, Y.; Deangelis, A.; Higgins, S.; Braun, A. A Nanocomposite
Photoelectrode Made of 2.2 eV Band Gap Copper Tungstate (CuWO4) and Multi-Wall
Carbon Nanotubes for Solar-Assisted Water Splitting. Int. J. Hydrogen Energy 2013, 38,
3166–3176.

(46)

Gao, Y.; Hamann, T. W. Quantitative Hole Collection for Photoelectrochemical Water
Oxidation with CuWO4. Chem. Commun. 2017, 53, 1285–1288.

(47)

Pyper, K. J.; Yourey, J. E.; Bartlett, B. M. Reactivity of CuWO4 in Photoelectrochemical
Water Oxidation Is Dictated by a Midgap Electronic State. J. Phys. Chem. C 2013, 117,
24726–24732.

(48)

Gao, Y.; Zandi, O.; Hamann, T. W. Atomic Layer Stack Deposition-Annealing Synthesis
of CuWO4. J. Mater. Chem. A 2016, 4, 2826–2830.

(49)

Hill, J. C.; Choi, K. Synthesis and Characterization of High Surface Area CuWO 4 and
Bi2WO6 Electrodes for Use as Photoanodes for Solar Water Oxidation. J. Mater. Chem. A
2013, 1, 5006–5014.

(50)

Bohra, D.; Smith, W. A. Improved Charge Separation via Fe-Doping of Copper Tungstate
Photoanodes. Phys. Chem. Chem. Phys. 2015, 17, 9857–9866.

(51)

Yourey, J. E.; Bartlett, B. M. Electrochemical Deposition and Photoelectrochemistry of
CuWO4, a Promising Photoanode for Water Oxidation. J. Mater. Chem. 2011, 21, 7651–
7660.

(52)

Ye, W.; Chen, F.; Zhao, F.; Han, N.; Li, Y. CuWO4 Nanoflake Array-Based Single-Junction
19

and Heterojunction Photoanodes for Photoelectrochemical Water Oxidation. ACS Appl.
Mater. Interfaces 2016, 8, 9211–9217.
(53)

Hu, D.; Diao, P.; Xu, D.; Xia, M.; Gu, Y.; Wu, Q.; Li, C.; Yang, S. Copper (II) Tungstate
Nanoflake Array Films: Sacrificial Template Synthesis, Hydrogen Treatment, and Their
Application as Photoanodes in Solar Water Splitting. Nanoscale 2016, 8, 5892–5901.

(54)

Gaillard, N.; Changa, Y.; DeAngelis, A.; Higgins, S.; Braun, A. A Nanocomposite
Photoelectrode Made of 2.2 eV Band Gap Copper Tungstate (CuWO4) and Multi-Wall
Carbon Nanotubes for Solar-Assisted Water Splitting Nicolas. Int. J. Hydrogen Energy
2013, 38, 3166–3176.

(55)

George, S. M. Atomic Layer Deposition: An Overview. 2009, 110, 111–131.

20

Chapter 2:
Atomic Layer Stack DepositionAnnealing Synthesis of CuWO4

Adapted with permission from:
Atomic Layer Stack Deposition-Annealing Synthesis of CuWO4, Yuan Gao, Omid Zandi and
Thomas W. Hamann, J. Mater. Chem. A, 2016, 4, 2826–2830. Copyright 2016 The Royal Society
of Chemistry.
21

2.1 Abstract
A stack deposition-annealing (SDA) approach was demonstrated for the synthesis of
CuWO4 via atomic layer deposition (ALD). Certain amount of CuO was deposited on WO3,
followed by annealing to produce CuWO4. The resulted CuWO4 shows a band gap of 2.3 eV. The
resulted CuWO4 shows an onset potential of 0.8 V vs RHE and a photocurrent density of 0.11 mA
cm-2 at 1.23 V vs RHE in pH 9 KBi buffer, which is promising activity as a photoanode for
photoelectrochemical (PEC) water splitting. The resulted photoelectrode shows great stability over
4 hours. This methodology can in principal, be utilized for the ALD of various ternary and
quaternary oxides for different applications.

22

2.2 Introduction
Atomic layer deposition (ALD) is an ideal material synthesis technique to fabricate planar
and nanostructured materials due to its self-limiting vapor-phase surface chemistry.1–3 These
unique properties allow ALD to offer angstrom-level thickness and dimensional control of
materials on high aspect ratio substrates with excellent reproducibility. Thus, ALD has become
widely utilized for the fabrication of microelectronic devices and materials for energy
applications.2,4–6 In typical ALD process of a binary oxide, the metal precursor and the oxidant are
alternately introduced into the reaction chamber to react with the substrate in a self-limiting fashion.
The simplified schematic illustration of the ALD instrument used in this study and ALD process
is shown in Figure 2-1. While the ALD of binary oxides is relatively straightforward, the ALD of
ternary oxides can be challenging. This is a major limitation since ternary oxides are widely used
as electrochromic and transparent conductive materials,7 electrocatalysts,8 and photocatalytic
materials.9

a)

b)

Figure 2-1. Schematic illustration of a) ALD reactor and b) one cycle of ALD process.

23

One method to prepare a ternary oxide via ALD is to use bimetallic single-source
precursors.10 Limitations of this method are lack of suitable bimetallic precursors and the fixed
stoichiometry defined by the molecular precursor, leaving little room for modification. Another
method is to use metal alkoxides as the oxygen source instead of H2O, O2 or O3, so that it provides
two of the three elements of the ternary oxide.11 This method also suffers from a lack of volatile
and reactive precursors. The most straightforward and common way to deposit ternary metal
oxides is alternating separate binary oxide deposition cycles, where the desired stoichiometry is
obtained by adjusting the pulsing ratios of two binary oxides.12,13 This approach requires the two
binary oxides to deposit at the same temperature, so an overlap of their ALD windows is necessary.
Unfortunately, the ALD windows of many binary oxides have small to no overlap. Another
limitation is that since ALD is a surface-sensitive process, the growth rate may vary on different
surfaces, thus it is difficult to control the stoichiometry based on the growth rates of binary
oxides.14,15 Further, some materials have surface-inhibited growth on certain substrates due to the
seeding layer in the initial deposition stage,15,16 which prevents the realization of ternary metal
oxides. As a consequence, there are no examples of the ALD of many promising ternary oxide
materials, such as copper tungstate (CuWO4), reported. In this work, we introduce a new
methodology for the ALD of ternary metal oxide thin films via a stack deposition-annealing (SDA)
approach and report the first ALD of CuWO4 to demonstrate the promise of this approach. This
approach of stacking layers followed by annealing has been widely used in material synthesis in
fields besides ALD,17,18 including a solution-solid phase reaction to make CuWO4,19 while ALD
provides a better control over the film thickness and composition. We note that a similar method
has been applied for synthesis of Cu2ZnSnS4 by stacking binary sulfides via ALD followed by
annealing.20 Ginestra et al. also utilized the same method, but their purpose was synthesis of

24

Y2O3/ZrO2 alloy.21 Yang et al. showed Mn-doping by depositing a thin layer of MnOx on TiO2
nanowires followed by annealing to allow for the solid state diffusion of Mn into TiO2.22 The
results reported herein, however, represent the first example of ALD of a pure phase ternary oxide
material deposited by the SDA approach.
2.3 Experimental
2.3.1 Film Deposition
Atomic layer deposition was performed using a Savannah 100 reactor (Cambridge
Nanotech Inc.) at a temperature of 250 °C. Fluorine-doped tin oxide (FTO) coated glass (Hartford
Glass, TEC 15, 12Ω cm-2), Si wafer (University Wafer) and quartz (Technical Glass Products, 1/16
inch thick) were used as the substrates. The substrates were sonicated in soap water, distilled water
and isopropanol, and blown dry with nitrogen prior to deposition. The precursors for the deposition
of WO3 and CuO were bis(tert-butylimido)bis(dimethylamido) tungsten (VI) ((tBuN)2(Me2N)2W)
(Strem Chemicals Inc., >97%) and copper(I)-N,N’-di-sec-butylacetamidinate ([Cu(sBu-amd)]2)
(Dow Chemical Co., >99.0%), respectively. The nitrogen carrier gas was kept at 5 sccm. WO3 was
deposited using a modified version of a reported procedure.23 The precursor was heated up to 75 °C
but the vapor pressure was too small to be detected by our pressure gauge. In each ALD cycle, the
tungsten precursor was pulsed for 2 s, followed by 10 s under exposure mode and 6 s nitrogen
purge. De-ionized water (millipore, 18 MΩ) was used as the oxidant and was pulsed for 0.5 s
followed by 15 s of exposure mode and 6 s of purge. 2000 cycles of WO3 (~160 nm) was deposited
for characterization. The CuO was deposited using a modified recipe suggested by Ultratech Inc.
(unpublished). The precursor was heated to 150 °C but the vapor pressure was not detectable by
the pressure gauge. The copper precursor was pulsed for 3 s in each ALD cycle. After purging for
6 s, the oxidation was performed. The oxidation was a 2 s ozone (~10% by weight O3 in ultrahigh

25

purity O2 produced by Yanco Industries ozone generator) pulse, followed by a 3 s purge to allow
for enough generation of ozone, and performed for 10 times. A total of ~1500 cycles of CuO (~150
nm) was deposited for the ease of characterization. These binary oxide films were sintered at
550 °C at a ramping rate of 10 °C/min and heated for 2 h. CuWO4 thin films were made by
depositing suitable cycles of CuO on top of WO3 followed by annealing at 600 °C in air for 30
min at a ramping rate of 2 °C/min.
2.3.2 Calculation of CuO Deposition Cycles
Equations used to calculate the thickness of CuO needed to produce desired 1:1
stoichiometry with WO3:

nWO3 = nCuO
rWO ´ Area ´ tWO
3

3

M WO3

=

tCuO = tWO3 ´

(2.1)

rCuO ´ Area ´ tCuO
M CuO

rWO M CuO
´
rCuO M WO
3

(2.2)

(2.3)

3

where nWO3 , tWO3 , rWO3 and M WO3 are the number of moles, thickness, density and molar mass of
WO3 (density 7.16 g/cm3, molar mass 231.84 g/mol), and nCuO , tCuO , rCuO and M CuO are the
corresponding parameters of CuO (density 6.315 g/cm3, molar mass 79.545 g/mol). Since they are
uniformly deposited on the same substrate, the area should be the same for both WO3 and CuO.
Thickness of WO3 is measured by spectroscopic ellipsometry, and thus the thickness of CuO
needed can be calculated.
2.3.3 Calculation of Film Thickness
Theoretical thickness of CuWO4 is calculated using the following equations:
mWO3  mCuO  mCuWO4

26

(2.4)

WO VWO  CuO VCuO  CuWO VCuWO
3

3

4

(2.5)

4

WO  A  tWO  CuO  A  tCuO  CuWO  A  tCuWO
3

3

4

4

(2.6)

WO  tWO  CuO  tCuO  CuWO  tCuWO

(2.7)

WO  tWO  CuO  tCuO
CuWO

(2.8)

3

tCuWO4 

3

4

3

3

4

4

rCuWO and tCuWO are the density and theoretical thickness of resulted CuWO4. The density of
4

4

CuWO4 we used herein is 7.790 g cm-3, obtained by theoretical calculations based on the crystal
structure in literature,24 which is close to experimental values.24,25
2.3.4 Film Characterization
The as-deposited and annealed films were characterized by X-ray diffraction (XRD), Xray photoelectron spectroscopy (XPS) and Raman spectroscopy. XRD was taken with Bruker
Davinci Diffractometer operating at 40 kV and 40 mA using Cu Kα radiation. XPS analysis was
made using Perkin Elmer Phi 5600 ESCA system with a magnesium Kα X-ray source at a takeoff
angle of 45°. Survey scans of 0-1100 eV binding energy and detailed scans for C 1s, O 1s, W 4f
and/or Cu 2p regions were scanned for all samples. Peaks were corrected with the C 1s peak set to
284.8 eV, and subsequent Shirley background subtraction was performed for fitting. Raman
spectra were recorded using LabRam Armis, Horiba Jobin Yvon instrument equipped with 532
nm laser and a ×50 microscope to focus the laser on the film surface. The thicknesses of binary
oxide films were obtained by ellipsometric measurements (Horiba Jobin Yvon, Smart-SE).
Thickness of the ternary film and the surface morphology of the films were examined by atomic
force microscopy (AFM) using MFP-3D from Asylum Research. Cross-section scanning electron
microscopy (SEM) (Carl Zeiss Microscopy) was also used to determine the film thicknesses, as
well as their morphology. Energy-dispersive X-ray (EDX) analysis spectra were collected using
27

Ametek-EDAX Apollo X. Accelarating voltages used were 15 kV and 20 kV, with 8 mm working
distance. Spectra were analyzed using TEAM EDS software. Cu K and W M emission lines were
analyzed for 15 kV beam voltage and Cu K and W L were analyzed for 20 kV beam voltage. The
error bar in the main text is the standard deviation of measurements on 10 different spots.
Absorbance measurements were made using a Perkin-Elmer Lambda 35 UV-vis spectrometer with
a Labsphere integrating sphere.
2.3.5 Photoelectrochemistry
Thin film electrodes were masked with a 60 μm Surlyn film (Solaronix) with a 0.28 cm2
hole to define the active area and to prevent scratching of the thin films which were clamped to a
custom-made glass electrochemical cell. Surlyn films were adhered to the electrodes by heating to
120 °C. A homemade saturated Ag/AgCl electrode was used as a reference electrode, and high
surface area platinum mesh was used as the counter electrode. All experiments shown used a pH
9 potassium borate (KBi) buffer containing 0.2 M KCl as a supporting electrolyte. The buffer was
prepared by adding an appropriate amount of KOH pellets into 0.1 M H3BO3 to adjust the pH. The
pH was determined using a Fisher Scientific Accumet pH meter. All photoelectrochemical
measurements were made with an Eco Chemie Autolab potentiostat coupled with Nova
electrochemical software. The light source was a 450 W Xe arc lamp. An AM 1.5 solar filter
(Sciencetech Inc.) was used to simulate sunlight at 100 mW cm-2 (1 sun). The J-V curves were
measured at a scan rate of 20 mV/s by shining light from the electrolyte-electrode (EE, front
illumination) interface. The monochromatic photocurrents (IPCE) were measured at 1.23 V vs
RHE using the same light source coupled with a grating monochromator with a 10 nm step.

28

2.4 Results and Discussion
WO3 was deposited using a modified literature procedure at a substrate temperature of
250 °C.23 According to the XPS data shown in Figure 2-2, the as-deposited film and annealed film
both show W 4f7/2 and W 4f5/2 peaks at 35.6 eV and 37.8 eV, consistent with W6+ peaks in
literature.26–29 The O 1s peak shows a main peak at 530.4 eV, corresponding to the metal oxide O
1s peak, with a small contribution from carbon bonded oxygen at 531.9 eV. According to the XRD
(Figure 2-3) and Raman spectra (Figure 2-4), the as-deposited films appear to be amorphous since
no peaks corresponding to WO3 are present. After annealing, however, diffraction peaks at 23.0°,
24.3° and 34.1° clearly show the presence of WO3. These peaks were assigned using the
International Union of Crystallography (IUCr) database. Raman peaks at 272, 325, 713 and 808
cm-1 also appeared with annealing, in agreement with Raman scattering from WO3.30,31 Therefore,
the as-deposited WO3 was amorphous, and after annealing in air at 550 °C for 2h, it turned
crystalline. As shown in Figure 2-5, ALD of WO3 reaches saturation when the metal precursor
pulse length is 2 s with a growth rate of 0.86 ± 0.03 Å/cycle. The growth rate is the average growth
rate of four batches, including two batches pulsing 2 s and two batches pulsing 3 s. The error is
calculated as the standard deviation of the four batches described above.

29

Figure 2-2. Experimental X-ray photoelectron spectroscopy (XPS) spectra of W 4f and O 1s peaks
(black scatters) with fitting (red solid lines and dots) of as-deposited (top) and annealed (bottom)
WO3.

Figure 2-3. X-ray diffraction (XRD) patterns of as-deposited (black) and annealed (red) WO3.

30

Figure 2-4. Raman spectra of as-deposited (black) and annealed (red) WO3.

Figure 2-5. Thickness of WO3 thin films vs number of deposition cycles using different metal
precursor pulse lengths.

31

CuO was deposited at a substrate temperature of 250 °C, followed by annealing in air at
550 °C for 2 h. In the XPS spectra shown in Figure 2-6, the as-deposited film and annealed film
both show Cu2+ peaks in the Cu 2p region. Cu 2p3/2 and Cu 2p1/2 peaks are located at 933.9 and
953.8 eV, with their satellite peaks at 940.9, 943.4 (Cu 2p3/2 satellite) and 962.0 eV (Cu 2p1/2
satellite), respectively, which is the distinguishing feature of the XPS spectra of Cu2+.32,33 The O
1s peak shows a main lattice oxide peak at 529.8 eV and a small contribution from carbon bonded
oxygen at 532.1 eV. Also, there is a huge deconvoluted peak at 531.2 eV, which can be attributed
to the defective oxygen CuO.34 The XRD pattern and Raman Spectra of CuO films are shown in
Figure 2-7 and 2-8. Both as-deposited and annealed CuO thin films are crystalline. All X-ray
diffraction peaks agree with CuO diffractions according to the IUCr database, including CuO
single crystals grown by different methods such as flux method35,36 and thermal decomposition of
gerhardtite.37 Raman spectra show three peaks which are all assigned to Raman scattering of
CuO.38–40 The saturation growth rate with 1 s Cu source pulse length is 0.45 ± 0.04 Å/cycle, and 3
s pulse length was used in following experiments just to be conservative. This growth rate was
calculated using the slope of the best fit line of the thickness vs number of cycles plot, but due to
the surface-enhanced behavior41,42 of CuO on Si substrate, the intercept with the thickness axis is
not zero (Fig 2-9). Thus, this growth rate is not reflecting the real growth rate of CuO on itself or
in our case, on WO3. This experiment is only to show the self-limiting growth behavior of CuO.
In the following demonstration of ALD of CuWO4, we are using the growth rate of CuO on WO3
for calculation.

32

Figure 2-6. Experimental XPS spectra of Cu 2p and O 1s peaks (black scatters) with fitting (red
solid lines and dots) of as-deposited (top) and annealed (bottom) CuO.

Figure 2-7. XRD patterns of as-deposited (black) and annealed (red) CuO.

33

Figure 2-8. Raman spectra of as-deposited (black) and annealed (red) CuO.

Figure 2-9. Thickness vs number of deposition cycles of CuO thin films with different metal
precursor pulse length.

Numerous attempts utilizing a variety of conditions to make the ternary oxide CuWO4 by
conventional approaches were unsuccessful. In order to determine the cause, we deposited WO3
34

on CuO films and CuO on WO3 films. Figure 2-10a shows the growth rate of WO3 on as-deposited
CuO. There is no significant increase in the total film thickness until after ~100 cycles of WO3.
This suggests that WO3 has substrate-inhibited growth16 (blue line) on the CuO surface. After ~100
cycles, when the CuO surface is fully covered with WO3, WO3 grows linearly as expected (red
line). This lack of growth of WO3 on the CuO surface explains the failure of the conventional
method to deposit CuWO4 by alternating WO3 and CuO deposition cycles. Figure 2-10b shows
that CuO grows linearly on WO3 (1.5 ± 0.1 Å/cycle), however, which means the WO3 surface does
not inhibit the deposition of CuO. While in principle, it may be possible to use a ~100:1 pulsing
ratio of WO3:CuO in each super cycle to deposit one layer of CuWO4, this is impractical since it
would waste 99% of the W precursor and a lot of time, and is not reproducible. Alternatively, new
precursors could be identified, but that is a very time intensive process and not a general solution.

35

Figure 2-10. Total thickness of a) WO3 on 100 cycles of CuO and b) CuO on 100 cycles of WO3
as a function of number of deposition cycles.

