Tesis Maria Isabel Diez Garcia
Cargado por
maximecontreras2Copyright:
Formatos disponibles
Tesis Maria Isabel Diez Garcia
Cargado por
maximecontreras2Derechos de autor
Formatos disponibles
Compartir este documento
Compartir o incrustar documentos
¿Le pareció útil este documento?
¿Este contenido es inapropiado?
Copyright:
Formatos disponibles
Tesis Maria Isabel Diez Garcia
Cargado por
maximecontreras2Copyright:
Formatos disponibles
PREPARATION AND STUDY OF TERNARY METAL OXIDE PHOTOCATHODES FOR SOLAR
ENERGY CONVERSION
María Isabel Díez García
DEPARTAMENTO DE QUÍMICA FÍSICA
FACULTAD DE CIENCIAS
PREPARATION AND STUDY OF TERNARY METAL OXIDE PHOTOCATHODES FOR SOLAR
ENERGY CONVERSION
MARÍA ISABEL DÍEZ GARCÍA
Tesis presentada para aspirar al grado de
DOCTORA POR LA UNIVERSIDAD DE ALICANTE
MENCIÓN DE DOCTORA INTERNACIONAL DEL PROGRAMA DE DOCTORADO EN
CIENCIA DE MATERIALES
Dirigida por:
ROBERTO GÓMEZ TORREGROSA
D. ROBERTO GÓMEZ TORREGROSA, catedrático de Química Física de la
Universidad de Alicante
CERTIFICA:
Que el trabajo con título “Preparation and study of ternary metal oxide
photocathodes for solar energy conversion”, presentado por María Isabel Díez
García, ingeniera química, para aspirar al grado de Doctora dentro del programa de
doctorado de Ciencia de Materiales, ha sido realizado en el Departamento de
Química Física de la Universidad de Alicante bajo mi supervisión.
Y para que así conste a los efectos oportunos, se firma el presente certificado en
Alicante, a 29 de noviembre de 2017.
Roberto Gómez Torregrosa
Dña. TERESA LANA VILLARREAL, secretaria del Departamento de Química
Física de la Universidad de Alicante
CERTIFICA:
Que la presente memoria “Preparation and study of ternary metal oxide
photocathodes for solar energy conversion”, presentada por María Isabel Díez
García para aspirar al grado de Doctora, ha sido realizada en el Departamento de
Química Física de la Universidad de Alicante, bajo la dirección del Dr. Roberto
Gómez Torregrosa.
Y para que así conste a los efectos oportunos, se firma el presente certificado en
Alicante, a 29 de noviembre de 2017.
Teresa Lana Villarreal
A mis padres,
"Hay que perseverar y, sobre todo, tener confianza en uno mismo"
Marie Curie
AGRADECIMIENTOS
En la elaboración de este trabajo, además de mucha dedicación y medios técnicos,
existe un gran grupo humano que te acompaña a lo largo de este camino. Para culminar
esta tesis me gustaría agradecer a todas aquellas personas que de una forma u otra han
estado presentes durante este periodo tan importante de mi vida.
En primer lugar, me gustaría agradecer a mi director de tesis, Roberto Gómez
Torregrosa, el haberme brindado la oportunidad de unirme a su grupo y empezar mi
investigación en un tema tan interesante como es la conversión de energía solar, así
como su gran ayuda durante estos años tanto en los aspectos teóricos como técnicos,
incluyendo la elaboración de esta memoria. Igualmente, el que haya intentado sacar lo
mejor de mí y de mi investigación.
Gracias a Teresa Lana, que también ha estado muy presente durante este trabajo, por
haberme instruido en el laboratorio y fuera de él, y por haber estado dispuesta a ayudar
cuando la he necesitado. De igual forma quiero agradecer a Pedro Bonete por su apoyo
y por haberme guiado durante mi trabajo fin de máster.
Agradezco también a todos mis compañeros del grupo de investigación. A Débora y
Dejan por todos los momentos compartidos y por hacer el trabajo más ameno, a Ana
Korina con la que he compartido muchos días de laboratorio. Igualmente, a Ainhoa,
Javi, Néstor M., Fran P., Carmen y Fran N. por crear un ambiente de trabajo tan
agradable. Por supuesto, a los ya doctores Irene, Milena, Néstor G., que han sido una
buena referencia a seguir, por sus consejos y los buenos ratos que me han brindado.
Aparte de las personas de mi grupo, a Paula y Nacho, y también a todos mis
compañeros de sala de trabajo.
Quiero agradecer en general a todo el Departamento de Química Física en el que he
conocido a muchas personas durante todos estos años. A las secretarias del
departamento Violeta y Marga. Al personal de los servicios técnicos, por intentar sacar
el mejor partido a los análisis, en especial a Cristina. También a Verónica S., por estar
siempre dispuesta a ayudar.
De mi estancia en Bristol guardo buenos recuerdos. Quiero agradecer a David J. Fermín
y Verónica C. la buena acogida que me dieron en su grupo, realmente me hicieron sentir
parte de su equipo. También quiero recordar de esta etapa a Thomas, Jo y Daniela.
Quisiera hacer una mención especial al profesor José González, a quién agradezco
haberme permitido colaborar en su grupo en mis comienzos en investigación.
Igualmente quiero destacar a Desi, quien estuvo a mi lado en mis primeros
experimentos y a Nacho, con quien también compartí laboratorio.
A Blanca, amiga y compañera de carrera desde el primer día, por transmitirme su
positividad y por sus ánimos durante todo este tiempo. Y por supuesto, a mis padres, y a
mi hermano por su incondicional apoyo, por estar siempre ahí cuando les he necesitado.
Sin duda, los valores que me han transmitido han hecho posible este trabajo. También
otras personas, que desde fuera de la universidad me han acompañado en esta etapa.
OBJECTIVES
This PhD thesis mainly focuses on the preparation and study of ternary oxide materials
as photoelectrodes for photoelectrochemical water splitting cells together with the
investigation of different strategies for improving their photoresponse. An exhaustive
physical and photoelectrochemical characterization of the materials is also pursued.
The objectives of this thesis can be summarized as:
To prepare and study ternary oxide materials for water splitting devices,
particularly those having p-type character. For this goal, different synthetic
routes are investigated to obtain pure phase materials and different electrode
architectures are examined, focusing mainly on compact and nanoparticulate
electrodes.
To accomplish a comprehensive characterization of ternary oxide electrodes by
different spectroscopic and microscopic techniques (XRD, electron microscopy,
XPS, UV/vis spectroscopy, etc.) with the aim of examining the material phase,
morphology, surface composition, light absorption properties, etc.
To characterize the synthesized ternary oxide electrodes by using
(photo)electrochemical methods, such as cyclic voltammetry,
chronoamperometry and linear sweep voltammetry under transient illumination.
The goal is the evaluation of both the stability and the potential efficiency for
hydrogen evolution (or oxygen evolution in case of n-type semiconductors) of
the photoelectrodes.
To investigate different strategies for improving the photoelectrochemical
efficiency for hydrogen evolution (or oxygen evolution in case of n-type
semiconductors) of the ternary oxide materials, such as electrochemical doping,
metal doping, introducing charge extracting layers, or cocatalysts.
To gain insight into the photoelectrochemical H2 evolution reaction by means of
electrochemical techniques, especially by using photoelectrochemical
impedance spectroscopy (PEIS) to analyze the factors limiting the
photoelectrochemical response.
To evaluate the prospect of the implementation of the synthesized
photoelectrodes in photoelectrochemical devices for artificial photosynthesis.
TABLE OF CONTENTS
Chapter 1. Introduction ........................................................................... 1
1.1. The global energetic problem. Photoelectrochemical water splitting as a clean
alternative for H2 production .............................................................................................3
1.2. Basics of semiconductors physics ..........................................................................7
1.2.1. Density of states and charge carrier density ....................................................8
1.2.2. Surface states .................................................................................................14
1.2.3. Light absorption by semiconductors .............................................................15
1.3. The semiconductor-liquid interface ......................................................................18
1.3.1. Thermodynamics of semiconductor/electrolyte interfaces ...........................18
1.3.2. Double layer at a semiconductor-liquid interface .........................................22
1.3.3. Application of a potential ..............................................................................27
1.3.4. Illuminated semiconductor/electrolyte interface ...........................................29
1.3.5. Carrier collection under illumination ............................................................33
1.4. Nanocrystalline electrodes ....................................................................................35
1.4.1. Chemical capacitance of nanostructured semiconductor films .....................35
1.4.2. Carrier collection under illumination ............................................................39
1.5. Photoelectrochemical cells ...................................................................................42
1.5.1. Photoelectrochemical cell configurations .....................................................43
1.5.2. Photocathode requirements and challenges ...................................................46
1.6. Ternary oxides as photocathodes..........................................................................48
1.6.1. Spinels ...........................................................................................................48
1.6.2. Delafossites ...................................................................................................51
1.6.3. Perovskites ....................................................................................................54
1.6.4. Other ternary oxides ......................................................................................55
1.7. Strategies for improving photoelectrochemical performance ..............................58
1.7.1. Improving light absorption by doping ...........................................................60
1.7.2. Improving separation of charge carriers by heterojunctions .........................61
1.7.3. Improving charge collection by doping ........................................................64
1.7.4. Decreasing surface recombination by passivation layers ..............................67
1.7.5. Improving interfacial reaction kinetics by cocatalysts ..................................68
1.8. Bibliography .........................................................................................................73
i
Chapter 2. Experimental methods......................................................... 89
2.1. Preparation of ternary oxide electrodes ................................................................91
2.1.1. Methods based on pre-synthesized particles .................................................93
2.1.2. Methods based on direct growth on the substrate .........................................98
2.2. Photoelectrochemical methods ...........................................................................101
2.2.1. Linear and cyclic voltammetry ....................................................................104
2.2.2. Chronoamperometry....................................................................................107
2.2.3. Chronopotentiometry and photopotential....................................................109
2.2.4. (Photo)electrochemical impedance spectroscopy (PEIS) ...........................111
2.2.4.1. Constant frequency (variable potential), Mott-Schottky representation ........ 114
2.2.4.2. Constant potential (variable frequency) Nyquist representation ................... 116
2.2.5. Electrochemical quartz-crystal microbalance (EQCM) ..............................118
2.3. Spectroscopic methods .......................................................................................119
2.3.1. UV-visible spectroscopy .............................................................................119
2.3.2. X-ray Diffraction (XRD) .............................................................................123
2.3.3. X-ray photoelectron spectroscopy (XPS) ....................................................126
2.4. Microscopic methods..........................................................................................129
2.4.1. Transmission electron microscopy (TEM) ..................................................131
2.4.2. Scanning electron microscopy (SEM).........................................................133
2.5. Bibliography .......................................................................................................136
Chapter 3. CuFe2O4 photocathodes: improving photoactivity by
electrochemical pretreatment .............................................................. 143
3.1. Introduction ........................................................................................................145
3.2. Experimental section ..........................................................................................146
3.2.1. CuFe2O4 electrode preparation ....................................................................146
3.2.2. Electrode characterization ...........................................................................146
3.3. Nanostructured CuFe2O4 films from commercial nanoparticles ........................148
3.3.1. Physical and optical characterization ..........................................................148
3.3.2. Photoelectrochemical response of the as-prepared electrodes ....................151
3.3.3. Electrochemical pretreatment ......................................................................153
3.3.4. Mechanism of the electrochemical pretreatment ........................................161
3.3.5. Photoelectrochemical characterization of pretreated p-type CuFe2O4
electrodes ...................................................................................................................166
3.4. Conclusions ........................................................................................................172
ii
3.5. Bibliography .......................................................................................................173
Chapter 4. Investigating water splitting with CaFe2O4 photocathodes
by electrochemical impedance spectroscopy ....................................... 179
4.1. Introduction ........................................................................................................181
4.2. Experimental section ..........................................................................................182
4.2.1. CaFe2O4 photoelectrode preparation ...........................................................182
4.2.2. Electrode characterization ...........................................................................182
4.3. PEIS mathematical model for hydrogen reduction.............................................183
4.3.1. Mechanism and theoretical approach ..........................................................183
4.3.1. Simulation of the impedance function ........................................................193
4.4. CaFe2O4 compact film electrodes .......................................................................198
4.4.1. Physical and chemical characterization.......................................................199
4.4.2. Investigation on the mechanism of photoelectrochemical water
reduction ....................................................................................................................200
4.5. Conclusions ........................................................................................................211
4.6. Bibliography .......................................................................................................212
Chapter 5. LaFeO3 photocathodes: metal doping as way to enhance the
photoelectrochemical behavior ............................................................ 217
5.1. Introduction ........................................................................................................219
5.2. Experimental section ..........................................................................................220
5.2.1. LaFeO3 thin film preparation ......................................................................220
5.2.2. Electrode characterization ...........................................................................221
5.3. Photoelectrochemical characterization of pristine and doped LaFeO3 thin film
electrodes .......................................................................................................................222
5.3.1. Physical characterization of the electrodes .................................................222
5.3.2. Effect of Mg and Zn doping in the PEC response .......................................226
5.3.3. Investigating the role of doping ..................................................................235
5.4. Conclusions ........................................................................................................240
5.5. Bibliography .......................................................................................................240
Chapter 6. Photoelectrochemistry of compact and nanostructured
YFeO3 photocathodes ........................................................................... 245
6.1. Introduction ........................................................................................................247
iii
6.2. Experimental section ..........................................................................................248
6.2.1. Synthesis of doped and undoped nanoparticulate thin film
electrodes (NP)...........................................................................................................248
6.2.2. Synthesis of compact thin film electrode (C) synthesis ..............................248
6.2.3. Instrumentation............................................................................................249
6.3. Photoelectrochemical response of YFeO3 photoelectrodes ................................249
6.3.1. Comparison between the two electrode fabrication routes .........................249
6.3.2. Effect of Mg-doping in NP-film electrodes ................................................262
6.4. Conclusions ........................................................................................................266
6.5. Bibliography .......................................................................................................267
Chapter 7. Enhancing water photooxidation on nanostructured NiTiO3
and anatase electrodes by electrochemical doping ............................. 271
7.1. Introduction ........................................................................................................273
7.2. Experimental section ..........................................................................................275
7.4.1. Preparation of NiTiO3 Nanorods .................................................................275
7.4.2. Electrode Preparation ..................................................................................276
7.4.3. Electrode Characterization ..........................................................................276
7.3. Effect of the ultrasound on the synthesis of NiTiO3 nanorods ...........................277
7.4. Electrochemical doping of NiTiO3 and comparison with the anatase case ........279
7.4.1. Physical characterization .............................................................................279
7.4.2. Photoelectrochemical characterization of pristine NiTiO 3..........................282
7.4.3. Effect of the electrochemical doping in NiTiO3 electrodes ........................284
7.4.4. Comparison of the electrochemical doping in NiTiO3 with anatase
electrodes ...................................................................................................................286
7.5. Conclusions ........................................................................................................298
7.6. Bibliography .......................................................................................................300
Chapter 8. Conclusions ........................................................................ 305
Resumen .............................................................................................. 311
List of symbols, abbreviations and physical constants ....................... 333
List of publications ............................................................................... 343
iv
Chapter 1
Introduction
1.1. The global energetic problem. Photoelectrochemical water splitting as a
clean alternative for H2 production .................................................................................... 3
1.2. Basics of semiconductors physics ............................................................................. 7
1.2.1. Density of states and charge carrier density .................................................... 8
1.2.2. Surface states .................................................................................................... 14
1.2.3. Light absorption by semiconductors .............................................................. 15
1.3. The semiconductor-liquid interface ....................................................................... 18
1.3.1. Thermodynamics of semiconductor/electrolyte interfaces ........................... 18
1.3.2. Double layer at a semiconductor-liquid interface ......................................... 22
1.3.3. Application of a potential ................................................................................ 27
1.3.4. Illuminated semiconductor/electrolyte interface ........................................... 29
1.3.5. Carrier collection under illumination ............................................................ 33
1.4. Nanocrystalline electrodes ...................................................................................... 35
1.4.1. Chemical capacitance of nanostructured semiconductor films ................... 35
1.4.2. Carrier collection under illumination ............................................................ 40
1.5. Photoelectrochemical cells ...................................................................................... 42
1.5.1. Photoelectrochemical cell configurations ....................................................... 43
1.5.2. Photocathode requirements and challenges ................................................... 46
1.6. Ternary oxides as photocathodes ........................................................................... 48
1.6.1. Spinels ................................................................................................................ 48
1.6.2. Delafossites ........................................................................................................ 51
1.6.3. Perovskites ........................................................................................................ 54
1.6.4. Other ternary oxides ........................................................................................ 55
1.7. Strategies for improving photoelectrochemical performance ............................. 58
1.7.1. Improving light absorption by doping ........................................................... 60
1.7.2. Improving separation of charge carriers by heterojunctions ...................... 61
1.7.3. Improving charge collection by doping .......................................................... 64
1.7.4. Decreasing surface recombination by passivation layers ............................. 67
1.7.5. Improving interfacial reaction kinetics by cocatalysts ................................. 68
1.8. Bibliography ............................................................................................................. 73
Introduction
1.1. The global energetic problem. Photoelectrochemical water
splitting as a clean alternative for H2 production
Since the advent of the industrial revolution, the global energetic consumption has been
continuously increasing (see Figure 1.1). As a result, the energetic problem is one of the
great challenges of the twenty-first century. The preservation of the current living
standards in our society will require an important change in energy supply. A global
population of 6000 million people required 15 TW of average power in 2006. The
continuous population growth together with the increment of industrialization in
developing countries will increase the energetic demand, estimated to be 30 TW in
2050. Fossil fuels currently provide about 85% of the total global energy supply,1
however, this source cannot sustain the future energy demand for a long time. In fact, if
we maintain the present rate of consumption, available reserves will only last for 50
years more.
Figure 1.1. Global energy consumption from 1800 to 2013 split according to the different
sources. Reprinted with permission from Fractional Flow, reference 2.
Together with the above mentioned issue of their depletion, the impact in the
environment derived from the use of fossil fuels is another serious concern. Their
3
Chapter 1
combustion produces polluting gases such as CO2, NOx, SOx, etc. Besides the harmful
effects of the emitted gases on human health, the additional problem of the
greenhouse effect has reached enormous importance because of the latent global
climate change experienced by the planet. The latter effect is mainly attributed to CO2
emission, although other gases, such as CH4 or N2O also contribute to a minor extent. In
1896 Svante Arrhenius was the first to propose a relationship between atmospheric CO2
and the Earth temperature, being aware that fossil fuel combustion contributes
significantly to the atmospheric CO2 concentration.3 Later developments in infrared
spectroscopy demonstrated a relation between the increase in the amount of atmospheric
carbon dioxide and the existence of more absorption of infrared radiation. The sharp
increase of the temperatures in the 1980s made the global warming theory to gain
ground. In fact, in 1988, the Intergovernmental Panel on Climate Change (IPCC)4 was
founded by the United Nations Environmental Programme and the World
Meteorological Organization, to advise governments by predicting the impact of the
greenhouse effect and by collecting data of related studies.
The access to clean, carbon-free energy technologies that could compete with the price
of fossil fuels, is expected to be a key factor for sustaining the energy requirements that
will result from the population growth. In Table 1.1, the main technology options and
their estimated global power generation capabilities are summarized. Neither of them
could cover all the predicted future demand, so the combination of them (and/or others)
seems to be the most viable possibility. Nowadays, nuclear energy provides 30% of the
electricity in Europe5 and 20% in USA.6 Although this technology may be classified as
low carbon power generation, it has important problematic issues associated to it: (i)
possible accidents would produce the liberation of large amounts of radioactive
materials with serious consequences for human health and the environment; (ii) high
level of radioactivity involved in the different steps of the nuclear cycle and the
accumulation of radioactive fission products (that remain active for millions of years);
(iii) finite reserves of the radioactive raw materials, i.e uranium. Biomass has also
become an attractive alternative because of its renewable nature; however, large extents
of land are required to produce it.