Therefore, a stack deposition approach was utilized to grow CuWO4 (Fig 2-11). A
stoichiometric amount of CuO was deposited on WO3, followed by annealing. The advantages of
this approach are numerous. Instead of empirically optimizing the pulse sequence, the precise
number of cycles to produce a 1:1 stoichiometry of W:Cu for a given thickness is easily determined.
36

Since the materials are deposited sequentially, they can be grown at different temperatures and
their ALD temperature windows do not have to overlap. This method should be general and
applicable to a wide variety of ternary and potentially quaternary metal oxide materials.

Figure 2-11. Sequential ALD of WO3/CuO and post-ALD annealing to make CuWO4 thin films
via stack deposition-annealing (SDA) approach.

1000 ALD cycles of WO3 was deposited on Si, fluorine-doped tin oxide (FTO) coated glass
and quartz substrates. The thickness of the WO3 thin film was 65 nm, measured by spectroscopic
ellipsometry and cross-section scanning electron microscopic (SEM) image (Figure 2-12a). To
match the stoichiometry, the desired thickness of CuO was determined by considering the ratios
of density and molar masses of WO3 and CuO (see experimental for calculation details). Based on
this calculation, 170 cycles of CuO was deposited on top of the 65 nm WO3 film. The as-deposited
films showed a brownish color, which is essentially the color of CuO. After annealing in air at 600
o

C for 30 min, to allow for the solid-state reaction between CuO and WO3, the thin film turned

pale yellow, which agrees with the expected color of CuWO4.43,44 The thickness of the resulted

37

film was 83 nm measured by atomic force microscopy (AFM). The AFM image in Figure 2-13c
shows that the annealed CuWO4 film has an RMS roughness of 4 nm. Comparing this value with
the roughness of as-deposited WO3 and as-deposited CuO on WO3 in Figure 2-13a and Figure 213b, which is 0.4 nm and 3 nm, respectively, we concluded that while the annealing process
resulted in a slight increase in the roughness of our films, the resulted CuWO4 film is overall
smooth. Cross-section SEM image confirmed that the resulted film has a thickness of ~80 nm
(Figure 2-12b). Considering the densities of WO3, CuO and CuWO4 and the thicknesses (or
number of cycles) of WO3 and CuO deposited, we should in principle obtain 80 nm of crystalline
CuWO4 (see experimental for details).

38

Figure 2-12. Cross-section SEM images of a) as-deposited WO3 and b) annealed CuWO4 thin
films on Si substrate.

39

Figure 2-13. AFM surface morphology image of a) as-deposited WO3, b) as-deposited CuO on
WO3, and c) annealed CuWO4 on Si substrate.

Energy-dispersive X-ray (EDX) analysis spectra (see Appendix) gives W:Cu atomic ratio
of 1.00 ± 0.05 using an accelerating voltage of 20 kV. A 15 kV beam voltage was also used to
minimize the contribution from the underlying substrate, and the determined W:Cu ratio is 0.99 ±
0.02. This agreement between the experimental and theoretical thicknesses/stoichiometry indicates
that using this new SDA method, we can precisely control the thickness and composition of
CuWO4 thin films by varying the number of cycles deposited, which is essentially one benefit of
ALD.
40

The X-ray photoelectron spectroscopy (XPS) survey scans of as-deposited and annealed
films (Fig 2-14) show W 4f, Cu 2p and O 1s peaks. Each of the peaks was further examined to
determine the oxidation states of the metals (Fig 2-15). Both films showed W 4f7/2 and W 4f5/2
peaks at binding energy of 35.2 and 37.4 eV, consistent with W6+.27,28 Spectra showed the Cu 2p3/2
peak at 934.2 eV with its satellite peaks at 941.1 and 943.5 eV, and the Cu 2p1/2 peak at 954.1 eV
with its satellite at 962.1 eV, respectively, which are unique to Cu2+.32,33 Figure 2-15b and Figure
2-15c shows the O 1s peaks of the as-deposited and annealed films, respectively, and both of them
are fitted with three peaks. The peaks at 530.1 eV belong to the main lattice oxide peak. The peaks
at 532.0 and 531.2 eV of as-deposited and annealed films are assigned to the defective oxygen.34
Finally, the peaks at 533.2 and 532.2 eV in the as-deposited and annealed films are attributed to
carbon-bonded oxygen.

Figure 2-14. XPS survey spectra of as-deposited (bottom) and annealed (top) film of CuO on top
of WO3.

41

Figure 2-15. Experimental XPS spectra of a) Cu 2p and W 4f peaks (black scatters) with fitting
(red solid lines and dots) of CuO on top of WO3 as-deposited (top) and annealed (CuWO4) (bottom)
and O 1s peaks (black scatters) with fitting (red solid lines and dots) of CuO on top of WO3 b) asdeposited and c) annealed (CuWO4).
42

X-ray diffraction (XRD) patterns and Raman spectra of the as-deposited stack and annealed
films are shown in Figure 2-16. The XRD pattern of the as-deposited film does not show any
diffraction peaks except those attributed to the FTO substrate. The diffraction peaks of CuO are
not observed in XRD of the as-deposited stack because the amount of CuO at the surface is below
its detection limit. The annealed film shows diffraction peaks of crystalline CuWO4 well matched
with literature.24,45 Raman spectra of the as-deposited films show a small peak around 290 cm-1,
which is assigned to the main Raman scattering peak of CuO. The small peak at 244 cm -1 and
humps around 600, 800 and 950 cm-1 are due to the background scattering from the FTO substrate.
This agrees with our results of binary oxides WO3 and CuO, indicating that the underlying WO3
is amorphous, while CuO is crystalline after deposition. Since only a thin layer of CuO (25 nm)
was deposited, the other two peaks of CuO are not observed because of their low intensity. All the
peaks from the annealed films match with Raman scattering from the crystal structure of CuWO 4
previously reported,45,46 except those from the substrate. The CuWO4 Raman scattering peak at
550 cm-1 overlaps with the wide FTO scattering peak around 600 cm-1, but comparing this hump
with that of the as-deposited film, this peak is sharper and more intense, with the peak position
shifted toward smaller wavenumber, distinct from the hump from the substrate. Therefore, the
peak at 550 cm-1 is assigned to CuWO4. Thus, we conclude that the as-deposited film essentially
consists of a thin layer of crystalline CuO on top of amorphous WO3, which is different from metal
sulfide multilayers that can mix even at low deposition temperature (135 oC)20; after annealing,
the solid-state reaction between CuO and WO3 produces pure crystalline CuWO4.

43

Figure 2-16. a) X-ray diffraction patterns and b) Raman spectra of CuO on WO3 as-deposited
(black) and annealed (CuWO4) (red). Peak positions of CuWO4 from literature are plotted as
vertical dashed gray lines.

Absorbance and AFM thickness measurements of CuWO4 films with different thicknesses
were further conducted to show the fine control of thickness using our SDA approach. In Figure
44

2-17 shows different CuWO4 thin films deposited on FTO substrates. The thicknesses of resulted
CuWO4 films displayed in Figure 2-18 show a linear relationship with ALD cycles of WO3, with
a growth rate of 0.85 Å of CuWO4/cycle of WO3. The absorbance was calculated by correcting for
the reflectance from the substrate using previously reported method.47 Figure 2-19a shows that the
absorbance increases with increasing deposition cycles, in agreement with the trend of color shown
in Figure 2-17. Figure 2-19b and Figure 2-19c indicate a linear trend of absorbance of CuWO4 at
400 nm and 450 nm with ALD cycles, confirming the consistent growth rate of CuWO4 using our
SDA method. Considering the thicknesses of these films shown in Figure 2-18, the absorption
coefficients of CuWO4 at 400 nm and 450 nm are calculated to be 58370 cm-1 and 29058 cm-1,
respectively. These values are higher than some reported values,43,45,48 and close to those measured
with cosputtered films (~20000 cm-1 at 450 nm) but still slightly higher.49 One possible reason is
that our film is more compact and uniform compared with films synthesized via other methods.

Figure 2-17. Images of CuWO4 thin films deposited on FTO substrates with varying ALD cycles
of WO3 of 0, 500, 1000, 1500, 2000, and 2500.

45

Figure 2-18. Thickness of CuWO4 films with different deposition cycles of WO3 measured with
AFM (black squares). Error bars represent the standard deviation of 10 different measurements on
each film. Red dash shows the linear fit of thickness vs. number of cycles.

46

Figure 2-19. Top: a) Absorbance of CuWO4 thin films with ALD cycles of WO3 of 500 (red solid),
1000 (orange dash), 1500 (yellow short dash), 2000 (green dash dot), and 2500 (blue dot). Bottom:
Absorbance of CuWO4 thin films at wavelength of b) 400 nm and c) 450 nm varying with different
ALD cycles of WO3. Error bars represent standard deviation of independent measurements on
three different spots for each thickness.

UV-vis of CuWO4 electrodes was measured by illuminating from the FTO side. The
absorptance of the films was calculated using absorptance = Tsub + Rsub – Tfilm – Rfilm, where Tsub,
47

Rsub, Tfilm, and Rfilm are the measured transmittance and reflectance of the quartz substrate and
quartz coated with CuWO4 thin film, respectively. As seen in Figure 2-20a, the absorption around
500 nm is weak due to the indirect bandgap of the semiconductor.50 The absorbance was calculated
by correcting for the reflectance from the substrate using previously reported method from which
absorptivity and band gap are extracted.47 The Tauc plot of resulted CuWO4 (Fig 2-20b) shows an
indirect band gap of 2.3 eV, in agreement with the other reports.43,45,50 The absorption coefficient
shown in Figure 2-20c is slightly higher than previously reported values.43,48,49

Figure 2-20. a) IPCE (red scatters) in pH9 KBi buffer at 1.23 V vs RHE and absorptance spectrum
(black solid line), b) Tauc plot, and c) absorption coefficient of CuWO4 thin film.

48

The photoelectrochemistry of the CuWO4 films were examined under PEC water oxidation
conditions. The current density, J, vs. applied voltage, V, curve of CuWO4 under 1 sun illumination
is shown in Figure 2-21. The CuWO4 thin film electrodes exhibit excellent water oxidation activity
with the photocurrent onset potential of 0.8 V vs RHE and photocurrent of 0.11 mA cm-2 at 1.23
V vs RHE (the thermodynamic potential for water oxidation). The J-V curve under chopped light
matches well with both dark current and the curve under illumination, except that at potentials
close to the onset. These small spikes in the photocurrent transients suggest that surface
recombination occurs at these potentials.51 To confirm that the J-V curve is attributed strictly to
CuWO4, it was compared to pure CuO (35 nm) and WO3 (100 nm) films (Fig 2-22). The J-V curve
of CuO under illumination shows a cathodic wave at potentials negative of 0.8 V vs RHE, which
is due to the reduction of CuO to Cu2O, and an anodic peak at 0.9 V vs RHE, which is assigned to
the re-oxidation of Cu2O.52 The oxidation of Cu2O is a dark reaction because it is also observed in
the dark J-V curve of the CuO electrode. The dark J-V curve of WO3 shows no current within the
applied potential range, and under 1 sun illumination, there is a hysteresis in the reverse scan,
which might be attributed to the instability of the WO3 under the operating conditions. While the
initial J-V response of the WO3 appears similar to the CuWO4 electrode, the photocurrent quickly
decays. The photocurrent at 1.23 V was measured as a function of time to assess the stability of
CuWO4 and WO3 films in a pH 9 solution under 1 sun illumination (Figure 2-21b). Remarkable
stability was observed for the CuWO4 electrode over the 4 h experiment, while the current of WO3
degraded quickly in less than an hour, likely due to the photo-corrosion of WO3 in the basic
electrolyte.53,54 Incident-photon-to-current efficiency (IPCE) measurements of the CuWO4
electrode exhibited an onset of photocurrent around 500 nm, which agrees with the absorptance
measurements (Figure 2-20a). We note that the IPCE at wavelengths of 500–540 nm is very low

49

as a result of the minimal light absorption at these wavelengths due to the very thin film measured.
In any case, we confirm that the photocurrent is strictly due to CuWO4, not from either residual
CuO or WO3 impurities. The photocurrent from our CuWO4 is comparable to many results in
literature,44,55–57 but is lower than some reported by Bartlett et al.43,58 and Gaillard et al.48 The
electrodes examined herein are thinner than previously reported electrodes (80 nm vs 200 nm to
several microns), so not as much light was absorbed. In addition, previously reported films were
porous or had nanostructures, which possess a higher surface area that can facilitate overcoming
the charge separation limitations. The comparable PEC results of our CuWO4 film with previously
reported thicker films with higher surface area suggest the promise of this synthetic method.

50

Figure 2-21. a) J-V curves of an 80 nm thick CuWO4 electrode in the dark (black dash), under 1
sun illumination (blue solid) and chopped illumination (red solid). The vertical gray dash shows
the thermal dynamic potential of water oxidation. b) Stability test of CuWO4 (red) and WO3 (black)
at 1.23 V vs RHE under 1 sun illumination. Experiments were made in 0.2 M KCl in pH9 KBi
buffer.

51

Figure 2-22. J-V curve of a) CuO and b) WO3 electrode in the dark (black) and under 1 sun
illumination (red) measured in 0.2 M KCl in pH9 KBi buffer.

2.5 Conclusion
Overall, the PEC performance shows that the SDA method produces high quality
crystalline pure phase CuWO4 films with promising photocatalytic performance. This advance
enables fundamental studies of the limitations of such ternary oxide materials, such as charge
carrier diffusion length, as well as strategies to overcome them. For example, optimized
52

nanostructured CuWO4 photoanodes can in principle overcome a mismatch in light absorption and
charge carrier diffusion length. This work will in theory be enabled by the new materials synthesis
methodology described here.

53

APPENDIX

54

Figure A2-1. An example of an EDX spectrum of CuWO4 on FTO-coated glass.

55

REFERENCES

56

REFERENCES

(1)

George, S. M. Atomic Layer Deposition: An Overview. Chem. Rev. 2010, 110, 111–131.

(2)

Knez, M.; Nielsch, K.; Niinistö, L. Synthesis and Surface Engineering of Complex
Nanostructures by Atomic Layer Deposition. Adv. Mater. 2007, 19 (21), 3425–3438.

(3)

Puurunen, R. L. Surface Chemistry of Atomic Layer Deposition: A Case Study for the
Trimethylaluminum/water Process. J. Appl. Phys. 2005, 97, 121301.

(4)

Leskelä, M.; Ritala, M. Atomic Layer Deposition Chemistry: Recent Developments and
Future Challenges. Angew. Chemie - Int. Ed. 2003, 42, 5548–5554.

(5)

Wang, T.; Luo, Z.; Li, C.; Gong, J. Controllable Fabrication of Nanostructured Materials
for Photoelectrochemical Water Splitting via Atomic Layer Deposition. Chem. Soc. Rev.
2014, 43 (22), 7469–7484.

(6)

O’Neill, B.; Jackson, D. H. K.; Lee, J.; Canlas, C.; Stair, P. C.; Marshall, C. L.; Elam, J. W.;
Kuech, T. F.; Dumesic, J. A.; Huber, G. W. Catalyst Design with Atomic Layer Deposition.
ACS Catal. 2015, 1804–1825.

(7)

Granqvist, C. G. Electrochromics for Smart Windows: Oxide-Based Thin Films and
Devices. Thin Solid Films 2014, 564, 1–38.

(8)

Zhang, C.; Fagan, R. D.; Smith, R. D. L.; Moore, S. a.; Berlinguette, C. P.; Trudel, S.
Mapping the Performance of Amorphous Ternary Metal Oxide Water Oxidation Catalysts
Containing Aluminium. J. Mater. Chem. A 2015, 3 (2), 756–761.

(9)

Kim, T. W.; Choi, K.-S. Nanoporous BiVO4 Photoanodes with Dual-Layer Oxygen
Evolution Catalysts for Solar Water Splitting. Science 2014, 343, 990–995.

(10) Vehkamäki, M.; Ritala, M.; Leskelä, M.; Jones, A. C.; Davies, H. O.; Sajavaara, T.; Rauhala,
E. Atomic Layer Deposition of Strontium Tantalate Thin Films from Bimetallic Precursors
and Water. J. Electrochem. Soc. 2004, 151 (4), F69.
(11)

Ritala, M. Atomic Layer Deposition of Oxide Thin Films with Metal Alkoxides as Oxygen
Sources. Science 2000, 288 (2000), 319–321.

(12)

Pickrahn, K. L.; Garg, A.; Bent, S. F. ALD of Ultrathin Ternary Oxide Electrocatalysts for
Water Splitting. ACS Catal. 2015, 5, 1609–1616.

(13)

Mullings, M. N.; Hägglund, C.; Tanskanen, J. T.; Yee, Y.; Geyer, S.; Bent, S. F. Thin Film
Characterization of Zinc Tin Oxide Deposited by Thermal Atomic Layer Deposition. Thin
Solid Films 2014, 556, 186–194.
57

(14)

Satta Alessandra, a.; Schuhmacher, J.; Whelan, C. M.; Vandervorst, W.; Brongersma, S.
H.; Beyer, G. P.; Maex, K.; Vantomme, a.; Viitanen, M. M.; Brongersma, H. H.; et al.
Growth Mechanism and Continuity of Atomic Layer Deposited TiN Films on Thermal SiO2.
J. Appl. Phys. 2002, 92, 7641–7646.

(15)

Green, M. L.; Ho, M. Y.; Busch, B.; Wilk, G. D.; Sorsch, T.; Conard, T.; Brijs, B.;
Vandervorst, W.; Räisänen, P. I.; Muller, D.; et al. Nucleation and Growth of Atomic Layer
Deposited HfO2 Gate Dielectric Layers on Chemical Oxide (Si-O-H) and Thermal Oxide
(SiO2 or Si-O-N) Underlayers. J. Appl. Phys. 2002, 92, 7168–7174.

(16)

Puurunen, R. L.; Vandervorst, W. Island Growth as a Growth Mode in Atomic Layer
Deposition: A Phenomenological Model. J. Appl. Phys. 2004, 96, 7686–7695.

(17)

Kaigawa, R.; Funahashi, K.; Fujie, R.; Wada, T.; Merdes, S.; Caballero, R.; Klenk, R.
Tandem Solar Cells with Cu(In,Ga)S2 Top Cells on ZnO Coated Substrates. Sol. Energy
Mater. Sol. Cells 2010, 94 (11), 1880–1883.

(18)

Kaigawa, R.; Ohyama, a.; Wada, T.; Klenk, R. Electrical Properties of Homogeneous
Cu(In,Ga)S2 Films with Varied Gallium Content. Thin Solid Films 2007, 515, 6260–6264.

(19)

Zheng, J. Y.; Song, G.; Kim, C. W.; Kang, Y. S. Facile Preparation of P-CuO and P-CuO/nCuWO4 Junction Thin Films and Their Photoelectrochemical Properties. Electrochim. Acta
2012, 69, 340–344.

(20)

Thimsen, E.; Baryshev, S. V; Martinson, A. B. F.; Elam, W.; Veryovkin, I. V; Pellin, M. J.
Interfaces and Composition Profiles in Metal-Sulfide Nanolayers Synthesized by Atomic
Layer Deposition. Chem. Mater. 2013, 25, 313-319.

(21)

Ginestra, C. N.; Sreenivasan, R.; Karthikeyan, A.; Ramanathan, S.; Mcintyre, P. C. Atomic
Layer Deposition of Y2O3/ZrO2 Nanolaminates A Route to Ultrathin Solid-State Electrolyte
Membranes. 2007, 10, 161–165.

(22)

Resasco, J.; Dasgupta, N. P.; Rosell, J. R.; Guo, J.; Yang, P. Uniform Doping of Metal
Oxide Nanowires Using Solid State. J. Am. Chem. Soc. 2014, 136, 10521–10526.

(23)

Liu, R.; Lin, Y.; Chou, L. Y.; Sheehan, S. W.; He, W.; Zhang, F.; Hou, H. J. M.; Wang, D.
Water Splitting by Tungsten Oxide Prepared by Atomic Layer Deposition and Decorated
with an Oxygen-Evolving Catalyst. Angew. Chemie Int. Ed. 2011, 50, 499–502.