4
Introduction
Table 1.1. Overview of global power generating capabilities of sustainable energy sources.
Adapted from reference 7.
Energy source Power (TW)
Wind 4
Hydroelectric 1-2
Tidal and ocean currents <2
Geothermal 12
Biomass 10
Nuclear 10
Solar >20
Solar energy, is among the sustainable energy sources in Table 1.1, that with the highest
potential to meet most of the energetic needs. The Earth´s irradiation sketch shown in
Figure 1.2 depicts the resulting energy fluxes. Sun irradiates the Earth with 341 W m−2,
approximately 30% is reflected in the atmosphere, the clouds, and the Earth surface. A
23% of the incoming energy is absorbed by the atmosphere and a 47% is absorbed by
land and oceans. Importantly, the energy irradiated to the Earth in one hour is larger
than the annual energetic demand. Exploiting a fraction of the 3.85 · 1024 J year−1 of
solar energy absorbed by the atmosphere, the land and the oceans could potentially be a
renewable solution to the current energetic global problem. It has been estimated that a
land coverage of 0.16% with 10% efficient solar conversion systems would provide 20
TW of power.8
Solar energy conversion is commonly divided into two types. Thermal solar energy
corresponds to the transformation of solar energy into heat. In solar devices, a fluid
circulating through a solar thermal collector is heated by sun irradiation. In photovoltaic
solar energy, photovoltaic cells convert solar energy into electricity.9 The existing
photovoltaic cells reach very high efficiencies (more than 40%) on a laboratory scale.10
However, although the cost of Si-based solar panels has decreased rapidly,11 the
technology is still too expensive to compete with fossil fuel systems.
5
Chapter 1
Figure 1.2. Earth´s global mean energy fluxes for 2000-2005. The arrows represent the flow
and their thickness is proportional to their importance. Reprinted with permission from
reference 12.
The large-scale implementation of photovoltaic cells would require grid-based storage
capacities because of the intermittent nature of solar power. An interesting alternative is
storing the solar electromagnetic energy into chemical bonds contained in fuels such as
H2, CH4, CH3OH. Particularly, the production of H2 is attractive because it is basically
obtained from water, an abundant and cheap compound and, importantly, their
combustion is free of carbon, avoiding the problem of the greenhouse effect. In this
context, photoelectrochemical (PEC) water splitting is a promising alternative, which
consists in the conversion of water into H2 and O2 using semiconductor materials as
photoelectrodes to convert sunlight energy into chemical energy (with or without the
assistance of applied bias in the photoelectrochemical device). However, at present,
95% of the H2 production results from steam reforming and partial oxidation of CH4
(from fossil fuel).13 The office of Energy Efficiency and Renewable Energy (EERE)14 of
the USA includes the PEC approach as a long-term technology pathway with low or no
greenhouse emissions. Market viability will depend on future research in low-cost,
long-lifetime and highly efficient materials for energy conversion. This includes
development of strategies and materials that stabilize and improve the
photoelectrochemical performance.
6
Introduction
Likewise, the European Union developed the European Strategic Energy Technology
Plan (SET-Plan) to accelerate and help finance research projects oriented to low-carbon
technologies. Specifically, Annex I (Research and innovation actions), considers
artificial photosynthesis (including H2 production) as a subject of research establishing
that “For artificial photosynthesis, the most critical hurdle to overcome is design and
synthesis of catalysts for the oxidation of water at sufficient rates. A central research
goal is also the assembly and optimization of different parts to construct half cells and
thereafter functional integrated systems. The catalysts for water-oxidation and fuel
production can be molecular and/or solid-state or nanomaterials. They should also be
made from abundant and environmentally benign elements.” 15 This document highlights
the relevance of the design of suitable materials and their optimization, which is the
main goal of this thesis.
1.2. Basics of semiconductors physics
The understanding of semiconductor properties is critical for improving the efficiency
of photoelectrochemical cells. Here, the most important characteristics related with
semiconductor physics are explained; especially those connected to the phenomena
occurring during illumination.
In general terms, semiconductors are materials showing an electronic conductivity with
a value intermediate between that of metals (106 – 104 ohm−1 cm−1) and that of
dielectrics (< 10−10 ohm−1 cm−1).16 Intrinsic semiconductors are formed by
stoichiometric materials. In this case, conductivity is due to the thermal excitation of
some electrons from the valence band (VB) to the conduction band (CB) at T > 0,
together with the breaking of the corresponding chemical bonds. Figure 1.3A depicts the
basic bond representation of intrinsic Si, which is very pure and is ideally free of any
impurity. In contrast, extrinsic semiconductors are characterized by the presence of
impurities in the lattice, and conductivity involves a different mechanism, linked with
the presence of donor or acceptor impurities. For instance, Figure 1.3B shows Si doped
with P, in which a P5+ species substitutes a Si4+ atom and one electron is donated to the
lattice in the CB. Phosphorus can be said to be a donor impurity in the Si lattice and that
7
Chapter 1
the material is n-type. Similarly, a B3+ atom in the Si lattice creates a hole in the VB and
the material is p-type (Figure 1.3C), thus boron is an acceptor impurity in the Si lattice.
Donor impurities create electronic conductivity, while acceptor impurities generate hole
conductivity.
free electron free hole
A. B. C.
Si Si Si Si Si Si Si Si Si
Si Si Si Si P+ Si Si B− Si
Si Si Si Si Si Si Si Si Si
Figure 1.3. Schemes showing atoms and bonds in (A) intrinsic, (B) P doped (n-type) and (C) B
doped (p-type) Si.
1.2.1. Density of states and charge carrier density
When a semiconductor is in equilibrium, i.e. in the absence of illumination or
application of an electric field, a charge carrier density is only generated by thermal
excitation. It is important to discuss the density of state function in the semiconductor,
which is defined as the number of quantum states per unit volume and per unit
energy.16–18 The free electron model relates the kinetic energy of a free electron and the
k wave vector. The energy versus momentum plot has a parabolic shape, given by = h2
k2 /(82m), where m is the electron mass.16,18 The resolution of the Schrodinger
equation for a semiconductor also defines a parabolic shape of the CB edge, at least near
the minimum of energy, but with an important deviation with respect to the above
equation. In order to correct such a deviation, electron mass in the free electron model
equation is substituted by an “effective mass” that can deviate up to one/two orders of
magnitude from the free electron mass:18
h (1.1)
m
8
Introduction
The density of states in the conduction (DCB) and valence bands (DVB) derived when
Equation (1.1) holds, follows the expressions:16
me (1.2)
mh (1.3)
A deduction of these equations can be found in the literature.16,18 The density of states
(DOS) of the conduction (valence) band depends on the effective mass of the electron
(hole), me (mh), and the corresponding energy of the edge of the conduction (valence)
band, C (V). Figure 1.4 illustrates the density of states given by Equations (1.2) and
(1.3).
Energy
Valence Band
C
V
Conduction Band
DOS
Figure 1.4. Schematic plot of the DOS (DCB() and DCV()) in a semiconductor given by
Equations (1.2) and (1.3).
Analytical expressions (1.2) and (1.3) are deduced under idealized conditions and they
are usually not accurate enough for most metal oxides, particularly far from the CB
minimum and VB maximum. The real DOS are typically obtained from spectroscopic
measurements or quantum calculations. In this respect, density functional theory (DFT)
calculations are the most popular and versatile method for this aim. Nevertheless,
9
Chapter 1
analytical expressions (1.2) and (1.3) are extensively employed as an approximation to
the semiconductor behavior.
Aside from the density of states, the probability of occupation of a particular state must
be considered. In this respect, the probability in the case of fermions, which are particles
with a half-integral spin (following the Pauli exclusion principle), is described by the
Fermi-Dirac statistics:
f (1.4)
e
This equation is applicable to electrons, as they are fermions. f() represents the
probability that a state with energy is occupied by an electron and F is the “ ermi
level”, which corresponds to the electrochemical potential of the electrons in the
solid.16–20 The Fermi level is formally defined as the energy of the quantum state for
which the probability of finding an electron is 0.5. Figure 1.5 illustrates the Fermi-Dirac
distribution at different temperatures.
T3 > T2 > T1
Energy
T=0 K
T1
T2
T3
0.0 0.2 0.4 0.6 0.8 1.0
f(E)
Figure 1.5. The Fermi-Dirac distribution functions at different temperatures.
10
Introduction
The equilibrium concentration of electrons in an interval of energy d in the CB can be
determined as the product of the occupancy probability and the density of the states in
this region: dn = f()·D()d. The total density of free electrons in the conduction band
can be obtained by integrating this expression from the edge of the conduction band to
infinity. The same treatment can be applied for the free hole density in the valence band:
n f d (1.5)
f d (1.6)
The introduction of Equations (1.4) and (1.2) into (1.5), and (1.4) and (1.3) into (1.6)
results in an integral that cannot be solved analytically. As the concentration of
thermally excited electrons (holes) is usually small, it can be assumed that ( F)/kT >>
1, and the Fermi-Dirac function approximates the Boltzmann distribution. Figure 1.6
illustrates Boltzmann and Fermi-Dirac distributions at a given temperature. Both
functions overlap at low values of f().
Energy
Boltzmann
Fermi-Dirac
0.0 0.2 0.4 0.6 0.8 1.0
f()
Figure 1.6. Comparison between Fermi-Dirac and Boltzmann distributions at a given
temperature.
11
Chapter 1
Considering the Boltzmann approximation, integrals in Equations (1.5) and (1.6) can be
solved as reported,19 leading to the electron and hole equilibrium density expressions
under the above mentioned conditions:
n e (1.7)
e (1.8)
NC and NV are called “effective densities of states”, defined as the densities of states
within a few kT above the conduction band edge and below the valence band edge,
respectively. They are given by the following expressions:
me
(1.9)
h
mh
(1.10)
h
Intrinsic semiconductor
In intrinsic semiconductors, the product of the density of electrons and holes can be
obtained by multiplying Equations (1.7) and (1.8) and assuming that both densities are
equal. To preserve charge neutrality in an intrinsic semiconductor, it must hold that n0 =
p0 = ni:
g
n e ni (1.11)
ni is known as the intrinsic density of electrons (and holes). Then, the Fermi level
position can be obtained by combining Equations (1.7) and (1.8):
mh
ln ln (1.12)
me
Additionally, the substitution of NC and NV derives in an expression that is dependent
on the effective mass of electrons and holes. Specifically, the Fermi level in an intrinsic
12
Introduction
semiconductor is localized close to the middle of the band gap. If me mh, F is located
exactly in the middle of the band gap, as illustrated in Figure 1.7.
Intrinsic n-type p-type
Energy
CB CB CB
F
D
F
A F
VB VB VB
Figure 1.7. Band structure scheme showing the approximate position of the Fermi level in
intrinsic, and extrinsic n-type and p-type semiconductors. D and A represent the energy levels
of donor (D) and acceptor (A) impurities.
Doped semiconductor
A donor impurity is considered as neutral if filled by an electron (D), and positive if
empty (ionized, D+). Conversely, an acceptor impurity is considered as neutral if empty
(A) and negative if filled by an electron (ionized, A−). When impurities are introduced
in a certain density (ND and NA, respectively) in the intrinsic semiconductor, their
energy levels usually lie within the band gap. At a relatively high temperature, most
donor and acceptor are ionized. Likewise, the total density of negative charges
(electrons and ionized acceptors) must be equal to the total density of positive charges
(holes and ionized donors)16,19:
n (1.13)
Expression (1.11) for intrinsic semiconductors still applies (until degeneracy), and the
p0n0 product is independent of the added impurities. In an n-type semiconductor, ND
impurities have been intentionally added to the crystal of the intrinsic semiconductor,
and ND >> NA. Under this assumption, p0 ≈ ni while n0 comprises the contributions of
the intrinsic electron density and of the free electrons provided by the ionization of the
D impurities. Additionally, if ND >> ni, then ND >> p0 and Equations (1.13) and (1.11)
simplify to the following expressions:
13
Chapter 1
n ≈
n type (1.14)
≈ ni
and the electron density equals the density of ionized impurities. Analogous expressions
can be deduced for a p-type semiconductor:
≈
p type (1.15)
n ≈ ni
Similarly to the case of an intrinsic semiconductor, the Fermi level positions can be
obtained from Equations (1.7) and (1.8), by substituting n0 and p0 according to
Equations (1.14) and (1.15).
ln ln (1.16)
n
ln ln (1.17)
In the above expressions all the impurities are considered to be ionized: and
. In view of these equations, the Fermi level must be located in the proximity of
either the conduction band edge in n-type materials or the valence band edge in p-type
materials, but within the band gap, as schematized in Figure 1.7B and C. Note that when
the density of impurities is equal to the effective density of states, the Fermi level is
located at the corresponding band edge (conduction band for donor impurities or
valence band for acceptor impurities).
1.2.2. Surface states
Surface states are produced by a discontinuity in the periodic potential of the crystal at
the surface,21 which causes the appearance of additional energy levels together with
those related with the bulk of the material. In these states, electrons are localized at the
surface of the semiconductor (at the interface). The presence of surface states enables
charge carriers to be trapped at the surface (storing charge), which influences the
14
Introduction
equilibrium properties of the semiconductor. Importantly, not only they can act as
recombination centers, but they can also mediate in charge transfer processes to the
electrolyte. In the real surface of a semiconductor, several types of electronic surface
states can exist. They can be both donors or acceptors, and normally their density does
not exceeds 1015 cm 2
for flat surfaces,16 value that approximates the number of lattice
points per unit surface area of the semiconductor. Aside from the intrinsic states,
extrinsic surface states can also be present as a result of the interaction of the
semiconductor surface with the external medium. Typically, they are formed by
strongly adsorbed species.
1.2.3. Light absorption by semiconductors
The light absorption capability of a semiconductor is primarily defined by the band gap,
Eg. When a photon of energy (h) higher than Eg impacts the semiconductor, it is
absorbed generating an electron-hole pair. The irradiance of the solar spectrum as a
function of the wavelength is depicted in Figure 1.8. The maximum values of irradiance
are located in the visible and infrared regions. Hence, in the context of
photoelectrochemical cells and from the point of view of light absorption, low band gap
semiconductors are desirable in order to absorb most of the solar spectrum.
The absorption coefficient, in cm 1, in a semiconductor determines the penetration
depth of the light at a given wavelength that coincides with 1/. Thus, materials with
high values of , absorb the incident light within a very small thickness.
15
Chapter 1
Figure 1.8. Solar spectrum showing the solar irradiance with and without atmospheric
absorption. Reprinted with permission from reference 22.
Basically, there are three different types of electronic transitions upon light excitation as
depicted in Figure 1.9.16,18 The first is the absorption by free carriers within the bands,
called intraband transitions. In this case, the light radiation is converted, eventually,
into heat. The second is the absorption by charge carriers localized in impurity atoms or
structural imperfections in the lattice (impurity absorption). The third is the most
important absorption in photoelectrochemistry (called fundamental absorption),
corresponding to an interband transition in which the absorbed radiation produces a
delocalized electron in the CB and a delocalized hole in the VB.
16
Introduction
e-
(1)
e- e- e- e-
(5)
C
A
Energy
(2)
(6)
(4)
D
(3)
h+ h+ h+ V
Figure 1.9. Scheme showing the main types of optical transitions in a semiconductor: (1)
intraband transition, (2,4) transitions with the participation of impurity levels, (6) interband
transition. Processes (3) and (5) occurs a result of thermal excitation.
The interband electron transitions are subdivided into direct and indirect.7,16,18 In the
direct transition, the top of the valence band and the bottom of the conduction band, are
both located at the same value of the k-vector. Then, the transition occurs without a
change in the crystal momentum (Figure 1.10A). In contrast, in the indirect transition,
the absorption of a photon is accompanied by a change in the crystal momentum. Since
photons carry very little momentum, in an indirect transition, the absorption or emission
of a phonon (a lattice vibration) is required. Indirect transitions are more unlikely to
occur, and the materials possessing such behavior typically present lower absorption
coefficients.
Considering a parabolic shape for both the bottom of the CB and the top of the VB (as
described previously by Equation (1.1)), the optical absorption coefficient, , for an
interband transition is given by:16,23
m
h g (1.18)
h
where B is a constant, and m is an exponent that depends on the nature of the optical
transition, m = ½ for direct transitions and m = 2 for indirect transitions. This
expression is commonly used to determine the optical band gap of the material from the
representation of h vs h for direct and h vs h for indirect optical
transitions. Experimental optical measurements are detailed in Chapter 2, section 2.3.1.
17
Chapter 1
A. B.
h
h
h
k k
Figure 1.10. Schematic direct (A) and indirect (B) optical transitions in a semiconductor. The
indirect transitions require assistance of a phonon with energy h.
1.3. The semiconductor-liquid interface
The semiconductor/solution interface constitutes a contact between two conductors with
different types of conductivity, one electronic, the other ionic; and typically with
different states of aggregation of the contacting media: one solid, the other typically
liquid. The contact allows the exchange of particles between both phases until an
equilibrium state is reached. The inherent important features of the interface are
analyzed in the following, firstly in the dark, and then under illumination conditions.
1.3.1. Thermodynamics of semiconductor/electrolyte interfaces
In an electrolyte containing a redox couple, the electrochemical potential of electrons
16,18,23–25
stems from the potential of the redox couple in solution:
e ed (1.19)
“O ” is the o idized and “Red” is the reduced species, and n the number of electrons
involved in the redox process. Eredox is the reduction potential defined by the Nernst
equation:
18
Introduction
ao
redo redo ln (1.20)
red
where redo is the standard potential, which is the potential of the half reaction
corresponding to activity values of one for the dissolved species and 1 bar fugacity for
the gaseous species; aox and ared are the activities of oxidized and reduced species in the
electrolyte, respectively. In ideal solutions, the activities are equal to the concentrations
of the corresponding species.
At this point of the discussion, it is important to mention the different energy scales
used in semiconductor electrochemistry. In solid-state physics, the electron energy in
vacuum is used as a reference. However, electrochemists have adopted the potential of
the standard hydrogen electrode (SHE) as the zero of the electrode potential scale. In
order to compare the different values in both scales, the potential of the SHE with
respect to the vacuum level has been estimated. Although there exists a variability in the
reported values, a value of −4.44 e is considered here, which is that recommended by
the IUPAC.26 Equation (1.21) relates the energy scale vs vacuum with the electrode
potential scale vs SHE.
e 4.44 e e (1.21)
where ESHE is the potential vs SHE.
Similarly to a solid semiconductor, the notion of electron energy levels also applies to a
redox electrolyte. In a redox couple, such as Ox(aq)/Red(aq) dissolved, for instance, in
water, Ox(aq) represents the unoccupied redox electron energy level and Red(aq)
represents occupied states. The Ox and Red ions are surrounded by different solvation
shells because the ions of each species interact in a different way with the solvent.
Consequently, an electron transfer between Ox(aq) and Red(aq) involves energy levels
that differ from the thermodynamic value redox as schematized in Figure 1.11. The total
energy of the redox system in its reduced state is represented by Redsolv,red, in which the
subscri t “solv,red” indicates that the solvation shell is that typical of the reduced
species. Then, the reduced species can transfer an electron from Red solv,red into the
vacuum, leading to the Ox species. The electron transfer is faster than the process of
reorganization of the solvation shell that occurs when the oxidized species is formed.
Then, additionally to the energy “red” needed to form the Oxsolv,red, an energy connected
19
Chapter 1
to the solvent dipole reorganization is also exchanged to reach the equilibrium
configuration of the oxidized species, Oxsolv,ox. Such an energy is the so-called solvent
reorganization energy, , according to the electron transfer theory.27
Oxsolv,red + evac
Energy of the redox system
Oxsolv,ox + evac
red redox ox
Redsolv,ox
Redsolv,red
Figure 1.11. Scheme showing the energy of a redox system in its oxidized and reduced states.