(24)

Kihlborg, L.; Gebert, E. CuWO4, a Distorted Wolframite-Type Structure. Acta Crystallogr.
Sect. B Struct. Crystallogr. Cryst. Chem. 1970, 26, 1020–1026.

(25)

Pouchard, M.; Krussanova, M. A Photoelectrochemical Study of CuWO4 Single Crystals.
Phys. Stat. Sol. 1984, 82, 285–294.

58

(26)

Biloen, P.; Pott, G. T. X-Ray Photoelectron Spectroscopy Study of Supported Tungsten
Oxide. J. Catal. 1973, 30, 169–174.

(27)

Katrib, A.; Hemming, F.; Wehrer, P.; Hilaire, L.aire, G. Journal of Electron Spectroscopy
and Related Phenomena. J. Electron Spectros. Relat. Phenomena 1995, 76, 195–200.

(28)

Salvatl, L.; Makovsky, L. E.; Stencel, J. M.; Brownls, F. R.; Hercules, D. M. Surface
Spectroscopic Study. 1981, 104 (1977), 3700–3707.

(29)

Diaz-Reyes, J. Optical and Structural Properties of WO3 as a Function of the Annealing
Temperature. Adv. Sensors, Signals Mater. 2010, 99–104.

(30)

Desbat, B.; Lassegues, J. C. Infrared and Raman Study of WO3 Tungsten Trioxides and
WO3, xHzO Tungsten Trioxide Tydrates. J. Solid State Chem. 1987, 247, 235–247.

(31)

Zhang, J.; Zhang, W.; Yang, Z.; Yu, Z.; Zhang, X.; Chang, T. C.; Javey, A. Sensors and
Actuators B : Chemical Vertically Aligned Tungsten Oxide Nanorod Film with Enhanced
Performance in Photoluminescence Humidity Sensing. Sensors Actuators B. Chem. 2014,
202, 708–713.

(32)

Biesinger, M. C.; Payne, B. P.; Grosvenor, A. P.; Lau, L. W. M.; Gerson, A. R.; Smart, R.
S. C. Resolving Surface Chemical States in XPS Analysis of First Row Transition Metals,
Oxides and Hydroxides: Cr, Mn, Fe, Co and Ni. Appl. Surf. Sci. 2011, 257 (3), 2717–2730.

(33)

Tahir, D.; Tougaard, S. Electronic and Optical Properties of Cu, CuO and Cu2O Studied by
Electron Spectroscopy. J. Phys. Condens. Matter 2012, 24, 175002.

(34)

Biesinger, M. C.; Lau, L. W. M.; Gerson, A. R.; St, R.; Smart, C. Applied Surface Science
Resolving Surface Chemical States in XPS Analysis of First Row Transition Metals ,
Oxides and Hydroxides : Sc , Ti , V , Cu and Zn. Appl. Surf. Sci. 2010, 257 (3), 887–898.

(35)

Åsbrink, S.; Norrby, L. J. A Refinement of the Crystal Structure of copper(II) Oxide with a
Discussion of Some Exceptional E.s.d.’s. Acta Crystallogr. Sect. B Struct. Crystallogr.
Cryst. Chem. 1970, 26 (1), 8–15.

(36)

Asbrink, S.; Waskowska, A. CuO: X-Ray Single-Crystal Structure Determination at 196 K
and Room Temperature. J. Phys. Condens. Matter 1991, 3, 8173–8180.

(37)

Langford, J. I.; Louer, D. High-Resolution Powder Diffraction Studies of copper(II) Oxide.
J. Appl. Crystallogr. 1991, 24, 149–155.

(38)

Xu, J. F.; Ji, W.; Shen, Z. X.; Li, W. S.; Tang, S. H.; Ye, X. R.; Jia, D. Z.; Xin, X. Q. Raman
Spectra of CuO Nanocrystals. J. Raman Spectrosc. 1999, 30, 413–415.

(39)

Hagemann, H.; Bill, H.; Sadowski, W.; Walker, E.; François, M. Raman Spectra of Single
Crystal CuO. Solid State Commun. 1990, 73 (6), 447–451.
59

(40)

Goldstein, H. F.; Kim, D. S.; Yu, P. Y.; Bourne, L. C.; Chaminade, J. P.; Nganga, L. Raman
Study of CuO Single Crystals. Phys. Rev. B 1990, 41 (10), 7192–7194.

(41)

Dolgushev, N. V.; Malkov, a. a.; Malygin, a. a.; Suvorov, S. a.; Shchukarev, a. V.; Beljaev,
a. V.; Bykov, V. a. Synthesis and Characterization of Nanosized Titanium Oxide Films on
the (0001) α-Al2O3 Surface. Thin Solid Films 1997, 293 (1–2), 91–95.

(42)

Kim, H.; Rossnagel, S. M. Growth Kinetics and Initial Stage Growth during PlasmaEnhanced Ti Atomic Layer Deposition. J. Vac. Sci. Technol. A Vacuum, Surfaces, Film.
2002, 20 (2002), 802.

(43)

Yourey, J. E.; Bartlett, B. M. Electrochemical Deposition and Photoelectrochemistry of
CuWO4, a Promising Photoanode for Water Oxidation. J. Mater. Chem. 2011, 21, 7651–
7660.

(44)

Bohra, D.; Smith, W. A. Improved Charge Separation via Fe-Doping of Copper Tungstate
Photoanodes. Phys. Chem. Chem. Phys. 2015, 17, 9857–9866.

(45)

Ruiz-Fuertes, J.; Errandonea, D.; Segura, A.; Manjón, F. J.; Zhu, Z.; Tu, C. Y. Growth,
Characterization, and High-Pressure Optical Studies of CuWO4. High Press. Res. 2008, 28,
565–570.

(46)

Ruiz-Fuertes, J.; Sanz-Ortiz, M. N.; González, J.; Rodríguez, F.; Segura, A; Errandonea, D.
Optical Absorption and Raman Spectroscopy of CuWO4. J. Phys. Conf. Ser. 2010, 215,
12048.

(47)

Klahr, B. M.; Martinson, A. B. F.; Hamann, T. W. Photoelectrochemical Investigation of
Ultrathin Film Iron Oxide Solar Cells Prepared by Atomic Layer Deposition. Langmuir
2011, 27, 461–468.

(48)

Gaillard, N.; Chang, Y.; Deangelis, A.; Higgins, S.; Braun, A. A Nanocomposite
Photoelectrode Made of 2.2 eV Band Gap Copper Tungstate (CuWO4) and Multi-Wall
Carbon Nanotubes for Solar-Assisted Water Splitting. Int. J. Hydrogen Energy 2013, 38,
3166–3176.

(49)

Chang, Y.; Braun, A.; Deangelis, A.; Kaneshiro, J.; Gaillard, N. Effect of Thermal
Treatment on the Crystallographic, Surface Energetics, and Photoelectrochemical
Properties of Reactively Cosputtered Copper Tungstate for Water Splitting. J. Phys. Chem.
C 2011, 115, 25490–25495.

(50)

Lacomba-Perales, R.; Ruiz-Fuertes, J.; Errandonea, D.; Martinez-Garcia, D.; Segura, A.
Optical Absorption of Divalent Metal Tungstates: Correlation between the Band-Gap
Energy and the Cation Ionic Radius. 2008, 83, 37002.

(51)

Klahr, B.; Gimenez, S.; Fabregat-Santiago, F.; Bisquert, J.; Hamann, T. W. Electrochemical
60

and Photoelectrochemical Investigation of Water Oxidation with Hematite Electrodes.
Energy Environ. Sci. 2012, 5, 7626–7636.
(52)

Nakaoka, K.; Ueyama, J.; Ogura, K. Photoelectrochemical Behavior of Electrodeposited
CuO and Cu2O Thin Films on Conducting Substrates. J. Electrochem. Soc. 2004, 151, C661.

(53)

Heumann, T.; Stolica, N. The Electrochemical Behaviour of tungsten—I. The Dissolution
of Tungsten and Tungsten Oxides in Buffer Solutions. Electrochim. Acta 1971, 16 (5), 643–
651.

(54)

Heumann, T.; Stolica, N. The Electrochemical Behaviour of tungsten—II. The Dissolution
of Tungsten in NaOH Solutions. Electrochim. Acta 1971, 16 (10), 1635–1646.

(55)

Hill, J. C.; Choi, K. Synthesis and Characterization of High Surface Area CuWO4 and
Bi2WO6 Electrodes for Use as Photoanodes for Solar Water Oxidation. J. Mater. Chem. A
2013, 1, 5006–5014.

(56)

Pyper, K. J.; Yourey, J. E.; Bartlett, B. M. Reactivity of CuWO4 in Photoelectrochemical
Water Oxidation Is Dictated by a Midgap Electronic State. J. Phys. Chem. C 2013, 117,
24726–24732.

(57)

Valenti, M.; Dolat, D.; Biskos, G.; Smith, W. A. Enhancement of the Photoelectrochemical
Performance of CuWO4 Thin Films for Solar Water Splitting by Plasmonic Nanoparticle
Functionalization. J. Phys. Chem. C 2015, 119, 2096-2104.

(58)

Yourey, J. E. J.; Pyper, K. K. J.; Kurtz, J. B.; Bartlett, B. M. Chemical Stability of CuWO4
for Photoelectrochemical Water Oxidation. J. Phys. Chem. C 2013, 117, 8707–8718.

61

Chapter 3:
Quantifying
Hole
Collection
Efficiency of CuWO4 for Water
Oxidation

Adapted with permission from:
Quantitative hole collection for photoelectrochemical water oxidation with CuWO4, Yuan Gao
and Thomas W. Hamann, Chem. Comm., 2017, 53, 1285-1288. Copyright 2017 The Royal
Chemistry Society.
62

3.1 Abstract
The hole collection efficiency of water oxidation was evaluated for CuWO4 electrodes
from comparisons of the photocurrent of H2O2 and Na2SO3 oxidation as well as intensity
modulated photocurrent spectroscopy (IMPS) measurements. We found current multiplication
using H2O2, however use of Na2SO3 and IMPS revealed quantitative water oxidation at 1.23 V vs
RHE.

63

3.2 Introduction
In chapter 1, copper tungstate (CuWO4) was introduced as a promising photoanode
material for water splitting due to its suitably low band gap (2.3 eV), stability in neutral and slightly
basic solutions under illumination and its earth abundant composition.1–5 Previous studies have
indicated that the water oxidation efficiency of CuWO4 is limited by both charge separation in the
bulk of the electrode as well as hole collection via water oxidation at the electrode surface.4–7
Efforts to nanostructure the electrode have been made to overcome the inefficient charge
separation of this weakly absorbing material.8,9 Even if such strategies are successful, however, all
photogenerated holes that reach the electrode surface must selectively drive the water oxidation
reaction instead of other side or recombination reactions.
The water oxidation kinetics at the CuWO4/electrolyte interface has been suggested to be
sluggish with competing surface recombination reactions resulting in a loss in efficiency.3,4,6
Smith’s group compared the photocurrent measured in H2O and in an aqueous electrolyte
containing H2O2 as a sacrificial hole scavenger and extracted a hole collection efficiency  HC ,
from the ratio of the photocurrent densities: 7

HC  J H O J scavenger
2

(3.1)

These results indicated that the hole collection efficiency is only 60 % at 1.23 V vs RHE, the
thermodynamic water oxidation potential. This is the only report that has attempted to quantify the
hole collection efficiency of water oxidation with CuWO4 electrodes, although others have used
H2O2 to make semi-quantitative evaluations of water oxidation with CuWO4 electrodes.4,10 This
result suggests that surface modification strategies, such as adding a water oxidation co-catalyst to
the surface, can potentially greatly improve the water oxidation performance of CuWO4.

64

Interestingly, however, only a manganese phosphate catalyst has been reported to somewhat
improve the performance.11,12
In this chapter, we will examine the use of various hole scavengers to determine the water
oxidation efficiency of bare CuWO4 electrodes, and compare this to results of intensity modulated
photocurrent spectroscopy (IMPS) measurements. These results will allow us to quantify the extent
to which surface recombination inhibits water oxidation on the surface of CuWO4 and elucidate
possible causes of the lack of improvement observed from adding known water oxidation catalysts
to the surface.
3.3 Experimental
3.3.1 Film Preparation and Characterization
Copper tungstate (CuWO4) thin films were deposited on fluorine-doped tin oxide (FTO)
coated glass (Hartford Glass, 12 Ω cm-2) and Si wafer (University Wafer) by an atomic layer stack
deposition-annealing (SDA) method via atomic layer deposition (ALD) (Savannah 100,
Cambridge Nanotech Inc.) using a procedure described in chapter 2.5 ~170 ALD cycles of CuO
was deposited on 1000 cycles of WO3 according to calculations described in chapter 25 to match
1:1 ratio between the two metals. After deposition, the stack of binary oxides was annealed at
550 °C in air for 30 min with a ramping rate of 2 °C/min. The slight adjustment of annealing
temperature is explained in Appendix.
3.3.2 Characterization
The thickness of as-deposited WO3 films were measured to be ~65 nm thick on Si by
ellipsometry measurements (Horiba Jobin Yvon, Smart-SE). The thickness of the CuWO4 thin film
was calculated to be ~80 nm using the method described in chapter 2,5 which agrees with the
measured growth rate and previous measurement using AFM and SEM in chapter 2.5 The annealed

65

films were characterized by X-ray diffraction (XRD), Raman spectroscopy and scanning electron
microscopy (SEM) using the same instruments as described in chapter 2. The characterization
results are shown in Appendix.
3.3.3 Photoelectrochemical Measurements
CuWO4 electrodes were assembled and measured in the same three-electrode system as
described in chapter 2, with an active area of 0.28 cm2. The CuWO4 electrodes were examined in
contact with aqueous pH 9.0 1 M KBi buffer with or without 0.5 M Na2SO3 or H2O2. All potentials
measured in aqueous solutions were converted to the reversible hydrogen electrode (RHE) scaled
by the equation VRHE = VAg/AgCl + 0.197 + pH(0.059). Current density-voltage (J-V) curves were
measured using a scan rate of 20 mV s-1. The white light condition was the same as described in
chapter 2. All photoelectrochemical measurements were performed by shining light on the CuWO4
electrode through the electrolyte. Photocurrent densities (Jphoto) are obtained by subtracting the
dark current density (Jdark) from the current density under illumination (Jtotal). Intensity modulated
photocurrent spectroscopy (IMPS) was gathered using 10% of the fixed LED driving current as
the ac perturbation of the incident light intensity from 10000 to 0.1 Hz. The incident light of IMPS
measurements was provided by a 470 nm LED with an LED driver from Metrohm, and the light
intensity was calibrated using the current measured with a photodiode from the same manufacturer.
The constant light intensity was adjusted to 116 mW cm-2 so that the J-V curve is close to that
under 1 sun illumination. This ensures that the flux of holes reaching the surface is about the same
as 1 sun condition, so that the surface dynamics (rate of hole collection and recombination) is
comparable.

66

3.4 Results and Discussion
Current density, J, vs applied voltage, V, curves were measured in the dark and under
simulated AM 1.5 (1 sun) illumination for CuWO4 electrodes in contact with an aqueous
electrolyte and the same electrolyte containing two different sacrificial hole scavengers: 0.5 M
Na2SO3 and 0.5 M H2O2. Figure 3-1 shows plots of the J-V curves where the photocurrent density,
Jphoto, is determined by subtracting the dark current density, Jdark, from the total current density,
Jtotal, measured under 1 sun illumination, since CuWO4 has significant dark current when in contact
with the hole scavengers. The corresponding Jdark-V and Jtotal-V curves are shown in Figure A3-3.
The Jphoto-V curve measured in contact with H2O shows an onset potential at ~0.85 V vs RHE and
a photocurrent density of 0.15 mA cm-2 at 1.23 V vs RHE, which is comparable to the results in
chapter 2 and many other studies.3–6 Na2SO3 and H2O2 have both been widely used as hole
scavengers to study the hole collection efficiency of water oxidation using different semiconductor
materials assuming that their hole collection is quantitative.13–17 The photocurrent onset potential
measured with either hole scavenger shows a negative shift of 150 - 200 mV compared to the water
oxidation onset potential. This shift can be attributed to the limited  HC of water oxidation at the
CuWO4 surface, due to surface recombination kinetically competing with water oxidation.3,4,6,18,19

67

Figure 3-1. J-V curves of a CuWO4 electrode immersed in a pH 9.0 KBi electrolyte in contact with
H2O (black dashed curve), Na2SO3 (red solid curve) and H2O2 (blue dash dot) under 1 sun
illumination. The vertical gray dash represents the thermodynamic potential of water oxidation
(1.23 V vs RHE).

Interestingly, the J-V curves measured with CuWO4 electrodes immersed in Na2SO3 and
H2O2 solutions show very different photocurrents. The photocurrent measured in a Na2SO3
electrolyte is higher than that measured in H2O at low potentials, but when the potential passes
1.23 V vs RHE, the photocurrent in both electrolytes are essentially identical. However, the
photocurrent measured in a H2O2 electrolyte is 1.5 - 2 times of that in Na2SO3 or H2O at all
potentials. This is the first report of the J-V behavior of CuWO4 with Na2SO3, however. The higher
photocurrent measured in H2O2 compared to Na2SO3 can result from either 1) a less than unity

 HC for CuWO4 in oxidizing Na2SO3 or 2) a current multiplication effect due to H2O2 oxidation
with CuWO4. In order to test the first possible cause of the discrepancy, we measured the
photocurrent of CuWO4 with different concentrations of Na2SO3. Figure 3-2a shows that Jphoto
68

initially increases with Na2SO3 concentration and reaches saturation at 0.05 M, indicating that
when the diffusion limitation of Na2SO3 is eliminated, the photocurrent is independent of its
concentration.20 Current transient measurements were also conducted to evaluate the extent of
surface recombination. Figure 3-2b shows that when CuWO4 is immersed in 0.5 M Na2SO3, after
turning the light on, the instantaneous current remains constant, and the steady state photocurrents
extracted from the current transient measurements are consistent with those from the J-V curve.
These results suggest that the hole collection is not limited by the recombination of charge carriers
at CuWO4 surface when in contact with Na2SO3 and thus that  HC is quantitative. Control
experiments conducted in H2O are shown in Figure 3-3, where the current transient shows an
instantaneous spike when the light is turned on which quickly decays to a lower steady state value.
This indicates recombination of photogenerated holes at CuWO4 surface with conduction band
electrons in the case of water oxidation at low potentials. The decrease in the instantaneous current
appears less obvious at higher potentials, suggesting less severe surface recombination at these
potentials, which coincides with the similar photocurrent of water oxidation and oxidation of hole
scavenger Na2SO3 after 1.23 V vs RHE.

69

Figure 3-2. a) Photocurrent densities measured with a CuWO4 electrode in an aqueous electrolyte
containing Na2SO3 with various concentrations at 0.88 (red squares), 1.23 (green triangles) and
1.53 (blue circles) V vs RHE under 1 sun illumination. b) Current transients measured under 1 sun
illumination in 0.5 M Na2SO3 electrolyte at 0.88 (red solid), 1.23 (green dash) and 1.53 (blue dots)
V vs RHE.

70

Figure 3-3. Current transients measured with CuWO4 in an aqueous electrolyte (H2O) under 1 sun
illumination at 0.88 (red solid), 1.23 (green dash) and 1.53 (blue dots) V vs RHE.