Both values of the reorganization energies are considered to be equal.
The energy diagram in Figure 1.11 is a simplified adaptation to discuss the process. The
Gerischer28 model is that used for defining the thermal fluctuations of the energy levels
for both the oxidized species, Wox, and the reduced species, Wred. They are described by
Gaussian-type distributions:
redo
o e (1.22)
4
redo
red e (1.23)
4
The energy levels corresponding to Equations (1.22) and (1.23) are illustrated in Figure
1.12, in which now, ox and red represent the average (most probable) energies of the
oxidized and reduced species, respectively.
Once the energy levels of the electrolyte species defined, let us describe the processes
occurring when a semiconductor is brought in contact with an electrolyte. In
equilibrium, the electrochemical potential of electrons at both sides of the interface
SC/electrolyte must be equal. In other words, the Fermi level in the semiconductor
20
Introduction
equilibrates with that in solution (equilibrium potential of the redox couple): F = redox.
For the equilibration of both phases an electron transfer must occur between them,
generating a band bending of the semiconductor bands at the interfacial region within
the semiconductor.
A. B.
CB e- e- e-
ox ox
redox redox
ESHE / V
/ eV
red red
VB
SC n-type SC n-type
C. D.
CB
ox ox
redox redox
ESHE / V
/ eV
red
red
h+ h+ h+
VB
SC p-type SC p-type
Figure 1.12. Band diagram of the semiconductor and density of states in the electrolyte before
(A and C) and after (B and D) the semiconductor and the electrolyte phases are brought into
contact for (A and B) n-type and (C and D) p-type semiconductors. In the electrolyte side,
empty areas correspond to the unoccupied levels and shaded areas to the filled states.
In the case of an n-type semiconductor (Figure 1.12A and B), F is typically higher than
redox. Then, when both phases are brought into contact, free carriers (electrons) in the
semiconductor are initially transferred to unoccupied states of the electrolyte, until F
equals redox. As a consequence, a depletion of electrons (majority carriers) occurs near
the semiconductor surface, creating a region called “s ace charge layer”. The built-up
voltage generated in this region results from the fact that the bands bend upward. An
analogous situation occurs for p-type semiconductors, where F is commonly lower than
redox (Figure 1.12C and D). In this case, holes in the valence band are transferred to the
occupied levels of the electrolyte, and the bands bend downward.
21
Chapter 1
1.3.2. Double layer at a semiconductor-liquid interface
As discussed above, when the solid phase is brought into contact with the liquid phase a
charge reorganization occurs. The semiconductor-liquid interface is modeled in the
classical theory by an electric double layer composed of three different regions16,18 as
illustrated in Figure 1.13: first, the space charge region in the semiconductor side,
second the Helmholtz layer and third, the Gouy-Chapman region in the solution side. In
the semiconductor side, the charge is produced by the redistribution of electrons and
holes near the surface, while in the electrolyte there exists a redistribution of ions. The
knowledge of the electrical double layer characteristics is critical as it controls the
electric and chemical properties of the surface (carrier recombination, photovoltaic
effects, electrode reactions, etc.). The presence of surface states and how they affect the
double layer will also be discussed in the present section.
22
Introduction
Space Helmholtz
Semiconductor bulk Charge Gouy-Chapman Electrolyte bulk
− + − + + − − − + − + −
−
− − − − −
− + + + −
− +
−
+ −
+ + +
− − − +
− −
+
− + − −
−
+
+
+ + + − +
−
+
− − −
−
− + + + − − − +
−
+
IHP OHP
SC
H
GC
Figure 1.13. Scheme of the electric double layer for an n-type electrode in contact with an
electrolyte in depletion regime and the corresponding drop of potential across the different
layers. Charge carriers are not represented at scale.
The Helmholtz layer
The Helmholtz layer or compact layer is the region in the electrolyte in contact with the
semiconductor surface. Two regions can be distinguished in this layer. The first is
delimited by the electrode surface and the inner Helmholtz plane (IHP), defined by the
ionic species specifically adsorbed at the solid surface and by solvent molecules. The
second region is located between the IHP and the so-called outer Helmholtz plane
(OHP). At the OHP solvated ions of a charge opposite to that in the solid are present,
attracted electrostatically by the charge at the electrode surface. The total potential drop
over the region from the semiconductor surface to the OHP is denoted by H. This
region can be modeled as formed by two planar sheets of charge. Its thickness is nearly
constant and independent of the nature of the solid, being 1.5-2 times the thickness of a
23
Chapter 1
mono-molecular water layer (0.3 0.5 nm).17 The capacitance of the Helmholtz layer can
be calculated as:
(1.24)
The capacitance of the Helmholtz layer only depends on the permittivity of the medium,
and the total thickness of the layer, x H. Considering a dielectric constant of = 20, and
xH = 3 Å,20 the Helmholtz capacitance would be about 60 F cm 2. Experimental values
are frequently close to this value. Typically, Helmholtz capacitance is not strongly
dependent on applied potential.
The Gouy-Chapman layer
The Gouy-Chapman layer is a region in the electrolyte beyond the OHP, in which an
excess of charge persists. It appears when the electrolyte concentration is low,
generating a diffuse region away from the Helmholtz layer. It is characterized by a
capacitance Cd. The contribution of the diffuse layer to the overall drop of potential
across the interface is not negligible for electrolyte concentrations below 10−3 M.18 In
semiconductor electrochemistry, the contribution of the diffuse layer is commonly
avoided by using electrolyte concentrations above 10−2 M, in order to simplify the
analysis.
The space charge layer
There is a characteristic region at the semiconductor interphase in which part of the free
charge has been removed by an equilibration process (section 1.3.1). The potential and
charge distribution7,16,18,29 in the space charge layer can be described by means of the
Poisson equation, which relates the potential with the net charge density:
d
(1.25)
d
24
Introduction
The volume charge density, (x), as an implicit function of the distance from the
surface, x, can be expressed as:
e n (1.26)
When a space charge layer is generated, the electron and hole densities change with the
distance to the surface (x) according to:
b e
n n e n e (1.27)
b e
e e (1.28)
b b
and are the conduction and valence band energies in the bulk, respectively, and n
and are the carrier densities in the bulk, which are given by Equations (1.7) and (1.8).
b
has been defined as e.
The electron density at the surface is given by Equations (1.29) and (1.30):
s b e
ns n e n e (1.29)
s b e
s
e e (1.30)
The electron density is a function of the band bending across the space charge layer. In
equilibrium, ns0 (for n-type materials) and ps0 (for p-type materials) are the minority
carrier densities at the surface.
By using the above equations in the case of an n-type material, and by considering that
donors are completely ionized ( ) and that the hole density is negligible as
compared with that of electrons (p(x) << n(x)), one may work out for the space charge
layer:29
25
Chapter 1
e
d n n e
sc (1.31)
e e
d
By considering n0 ≈ D, as deduced in section 1.2.1 for an n-type material, Equation
(1.31) simplifies to:
d sc e e sc
e (1.32)
d
An analytic solution for this equation leads to the Mott-Schottky equation for an n-type
semiconductor, which connects the differential capacitance, CSC = dq/dSC, and the
potential drop, sc
across the space charge layer:
sc (1.33)
e e
SC represents the band bending. Analogously, the Mott-Schottky equation for a p-type
material can be obtained by considering the acceptor impurity density (NA) and the
adequate sign reversal:
sc (1.34)
e e
By considering the space charge layer as an ideal parallel plate capacitor, its thickness
can be derived from the relation LSC = 0/CSC and Equations (1.33) and (1.34) for n-
type and p-type electrodes, respectively:
e (1.35)
e ,
This equation applies for both n-type (ND) and p-type electrodes (NA).
The presence of surface states affects the interfacial double layer of the semiconductor
electrode. In the absence of surface states, the double layer can be modeled as two
capacitors in series: the capacitance of the space charge layer and that of the Helmholtz
layers (supposing a negligible potential drop across the Gouy-Chapman layer). When
26
Introduction
surface states are present, they are represented as a third capacitance (CSS) in parallel
with CSC.
CSC CH Cd
CSS
Figure 1.14. Equivalent circuit for the double layer including a surface state capacitance.
The full charge on the semiconductor side equals QSC+QSS, where QSC is the charge
stored in the space charge region and the differential capacitance at this side is C SS+CSC.
When the magnitude of CSS is larger than that of CSC, the influence of surface states in
the characteristics of the double layer will be significant.
1.3.3. Application of a potential
In section 1.3.1, the dynamic equilibrium between a semiconductor (n-type or p-type)
and an electrolyte was described. In the present section, the processes that occur under
perturbation of such equilibrium by application of an external potential will be
discussed.7,16,23 For simplicity, in the following discussion the presence of surface states
is not considered.
In an n-type material in equilibrium, there is no net current flowing across the interface
(Figure 1.15A). When CSC << CH, and a potential is applied to the electrode, this
potential is invested in changing the potential drop across the space charge layer,
modifying the extent of band bending. In this case, the potential drop across the
Helmholtz layer is unaltered and it can be established that the edges of the bands are
pinned at the surface. This situation is called Band Edge Pinning (BEP). Thus, when a
potential such that F < redox (reverse bias potential) is applied, band bending increases
(Figure 1.15B) and the electron density at the surface decreases with respect to the
27
Chapter 1
equilibrium state (ns < ns0). On the contrary, if the applied potential makes F > redox,
the bending of the bands decreases with respect to the equilibrium situation until a
potential in which the bands become flat is reached (Figure 1.15C). Note that this
potential denotes the position of the Fermi level in the semiconductor, which is close to
the conduction band edge in an n-type material (and close to the valence band edge for a
p-type material) and it is called “ lat and Potential” (Efb). If F increases above the
energy corresponding to the flat band potential, electron density at the surface (ns)
increases with respect to the equilibrium density (ns0), i.e. ns > ns0 (accumulation
regimen) and a faradaic cathodic current can flow across the interface as long as the
redox couple efficiently accepts electrons. An analogous treatment can be applied for p-
type materials.