We thus speculate that the higher photocurrent in the presence of H2O2 may result from a
current multiplication effect of H2O2 at the CuWO4 surface. H2O2 is well known to give rise to
current doubling in the cathodic region on p-type semiconductors under illumination.21,22 There
has also been one report of current doubling on n-type InP with H2O2 in its anodic region under
illumination.23 While it is not common to observe current doubling in the anodic region with H2O2,
many other species such as methanol, formic acid, sodium formate, etc. have been reported to
show current multiplication effect in the anodic region.24–26 One possible mechanism of current
doubling with H2O2 in the anodic region involves the injection of an electron from the radical
intermediate of H2O2 oxidation into the conduction band, where the normal 2-hole oxidation
process becomes a 1-hole process (see Appendix).23 Thus, the resulted current multiplication factor
would vary from 1 to 2 (i.e. current doubling) since the two oxidation mechanisms compete with
each other. Such current multiplication effects can be detected using intensity modulated

71

photocurrent spectroscopy (IMPS), since IMPS can provide information about quantum efficiency
(QE) of charge transfer processes at an interface. Therefore, IMPS measurements were conducted
in H2O, Na2SO3 and H2O2 to determine the extent of current multiplication.
IMPS was measured over a potential range from 0.73 to 1.73 V vs RHE. The output signal
is expressed by:
Hˆ  Iˆphoto / IˆLED

(3.2)

where Îphoto is the photocurrent, ÎLED is the driving current for LED, and Ĥ represents the transfer
function of the output. Therefore, the horizontal and vertical axes of a Nyquist plot represent the
real and imaginary parts of this output. The Nyquist plots of water oxidation are shown in Figure
3-4. In the upper quadrant of the Nyquist plot, a semicircle provides information about charge
transfer and recombination at the surface. In the low frequency limit, the illumination condition is
very close to steady state. Thus, the low frequency intercept on the real axis, HLF, is related to the
actual water oxidation photocurrent. When the frequency of the modulated light increases, the
recombination at the surface is “frozen out” at the high frequency limit, HHF, so that it reflects the
total flux of holes reaching the surface. The semicircle at higher frequency in the lower quadrant
is assigned to the RC attenuation of the electrochemical cell, which is beyond the scope of our
discussion.

72

Figure 3-4. Nyquist plots of IMPS measured with CuWO4 in H2O under 470 nm monochromatic
light with intensity of 116 mW cm-2. Potential ranges from 0.73 to 1.73 V vs RHE with an interval
of 0.1 V corresponding to the red, orange, yellow, dark yellow, green, dark green, dark cyan, blue,
navy, purple and wine colored circles.

If first order kinetics with respect to the hole concentration at surface is assumed as in
previous reports, 27–29 the quantum efficiency, QE, which is equivalent to  HC in the case of water
oxidation, can be extracted from IMPS data by:

 HC  H LF H HF

(3.3)

HHF should be identical in different electrolytes. If current multiplication occurs, two semicircles
in the lower quadrant are expected, with the low frequency part representing the charge transfer
process and the high frequency part representing RC attenuation, and the QE should be between 1
and 2.23,26,27 This behavior is exactly what we observe in the presence of H2O2, as shown in Figure
3-5; the Nyquist plots show two semicircles in the lower quadrant with HLF greater than HHF,
73

indicating a QE greater than unity. The calculated current multiplication factor, represented by QE,
is shown in Figure 3-6, where the QE is ~1.5 at lower potentials and reaches ~1.9 after 1.1 V vs
RHE. The lower quantum efficiency measured at lower potentials agrees with literature reports,
where current multiplication has to compete with recombination due to lower degree of band
bending.23,26 We noticed that the Nyquist plots measured in the presence of Na2SO3 look similar
as that of water oxidation, except that in the photocurrent onset potential region, the recombination
semicircle is smaller, indicating less surface recombination at low potentials.

74

Figure 3-5. Nyquist plots of IMPS measurements under 116 mW cm-2 light at 470 nm measured
in H2O (black squares), Na2SO3 (red circles), and H2O2 (blue triangles) electrolyte at a) 0.93 V and
b) 1.23 vs RHE.

75

Figure 3-6. Current multiplication factor of CuWO4 in contact with H2O2 represented as quantum
efficiency (QE) extracted from IMPS data.

It can be seen in Figure 3-7a that  HC extracted from IMPS matches perfectly with that
calculated using equation 3.3 when using Na2SO3 as a hole scavenger. Water oxidation with
CuWO4 shows a limited  HC around its onset potential region, but is essentially quantitative after
1.23 V vs RHE. However, as shown in Figure 3-7b, “  HC ” calculated using H2O2 only has a
maximum of ~60% which does not agree with the IMPS results, but is in excellent agreement with
the report by Smith et al.7 This discrepancy is consistent with the current multiplication effect of
H2O2 with CuWO4, which would produce an artificially low hole collection efficiency; thus, H2O2
cannot be used to determine the hole collection efficiency of CuWO4 electrodes.

76

Figure 3-7. Calculated  HC of water oxidation with CuWO4 electrodes extracted from IMPS data
at 116 mW cm-2 with 470 nm LED (orange triangles) and from comparisons of J-V curves of water
oxidation and oxidation of a) Na2SO3 (purple solid) and b) H2O2 under 1 sun illumination (green
solid).

77

3.5 Conclusion
In conclusion, while prior reports have suggested that sluggish water oxidation on CuWO4
limits the hole collection efficiency to ~60% at 1.23 V vs RHE via comparisons with H2O2
oxidation, we showed that such use of H2O2 can be problematic as it gives rise to current
multiplication. The combined PEC and IMPS measurements presented here clearly show that
Na2SO3 does act as an ideal sacrificial hole scavenger for CuWO4 and can be used to accurately
evaluate the hole collection efficiency of water oxidation with CuWO4 electrodes. Interestingly,
we found that the water oxidation hole collection efficiency is quantitative at 1.23 V vs RHE. Thus,
it is not feasible to increase the photocurrent density at this potential through surface modification
strategies such as addition of water oxidation co-catalysts. This finding explains the recently
reported challenge in identifying effective cocatalysts.12 Comparison of the J-V curves for Na2SO3
vs water oxidation (Figure 3-1) shows that an improvement of the onset potential by up to 200
mV is feasible, however, due to surface recombination.6 Interestingly, there has been only one
catalyst reported12,30 to be integrated with CuWO4 photoanode that shows moderate improvement
on its water oxidation performance, which is also consistent with our findings. Thus, CuWO4
appears to have excellent surface properties, with only limited further possible improvement in ηhc,
and thus must primarily be limited by charge separation in the bulk. Finally, we note that the
findings reported here may depend somewhat on the preparation route, as we have recently shown
for hematite photoelectrodes, which is worth further investigation.31

78

APPENDIX

79

a)

b)

c)

Figure A3-1. SEM images of a) as-deposited CuO on WO3, and CuWO4 annealed at b) 550 °C
and c) 600 °C.

80

Figure A3-2. Raman spectra of a) CuWO4 annealed at 550 °C and b) FTO. c) X-ray diffraction
pattern of CuWO4 annealed at 550 °C. Vertical gray dashed lines represent the peak positions of
CuWO4 recorded in literature32–34.

81

The SEM images in Figure A3-1 show that the as-deposited CuO on WO3 makes a uniform
film. After annealing at 550 °C, the grains are sharpened but the film is still uniform. However,
after annealing at 600 °C, although Raman and XRD show a pure composition of CuWO45, severe
aggregation occurs at the surface of the film, while underneath the top aggregated layer there are
smaller grains. We adjusted the annealing temperature to 550 °C for the ease of fundamental
investigations especially on-going charge carrier collection length studies in our laboratory.
Raman and XRD spectra in Figure A3-2 show that the peaks of the resulted film match with
CuWO4 peaks in literature32–34, except for those from the underlying FTO substrate. The XRD
pattern of FTO can be found in chapter 25. X-ray photoelectron spectroscopic (XPS) results are
also shown in chapter 2 to determine the composition and chemical state of our CuWO4 film.5

82

Figure A3-3. Jdark-V (black dash) and Jtotal-V (red solid) curves of CuWO4 electrode measured in
a) H2O, b) 0.5 M Na2SO3 and c) 0.5 M H2O2 electrolytes in the dark and under 1 sun illumination.

83

Current multiplication mechanism of H2O2 in the anodic region proposed by previous
studies23 (the reaction is considered in a basic electrolyte):

H 2O2  hVB
 OH   HO2•  H 2O

(A3.1)


HO2•  hVB
 OH   O2  H 2O

(A3.2)


HO2•  OH   O2  H 2O  eCB

(A3.3)

H2O2 captures a hole from the valence band (VB) of the n-type semiconductor to form a radical
intermediate (equation A3.1). The radical either captures another hole from VB, which is the route
of normal photoelectrochemical oxidation of H2O2 (equation A3.2), or inject an electron to the
conduction band (CB) of the semiconductor (equation A3.3). The second pathway contributes to
the current doubling effect, since the original two-hole process becomes a one-hole process.

84

REFERENCES

85

REFERENCES

(1)

Benko, F. A.; MacLaurin, C. L.; Koffyberg, F. P. CuWO4 and Cu3WO6 as Anodes for the
Photoelectrolysis of Water. Mat. Res. Bull. 1982, 17, 133–136.

(2)

Lalić, M. V.; Popović, Z. S.; Vukajlović, F. R. Electronic Structure and Optical Properties
of CuWO4: An Ab Initio Study. Comput. Mater. Sci. 2012, 63, 163–167.

(3)

Yourey, J. E.; Bartlett, B. M. Electrochemical Deposition and Photoelectrochemistry of
CuWO4, a Promising Photoanode for Water Oxidation. J. Mater. Chem. 2011, 21, 7651–
7660.

(4)

Hill, J. C.; Choi, K. S. Synthesis and Characterization of High Surface Area CuWO4 and
Bi2WO6 Electrodes for Use as Photoanodes for Solar Water Oxidation. J. Mater. Chem. A
2013, 1, 5006–5014.

(5)

Gao, Y.; Zandi, O.; Hamann, T. W. Atomic Layer Stack Deposition-Annealing Synthesis
of CuWO4. J. Mater. Chem. A 2016, 4, 2826–2830.

(6)

Pyper, K. J.; Yourey, J. E.; Bartlett, B. M. Reactivity of CuWO4 in Photoelectrochemical
Water Oxidation Is Dictated by a Midgap Electronic State. J. Phys. Chem. C 2013, 117,
24726–24732.

(7)

Bohra, D.; Smith, W. A. Improved Charge Separation via Fe-Doping of Copper Tungstate
Photoanodes. Phys. Chem. Chem. Phys. 2015, 17, 9857–9866.

(8)

Ye, W.; Chen, F.; Zhao, F.; Han, N.; Li, Y. CuWO4 Nanoflake Array-Based SingleJunction and Heterojunction Photoanodes for Photoelectrochemical Water Oxidation. ACS
Appl. Mater. Interfaces 2016, 8, 9211–9217.

(9)

Hu, D.; Diao, P.; Xu, D.; Xia, M.; Gu, Y.; Wu, Q.; Li, C.; Yang, S. Copper (II) Tungstate
Nanoflake Array Films: Sacrificial Template Synthesis, Hydrogen Treatment, and Their
Application as Photoanodes in Solar Water Splitting. Nanoscale 2016, 8, 5892–5901.

(10)

Gaillard, N.; Chang, Y.; Deangelis, A.; Higgins, S.; Braun, A. A Nanocomposite
Photoelectrode Made of 2.2 eV Band Gap Copper Tungstate (CuWO4) and Multi-Wall
Carbon Nanotubes for Solar-Assisted Water Splitting. Int. J. Hydrogen Energy 2013, 38,
3166–3176.

(11)

Nam, K. M.; Cheon, E. A.; Shin, W. J.; Bard, A. J. Improved Photoelectrochemical Water
Oxidation by the WO3/CuWO4 Composite with a Manganese Phosphate Electrocatalyst.
Langmuir 2015, 31 (39), 10897–10903.

(12)

Lhermitte, C. R.; Bartlett, B. M. Advancing the Chemistry of CuWO4 for
86

Photoelectrochemical Water Oxidation. Acc. Chem. Res. 2016, 49, 1121–1129.
(13)

Seabold, J. a.; Choi, K. S. Efficient and Stable Photo-Oxidation of Water by a Bismuth
Vanadate Photoanode Coupled with an Iron Oxyhydroxide Oxygen Evolution Catalyst. J.
Am. Chem. Soc. 2012, 134, 2186–2192.

(14)

Chang, X.; Wang, T.; Zhang, P.; Zhang, J.; Li, A.; Gong, J. Enhanced Surface Reaction
Kinetics and Charge Separation of P-N Heterojunction Co3O4/BiVO4 Photoanodes. J. Am.
Chem. Soc. 2015, 137, 8356–8359.

(15)

Mali, M. G.; Yoon, H.; Kim, M. W.; Swihart, M. T.; Al-Deyab, S. S.; Yoon, S. S.
Electrosprayed Heterojunction WO3/BiVO4 Films with Nanotextured Pillar Structure for
Enhanced Photoelectrochemical Water Splitting. Appl. Phys. Lett. 2015, 106 (15), 151603.

(16)

Duret, A.; Grätzel, M. Visible Light-Induced Water Oxidation on Mesoscopic α-Fe2O3
Films Made by Ultrasonic Spray Pyrolysis. J. Phys. Chem. B 2005, 109, 17184–17191.

(17)

Dotan, H.; Sivula, K.; Gr, M.; Warren, S. C.; Grätzel, M.; Rothschild, A.; Warren, S. C.;
Gr, M.; Warren, S. C. Probing the Photoelectrochemical Properties of Hematite (α-Fe2O3)
Electrodes Using Hydrogen Peroxide as a Hole Scavenger. Energy Environ. Sci. 2011, 4,
958–964.

(18)

Klahr, B.; Gimenez, S.; Fabregat-santiago, F.; Hamann, T.; Bisquert, J. Water Oxidation
at Hematite Photoelectrodes: The Role of Surface States. J. Am. Chem. Soc. 2012, 134,
4294–4302.

(19)

Klahr, B.; Gimenez, S.; Fabregat-Santiago, F.; Bisquert, J.; Hamann, T. W.
Electrochemical and Photoelectrochemical Investigation of Water Oxidation with
Hematite Electrodes. Energy Environ. Sci. 2012, 5, 7626–7636.

(20)

Klahr, B. M.; Hamann, T. W. Current and Voltage Limiting Processes in Thin Film
Hematite Electrodes. J. Phys. Chem. C 2011, 115, 8393–8399.

(21)

Memming, R. Mechanism of the Electrochemical Reduction of Persulfates and Hydrogen
Peroxide. J. Electrochem. Soc. 1969, 116 (6), 785–790.

(22)

Minks, B. P.; Vanmaekelbergb, D.; Kelly, J. J. Current-Doubling , Chemical Etching and
the Mechanism of Two-Electron Reduction Reactions at GaAs Part 2. A Unified Model. J.
Electroanal. Chem. 1989, 273, 133–145.

(23)

Theuwis, A.; Vermeir, I. E.; Gomes, W. P. Chemical and Electrochemical Interaction of
Acidic H202 Solutions with (100) InP. J. Electroanal. Chem. 1996, 410, 31–42.

(24)

Hykaway, N.; Sears, W. M.; Morisaki, H.; Morrison, S. R. Current-Doubling Reactions on
Titanium Dioxide Photoanodes. J. Phys. Chem. 1986, 90 (25), 6663–6667.

87

(25)

Fujishima, A.; Kato, T.; Maekawa, E.; Honda, K. Mechanism of the Current Doubling
Effect. I. The ZnO Photoanode in Aqueous Solution of Sodium Formate. Bull. Chem. Soc.
Jpn. 1981, 54, 1671–1674.

(26)

Herrasti, P.; Peter, L. Photocurrent Doubling during the Oxidation of Formic Acid at NCdS : An Investigation by Intensity Modulated Photocurrent Spectroscopy. J. Electroanal.
Chem. 1991, 305, 241–258.

(27)

Peter, L. M.; Vanmaekelbergh, D. Time and Frequency Resolved Studies of
Photoelectrochemical Kinetics. In Advances in Electrochemical Science and Engineering,
Volume 6; Wiley Online Library, 1999; pp 125–131.

(28)

Peter, L. M.; Ponomarev, E. A.; Fermin, D. J. Intensity-Modulated Photocurrent
Spectroscopy : Reconciliation of Phenomenological Analysis with Multistep Electron
Transfer Mechanisms. J. Electroanal. Chem. 1997, 427, 79–96.

(29)

Peter, L. M.; Upul, K. G.; Tahir, A. A. Kinetics of Light-Driven Oxygen Evolution at αFe2O3 Electrodes. Faraday Discuss. 2012, 155, 309–322.

(30)

Nam, K. M.; Cheon, E. A.; Shin, W. J.; Bard, A. J. Improved Photoelectrochemical Water
Oxidation by the WO3/CuWO4 Composite with a Manganese Phosphate Electrocatalyst.
Langmuir 2015, 31 (39), 10897–10903.

(31)

Zandi, O.; Schon, A. R.; Hajibabaei, H.; Hamann, T. W. Enhanced Charge Separation and
Collection in High-Performance Electrodeposited Hematite Films. Chem. Mater. 2016, 28
(3), 765–771.

(32)

Kihlborg, L.; Gebert, E. CuWO4, a Distorted Wolframite-Type Structure. Acta
Crystallogr. Sect. B Struct. Crystallogr. Cryst. Chem. 1970, 26, 1020–1026.

(33)

Ruiz-Fuertes, J.; Errandonea, D.; Segura, A.; Manjón, F. J.; Zhu, Z.; Tu, C. Y. Growth,
Characterization, and High-Pressure Optical Studies of CuWO4. High Press. Res. 2008,
28, 565–570.

(34)

Ruiz-Fuertes, J.; Sanz-Ortiz, M. N.; González, J.; Rodríguez, F.; Segura, a; Errandonea,
D. Optical Absorption and Raman Spectroscopy of CuWO4. J. Phys. Conf. Ser. 2010, 215,
12048.

88

Chapter 4:
Elucidation of Role of CuWO4 Surface
States During Photoelectrochemical
Water Oxidation

Adapted with permission from:
Elucidation of CuWO4 Surface States During Photoelectrochemical Water Oxidation, Yuan Gao
and Thomas W. Hamann, J. Phys. Chem. Lett. 2017, 8, 2700-2704. Copyright 2016 American
Chemistry Society.
89

4.1 Abstract
Electrochemical, photoelectrochemical and impedance spectroscopy measurements were
performed to investigate the role of CuWO4 surface states during water oxidation. We found that
a capacitive feature related to a surface state is clearly observable under water oxidation conditions.
The magnitude of the surface state capacitance is light intensity-dependent, with a peak potential
that coincides with the water oxidation onset potential region. The surface state is not observed in
the dark nor in contact with non-aqueous solvents. These results strongly support our assignment
of this surface state as the build-up of water oxidation intermediate species at the surface of CuWO4
photoanodes, not a permanent or intrinsic state as previously reported. We suggest this is a general
feature that controls the behavior and efficiency of solar water splitting reactions.