A. B. C.
e SC = 0
V V V
e SC F
Eredox e SC Eredox Eredox
F
ESHE / V
Efb
/ eV
Eeq Eref F Eref Eref
E
C C C
D. E. F.
e SC = 0
V V V
Eref Eref Eref
ESHE / V
E
/ eV
F
Eeq Eredox F Eredox Efb
Eredox
e SC
C
e SC
C F C
Figure 1.15. Band energy diagrams for an n-type (A, B, C) and a p-type (D, E, F)
semiconductor under band edge pinning (A and D) at equilibrium, (B and E) under reverse bias
potential (C and F) at a potential such as E = Efb.
By considering the opposite case, in which CSC >> CH, the application of a bias
potential will change the potential drop across the Helmholtz layer. As a consequence,
the energy levels at the surface would shift with respect to the energy levels in solution
without altering the band bending. This is the condition of Fermi Level Pinning (FLP)
at the interface, which is illustrated in Figure 1.16A and B.
28
Introduction
A. B.
e SC
Eredox Eredox
ESHE / V F e SC
/ eV
Eeq Eref F Eref
E
e H
e H
OHP OHP
C. D.
Eref Eref
ESHE / V
E
/ eV
F
Eeq Eredox F Eredox
e SC
e SC
e H
e H
OHP OHP
Figure 1.16. Band energy diagrams for an n-type (A and B) and p-type (C and D)
semiconductors under Fermi Level Pinning (A and C) at equilibrium, (B and D) under reverse
bias potential.
In real systems, the SC/electrolyte interface is typically governed by an intermediate
situation between the BEP and FLP cases, and a simultaneous change in the potential
drop in the space charge region and in the Helmholtz layer occurs when a potential is
applied.
1.3.4. Illuminated semiconductor/electrolyte interface
Up to now, the introduction has focused on a discussion about the semiconductor
electrodes in the dark. Under thermal equilibrium, the Fermi level has a value F, as
shown in Figure 1.17. When the semiconductor is illuminated with photons of energy
higher than the band gap, their energy is absorbed to generate excited electron-hole
pairs (the light absorption properties of semiconductors were described in section 1.2.3).
Consequently, if the thermal equilibrium existing in the dark is disturbed by light
29
Chapter 1
excitation, an increase in both the free electron density in the conduction band, and the
7,16,24
free hole density in the valence band occurs.
Upon illumination, the excited charge carriers are no longer in equilibrium, although the
photoexcited carriers can reach a steady state (but not a formal equilibrium state) when
the processes of generation and recombination within the semiconductor are balanced.
This leads to a stationary density of electrons and holes (n* and p*). In this case, the
electrochemical potentials for photoexcited electrons and holes correspond to the quasi-
Fermi levels of electrons (F,n*), and quasi-Fermi level of holes (F,p*).17 The carrier
concentration under illumination contains the contribution of both the carriers generated
through thermal excitation (n and p), and additional electrons (n*) and holes (p*)
produced by photoexcitation, being n* = p*:
n n n (1.36)
(1.37)
The quasi-Fermi levels of electrons and holes can be defined by Equations (1.16) and
(1.17) but considering the charge carrier density under photoexcitation instead of the
density values in the dark:
,n ln ln (1.38)
n n n
, ln ln (1.39)
The combinations of Equation (1.16) with Equation (1.38), and Equation (1.17) with
Equation (1.39) result in the following expressions:
n
,n ln (1.40)
n n
, ln (1.41)
In a doped n-type semiconductor, the density of electrons in thermal equilibrium is high,
so n* << n and n* ≈ n. Thus the quasi-Fermi level of electrons coincides with the
30
Introduction
Fermi level in the dark, ,n , as illustrated in Figure 1.17. On the other hand, the
hole density is quite low, and p* >> p.
n-type p-type
Energy
CB CB
e- e- e- e- e-
F F,n*
F,n*
F,p*
F F,p*
h+ h+ h+ h+ h+
VB VB
Figure 1.17. Band structure diagram showing the splitting of the Fermi level in the dark (F)
and the quasi-Fermi levels of electrons (F,n*) and holes (F,p*) when the semiconductor is under
illumination for n-type and p-type semiconductors.
In a p-type material, the opposite situation is produced, and , (see Figure 1.17).
Equations (1.40) and (1.41) simplify to:
,n
n type (1.42)
, ln
n
,n ln (1.43)
n p type
,
Now, let us address the case in which a space charge layer is present at the
semiconductor interface (see section 1.3.2) in equilibrium with the redox couple in the
electrolyte and without the application of an external bias, as shown in Figure 1.18A
and C. Upon illumination, in this region the photoexcited electrons and holes are
separated by the electric field and they move in opposite directions. This migration of
the photoexcited carriers induces an inverse potential within the electrode, reducing the
potential drop (SC) across the space charge layer, as illustrated in Figure 1.18B and D.
Consequently, the band bending decreases and the quasi-Fermi level of the majority
31
Chapter 1
carriers varies. The difference between the Fermi level in the dark and the quasi-Fermi
level of the majority charge carriers under illumination is equivalent to the
photopotential (Eph) measured as the change in open circuit potential when the electrode
is illuminated (see Chapter 2, section 2.2.3).16,17 In Figure 1.18B and D, it can be
observed that the quasi-Fermi levels of the minority carriers are shifted from the value
in the dark only within a limited depth at the semiconductor solution interface,
coinciding with the distance that the photons can penetrate.
A. B.
e- *
F,n
ESHE / V
Eredox Eph Eredox
/ eV
F
*
F,p
h+
photoanode electrolyte photoanode electrolyte
C. D.
e-
*
Eredox F,n Eredox
ESHE / V
F
/ eV
Eph *
F,p
h+
photocathode electrolyte photocathode electrolyte
Figure 1.18. Band energy diagrams for an n-type semiconductor in depletion regime showing
the location of the Fermi and quasi-Fermi levels for electrons and holes (A and C) at
equilibrium and (B and D) under illumination.
On the other hand, it can be concluded that the illumination of the electrode allows for
producing reactions that are not thermodynamically spontaneous in the dark as a
consequence of the splitting of the Fermi level into the quasi-Fermi levels. In this
respect, the location of the quasi-Fermi levels is critical as the difference between them
and the equilibrium potential of the redox reaction of interest will determine the
thermodynamic driving force for such electrochemical reactions to occur.
32
Introduction
1.3.5. Carrier collection under illumination
In photoelectrochemistry, the flux of carriers collected in the external circuit is
16,23,25,30
measured and recorded as a current value. The maximum current corresponds
to the rate of photon absorption in the absence of current multiplication mechanisms;
however, there is a loss by recombination processes in the semiconductor bulk, in the
space charge region and at the surface. The light is considered to be absorbed across a
distance 1/ from the semiconductor surface (see Figure 1.19). Commonly, the carriers
are generated both in the space charge region (characterized by the presence of an
electric field) and in the bulk (field-free region). In the region of the semiconductor with
depletion regime, the space charge layer extends across a distance LSC from the surface
as given by Equation (1.35). In an n-type (p-type) SC, the electric field in the space
charge region drives the holes (electrons) to the surface where they react with
electrolytic species, and the electrons (holes) to the rear ohmic contact. Holes
(electrons) photogenerated in the bulk are transported by diffusion to the space charge
region. The characteristic diffusion length of the minority carriers, Ln,p (see Figure
1.19), defines the zone in the semiconductor bulk within which the photogenerated
minority carriers can effectively reach the space charge region:
n, (1.44)
n, n, n, n,
e
This equation is valid either for n-type (Lp, Dp, p, p) or p-type (Ln, Dn, n, n)
materials. Dn,p is the diffusion coefficient, n,p the life-time and n,p the mobility of the
minority carriers.
In a simplified case, it can be assumed that the photocurrent is the sum of the current
associated to minority carriers photogenerated in the depletion layer and in the bulk as
long as they can reach the depletion layer by diffusion. It is supposed that all the
photogenerated minority carriers in the space charge region reach the electrode surface
and subsequently contribute to the photocurrent. With these assumptions, Gärtner
developed a simple model in which the photocurrent is expressed as a function of the
incident photon flux (0), the absorption coefficient (), LSC and the diffusion length of
the minority carriers16,31,32:
33
Chapter 1
e
h
e (1.45)
n,
For large values of band bending (high SC) the Gärtner model has been found to be
satisfactory, but significant deviations occur at potentials near the flat band
potential.23,32 This model only considers charge carrier recombination in the material
bulk. For instance, the role of surface states as alternative charge recombination and/or
transfer pathways is not considered.
Figure 1.19 illustrates the different characteristic lengths given by the Gärtner model in
an illuminated n-type semiconductor. In a general vein, the electron-hole pairs
photogenerated beyond Lp+LSC (for an n-type semiconductor), will tend to recombine.
Then, the larger the Lp and LSC values, the larger the quantum yield of the
semiconductor electrode.
LP + LSC
LSC
Energy
e- e- e- e-
CB
h
VB h+
h+ h+ h+
1/
Figure 1.19. Scheme for the photogeneration of electron-hole pairs in an n-type semiconductor
under illumination from the electrolyte-semiconductor side, considering the penetration depth of
light (1/), the space charge layer width (LSC) and the diffusion length (Lp) of holes. The
minority carriers generated in the red area (Lp + LSC) are those contributing to the photocurrent
according to the Gärtner model.
34
Introduction
1.4. Nanocrystalline electrodes
Nanocrystalline electrodes consist of a porous tridimensional network of nanoparticles
deposited on a conducting substrate.33 The three-dimensional geometry contrasts the
“flat” or lanar nature of single crystals or olycrystalline electrodes, to which the
above theoretical description refers. Nanocrystalline electrodes are characterized by
smaller crystallite sizes (in the nm scale) and by the allowed permeation of the
electrolyte through the solid network. They are also termed as nanoporous electrodes,
mesoporous electrodes or nanostructured electrodes.