90

4.2 Introduction
In chapter 3, it was determined that the hole collection efficiency of CuWO4 in water
oxidation is actually quantitative at potentials higher than 1.23 V vs. RHE.1 This finding required
use of Na2SO3 as a suitable sacrificial hole scavenger, whereas we found that H2O2 gives rise to
current multiplication with CuWO4.2 Thus, poor hole collection via water oxidation with CuWO4
only occurs in the low potential region. There is only one report to date that has investigated the
nature of the surface states of CuWO4, where it was concluded to be an intrinsic state.3 This
interpretation is different from what we have now established with hematite electrodes, where a
surface state develops as a water oxidation intermediate species.4–7
In this chapter, we aim to elucidate the nature of surface states in CuWO4 and better
understand water oxidation on semiconductors in general.
4.3 Experimental
4.3.1 Film Preparation
1) CuWO4 by Atomic Layer Deposition
CuWO4 thin films were deposited by an atomic layer stack deposition-annealing (SDA)
method using a procedure described in chapter 3.8 1000 ALD cycles of WO3 was deposited on
fluorine-doped tin oxide (FTO) coated glass substrate followed by ~170 ALD cycles of CuO to
match 1:1 ratio between the two metals. After deposition, the stack of binary oxides was annealed
at 550 °C in air for 30 min with a ramping rate of 2 °C / min. The thickness of the CuWO4 thin
film was calculated to be ~80 nm.
2) CuWO4 by spray pyrolysis (SP-CuWO4)
CuWO4 was also synthesized by spray pyrolysis using a modified procedure from
literature.9 An aqueous precursor solution containing 0.01 M of Cu (II) and 0.01 M of W (II) was

91

prepared from CuCl•2H2O (Sigma-Aldrich, ≥ 99.0%,) and (NH4)6H2W12O40 (Alfa Aesar, metal
impurity ≤ 0.1%). FTO coated glass substrate was kept at 275 °C at a distance of 31 cm from the
nozzle. The precursor solution was sprayed onto the substrate for 1 s and the wait time was 5 s.
The as-deposited film was then annealed at 550 °C in air for 1 hr with 1 hr ramping time.
4.3.2 Film Characterization
The composition and structure of the film synthesized via spray pyrolysis was
characterized using Raman spectroscopy and X-ray diffraction using the same instruments as in
chapter 2. The thickness of the resulted film was determined to be 2 µm by cross-section scanning
electron microscopy (SEM) (Carl Zeiss Microscopy).
4.3.3 Electrochemical / Photoelectrochemical Measurements
All electrochemical and photoelectrochemical measurements were made with an Eco
Chemie Autolab potentiostat coupled with Nova electrochemical software. Current density-voltage
(J-V) curves were measured using a scan rate of 20 mV s-1. The white light source was a 450 W
Xe arc lamp (Horiba Jobin Yvon). An AM 1.5 solar filter was used to simulate sunlight at 100 mW
cm-2 (1 sun). A neutral density filter was used to reduce the light intensity to 33 (0.33 sun) or 10
mW cm-2 (0.1 sun). All photoelectrochemical measurements in aqueous solutions were performed
by shining light on the CuWO4 electrode through the electrolyte. Photocurrent densities (Jphoto)
were obtained by subtracting the dark current density (Jdark) from the total current density under
illumination (Jtotal).
Cyclic voltammogram (CV) surface state measurements were performed by applying a
positive potential under 1 sun illumination or in the dark for 60 seconds, and then scanning toward
the cathodic direction over a potential range in the dark.
EIS data were gathered using a 10 mV amplitude perturbation of between 10000 and 0.04

92

Hz in the dark or under illumination of 0.1 sun, 0.33 sun or 1 sun. Data were fitted using Zview
software (Scribner Associates).
4.3.4 Measurements in Aqueous Electrolytes
CuWO4 electrodes measured in aqueous electrolytes were in the same three-electrode setup
as described in chapter 2, with a photoanode active area of 0.28 cm2. The working electrodes were
examined in contact with aqueous pH 9.0 1 M KBi buffer with or without 0.5 M Na2SO3. Ag/AgCl
was used as the reference electrode and a high surface area Pt mesh was used as the counter
electrode. All potentials measured in aqueous solutions were converted to the reversible hydrogen
electrode (RHE) scaled by the equation VRHE = VAg/AgCl + 0.197 V + pH(0.059) V.
4.3.5 Measurements in Non-aqueous Electrolytes
When in contact with a non-aqueous electrolyte, the active area of the photoanode was
defined by the o-ring which is 0.64 cm2. A custom-made glass electrochemical cell without a
quartz window was used in the case of a non-aqueous electrolyte to avoid dissolution of the glue.
0.01 M AgNO3 was dissolved in acetonitrile (MeCN) with 0.1 M tetrabutylammonium
hexafluorophosphate (TBAPF6) to form Ag/AgNO3 as the reference electrode. The working
electrodes were examined in contact with anhydrous MeCN or dichloromethane with 0.2 M
TBAPF6 as the supporting electrolyte. Measurements in anhydrous solutions were conducted in
the glove box.
4.4 Results and Discussion
The Jphoto-V curves of CuWO4 in contact with H2O and with a sacrificial hole scavenger,
Na2SO3, are compared in Figure 4-1. The raw data of Jtotal-V and Jdark-V are shown in Figure A41. Na2SO3 has been shown to be a suitable sacrificial hole scavenger which produces quantitative
hole collection with CuWO4.1 The photocurrent onset potential of water oxidation is ~0.85 V vs.

93

RHE. When Na2SO3 is in the electrolyte, however, the onset potential is shifted negatively to 0.6
V vs. RHE. The discrepancy in the onset potential, as well as the photocurrent in the onset potential
region, between water oxidation and oxidation of Na2SO3 is due to inefficient hole collection via
water oxidation at the CuWO4 surface.1

Figure 4-1. Jphoto-V curves of a CuWO4 electrode in contact with a pH 9.0 KBi electrolyte in the
absence (black solid curve) and presence (red dashed curve) of 0.5 M Na2SO3 under 1 sun
illumination. The vertical gray dash represents the thermodynamic potential of water oxidation
(1.23 V vs. RHE).

EIS measurements were performed on nominally identical CuWO4 electrodes in contact
with H2O and Na2SO3 electrolytes to elucidate the processes that control the J-V behavior.4 Figure
4-2 displays Bode plots which indicate capacitive elements that appear in the EIS spectra. The
Bode phase plots are presented instead of the Nyquist plots since in some cases the two semicircles
are merged together, however the corresponding Nyquist plots are shown in Figure A4-2. In the
94

dark, the Bode phase plots show only one peak at all potentials, in contact with either H2O or
Na2SO3 electrolyte, which is assigned to the space charge capacitance. This suggests that the
surface state capacitance is not observable in the dark. Under 1 sun illumination, in the potential
range of 0.68-1.23 V vs. RHE, two capacitive peaks are clearly observed in the Bode phase plot
for CuWO4 electrodes immersed in the H2O electrolyte, however at higher potentials only one
peak is seen. On the other hand, only one peak could be observed in the Na2SO3 electrolyte at all
potentials under illumination.

95

Figure 4-2. Bode phase plots of CuWO4 electrode at 0.93 V vs. RHE a) in the dark and b) under
1 sun illumination and c) at 1.53 V vs. RHE under 1 sun illumination. Data measured in H2O (black
squares) and 0.5 M Na2SO3 (red circles) are both shown. Fitting curves are shown as black dash
for H2O and red dots for Na2SO3. The resulted parameters from fitting are listed in Table A4-1 to
Table A4-5.

The EIS data with two capacitive features were fit to a model shown in Figure 4-3a, which
includes the impedance from a surface state, which is adopted from the model established with
hematite.4,10 All the fitting results are shown in Table A4-1 to Table A4-5. Here Rs represents the
series resistance from the bulk of the semiconductor, electrical contacts, and electrolyte, Cbulk
96

describes the bulk capacitance, which in this potential region is dominated by space charge
capacitance, Rtrap is the resistance of trapping of holes in the surface states. Css is the surface state
capacitance, and Rct,ss describes the charge transfer resistance via the surface states. For EIS data
in the dark and under illumination with only one capacitive feature, the Randles circuit is used (Fig
4-3b).

Figure 4-3. a) Equivalent circuit for EIS data with two capacitive features. b) Randles circuit.

Values of Rtrap, Rct,ss, Cbulk and Css extracted from fitting EIS data to the equivalent circuits
displayed in Figure 4-3 are plotted in Figure 4-4. The only spectra that exhibited two capacitive
features, as shown in Figure 4-2, are measured under illumination with an H2O electrolyte - i.e.
under PEC water oxidation conditions. The magnitude of the fitted Css peak in Figure 4-4a
decreases when the light intensity is decreased from 1 sun to 0.1 sun. The peak in Css at ~1.0 V vs.
RHE is associated with a dip in Rct,ss (Fig 4-4b), which is close to the water oxidation onset
potential. The peak position of Css also matches the charge stored at the surface determined by
integration of photocurrent transient measurements shown in Figure 4-6c. These results strongly

97

support the assignment of this peak to a water oxidation intermediate species as opposed to an
intrinsic state.

Figure 4-4. Plots of equivalent circuit parameters obtained from fitting EIS data in the dark (open
squares), under 0.1 sun (blue triangles) and 1 sun (red circles) illumination in H2O electrolyte. a)
Surface state capacitance; b) Charge transfer resistance; c) Bulk capacitance; d) Resistance of
trapping of holes in surface states.

98

The fitted Cbulk values decrease with applied voltage, but are essentially invariant at a given
potential in the dark and under different light intensities (Fig 4-4c). The flat band potential, Vfb,
and the dopant density, ND, can be calculated from the following equation:
2
Cbulk



kT 
 V  V fb 

q 0 N D 
q 
2

(4.1)

where q is the elementary charge (1.6 × 10-19 C), ε is the dielectric constant of CuWO4 (83)11, ε0 is
the vacuum permittivity (8.854 × 10-12 F m-1), V is the applied potential, k is the Boltzmann
constant (1.38 × 1023 m2 kg s-2 K-1), T is the room temperature (300 K). The Mott-Schottky plots
displayed in Figure 4-5a measured in the dark, under 0.1 sun and 1 sun conditions overlap with
each other. The flat band potential (Efb) of CuWO4 was calculated to be 0.50 ±0.02 V vs. RHE,
and the dopant density (Nd) is 6.3 ± 0.1 × 1019 cm-3 based on the data measured in the dark. The
calculated results under different illumination conditions are compared in Table 4-1. This flat band
potential is comparable to literature values of ~0.42 V vs. RHE.12–14 In contrast to a prior report by
the Bartlett group3, no indication of Fermi level pinning is observed herein for CuWO4 in contact
with H2O, nor with the hole scavenger, Na2SO3 (Fig 4-5b).

99

Figure 4-5. Mott-Schottky plots of CuWO4 immersed in a) aqueous electrolyte (H2O), measured
in the dark (open squares), under 0.1 sun (blue triangles) and 1 sun (red circles) illumination; and
in b) 0.5 M Na2SO3, measured in the dark (open squares) and under 1 sun (red circles) illumination.

100

Illumination / sun

Vfb / V vs. RHE

Nd / cm-3

0

0.50  0.02

6.3 (  0.1) × 1019

0.1

0.47  0.04

8.0 (  0.3) × 1019

1

0.51  0.02

6.7 (  0.1) × 1019

Table 4-1. Flat band potential and dopant density calculated from Mott-Schottky plots in the dark
and under illumination.

Current transient measurements were taken on CuWO4 to study the charge storage behavior
of its surface states. Figure 4-6a shows that a spike in current appears upon turning the light on,
which decays to a lower steady state current. This indicates that upon illumination, photogenerated
holes accumulate at the surface, resulting in the initial transient current. Afterwards, only a portion
of these photogenerated holes undergo charge transfer (water oxidation) at the electrode/electrolyte
interface, which contributes to the steady state photocurrent. When the light is switched off, as in
Figure 4-6b, the accumulated holes at the surface recombine with conduction band (CB) electrons,
which causes the cathodic spike, and eventually the current reaches its steady state level (dark
current). We note that the scale of the transient current first increases with potential, and after it
reaches a maximum level, it decreases with potential. Each cathodic transient current is integrated
in order to calculate the charge stored in the surface state. The Gaussian behavior of the charge as
a function of potential shown in Figure 4-6c agrees well with the behavior of a chemical
capacitance.15,16 Interestingly, the peak of accumulated charge appears around 1.0 V vs. RHE,
which is at the same potential as Css peak in EIS measurements. This confirms that the second
capacitive feature observed in EIS measurements is related to the charging of the surface state of
CuWO4.
101

Figure 4-6. Current transients of CuWO4 in H2O after turning the 1 sun illumination a) on and b)
off. The applied potential varies from 0.68, 0.73, 0.83, 1.03, 1.23 and 1.43 V vs. RHE
corresponding to red solid, orange dash, yellow dots, green dash dots, blue short dash and purple
short dots. c) Charge stored at CuWO4 surface calculated from current transients after turning off
the light (black squares) and Gaussian fitting of the charge (red dash) at 0.68, 0.73, 0.83, 0.93,
1.03, 1.13, 1.23, 1.33 and 1.43 V vs. RHE.

Cyclic voltammetry (CV) was also utilized to examine the charging / discharging of the
surface state.17 A pretreatment of holding a positive potential of 2.3 V vs. RHE either under

102

illumination or in the dark for 60 s was performed, followed by a fast potential scan in the negative
direction. We note this potential is sufficiently positive to oxidize the surface state even in the dark.
As seen in Figure 4-7a and Figure A4-3, a cathodic peak appears around 0.9 V vs. RHE, indicating
the reduction of the surface state, which is equivalent to recombination of surface-trapped holes
with CB electrons. The intensity of the peaks increases with scan rate, which confirms the
capacitive nature. The peak position is very close to the Css peak at 1.0 V vs. RHE we observed in
EIS measurements, as well as the peak position of the stored charge in current transient
measurements, although in CV it is slightly shifted towards the negative potential due to the nonsteady state conditions at the fast scan rates. This further suggests that the peak of Css is associated
with charging the surface state, which is related to water oxidation.

Figure 4-7. a) CVs taken after holding potential at 2.3 V vs. RHE for 60 s under 1 sun illumination
at scan rates from 200, 400, 600, 800, 1000, 1200 and 1500 mV s-1 corresponding to red, orange,
dark yellow, green, dark green, blue and purple solid. b) CV after holding potential at ~1.8 V vs.
Vfb for 60 s in the dark, measured in H2O (black solid) and MeCN (red dash). Scan rate is 1500
mV s-1.

103

Analogous fast scan CVs were also performed in non-aqueous solvents, to determine if the
generated surface state is consistent with a water oxidation intermediate species. These
experiments were conducted in a sealed electrochemical cell in the glovebox to minimize the
influence of trace amounts of water. The CV curves measured in anhydrous acetonitrile (MeCN)
are plotted vs. Vfb of CuWO4 measured in the same electrolyte (see Fig A4-4 for Mott-Schottky
plots). No corresponding peak was observed even with fast scan rate (1500 mV s-1) as shown in
Figure 4-7b. We added 2% of H2O to the MeCN electrolyte and conducted the same CV
measurement. A small peak appeared at 0.4 V vs. Vfb, which is the same potential where we
observed a peak in aqueous electrolytes (Fig A4-5). Similar CV measurements were also
performed in a non-coordinating solvent, dichloromethane, to eliminate the influence of any
coordination of electrolyte species to the electrode surface. Figure 4-8 shows that no peak is under
these conditions either. These results indicate that the surface state is available only in the presence
of water, with sufficient applied bias or illumination to oxidize water which confirms our argument
that the surface state of CuWO4 is related to water oxidation.

104

Figure 4-8. CV after holding potential at ~1.8 V vs. Vfb for 60 s in the dark, measured in 0.2 M
TBAPF6 in dichloromethane in the glovebox. Scan rates varies from 200, 800, 1500 and 2000 mV
/ s corresponding to pink, yellow, cyan and purple solid lines. Initial scan is in the cathodic
direction.

Electrodes prepared via ALD were utilized since this produces reproducible, uniform and
compact films which allows a more straightforward interpretation of EIS data compared to
nanostructured, porous films. Since differences in synthetic methods may potentially lead to
different surface properties, we also carried out a complete set of control experiments with CuWO4
synthesized via spray pyrolysis (SP-CuWO4). As shown in Appendix, the Css obtained from fitting
EIS data increased with increasing illumination intensity corresponding to 0.1, 0.33 and 1 sun,
consistent with results discussed above of the ALD films. The Mott-Schottky plots of the SPCuWO4 in the dark show no indication of Fermi level pinning. Thus, in all cases, the electrodes
prepared by SP behaved in nearly identical fashion as electrodes prepared by ALD. These

105

combined results indicate that photogenerated surface states we observe for CuWO4 is general to
this material.
In conclusion, a surface state is clearly observed on CuWO4 via several electrochemical
methods. Importantly, we are only able to observe this state under water oxidation conditions, at a
potential closely related to the water oxidation onset potential. These results strongly support an
assignment of this surface state as a water oxidation intermediate species. This assignment
contradicts prior reports who identified an intrinsic Cu-centered surface state.3 Since the results
did not depend on the synthetic method of preparing CuWO4 electrodes, the assignment of the
surface state to water oxidation intermediate species appears to be general for CuWO4 photoanodes.
It is also striking how similar CuWO4 behaves compared to hematite, where it is established that
a surface state arises from the first step of water oxidation. Recombination of this intermediate
species with conduction band electrons gives rise to the shift in photocurrent onset potential since
an additional potential needs to be applied to reduce surface electron concentration and thus
recombination. Thus, the results presented here offer the first indication that this behavior of photoinduced surface state recombination is general for water oxidation with metal oxide
semiconductors with suggests a fundamental limitation to solar water splitting with bare electrodes.
Addition of co-catalysts, such as Co-Pi or NiFeOx to the electrode surface, however, offers a route
to circumvent this recombination path since photogenerated holes at the electrode surface can
oxidize the catalyst preferentially to the electrode surface. Work is ongoing to elucidate the detailed
charge transfer pathways and kinetics of the complex semiconductor / catalyst interface during
water oxidation.

106

APPENDIX

107

Figure A4-1. Jdark-V (black dash) and Jtotal-V (red solid) curves of CuWO4 electrode measured in
a) H2O and b) 0.5 M Na2SO3 electrolytes in the dark and under 1 sun illumination.

108

Figure A4-2. Nyquist plots of CuWO4 electrode at 0.93 V vs. RHE a) in the dark and b) under 1
sun illumination and c) at 1.53 V vs. RHE under 1 sun illumination. Data measured in H2O (black
squares) and 0.5 M Na2SO3 (red circles) are both shown. Fitting curves are shown as black dash
for H2O and red dots for Na2SO3.

109

V / V vs. Ag/AgCl, pH 9
-0.20
-0.15
-0.10
-0.05
0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
0.45
0.50
0.55
0.60
0.65
0.70
0.75
0.80
0.85
0.90
0.95
1.00

Rs / Ω
147.9
147.5
146.7
145.5
146.7
147.1
147.2
147.5
148.0
148.1
148.5
148.7
148.2
147.6
147.4
147.4
147.4
147.1
146.8
146.7
146.3
146.0
145.4
144.6
144.0

V / V vs. RHE
0.53
0.58
0.63
0.68
0.73
0.78
0.83
0.88
0.93
0.98
1.03
1.08
1.13
1.18
1.23
1.28
1.33
1.38
1.43
1.48
1.53
1.58
1.63
1.68
1.73

Rct / Ω
3.31E+05
6.43E+05
7.83E+05
7.46E+05
2.27E+05
2.09E+05
3.56E+05
1.25E+06
1.12E+06
4.99E+05
1.61E+05
57352
25559
15293
11386
9561
8276
6738
5149
3810
2725
1912
1321
902.6
620.8

Cbulk / F
1.33E-05
8.73E-06
6.54E-06
5.62E-06
4.73E-06
4.21E-06
3.81E-06
3.62E-06
3.38E-06
3.21E-06
3.03E-06
2.84E-06
2.64E-06
2.49E-06
2.39E-06
2.33E-06
2.28E-06
2.22E-06
2.15E-06
2.08E-06
2.01E-06
1.96E-06
1.90E-06
1.85E-06
1.80E-06

Table A4-1. Parameters extracted from fitting EIS data in 0.5 M Na2SO3 electrolyte in the dark.

110

V / V vs. Ag/AgCl, pH 9
-0.20
-0.15
-0.10
-0.05
0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
0.45
0.50
0.55
0.60
0.65
0.70
0.75
0.80
0.85
0.90
0.95
1.00

Rs / Ω
147.8
143.8
149.6
150.5
150.7
153.5
155.7
155.7
155.7
155.7
154.8
154.3
154.2
153.1
153.6
153.9
153.3
152.4
152.0
150.4
149.6
149.3
148.7
147.9
146.9

V / V vs. RHE
0.53
0.58
0.63
0.68
0.73
0.78
0.83
0.88
0.93
0.98
1.03
1.08
1.13
1.18
1.23
1.28
1.33
1.38
1.43
1.48
1.53
1.58
1.63
1.68
1.73

Rct / Ω
1.53E+05
1.19E+05
56657
28487
16358
11906
9468
7792
7162
7319
7645
7276
6208
5126
4466
4042
3635
3139
2587
1992
1510
1134
826.9
594.4
431.9

Cbulk / F
1.73E-05
1.09E-05
7.54E-06
5.92E-06
4.89E-06
4.29E-06
3.98E-06
3.73E-06
3.49E-06
3.27E-06
3.11E-06
2.94E-06
2.79E-06
2.67E-06
2.60E-06
2.55E-06
2.49E-06
2.43E-06
2.37E-06
2.31E-06
2.24E-06
2.19E-06
2.14E-06
2.09E-06
2.05E-06

Table A4-2. Parameters extracted from fitting EIS data in 0.5 M Na2SO3 electrolyte under 1 sun
illumination.