According to their morphology, two types of nanocrystalline electrodes can be
distinguished: (i) random nanoparticulate electrodes (crystallites are deposited without
imposing any structural control) (ii) ordered nanocrystalline electrodes, where
crystallites form organized assemblies (they can be deposited or directly grown on the
substrate).
One of the main virtues of these electrodes is their large specific area, i.e. large
roughness factor, which is advantageous when it is convenient to adsorb electrolytic
species on the electrode surface. For instance, this type of electrodes are widely
employed in dye sensitized solar cells (DSCC),34,35 and also in quantum dot solar cells
(QDSC).36,37 Particular geometries may favor more efficient charge carrier separation,
as the minority carriers need to travel shorter distances before reaching the SEI. This is
especially notable in ordered nanocrystalline electrodes. 24
1.4.1. Chemical capacitance of nanostructured semiconductor films
In compact semiconductor electrodes, when the electrode is under depletion, the
capacitance in the electrode side is determined by the space charge layer, CSC, whose
value is deduced from the Poisson-Boltzmann equation, as described in section 1.3.2,
leading to the Mott-Schottky equation. Nanostructured electrodes display a different
behavior, as they are typically composed by a porous network of low-doped crystallites
in the nanometer range permeated by the electrolyte. The crystal size causes band
bending to be, in most cases, negligible. Commonly, the diameter of the particles is
significantly smaller than the width of the space charge region (LSC), especially for low-
35
Chapter 1
doped materials, as smaller ND or NA values lead to larger LSC values (Equation (1.35)).
If the Fermi level is located far below de CB edge, the electrode material behaves as an
insulator. Upon negative polarization (for n-type SCs), the nanostructured electrode
acquires a conductive behavior. The electrode is homogeneously charged and the Fermi
level approaches the CB edge as long as the band edges are pinned, Figure 1.20
illustrates the change in the Fermi level upon application of a forward bias potential. It
acquires the quasi-metallic state at sufficiently negative polarization. Electron
accumulation is accompanied by electrosorption of ions at the electrode internal surface
or by ion absorption/intercalation (usually occurring for small cations, such as H+ or
Li+).
As there is an absence of a well-defined space charge layer, electron transport within the
nanostructure is basically diffusive. This behavior contrasts that of the compact
electrodes, where a space charge layer is generated in the electrode in contact with the
electrolyte.
Fermi level control in nanostructured electrodes is linked to electrode characteristics
such as a small size of the building blocks, a low level of doping, good electrical contact
among particles and of particles with the substrate, and electrolyte permeation. Thus,
homogeneous electron accumulation throughout the film occurs together with partial
band unpinning. Accordingly, there exists a capacitance in the electrode side that is the
intrinsic film capacitance, C, and the capacitance of the electrode (Cfilm) is given by the
series capacitance of C and the Helmholtz layer capacitance:
film (1.46)
The chemical capacitance is often smaller than the Helmholtz layer capacitance, C <<
CH, and band edge pinning occurs at least in limited potential regions. Under this
condition, the conductivity of the film strongly depends on the electrode potential. The
contribution from the capacitance of the substrate/electrolyte interface (due to the
permeation of the electrolyte into the solid network) is commonly negligible, although it
can be significant at sufficiently positive applied potentials.
The chemical capacitance can be defined as:
36
Introduction
d dn
ed (1.47)
d d
where p is the porosity and d is the thickness of the film.
In turn, the electrochemical potential of electrons (Fermi level) depends on the inner
potential (or Galvani potential), , and the chemical potential of electrons, , as:
e (1.48)
For ideally polarizable nanostructured films in quasi-equilibrium, it can be assumed that
and are constant across the thickness of the film. If charge accumulation in the film
is considered to be accompanied by absorption/intercalation of cations from the
electrolyte within the solid, the inner potential () is unaltered when the electron density
increases. Then, dF = d, and there is a drop in the chemical potential of electrons in
the film when the Fermi level is altered.
A. B.
C C
Eredox F Eredox
F
ESHE / V
/ eV
E Eref
Eref
V V e H
e H
OHP OHP
Figure 1.20. Band energy diagrams for a nanostructured n-type semiconductor (A) in
equilibrium and (B) under the application of a potential (E).
In Equation (1.5), the electron density was expressed as a function of the density of
states and the Fermi-Dirac function. Assuming the zero-temperature approximation
(considering that f(-F) is a step function), dn/dF = DOS(F). Hence, Equation (1.47)
can be expressed then as:
de (1.49)
37
Chapter 1
where DOS(F) is the density of states at the Fermi level. Importantly, Equation (1.49)
reveals that a perturbation of dF (i.e. dE = d/e) gives rise to the filling of a slice of
states at the Fermi level. In the following, two cases will be considered: (i) chemical
capacitance associated to the filling of conduction-band states and (ii) chemical
capacitance associated to band-gap states.
Conduction band states
As presented further below (in section 1.2.1), the thermodynamic equilibrium density of
CB electrons per unit volume is given by Equation (1.5). In this expression the density
of states in the CB can be calculated by means of Equation (1.2) assuming a parabolic
band structure. The density of electrons in the CB may be described by Equations (1.7)
(or holes in the VB by Equation (1.8)).
For an n-type semiconductor, the chemical capacitance associated to the CB can be
obtained by introducing DCB(F) = DOS(F) from Equation (1.7) into (1.49):
e d
e (1.50)
with NC described by Equation (1.9). By substituting (F C)/kT by e(EC E)/kT,
Equation (1.50) can be transformed into its logarithmic form:
e d e e
ln ln (1.51)
The ln vs E plot should yield a straight line, with a slope
d log m decade. Equation (1.51) allows for the determination of the
CB edge energy, from the intercept of the straight line with the y-axis.
38
Introduction
Band gap states
Aside from the conduction and valence band states, band gap states can also contribute
to the total chemical capacitance. In the case of surface states just below the CB edge,
the following expression has been proposed for nanostructured anatase electrodes
assuming an exponential distribution of surface states:33,38
e d e
e (1.52)
c c
where is the chemical capacitance due to the presence of such surface states, is
the total volume density of traps and T c is a characteristic temperature. The last
parameter can also be described by the coefficient C = T/Tc. This equation has been
obtained by assuming a step shape of the Fermi-Dirac function, i.e. by assuming that kT
0. Equation (1.52) only differs from Equation (1.50) in the presence of the parameter
Tc, instead of the absolute temperature.
In addition to the surface states, monoenergetic states can also be present within the
band gap. They are associated to trap states proposed to be located at grain boundaries
(GB)33,39 generated by the contact between the particles in the solid network. The
capacitance of a monoenergetic state of total volume density NGB can be described by:
e d
f f (1.53)
where f is the average occupancy that can be described by the Fermi-Dirac distribution
as f = exp( e(E EGB)/kT)). In this equation EGB is the potential of the monoenergetic
state level of the grain boundaries.
The total chemical capacitance can be obtained as the sum of the three contributions
described above: that of the conduction band states, that of the surface states and that of
the monoenergetic states at GBs:
(1.54)
39
Chapter 1
1.4.2. Carrier collection under illumination
In the case of nanostructured electrodes, the mechanism for charge separation once the
electrode is illuminated depends on the semiconductor crystal size and architecture. In
the case of compact electrodes, a space charge layer is formed in the vicinity of the solid
surface in contact with the electrolyte. The electric field created causes the separation of
the photogenerated charge carriers and the carriers reaching the space charge layer by
diffusion contribute to the photoeffect, as described by the Gärtner model (Equation
(1.45)). Consequently, only one type of carriers reaches the SEI. In the case of a
nanostructured semiconductor electrode, the charge separation mechanism is different.
The transport in the nanostructured network is mainly governed by diffusion, as stated
above. As there is generally an absence of space charge layer, the different reactivity of
electrons and holes toward species in solution and/or at the interface determines the fate
of carriers. It is important to mention that in nanostructured electrodes where particles
are large enough, significant band bending could contribute to the transport and
separation of the charge carriers, although this case is not considered in the theoretical
framework of the present section.
When the charge carriers reach the surface, they can experience different transfer
processes from (to) the reactive sites at the particle surface to (from) dissolved or
absorbed species at the illuminated SEI. Charge transfer requires a large enough
difference between the potential levels of the photogenerated carriers (i.e. CB or VB
edges) and that of the reactant species either adsorbed or in solution (i.e. redox
potentials, defined above). For instance, nanoparticulate TiO 2 has been employed as a
photoanode for water and organics oxidations, because the photogenerated VB holes
have a very high oxidizing power (a very positive potential). It is important to mention
that under illumination, the theory described in section 1.3.4 also applies for
nanostructured electrodes, and the electrochemical potential is equal to the quasi-Fermi
level of electrons for n-type materials or of holes for p-type materials.
The concentration of charged species in a nanocrystalline electrode is given by the
continuity equation considering the three basic type of processes: transport, generation
and recombination (reaction). Electron concentration across the film depends on the
potential applied to the conducting substrate. When a potential is applied corresponding
to a Fermi level below redox, electron concentration in the back contact is lower than in
40
Introduction
the bulk, generating an electron concentration gradient. Thus, the application an
adequate potential can efficiently extract electrons at the back contact. The photocurrent
(defined in Chapter 2, Equation 2.5), is proportional to the charge carrier flux at the
back contact of the nanoporous film. Solving the continuity equation under steady state
conditions, that is (i.e. n ,t t ), an expression for the photocurrent can be
obtained. The boundary conditions depend on the illumination direction. 33 They are
commonly called as EE illumination when the electrode is illuminated from the
electrolyte-electrode side and SE when it is illuminated from the substrate–electrode
side, as depicted in Figure 1.21.
SE EE
FTO
Figure 1.21. Scheme of a cross section of a nanocrystalline electrode, illustrating the EE
(electrode-electrolyte) and SE (substrate-electrolyte) illumination directions.