111

V / V vs. Ag/AgCl, pH 9
-0.20
-0.15
-0.10
-0.05
0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
0.45
0.50
0.55
0.60
0.65
0.70
0.75
0.80
0.85
0.90
0.95
1.00

Rs / Ω
188.8
188.5
188.3
188.2
188.5
189.0
189.3
189.6
189.9
190.0
190.3
190.6
190.7
191.0
191.1
191.2
191.2
191.3
191.3
191.4
191.6
191.7
191.9
192.1
192.3

V / V vs. RHE
0.53
0.58
0.63
0.68
0.73
0.78
0.83
0.88
0.93
0.98
1.03
1.08
1.13
1.18
1.23
1.28
1.33
1.38
1.43
1.48
1.53
1.58
1.63
1.68
1.73

Rct / Ω
5363
25555
3.14E+05
2.98E+06
1.04E+07
1.65E+07
2.04E+07
2.23E+07
2.10E+07
1.79E+07
1.58E+07
1.61E+07
1.87E+07
2.26E+07
2.60E+07
2.80E+07
2.73E+07
2.46E+07
2.20E+07
2.27E+07
2.37E+07
2.25E+07
1.92E+07
9.27E+06
1.61E+06

Cbulk / F
1.39E-05
8.70E-06
5.61E-06
4.31E-06
3.75E-06
3.37E-06
3.06E-06
2.81E-06
2.62E-06
2.48E-06
2.37E-06
2.27E-06
2.19E-06
2.12E-06
2.06E-06
2.00E-06
1.95E-06
1.90E-06
1.86E-06
1.81E-06
1.77E-06
1.74E-06
1.72E-06
1.70E-06
1.68E-06

Table A4-3. Parameters extracted from fitting EIS data in aqueous electrolyte (pH9 KBi buffer)
in the dark.

112

V / V vs.
Ag/AgCl,
pH 9
-0.20
-0.15
-0.10
-0.05
0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
0.45
0.50
0.55
0.60
0.65
0.70
0.75
0.80
0.85
0.90
0.95
1.00

V / V vs.
RHE

Rs / Ω

Cbulk / F

Rtrap / Ω

Css / F

Rct,ss / Ω

0.53
0.58
0.63
0.68
0.73
0.78
0.83
0.88
0.93
0.98
1.03
1.08
1.13
1.18
1.23
1.28
1.33
1.38
1.43
1.48
1.53
1.58
1.63
1.68
1.73

239.2
236.1
233.4
230.4
231.0
231.5
232.4
232.6
232.4
232.6
232.9
233.3
233.5
231.6
231.4
231.7
231.7
234.2
234.1
234.1
234.3
234.8
235.1
234.9
235.6

1.70E-05
9.95E-06
6.25E-06
4.31E-06
3.80E-06
3.49E-06
3.27E-06
3.05E-06
2.86E-06
2.70E-06
2.58E-06
2.46E-06
2.36E-06
2.16E-06
2.06E-06
1.99E-06
1.94E-06
2.07E-06
2.02E-06
1.98E-06
1.95E-06
1.92E-06
1.90E-06
1.88E-06
1.84E-06

26731
25558
34145
49570
37032
28345
28631
36213
52497
69903
20136
15222
15319
18017
-

8.96E-07
7.59E-07
8.50E-07
1.42E-06
2.76E-06
5.39E-06
9.32E-06
1.12E-05
8.74E-06
4.05E-06
5.70E-07
4.65E-07
4.28E-07
4.14E-07
-

4040
19084
2.20E+05
2.09E+06
4.78E+06
3.01E+06
1.00E+06
3.02E+05
1.16E+05
58638
41131
39867
56107
1.41E+05
1.80E+05
2.05E+05
2.24E+05
2.49E+05
2.59E+05
2.73E+05
2.83E+05
2.92E+05
2.81E+05
1.83E+05
62570

Table A4-4. Parameters extracted from fitting EIS data in aqueous electrolyte (pH9 KBi buffer)
under 0.1 sun illumination.

113

V / V vs.
Ag/AgCl,
pH 9
-0.20
-0.15
-0.10
-0.05
0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
0.45
0.50
0.55
0.60
0.65
0.70
0.75
0.80
0.85
0.90
0.95
1.00

V / V vs.
RHE

Rs / Ω

Cbulk / F

Rtrap / Ω

Css / F

Rct,ss / Ω

0.53
0.58
0.63
0.68
0.73
0.78
0.83
0.88
0.93
0.98
1.03
1.08
1.13
1.18
1.23
1.28
1.33
1.38
1.43
1.48
1.53
1.58
1.63
1.68
1.73

235.1
231.3
230.4
230.1
230.3
231.3
231.8
231.7
231.4
231.1
230.7
230.5
229.6
229.2
229.1
230.0
229.7
229.4
228.5
228.5
228.3
228.4
228.5
228.5
230.4

1.44E-05
8.86E-06
5.80E-06
4.33E-06
3.85E-06
3.59E-06
3.33E-06
3.07E-06
2.83E-06
2.64E-06
2.49E-06
2.35E-06
2.22E-06
2.12E-06
2.06E-06
2.06E-06
1.97E-06
1.91E-06
1.85E-06
1.81E-06
1.76E-06
1.75E-06
1.71E-06
1.71E-06
1.69E-06

40293
13033
6187
3112
2044
1844
2086
2761
3765
4593
5202
6642
-

1.20E-06
1.37E-06
2.43E-06
4.51E-06
8.47E-06
1.36E-05
1.59E-05
1.57E-05
1.14E-05
5.71E-06
3.03E-06
5.21E-06
-

2176
12966
1.64E+05
1.56E+05
77806
34892
14666
6716
3662
2505
2055
1985
2366
3020
2753
10058
11139
11739
12260
12957
13218
14248
13686
14931
14752

Table A4-5. Parameters extracted from fitting EIS data in aqueous electrolyte (pH9 KBi buffer)
under 1 sun illumination.

114

Figure A4-3. CV taken after holding potential at 2.3 V vs. RHE for 60 s in the dark. Scan rates
varies from 200, 400, 600, 800, 1000, 1200 and 1500 mV s-1 corresponding to red, orange, dark
yellow, green, dark green, blue and purple solid. The peak intensity after the dark pretreatment is
slightly smaller compared to the light pretreatment.

115

Figure A4-4. Mott-Schottky plots (black scattered points) and linear fit (red dash) of CuWO 4 in
a) MeCN electrolyte, b) MeCN electrolyte containing 2% H2O, and c) dichloromethane electrolyte.

116

Figure A4-5. CV after holding potential at ~1.8 V vs. Vfb for 60 s in the dark, measured in H2O
(black solid), MeCN (red dash) and MeCN with 2% H2O (blue dot). Scan rate is 1500 mV s-1.

117

Figure A4-6. a) Raman spectrum and b) XRD pattern of CuWO4 synthesized by spray pyrolysis
and annealed at 550 °C for 1 hr (SP-CuWO4). Experimental data of CuWO4 are shown as black
curves and peak positions from literature are shown as vertical dark yellow dash. The XRD pattern
of FTO substrate is shown as blue curve as control experiment, and the diffraction peaks from FTO
are labeled as *.

As shown in Figure A4-6, the resulted SP-CuWO4 films show Raman and XRD peaks
corresponding to CuWO4 peaks in literature18–20, apart from those from the underlying FTO
substrate marked as * in XRD pattern. These results confirm the pure CuWO4 composition of the
resulted films.

118

Figure A4-7. J-V curves of SP-CuWO4 under 1 sun (red solid), 0.33 sun (green dash), 0.1 sun
(blue dots) illumination and in the dark (black dash dot) measured in pH 9 aqueous (H2O)
electrolyte.

119

Figure A4-8. Bode phase plots of SP-CuWO4 at 0.93V vs. RHE in the dark (black squares) and
under 1 sun illumination (red circles) measured in pH 9 aqueous (H2O) electrolyte. Fitting curves
are shown as black dash and red dots for dark and 1 sun measurements, respectively.

Figure A4-9. Css obtained from fitting EIS measured in pH 9 aqueous (H2O) electrolyte under 0.1
sun (blue triangles), 0.33 sun (green diamonds) and 1 sun (red circles) illumination with SPCuWO4.
120

Figure A4-10. Mott-Schottky plots of SP-CuWO4 measured in the dark in pH 9 aqueous (H2O)
electrolyte.

121

Figure A4-11. CVs taken after holding potential at 2.3 V vs. RHE for 60 s under 1 sun illumination
at scan rates from 200, 400, 600, 800, 1000, 1200 and 1500 mV s-1 corresponding to red, orange,
dark yellow, green, dark green, blue and purple. Experiment was conducted in pH 9 aqueous
electrolyte.

122

As shown in Figure A4-7, the J-V curves of SP-CuWO4 measured under illumination for
water oxidation show an onset potential of 0.85 - 0.90 V vs. RHE, depending on the light intensity.
The photocurrent at 1.23 V vs. RHE under 1 sun illumination is ~0.13 mA cm-2, which is similar
as reported values.3,8,9,12 In Figure A4-8, the EIS data show two capacitive features in the Bode
phase plot under 1 sun illumination, while only one is observed in the dark. The surface state
capacitance, Css, increases with increasing light intensity, as shown in Figure A4-9. These facts
suggest that the surface state of SP-CuWO4 is light-induced, rather than intrinsic. The MottSchottky plot measured in the dark in Figure A4-10 shows no Fermi level pinning, further
confirming that the surface state is not present in the dark. Lastly, the CV surface measurement in
Figure A4-11 shows a capacitive peak around 0.9 V vs. RHE, the peak position of which shifts
negatively with increasing scan rate, while its peak current increases. This peak position is close
to that of Css in Figure A4-9, and matches with water oxidation onset potential, which is essentially
the same observation with CuWO4 synthesized via ALD. Therefore, we conclude that the role of
surface state as water oxidation intermediate is general for CuWO4, rather than specifically limited
to our electrode made by ALD.

123

REFERENCES

124

REFERENCES

(1)

Gao, Y.; Hamann, T. W. Quantitative Hole Collection for Photoelectrochemical Water
Oxidation with CuWO4. Chem. Commun. 2017, 53, 1285–1288.

(2)

Bohra, D.; Smith, W. A. Improved Charge Separation via Fe-Doping of Copper Tungstate
Photoanodes. Phys. Chem. Chem. Phys. 2015, 17, 9857–9866.

(3)

Pyper, K. J.; Yourey, J. E.; Bartlett, B. M. Reactivity of CuWO4 in Photoelectrochemical
Water Oxidation Is Dictated by a Midgap Electronic State. J. Phys. Chem. C 2013, 117,
24726–24732.

(4)

Klahr, B.; Gimenez, S.; Fabregat-santiago, F.; Hamann, T.; Bisquert, J. Water Oxidation at
Hematite Photoelectrodes: The Role of Surface States. J. Am. Chem. Soc. 2012, 134, 4294–
4302.

(5)

Klahr, B.; Gimenez, S.; Fabregat-Santiago, F.; Bisquert, J.; Hamann, T. W. Electrochemical
and Photoelectrochemical Investigation of Water Oxidation with Hematite Electrodes.
Energy Environ. Sci. 2012, 5, 7626–7636.

(6)

Zandi, O.; Hamann, T. W. Determination of Photoelectrochemical Water Oxidation
Intermediates on Haematite Electrode Surfaces Using Operando Infrared Spectroscopy. Nat.
Chem. 2016, 8, 778–783.

(7)

Klahr, B.; Hamann, T. Water Oxidation on Hematite Photoelectrodes: Insight into the
Nature of Surface States through in Situ Spectroelectrochemistry. J. Phys. Chem. C 2014,
118, 10393–10399.

(8)

Gao, Y.; Zandi, O.; Hamann, T. W. Atomic Layer Stack Deposition-Annealing Synthesis
of CuWO4. J. Mater. Chem. A 2016, 4, 2826–2830.

(9)

Lhermitte, C. R.; Bartlett, B. M. Advancing the Chemistry of CuWO4 for
Photoelectrochemical Water Oxidation. Acc. Chem. Res. 2016, 49, 1121–1129.

(10)

Bertoluzzi, L.; Bisquert, J. Equivalent Circuit of Electrons and Holes in Thin Semiconductor
Films for Photoelectrochemical Water Splitting Applications. J. Phys. Chem. Lett. 2012, 3,
2517–2522.

(11)

Arora, S. K.; Mathew, T. Dielectric Studies of CuWO4 Crystals. Phys. Stat. Sol. 1989, 116,
405–413.

(12)

Yourey, J. E.; Bartlett, B. M. Electrochemical Deposition and Photoelectrochemistry of
CuWO4, a Promising Photoanode for Water Oxidation. J. Mater. Chem. 2011, 21, 7651–
7660.
125

(13)

Yourey, J. E. J.; Pyper, K. K. J.; Kurtz, J. B.; Bartlett, B. M. Chemical Stability of CuWO4
for Photoelectrochemical Water Oxidation. J. Phys. Chem. C 2013, 117, 8707–8718.

(14)

Hill, J. C.; Choi, K. Synthesis and Characterization of High Surface Area CuWO 4 and
Bi2WO6 Electrodes for Use as Photoanodes for Solar Water Oxidation. J. Mater. Chem. A
2013, 1, 5006–5014.

(15)

Bisquert, J. Chemical Capacitance of Nanostructured Semiconductors: Its Origin and
Significance for Nanocomposite Solar Cells. Phys. Chem. Chem. Phys. 2003, 5, 5360–5364.

(16) Bisquert, J. Physical Electrochemistry of Nanostructured Devices. Phys. Chem. Chem. Phys.
2008, 10, 49–72.
(17)

Bertoluzzi, L.; Badia-bou, L.; Fabregat-santiago, F.; Gimenez, S.; Bisquert, J. Interpretation
of Cyclic Voltammetry Measurements of Thin Semiconductor Films for Solar Fuel
Applications. J. Phys. Chem. Lett. 2013, 4, 1334–1339.

(18)

Kihlborg, L.; Gebert, E. CuWO4, a Distorted Wolframite-Type Structure. Acta Crystallogr.
Sect. B Struct. Crystallogr. Cryst. Chem. 1970, 26, 1020–1026.

(19)

Ruiz-Fuertes, J.; Errandonea, D.; Segura, A.; Manjón, F. J.; Zhu, Z.; Tu, C. Y. Growth,
Characterization, and High-Pressure Optical Studies of CuWO4. High Press. Res. 2008, 28,
565–570.

(20)

Ruiz-Fuertes, J.; Sanz-Ortiz, M. N.; González, J.; Rodríguez, F.; Segura, a; Errandonea, D.
Optical Absorption and Raman Spectroscopy of CuWO4. J. Phys. Conf. Ser. 2010, 215,
12048.

126

Chapter 5:
Investigations of Charge Separation
Limitations of CuWO4

127

5.1 Abstract
In this chapter, mesoporous CuWO4 electrodes of ~800 nm thick were synthesized by spray
pyrolysis. The pores of CuWO4 were filled by depositing Al2O3 using atomic layer deposition
(ALD) followed by removing the excess top Al2O3 layer using 1 M KOH solution. By comparing
the PEC performance with front and back illumination on the resulted compact film, it was
determined that the electrons are the limiting charge carriers of CuWO4. The wavelength
dependent behavior of incident photon-to-current efficiency (IPCE) suggests that the electron
collection length of CuWO4 is shorter than 577 nm.

128

5.2 Introduction
As discussed in chapter 1, CuWO4 in theory should produce a maximum photocurrent
density of ~9 mA cm-2, reaching a theoretical maximum solar-to-hydrogen (STH) efficiency of 11%
if paired with a suitable photocathode in a tandem cell configuration for photoelectrochemical
(PEC) water splitting. However, the actual photocurrent density that has been achieved with
CuWO4 so far has been low, which varies from 0.1 to 0.3 mA cm-2 at the thermodynamic potential
of water oxidation (1.23 V vs RHE). It was discussed in chapter 2 and chapter 3 that water
oxidation is limited by surface recombination, i.e., recombination of oxidized water oxidation
intermediate species at the surface with conduction band electrons, at low applied bias. However,
once it passes the photocurrent onset potential region, the hole collection efficiency is essentially
quantitative at potentials higher than 1.23 V vs RHE. Therefore, the huge discrepancy between the
actual photocurrent density and the theoretical value indicates that charge separation efficiency
(CS ) is not optimized. As was mentioned in chapter 1, due to the weak absorption of this indirect
semiconductor, to absorb 95% of the incident light, a thickness of several hundred nm to several
µm is required, depending on the wavelength. However, the charge carriers are not necessarily
able to travel that far. The fact that thin film CuWO4 photoelectrode (80 nm to 200 nm thick)1,2
produces similar photocurrent as much thicker films (2 to 3 µm thick)3,4 suggests that the charge
carrier collection length is limiting the performance of CuWO4. Efforts have been made by many
research groups to solve the contradiction between a long light penetration depth and short carrier
collection length by doping5 and nanostructuring6–9. Bohra et al.5 and Gaillard et al.8 compared the
performance of CuWO4 by illuminating from different directions and concluded that the electronic
transport properties are the major limiting factors based on the better performance by illuminating
from the substrate-electrode (back) side. Therefore, Bohra et al. employed Fe-doping for their

129

CuWO4 photoanode, which is the only example of doping this material for PEC water oxidation.5
As a result, their doped film shows 1.5 times the photocurrent as the bare one, which is 0.25 mA
cm-2 at 1.23 V vs RHE.5 Gaillard et al. used a nanocomposite of CuWO4 and multi-wall carbon
nanotubes (MWCNT) to reduce the bulk resistance of CuWO4 electrode. A photocurrent density
of ~0.20 mA cm-2 was achieved for this photoanode.8 Other nanostructured electrodes were also
pursued in order to improve CS , such as CuWO4 nanoflakes.6,7 The improvement, however, has
been limited.
Nanostructured electrodes (e.g. nanorods as shown in Figure 5-1b) are typically
encouraged to overcome the opposing properties of a long light absorption depth and short
minority carrier collection length. As shown in Figure 5-1a, for a thin film electrode deposited on
a flat substrate, the direction of charge separation is the same as light absorption. Therefore, if the
thickness of the film is made close to the light absorption depth, which is assumed longer than the
collection length of the minority charge carrier (which are holes for n-type semiconductors), the
holes generated close to the back contact will not be able to reach the semiconductor/electrolyte
interface to realize water oxidation. In Figure 5-1b, however, for a nanostructured electrode, the
light is absorbed along the direction of the nanorods, while the holes transport in the direction
perpendicular to that of light absorption. To optimize light absorption and charge separation at the
same time, the length and width of the nanorods can be made the same as light penetration depth
and hole collection length, respectively. However, this electrode morphology is helpful only for
reducing the distance that holes need to travel. The photogenerated electrons still need to travel
through the length of the nanorods to reach the back contact.

130

Figure 5-1. Scheme of charge separation in a) planar CuWO4 electrode and b) optimized
nanostructured CuWO4 photoanode.

In practice, the improvement with nanostructuring in reported works has been limited,6,7
with photocurrents still close to those observed with our 80 nm thin films. One possible reason is
that the length and width of the CuWO4 nanoflakes are not optimized. The other possible reason
is that even if the dimensions of the nanoflakes had been optimized, the electrons might still have
a short collection length, therefore limiting the bulk property of the photoanodes, which is
suggested by previous works.

5,8

This indicates that although doping, nanostructuring, or even

tuning the substrate could potentially improve the bulk properties of the semiconductors in some
cases, they have to be rationally designed based on the knowledge of the material. Therefore, this
motivated us to carry out fundamental investigations with the bulk properties of CuWO4.
The first question to answer about bulk CuWO4 is whether the limiting charge carriers are
electrons or holes. According to previous literature, the bulk of CuWO4 is limited by its electronic
transport properties.5,8 For a semiconductor, the conductivity, σ, is given by:

  qn n  q p p
131

(5.1)

where  n and  p are the mobilities of the electrons and holes, respectively; q is the elementary
charge; n and p are the concentrations of the electrons and holes, respectively. For a n-type
semiconductor, n >> p. Therefore, if electrons are the limiting charge carrier,  n <<  p . This is a
rare situation but is possible as for n-type BiVO4 photoanodes reported in previous literature.10,11
However, the CuWO4 electrodes in previous works are nanostructured or mesoporous,5,8 which
makes it insufficient to justify the argument that electrons are limiting. Compact films are preferred
instead to conduct these fundamental investigations.
Once the limiting charge carrier is understood, the knowledge of its collection length would
be helpful to direct design of the nanostructured electrode. Because of a lack of knowledge of this
length, all of the nanostructuring work (including nanoflakes and coat MWCNT with CuWO4)
have only achieved moderate improvement. Therefore, it is crucial to quantify the hole collection
length of CuWO4.
5.3 Approaches and Methods
To answer the first question, whether electrons or holes are the limiting charge carriers,
the PEC performance of CuWO4 by illuminating from different directions will be evaluated. As
illustrated in Figure 5-2, photogenerated charge carriers always have a higher concentration in the
thin film closer to the side of illumination (holes are presented but electrons are generated together
with holes). When the electrode is illuminated from the electrolyte-electrode (front) side, more
charge carriers are generated closer to the electrode/electrolyte interface. Therefore, electrons have
to travel for a longer distance to reach the back contact than holes to reach the interface. The
situation is the opposite when illuminating the electrode from the substrate-electrode (back) side.
Therefore, by comparing the performance with different illumination directions, one should be
able to determine whether electrons or holes are the limiting charge carriers.
132

Figure 5-2. Schematic illustration of hole generation and charge separation in CuWO4 photoanode
illuminated from a) electrolyte-electrode (front) side and b) substrate-electrode (back) side. The
red arrows represent hole collection and green arrows represent electron transport.

There are several considerations about this method. Firstly, the electrode should be thick
enough to generate significant difference in charge carrier concentration throughout the film.
Generation profile of charge carriers can be calculated using the following equation:

G

dI
  I 0 e  x
dx

(5.2)

G(, x)   ( ) I 0 ( )e (  ) x

(5.3)

G(, x) / I 0 ( )   ( )e ( ) x

(5.4)

where G represents the generation rate of charge carriers; I and I0 represent the intensity of the
transmitted light at a certain distance (x) into the electrode and incident light, respectively; α is the
absorption coefficient; λ is the wavelength of the incident light. Based on the α we obtained with
CuWO4 by atomic layer deposition (ALD) as shown in chapter 2, the drop in G through an 80 nm
thick film with 405 nm monochromatic light is 18.2 %, and is 11.1 % with 470 nm light. This
change in charge carrier concentration throughout the film is too small to observe significant
133

difference in PEC performance by shining the electrode from different directions. We also did the
calculation for a 1 µm thick CuWO4 film assuming same α. The drop in G is 94.0 % under 405 nm
illumination and 73.1 % under 470 nm illumination. ALD is ideal for depositing ultrathin films
with thickness control down to the nm or even Å level, but is inefficient for depositing films of
several hundred nm to even µm. Therefore, we decided to use spray pyrolysis to synthesize our
CuWO4 film for its bulk studies.
Another consideration is the morphology of the film. As mentioned earlier, a
nanostructured or mesoporous film could lead to inaccurate results. Therefore, we decided to adopt
a method similar as what was used by the Lewis group12 to fill the pores in CuWO4 with Al2O3 by
ALD and then remove the excess Al2O3 layer at the top. As illustrated in Figure 5-3, instead of
mechanically polishing the top Al2O3 layer, we used 1 M KOH to etch the excess Al2O3 at the top
so that CuWO4 photoanode is exposed. With this treatment, the direction of hole and electron
transport will be along the light absorption direction, which will provide insight into identifying
the limiting charge carriers.

Figure 5-3. Scheme of a) a nanostructured CuWO4 film, b) a nanostructured film with pores filled
with excess Al2O3 and c) after etching the top Al2O3 layer with 1 M KOH to expose CuWO4
electrode to the electrolyte.

134

The third consideration is optical control. When using a regular glass PEC cell to perform
measurements, we found that the quartz window and the glass wall of the cell would increase the
intensity of incident light, while the fluorine doped tin oxide (FTO) coated glass substrate would
absorb and reflect some light, resulting in lower incident light on CuWO4 electrodes. It is difficult
to calibrate the actual incident light when illuminating from the front side because of the irregular
geometry of the glass cell. Therefore, for this study, we decided to use a CuWO 4 PEC sandwich
cell which helps optical control.
Once we have determined that the electrons are the limiting charge carriers, the other
question to answer is the electron collection length. Analysis of the wavelength dependent
behavior of incident photon-to-current efficiency (IPCE) would allow us to extract information
about the electron collection length of CuWO4. According to the scheme shown in Figure 5-4a,
when CuWO4 is illuminated from the front side, if the wavelength of the incident light is so short
that it is absorbed outside the electron collection region, there should be minimal photocurrent. As
the wavelength increases, a higher portion of absorbed photons will be able to contribute to
photocurrent. Therefore, we should expect an increase in IPCE with increasing wavelength.
However, once the wavelength is too long for CuWO4 to absorb all of the photons, IPCE will
decrease. Therefore, we should expect to see a peak in IPCE where the light absorption depth is
close to the film thickness. Typically, the light penetration depth is defined as the depth at which
the intensity of the radiation inside the material drops to 1/e (~37%) of its original value just
beneath the surface, which should be 1/α. However, here we use 3/α which represents the depth at
which 95% of the photons are absorbed. Figure 5-4b shows the scheme of illuminating CuWO4
from the back side with monochromatic light of different wavelengths. When the light absorption
depth is shorter than the electron collection length, Ln, IPCE should essentially remain the same

135

when increasing the wavelength. Once the penetration depth of light is longer than Ln, not all the
photogenerated electrons will be able to reach the FTO/CuWO4 back contact. Therefore, one
should expect to see a decrease in IPCE when increasing the wavelength further. Based on these
theories, Ln of CuWO4 could be estimated.

Figure 5-4. Schematic illustration of light absorption in CuWO4 electrode by a) front and b) back
illumination. Ln is the electron collection length. The light red, orange, green, blue and purple
arrows represent illumination from relatively long wavelengths to short wavelengths.

5.4 Experimental
5.4.1 Film Preparation
CuWO4 was deposited on FTO coated glass substrate by spray pyrolysis using the
procedure described in chapter 4. After deposition, the film was annealed in air at 550 °C for 1 hr
allowing for 1 hr to ramp up and cool down. 1500 cycles of Al2O3 (~150 nm) was deposited using
ALD (Savannah 200, Cambridge Nanotech Inc.) to fill the pores in CuWO4 film.
136

Trimethylaluminum (TMA, Aldrich) was used as the Al source and H2O was the oxidant. The
substrates were kept at 230 °C and the two precursors were kept at room temperature. In each cycle,
the Al precursor and H2O were alternatively pulsed for 0.015 s and kept under exposure mode for
20 s to allow sufficient diffusion of the precursor vapor into the pores. Between each pulse, the
chamber was pumped for 10 s to remove excess precursor and by-product. The excess top Al2O3
layer was removed by etching the film with 1 M KOH aqueous solution for 40 min.
5.4.2 Film Characterization
The thickness of Al2O3 layer is ~150 nm based on ellipsometric measurement (Horiba
Jobin Yvon, Smart-SE) with 1500 cycles of Al2O3 on Si substrate, assuming it has the same growth
rate on CuWO4. The morphology and thickness of CuWO4 electrode was examined by top-down
and cross-section scanning electron microscopy (SEM) (Carl Zeiss Microscopy). Absorption of
electrolytes was measured using a Perkin-Elmer Lambda 35 UV-vis spectrometer. Absorbance
measurements of thin films and integrated sandwich cells were performed with the same UV-vis
spectrometer with a Labsphere integrating sphere.
5.4.3 Sandwich Cell Preparation
A platinized FTO (Pt-FTO) electrode was used as the counter electrode. Prior to
platinization, the FTO substrate was sonicated in soap water, water, 0.1 M hydrochloric acid in
ethanol, and isopropanol (IPA) for 15 min, respectively. The Pt-FTO was made by drop-casting
15 µL of a 5 mM chloroplatinic acid hydrate (≥99.9% trace metals basis, Aldrich) solution in IPA
on a 2×2 cm2 FTO coated glass substrate followed by spreading out the Pt precursor, drying in a
petri dish with holes on its cover for 10 min, and annealing in air at 380 °C for 20 min. The CuWO4
working electrode and counter electrode were attached by sandwiching a 25 µm thick Surlyn frame
(Solaronix) by applying light pressure at 145 °C to seal the cell. Electrolytes consisting of 1 M

137

KBi buffer (pH 9), 200 mM KI and 5 mM I2 were introduced by capillary force through the two
pre-drilled holes on Pt-FTO counter electrode, which were subsequently sealed with microglass
and Surlyn film. The holes were positioned apart from the cell active area to avoid undesired light
loss when the light was illuminated from the counter electrode side.
5.4.4 Photoelectrochemistry
All electrochemical and photoelectrochemical measurements were made with an Eco
Chemie Autolab potentiostat coupled with Nova electrochemical software. Current density-voltage
(J-V) curves were measured using a scan rate of 20 mV s-1. The white light source was a 450 W
Xe arc lamp (Horiba Jobin Yvon). An AM 1.5 solar filter was used to simulate sunlight at 100 mW
cm-2 (1 sun). 1 M KBi buffer (pH 9) with or without 200 mM KI / 5 mM I2 or 0.5 M Na2SO3 was
used as the electrolyte. Photoelectrochemical measurements of thin film electrodes in aqueous
solutions were performed by shining light on the CuWO4 electrode through the substrate.
Photocurrent densities (Jphoto) were obtained by subtracting the dark current density (Jdark) from
the total current density under illumination (Jtotal). IPCE of CuWO4 sandwich cell was measured
at short circuit (corresponding to 1.03 V vs RHE in the three-electrode configuration) using the
same light source coupled with a grating monochromator with a 10 nm step. Monochromatic J-V
response was measured under the illumination of a 470 nm LED with an LED driver from Metrohm,
and the light intensity was calibrated using the current measured with a photodiode from the same
manufacturer.
5.5 Results and Discussion
Absorption of bare CuWO4 deposited on FTO substrate was measured using UV-vis and
the results are shown in Figure 5-5. The absorption coefficient in Figure 5-5a was calculated using
the equation described in a previous work.12 The thickness of the film was determined to be ~800

138

nm by cross-section SEM as shown in Figure 5-10b. The absorption coefficient is nonzero at
wavelengths longer than 539 nm due to the side scatter of light in this thick and mesoporous film.
An indirect bandgap of ~2.3 eV was obtained according to the Tauc plot in Figure 5-5b if the
plateau region < 2.3 eV is corrected as the baseline. This bandgap is close to CuWO4 synthesized
by other methods.3,12 It has been suggested in chapter 4 that the composition of our CuWO4 by
spray pyrolysis is pure crystalline CuWO4.

Figure 5-5. a) Absorption coefficient and b) Tauc plot of CuWO4 indicating its indirect bandgap.
The red dashed line represents the fitting result of the linear region in Tauc plot. The blue dash
represents the corrected baseline according to the plateau region in Tauc plot.
139

The absorptance of CuWO4 ( A% ) on FTO was calculated by using equation
A%  (1  T %  R%)  (1  Tsub %  Rsub %)  Tsub %  Rsub %  T %  R%

(5.5)

where T % and R % are the total transmittance and reflectance of CuWO4 thin film on FTO, and
Tsub % and Rsub % are the transmittance and reflectance of FTO substrate. According to its bandgap

of 2.3 eV, CuWO4 should in theory absorb up to 539 nm. However, as shown in Figure 5-6a, A%
around 550 nm is still ~20%. This is due to the side scattering of light of the relatively thick and
mesoporous film. Therefore, the baseline of A% was corrected by subtracting the apparent A% at
539 nm, which is demonstrated in Figure 5-6b.

Figure 5-6. a) Calculated raw data and b) corrected absorptance of CuWO4.

In this work, 200 mM KI / 5 mM I2 (I-/I3-) was used as the hole scavenger to eliminate
surface recombination in order to investigate the bulk properties of CuWO4 instead of Na2SO3
since a redox couple was needed in the sandwich cell. As shown in Figure 5-7, bare CuWO4
generates comparable Jphoto-V response in oxidizing I-/I3- redox couple as oxidizing Na2SO3, which
indicates that it works as a suitable hole scavenger for CuWO4. Another advantage of using I-/I3-

140

instead of Na2SO3 is that with I-/I3- CuWO4 shows minimal dark current in our interested potential
region (Figure 5-8b) while Na2SO3 shows significant dark current (Figure 5-8a) which makes the
interpretation of the results less straightforward.

Figure 5-7. Jphoto-V response of CuWO4 in a pH 9 KBi electrolyte with 0.5 M Na2SO3 (blue dash)
and 200 mM KI / 5 mM I2 (red solid) under 1 sun illumination.

Figure 5-8. Dark J-V curves of CuWO4 measured in a) Na2SO3 and b) I-/I3- electrolyte.

141

Charge separation efficiency ( CS ) of CuWO4 under 1 sun illumination was calculated
using equation:

CS  J scavenger / J abs

(5.5)

J abs  q  539
330 ( ) A%d 

(5.6)

where J scavenger and J abs are the photocurrent with a hole scavenger and the maximum
photocurrent if all the photogenerated charge carriers are collected; q is the elementary charge;
( ) is the flux of incident light at a certain wavelength (  ). The calculated result is shown in

Figure 5-9. CuWO4 shows a poor CS which is 7.5% at 1.23 V vs RHE and < 5% at 1.03 V vs
RHE. This further confirms that the major limitation of CuWO4 is its poor bulk properties.

Figure 5-9. Charge separation efficiency of CuWO4 electrode.

The top-down morphology of the mesoporous CuWO4 synthesized by spray pyrolysis is

142

shown in Figure 5-10a with the pore size up to ~ 150 nm. Therefore, 1500 cycles (~150 nm) of
Al2O3 was deposited by ALD to fill these pores. After deposition, the Al2O3 insulates the contact
between CuWO4 electrode and I-/I3- electrolyte, as the back dashed J-V curve shown in Figure 511, which shows no photocurrent. The coated electrode was further etched in 1 M KOH aqueous
solution for different lengths of time, and the time dependent J-V curves under 1 sun illumination
are shown in Figure 5-11. From 30 min to 38 min, the photocurrent slowly increases, which is
associated with gradual removal of Al2O3. However, the photocurrent is low possibly due to only
islands of CuWO4 that are exposed. After 40 min of etching, the photocurrent increases drastically
but is still lower than that of bare CuWO4, indicating full removal of top Al2O3 without damaging
the Al2O3 that are in the pores. After this, the photocurrent increases even higher until it reaches
the same photocurrent as the bare electrode after 44 min of etching. This suggests that the Al2O3
in the pores are also removed. Therefore, all the electrodes with Al2O3 were etched for 40 min for
bulk studies. The morphology of these electrodes after Al2O3 deposition and KOH etching
treatment was examined using SEM. As shown in Figure 5-10c, the pores between grains of
CuWO4 are filled with Al2O3, and CuWO4 is exposed at the surface. In the cross-section SEM
image in Figure 5-10d, the thin film is compact with a thickness of ~800 nm.

143

Figure 5-10. SEM image of a) top-down and b) cross-section view of bare CuWO4 by spray
pyrolysis, c) top-down and d) cross-section view of CuWO4 deposited with 1500 cycles (~150 nm)
of Al2O3 and etched in 1 M KOH for 40 min.

144

Figure 5-11. J-V curves of CuWO4 coated with ~150 nm of Al2O3 (black dash) and etched in 1 M
KOH for 30 min (red), 32 min (orange), 34 min (yellow), 36 min (green), 38 min (dark cyan), 40
min (blue), 42 min (purple) and 44 min (wine). These J-V curves are also compared with that of
bare CuWO4 (pink dots). All measurements are done in pH 9 KBi electrolyte under 1 sun
illumination.

The post-treated CuWO4 electrodes were then made into sandwich cells. As a consideration
for optical control, when light is illuminated through the working electrode (back) side, the light
penetrates through the FTO substrate and then CuWO4 electrode as shown in Figure 5-12.
However, when the light is illuminated through the counter electrode side, it goes through Pt-FTO
and then electrolyte before it reaches CuWO4. Therefore, the actual light incident on CuWO4 as to
be calibrated. We firstly examined the absorption of Pt-FTO compared with bare FTO. The
absorptance calculated for FTO and Pt-FTO in Figure 5-13 shows that their absorption is close in
our interested wavelength region. Therefore, the absorption of Pt is negligible. The absorption of
the electrolyte was also analyzed. The major absorbing species in the electrolyte is I2. Therefore,
145

different concentrations of I2 was dissolved in pH 9 KBi buffer with 200 mM KI to examine its
extinction coefficient. The absorbance as a function of I2 concentration at different wavelengths
are presented in Figure 5-14. Based on the length of solution of 1 cm, the extinction coefficients
were calculated to be 61.3 (±0.2), 12.38 (±0.06), and 1.70 (±0.02) mM-1 cm-1 at 350nm, 405nm,
and 470nm. Since the thickness of the electrolyte inside the sandwich cell is 25 µm (thickness of
Surlyn film), the absorptance of I2 in the cell should be 54%, 14% and 2%. Therefore, it is
necessary to consider the absorption of the electrolyte, especially at shorter wavelengths.

Figure 5-12. Schematic illustration of CuWO4 sandwich cell and light illumination.

146

Figure 5-13. Absorptance of FTO-coated glass (black solid) and platinized FTO-coated glass
(blue dash). Both were measured by illuminating from the glass side.

147

Figure 5-14. Absorbance of different concentrations of I2 in pH9 KBi buffer with 200 mM KI at
a) 350 nm, b) 405 nm and c) 470nm.

The absorptance of the integrated cell with (red solid) and without (black dash) electrolyte
is compared by illuminating from the front side to examine the absorption of the electrolyte. As
shown in Figure 5-15a, when the cell is filled with electrolyte, the total absorptance is slightly
higher than without electrolyte, indicating the non-negligible absorption of the electrolyte species
when shining light from the front side. By substracting the absorptance of the unfilled cell from
the filled cell, the absorptance of the electrolyte was calculated and presented in Figure 5-15b.
Therefore, for monochromatic PEC measurements, the incident light intensity was corrected by
148

subtracting the absorptance of FTO substrate when illuminating from the back side, and by
subtracting the absorptance of FTO substrate and electrolyte when illuminating from the front side.

Figure 5-15. a) Absorptance of the integrated sandwich solar cell with (red solid) and without
(black dash) electrolyte by illuminating from the front side. b) Absorptance of the electrolyte inside
the sandwich cell.

The J-V response of the sandwich cell with post-treated CuWO4 under AM 1.5 1 sun
illumination are shown in Figure 5-16. The Jphoto-V curves are shown in Figure 5-16b by
subtracting dark current from the total current. As clearly shown in Figure 5-16b, the photocurrent
generated by back illumination is higher than front illumination, as was observed in previous
literature5,8. However, the intensity of the incident light was not corrected. Therefore, we further
analyzed the J-V behavior of this cell under monochromatic light with optical control.

149

Figure 5-16. a) J-V response of CuWO4 sandwich by illuminating from front (red dash) and back
(blue solid) side under AM 1.5 1 sun illumination and in the dark (black dots). b) Jphoto-V of CuWO4
sandwich cell illuminated from front (red dash) and back (blue solid) side.

In order to better quantify the photocurrent by correcting the light incident on CuWO4
electrode, monochromatic measurements were conducted. The raw data of J-V response are shown
in Figure 5-17a, and the corrected Jphoto-V and quantum efficiency (QE) curves are shown in Figure
5-17b and c. The corrected photocurrent is calculated using these equations:
J photo ( front , corrected )  J photo ( front , raw) / (1  AFTO %  AE %)

(5.7)

J photo (back , corrected )  J photo (back , raw) / (1  AFTO %)

(5.8)

where AFTO % and AE % are the absorptance of FTO substrate and electrolyte. QE was calculated
using equation:
QE  J photo (corrected ) / J flux

(5.9)

where J flux is the maximum possible photocurrent generated by the incident light. It was observed
that after optical correction, the photocurrent generated by back illumination appears higher than
that by front illumination. This indicates that for our compact CuWO4 with post treatment, the
150

electrons are the limiting charge carriers.

Figure 5-17. a) J-V response, b) Jphoto-V response and c) quantum efficiency of CuWO4 sandwich
cell with front (red dash) and back (blue solid) illumination and in the dark (black dots). Light
source was 470 nm LED with 218 mW cm-2 power density.

IPCE of CuWO4 sandwich cell was measured at short circuit (V=0 V vs Vs) under
monochromatic light in its bandgap range. The un-corrected IPCE data are shown in Figure 5-18a.
The IPCE data under 330 nm were eliminated due to the influence of the strong absorption of FTO

151

substrate at shorter wavelengths. The IPCE values were corrected by using equation (5.9), and are
shown in Figure 5-18b. Similar as what was observed in the monochromatic J-V response at 470
nm, IPCE is higher when the cell is illuminated from the back side. It was also noticed that QE
values at short circuit calculated in Figure 5-17c agree with the IPCE values in Figure 5-18b at
470 nm (3% for front illumination and 4% for back illumination) within reasonable experimental
error.

Figure 5-18. a) Original data and b) corrected IPCE of CuWO4 sandwich cell under front (red
circles) and back (blue triangles) illumination.

The IPCE results were further analyzed to extract information about the electron collection
length of CuWO4. As shown in Figure 5-18b, IPCE by front illumination shows a small peak at
~390 nm. Using the absorption coefficient that we calculated in Figure 5-5a, this wavelength
corresponds to a 3/α value of 947 nm, which is somewhat close to the ~800 nm thickness we
measured with SEM. IPCE by back illumination shows a monotonous decrease when increasing
the wavelength. This indicates that Ln should be shorter than the penetration depth (3/α) of 330 nm
light, which is 577 nm. However, this is the upper limit of electron collection length in our CuWO4
152

electrode. Since a plateau region is never observed in the back illuminated IPCE data before it
comes to a decrease, the actual electron collection length could be way smaller than this number.
The absorbed photon-to-current efficiency (APCE) can be calculated using equation:

APCE 

IPCE
A%

(5.10)

where A% is the absorptance calculated in Figure 5-6b. As shown in Figure 5-19, the low APCE
value indicates that the electron collection length is indeed much shorter than 577 nm.

Figure 5-19. Absorbed photon-to-current efficiency of CuWO4 sandwich cell under front (red
circles) and back (blue triangles) illumination.

5.6 Conclusions
In this chapter, it was shown that apart from non-optimized hole collection efficiency at
low applied bias due to recombination of surface-stored holes with conduction band electrons, the

153

major limitation of CuWO4 for water oxidation is its low charge separation efficiency. Mesoporous
CuWO4 electrodes with ~800 nm thickness were made by spray pyrolysis in order to investigate
its fundamental bulk properties. Al2O3 was deposited by ALD to fill the pores of CuWO4 in order
to understand its limiting charge carriers and relevant collection length. By comparing the PEC
performance with front and back illumination, it was determined that the electrons are the limiting
charge carriers of CuWO4. By analyzing the wavelength dependence of IPCE, the electron
collection length of CuWO4 was semi-quantified, which is shorter than 577 nm. This explains the
reason why thicker nanostructured CuWO4 electrodes (2 to 3 µm) has similar performance as
thinner electrodes (80 to 200 nm).1,3,4,13 However, further systematic thickness dependent studies
should be conducted to extract more accurate electron collection length in order to optimize the
film thickness. Meanwhile, analogous measurements should be performed with single crystal
CuWO4 to understand its intrinsic properties such as limiting charge carriers, charge carrier
mobility and diffusion length. These investigations will in theory provide some clue of overcoming
the opposing properties of weak light absorption and short electron collection length by depositing
CuWO4 thin film of a suitable thickness on a nanostructured transparent conductive oxide (TCO)
substrate. Other strategies such as doping or applying an under layer could also potentially improve
the charge separation efficiency.

154

REFERENCES

155

REFERENCES

(1)

Gao, Y.; Zandi, O.; Hamann, T. W. Atomic Layer Stack Deposition-Annealing Synthesis
of CuWO4. J. Mater. Chem. A 2016, 4, 2826–2830.

(2)

Pyper, K. J.; Yourey, J. E.; Bartlett, B. M. Reactivity of CuWO4 in Photoelectrochemical
Water Oxidation Is Dictated by a Midgap Electronic State. J. Phys. Chem. C 2013, 117,
24726–24732.

(3)

Yourey, J. E.; Bartlett, B. M. Electrochemical Deposition and Photoelectrochemistry of
CuWO4, a Promising Photoanode for Water Oxidation. J. Mater. Chem. 2011, 21, 7651–
7660.

(4)

Hill, J. C.; Choi, K. S. Synthesis and Characterization of High Surface Area CuWO4 and
Bi2WO6 Electrodes for Use as Photoanodes for Solar Water Oxidation. J. Mater. Chem. A
2013, 1, 5006–5014.

(5)

Bohra, D.; Smith, W. A. Improved Charge Separation via Fe-Doping of Copper Tungstate
Photoanodes. Phys. Chem. Chem. Phys. 2015, 17, 9857–9866.

(6)

Hu, D.; Diao, P.; Xu, D.; Xia, M.; Gu, Y.; Wu, Q.; Li, C.; Yang, S. Copper (II) Tungstate
Nanoflake Array Films: Sacrificial Template Synthesis, Hydrogen Treatment, and Their
Application as Photoanodes in Solar Water Splitting. Nanoscale 2016, 8, 5892–5901.

(7)

Ye, W.; Chen, F.; Zhao, F.; Han, N.; Li, Y. CuWO4 Nanoflake Array-Based SingleJunction and Heterojunction Photoanodes for Photoelectrochemical Water Oxidation. ACS
Appl. Mater. Interfaces 2016, 8, 9211–9217.

(8)

Gaillard, N.; Chang, Y.; Deangelis, A.; Higgins, S.; Braun, A. A Nanocomposite
Photoelectrode Made of 2.2 eV Band Gap Copper Tungstate (CuWO4) and Multi-Wall
Carbon Nanotubes for Solar-Assisted Water Splitting. Int. J. Hydrogen Energy 2013, 38,
3166–3176.

(9)

Zhang, H.; Yilmaz, P.; Ansari, J. O.; Khan, F. F.; Binions, R.; Krause, S.; Dunn, S.
Incorporation of Ag Nanowires in CuWO4 for Improved Visible Light-Induced
Photoanode Performance. J. Mater. Chem. A 2015, 3, 9638–9644.

(10)

Abdi, F. F.; Van De Krol, R. Nature and Light Dependence of Bulk Recombination in CoPi-Catalyzed BiVO4 Photoanodes. J. Phys. Chem. C 2012, 116, 9398–9404.

(11)

Park, Y.; McDonald, K. J.; Choi, K.-S. Progress in Bismuth Vanadate Photoanodes for
Use in Solar Water Oxidation. Chem. Soc. Rev. 2013, 42, 2321–2337.

(12)

Pala, R. a.; Leenheer, A. J.; Lichterman, M.; Atwater, H. a.; Lewis, N. S. Measurement of
156

Minority-Carrier Diffusion Lengths Using Wedge-Shaped Semiconductor
Photoelectrodes. Energy Environ. Sci. 2014, 7 (10), 3424–3430.
(13)

Lhermitte, C. R.; Bartlett, B. M. Advancing the Chemistry of CuWO4 for
Photoelectrochemical Water Oxidation. Acc. Chem. Res. 2016, 49, 1121–1129.

157

Chapter 6:
Conclusions and Future Directions

158

6.1 Conclusions
As discussed in chapter 1, as a promising semiconductor photoanode material for
photoelectrochemical (PEC) water splitting, CuWO4 has received moderate attention partially due
to its limited water oxidation performance. Only a handful of works have made effort to investigate
the limitations of CuWO4 and relevant water oxidation mechanisms with this material in order to
provide strategies to improve its efficiency and to understand general rules of water oxidation with
semiconductor materials. Systematic studies were thus performed in this dissertation in order to
better understand water oxidation with CuWO4.
1. A new atomic layer deposition (ALD) based stack deposition-annealing (SDA) method
was applied to deposit planar and uniform CuWO4 thin films with controllable thickness
(chapter 1).1 The resulted thin films show linear growth with increasing deposition cycles,
which is essentially one benefit of ALD. Compared with the widely used traditional ALD
route to deposit ternary or quaternary metal oxides, this SDA method has various
advantages: 1) the ALD temperature windows of the binary oxides do not have to overlap;
2) the problem of surface-inhibited growth of some materials (e.g. WO3 on CuO) is solved;
3) there is an easier control of stoichiometry if the growth rates of the binary oxides on
their relevant substrates are known. The resulted CuWO4 electrodes show comparable
water oxidation performance compared with other synthetic methods reported in previous
literature, suggesting the promise of this method.
2. CuWO4 electrodes were evaluated in contact with H2O, Na2SO3, and H2O2 (chapter 2)
by

photoelectrochemical

measurements

and

intensity

modulated

photocurrent

spectroscopy (IMPS) in order to understand its hole collection efficiency (  HC ).2 It was
shown that Na2SO3 is a suitable hole scavenger with quantitative hole collection at CuWO4

159

surface to quantify water oxidation  HC with CuWO4. Meanwhile, H2O2, which was used
in previous literature as a hole scavenger for CuWO4,3,4 has current doubling effect which
makes it problematic for quantifying  HC of water oxidation with CuWO4. With Na2SO3
as the suitable hole scavenger and using IMPS, identical results were obtained within
reasonable error for  HC from two different methods. CuWO4 shows non-optimal  HC at
low applied bias due to surface recombination resulted from insufficient band bending.
However, at higher potentials,  HC is essentially unity, suggesting quantitative hole
collection once it has passed the onset potential region.
3. Electrochemical impedance spectroscopy (EIS), cyclic voltammetry (CV) surface
measurements and photocurrent transient measurements were employed under different
conditions of applied potential, illumination, and electrolytes to understand the role of
surface states of CuWO4 in water oxidation (chapter 3).5 It was shown that a surface state
capacitance coincides with the turn-on of water oxidation, which is only available under
photoelectrochemical water oxidation conditions. This capacitance is correlated to the
charging of surface state measured from CV surface measurements, which is only
observable in aqueous electrolyte but absent in non-aqueous solvents. It was therefore
concluded that the surface state of CuWO4 is a water oxidation intermediate species.
4. Apart from surface recombination at low applied potentials, the major limitation of
CuWO4 for PEC water oxidation is its low CS (chapter 5). The PEC performance of a
sandwich cell made of a thick and compact CuWO4 film with pores filled with inert Al2O3
was examined by illuminating from different directions. It was concluded that electrons are
the limiting charge carriers of CuWO4. Incident photon-to-current efficiency (IPCE) was
further measured to semi-quantify the collection length of electrons in CuWO4. It was
160

determined that the electron collection length is smaller than 577 nm for our CuWO4 film.
These results partially explains why thicker nanostructured CuWO4 electrodes (2 to 3 µm)
show similar performance as thinner electrodes (80 to 200 nm).1,6–8
6.2 Future Directions
The surface state of CuWO4 has been identified as a water oxidation intermediate species
based on our complementary photoelectrochemical and EIS measurements. We hypothesize that
charging the surface state of CuWO4 with holes constitutes of the first oxidative step of water
oxidation. However, there is no direct evidence of this intermediate species. In attempts to identify
these species, operando ATR-IR measurements will be used in the future with CuWO4 electrodes
in contact with a D2O electrolyte in response to applied bias and illumination.9 Possible
intermediates that can potentially be detected are either oxo or peroxo groups such as W=O, W–
O–O–D, W–O–O–W, W–O–O–Cu, or Cu–O–O–Cu. We don’t expect a Cu oxo species to exist
since Cu is beyond the oxo wall. Oxygen isotope labeling will be utilized to distinguish between
oxo and peroxo species.9 There is a large body of literature that show W=O, W-O-W or W-O2 IR
vibrational peaks which will facilitate our determination of water oxidation intermediates (or
defects) with the proposed IR measurements.10–14 Isotope labeling (H vs D) will also be used to try
and distinguish M–O–O–M from M–O–O–D/H species. We note that it will be difficult to clearly
distinguish species such as Cu–O–O–D and W–O–O–D. We plan to address this by preparing
samples with W and Cu rich surfaces, which can be controlled by the subsequent ALD of the
desired metal oxide on the electrode surface which will help elucidate the small difference in
vibration expected. We will also perform complementary in-situ UV-vis spectroelectrochemistry
measurements, analogous to our prior measurements on hematite, where the absorption spectra
can help differentiate whether the hole is trapped as a Cu or W based intermediate.15

161

From these measurements, we expect to definitively determine whether the surface state is
actually a water oxidation intermediate species as we hypothesize. We will further gain
information about the nature of the surface state (acknowledging exact species may be hard to
resolve as noted above). This information is important in developing a mechanistic picture of water
oxidation on CuWO4 as well as obtaining a general understanding of performance controlling
processes of photo-induced water oxidation on semiconductor electrodes.
It was suggest by our work that the water oxidation performance of CuWO4 is limited by
surface recombination at low applied potentials. Therefore, the photocurrent onset potential in
theory can be reduced by applying a water oxidation catalyst. By comparing the onset potential of
water oxidation and oxidation of Na2SO3, one should expect a reduction in water oxidation onset
potential by 200 mV when a suitable catalyst is applied.2 However, no catalyst so far can reduce
the onset potential significantly.8,16 The effect of water oxidation catalyst on CuWO4 will be
investigated by integrating PEC, EIS, IMPS and operando spectroscopic measurements.
In chapter 5, it was suggested that apart from surface recombination at low applied bias,
the low CS is the major limitation of CuWO4. Since the measurements were taken with
polycrystalline CuWO4, the bulk properties might rely on the synthetic method. It would be
interesting if analogous measurements could be taken with single crystal CuWO4 electrodes in
order to get some intrinsic parameters such as charge carrier mobilities, diffusion length, etc. of
CuWO4. Thickness dependent measurement should also be performed to get more accurate value
for the electron collection length.
Interesting strategies such as intentional doping and using suitable substrates or underlayer
may benefit improving CS . Crystal orientation-dependent charge transport would also be
interesting to investigate in order to optimize its bulk and surface properties. Once the bulk and

162

surface properties are optimized, CuWO4 thin film with suitable thickness can be scaled up on
nanostructured transparent conductive oxide (TCO) substrates to overcome the opposing
properties of weak light absorption and limited electron collection length. Given the suitable
properties of bandgap, band positions, stability and earth abundance of CuWO4, as well as our
comprehensive understanding of its limiting factors in PEC water oxidation, CuWO4 has the
promise of enabling cost-effective and durable PEC water splitting in a tandem cell.

163

REFERENCES

164

REFERENCES

(1)

Gao, Y.; Zandi, O.; Hamann, T. W. Atomic Layer Stack Deposition-Annealing Synthesis
of CuWO4. J. Mater. Chem. A 2016, 4, 2826–2830.

(2)

Gao, Y.; Hamann, T. W. Quantitative Hole Collection for Photoelectrochemical Water
Oxidation with CuWO4. Chem. Commun. 2017, 53, 1285–1288.

(3)

Bohra, D.; Smith, W. A. Improved Charge Separation via Fe-Doping of Copper Tungstate
Photoanodes. Phys. Chem. Chem. Phys. 2015, 17, 9857–9866.

(4)

Hill, J. C.; Choi, K. Synthesis and Characterization of High Surface Area CuWO4 and
Bi2WO6 Electrodes for Use as Photoanodes for Solar Water Oxidation. J. Mater. Chem. A
2013, 1, 5006–5014.

(5)

Gao, Y.; Hamann, T. W. Elucidation of CuWO4 Surface States
Photoelectrochemical Water Oxidation. J. Phys. Chem. Lett. 2017, 8, 2700-2704.

(6)

Yourey, J. E.; Bartlett, B. M. Electrochemical Deposition and Photoelectrochemistry of
CuWO4, a Promising Photoanode for Water Oxidation. J. Mater. Chem. 2011, 21, 7651–
7660.

(7)

Hill, J. C.; Choi, K. S. Synthesis and Characterization of High Surface Area CuWO 4 and
Bi2WO6 Electrodes for Use as Photoanodes for Solar Water Oxidation. J. Mater. Chem. A
2013, 1, 5006–5014.

(8)

Lhermitte, C. R.; Bartlett, B. M. Advancing the Chemistry of CuWO4 for
Photoelectrochemical Water Oxidation. Acc. Chem. Res. 2016, 49, 1121–1129.

(9)

Zandi, O.; Hamann, T. W. Determination of Photoelectrochemical Water Oxidation
Intermediates on Haematite Electrode Surfaces Using Operando Infrared Spectroscopy. Nat.
Chem. 2016, 8, 778–783.

During

(10) Rice, C. A.; Kroneck, P. M. H.; Spence, J. T. Tungsten(V)-Oxo and Tungsten(VI)-Dioxo
Complexes with Oxygen, Nitrogen, and Sulfur Ligands. Electrochemical, Infrared, and
Electron Paramagnetic Resonance Studies. Inoganic Chem. 1981, 20 (7), 1996–2000.
(11) Sugimoto, H.; Tano, H.; Tajima, R.; Miyake, H.; Tsukube, H.; Ohi, H.; Itoh, S. In Situ
Generation of Oxo-sulfidobis(dithiolene)tungsten(VI) Complexes: Active-Site Models for
the Aldehyde Ferredoxin Oxidoreductase Family of Tungsten Enzymes. Inorg. Chem. 2007,
46 (21), 8460–8462.
(12) Zaki, M. I.; Mekhemer, G. A. H.; Fouad, N. E.; Rabee, A. I. M. Structure-Acidity
Correlation of Supported tungsten(VI)-Oxo-Species: FT-IR and TPD Studies of Adsorbed
165

Pyridine and Catalytic Decomposition of 2-Propanol. Appl. Surf. Sci. 2014, 308, 380–387.
(13) Hazarika, P.; Kalita, D.; Sarmah, S.; Islam, N. S. New Oxo-Bridged Peroxotungsten
Complexes Containing Biogenic Co-Ligand as Potent Inhibitors of Alkaline Phosphatase
Activity. Mol. Cell. Biochem. 2006, 284, 39–47.
(14) Aleeva, G. P.; Kotorlenko, L. A.; V. A. Lunenok-Burmakina. Study of the Structure of
Peroxo Compounds of Tungsten and Molybdenum by Infrared Spectroscopy. J. Struct.
Chem. 1972, 13 (4), 590–593.
(15) Klahr, B.; Hamann, T. Water Oxidation on Hematite Photoelectrodes: Insight into the
Nature of Surface States through in Situ Spectroelectrochemistry. J. Phys. Chem. C 2014,
118, 10393–10399.
(16) Nam, K. M.; Cheon, E. A.; Shin, W. J.; Bard, A. J. Improved Photoelectrochemical Water
Oxidation by the WO3/CuWO4 Composite with a Manganese Phosphate Electrocatalyst.
Langmuir 2015, 31 (39), 10897–10903.

166