Soedergren et al.40 were the first to obtain an expression for the EE and SE maximum
photocurrent under the assumptions that (i) no significant migration contribution exists
(ii) the diffusion coefficient is constant, (iii) only one type of carrier contributes to
charge transport, (iv) generation of carriers is governed by the Lambert-Beer law and
(v) recombination follows a first-order kinetics:
n e d d e d n
h
e tanh (1.55)
n n cosh d
n d en d
h
e tanh (1.56)
n n cosh d
41
Chapter 1
An analogous treatment to that described in the present section 1.4, can be applied for
nanostructured p-type semiconductor electrodes.
1.5. Photoelectrochemical cells
This section focuses on the description of photoelectrochemical cells for water splitting.
The idea of capturing solar light to produce electricity may be traced back to the
discovery of the photovoltaic effect in 1839 by E. Becquerel,41 consisting in the
observation of an electric current in an electrochemical cell with a photoelectrode of
AgCl exposed to light. In 1954, Bell Labs created the first practical high-power silicon
PV cell with a 6% efficiency.42 Later, H. Gerischer made important contributions to the
development of semiconductor electrochemistry through the development of models
and theories related with the energetics and kinetics of dark and photoinduced charge
transfer at the semiconductor/electrolyte interface. In 1972, in coincidence with the start
of the oil crisis in 1973, Fujishima and Honda reported the first photoelectrochemical
cell for water splitting43 using an n-type TiO2 rutile single crystal to drive oxygen
evolution by photogenerated holes, while hydrogen was generated at a platinum counter
electrode.
The overall reaction for water splitting is:
(1.57)
The half reactions in alkaline media are:
4 4e 4 . vs (1.58)
4 4h .4 vs (1.59)
while in acidic media, they are:
4 4e . vs (1.60)
42
Introduction
4h 4 . vs (1.61)
The standard Gibbs energy change for the conversion of one mol of H2O to H2 and ½O2
at 298 K is 237.2 kJ mol 1, which corresponds to E0 = 1.23 V. The reaction is thus not
spontaneous, and an energy supply is required. In the context of the PEC cells, this
additional energy can be provided by a photovoltage induced by illumination of the
semiconductor sometimes with the simultaneous application of a bias.7,44
1.5.1. Photoelectrochemical cell configurations
There exist several types of photoelectrochemical cells for water splitting.45–47 The basic
one is depicted in Figure 1.22, composed by a single light absorber material, concretely
an n-type (A) or a p-type (B) semiconductor together with a counter electrode. As
described in section 1.2.3, the semiconductor converts the incident photons into
electron-hole pairs. They are spatially separated by the presence of an electric field in
the space charge region and/or by diffusion processes. The generated minority charge
carriers, holes in n-type (Figure 1.22A) and electrons in p-type semiconductors (Figure
1.22B) are driven to the semiconductor-liquid interface, inducing there redox reactions.
In the case of water splitting, electrons are employed to reduce water, producing H2,
while holes are consumed in water oxidation, generating O 2. The majority carriers are
swept toward the back contact and are transported to the counter electrode (commonly a
metal) through the external circuit, producing the other half-reaction (O2 evolution for
cells based on p-type or H2 evolution for cells based on n-type electrodes).
43
Chapter 1
e-
A.
e-
*
F,n e-
H2/H+
* 1.23 V
F,p H2O/O2
h+
photoanode electrolyte metal
h+
B.
e-
H2/H+
*
F,n 1.23 V
H2O/O2
*
F,p h+
h+
photocathode electrolyte metal
Figure 1.22. PEC configurations composed of a photoanode/metal (A) and photocathode/metal
(B) showing the quasi-Fermi levels under illumination and schematizing the interfacial
reactions.
In the configurations depicted in Figure 1.22, the semiconductor must have adequate
conduction and valence band edges and absorb light with photon energies > 1.23 eV to
split water without applied bias. The overvoltages needed for the half-reactions increase
the minimum band gap value beyond 1.23 eV. The energy required for photoelectrolysis
using a single photoelectrode is reported to be 1.6−2.4 eV per photogenerated electron-
hole pair.47 Consequently, the band gap of a useful single absorber may be too high to
efficiently absorb the visible light spectrum. In addition, the band edges need to be
precisely located to drive water splitting under minimum band gap. A most practical
device for water photosplitting is a tandem cell composed of an adequate couple of
photocathode/photoanode as depicted in Figure 1.23. In this configuration, an additional
bias potential is generated by an additional absorber.
44
Introduction
e-
e-
H2/H+ *
e-
F,n
1.23 V
H2O/O2
* h+
F,p
h+
photoanode electrolyte photocathode
Figure 1.23. PEC tandem cell configuration showing the quasi-Fermi levels under illumination
and schematizing the interfacial reactions.
The most common approach to quantify the performance of a water splitting device is
the solar-to-hydrogen efficiency (STH), which is defined as the ratio of the chemical
energy output (by hydrogen generation) and the solar energy input. The STH can be
calculated as:
hoto
m cm .
(1.62)
Pin ut m cm
where hoto
is the photocurrent density, F is the faradaic efficiency for hydrogen
evolution and Pinput the incident illumination power density. When a bias potential (Ebias)
is applied to the cell (between the photoanode and the photocathode), the conversion
efficiency is defined by the ABPE (Applied bias photon-to-current efficiency), whose
expression is analogous to Equation (1.62), but substituting . by . − bias) V.44
Reported calculations have shown that a single absorber PEC system can reach a
maximum STH of 31%, while tandem system with two absorbers can reach a 40%
value.48,49 Figure 1.24 shows theoretical contour plots of the STH for a tandem cell as a
function of the band gap of the electrodes by considering a total overvoltage of 700 mV.
A maximum STH of 28.9% is reached for a combination of 0.99 eV and 1.64 eV band
gaps. It is worth noting that considering water absorption of the electrolyte through a
layer of 2 cm, the efficiency would drop to a maximum of 24.8%. 50 In view of Figure
1.24, it is deduced that efficiencies higher that 15% can be reached with one
45
Chapter 1
semiconductor with a band gap of 1 1.3 eV and other of 1.5 2.1 eV. Then, it is critical
finding semiconductor materials having a band gap comprised between these limits.
Figure 1.24. Contour plot of the theoretical tandem PEC efficiency limits as a function of the
top and the bottom band gap of the junctions for an overvoltage of 700 mV. Reprinted with
permission from reference 50.
1.5.2. Photocathode requirements and challenges
As mentioned above, the band gap of the selected photoelectrodes is crucial for an
enhanced photoelectrochemical water splitting performance (i.e. high STH). Aside from
this issue, the photoelectrodes have to meet additional requirements. The main ones for
optimum operation are:
o Adequate band edge locations for water reduction and oxidation reactions.
o High chemical stability in the dark and under illumination.
o Good charge transport (diffusion) across the semiconductor bulk.
o Low overpotentials for water oxidation and reduction reactions.
In addition, there are other requirements related with the viability of a practical device:
46
Introduction
o Low cost reagents for electrode synthesis.
o Low cost and scalable synthesis route.
o Materials composed of Earth abundant elements.
o Non-toxic and environmental friendly materials.
Since the seminal work of Fujishima and Honda,43 a large number of reports dealing
with the study and optimization of n-type materials for photoelectrochemical cells have
appeared in the literature. Up to now, TiO2 can be considered as the reference n-type
semiconductor material due to its high photoactivity, non-toxicity, inertness, low cost,
and abundance on Earth.33,51,52 Conversely, the drawback is that it can only harvest
photons in the UV range, as it possesses a band gap of 3−3.2 eV. This is the reason why
hematite52–56 and, more recently BiVO4,57–60 have become widely popular in the last
years.
On the other hand, in relation with p-type semiconductors, finding a high-efficiency,
low-cost, stable photocathode keeps on being a pending task. Non-oxide materials, such
as Si,61–65 SiC66–68 or InP69–71 are among the most extensively studied. Several recent
review papers deal with the state-of-the-art of photocathodes for water splitting.10,47,72–76
Among the p-type oxides composed of non-toxic and Earth-abundant elements, copper
oxides are the most studied.77–84 Cu2O has received the most attention, having a direct
band gap of 2 eV and a favorable CB edge position for hydrogen evolution. Although
high photoresponses have been reached with Cu2O, it has been found that it is unstable
in aqueous solutions under illumination77,78,80 and a current decay is observed during the
operation time, probably due to the reduction of Cu(I) to the metallic state. In the last
years, significant efforts have focused on avoiding copper reduction in copper oxides in
aqueous media under illumination, by using protective layers. Some strategies, as the
introduction of AZO layers and the loading of a cocatalyst have accomplished great
improvements relating both stability and efficiency.80,85,86 Despite these efforts, a copper
oxide photoelectrode for a commercially-viable solar hydrogen device has not been
achieved yet.
For reaching the goal of market viability of high-efficiency tandem cells for hydrogen
production, work is still needed. Currently, the main limitations for a practical device
47
Chapter 1
are related with the stability of the employed materials and the low efficiencies.
Unfortunately, it is found that the materials that exhibit the best efficiencies under
visible light are quite unstable under illumination. Thus, the more important efforts are
focused on (i) improving the durability of the known materials with the implementation
of protective surface coatings or (ii) finding new materials that present better stabilities.
Additionally, efficiencies must be improved by enhancing sunlight absorption and
finding better cocatalysts. Moreover, the cost of hydrogen production needs to be
considered in these strategies.
1.6. Ternary oxides as photocathodes
Many metal oxides, including those having a complex stoichiometry (i.e. ternary,
quaternary oxides) behave as semiconductors and present good stability in aqueous
environments, making them attractive candidates for water splitting devices. It is worth
noting that the number of reports dealing with oxide photocathode materials is much
scarcer than that dealing with photoanodes mainly because, in general, p-type
semiconductor metal oxides present limited corrosion resistance and low carrier
mobility. The limited number of useful binary metal oxides in electrochemistry makes
difficult finding a photoelectrode with adequate properties as a photocathode. In the last
years, some works have been devoted to ternary/quaternary oxides based on Earth
abundant metals. Ternary oxides have been presented as good candidates for water
splitting photocathodes because many of them are able to absorb visible light and have
suitable band edge positions for water reduction.72,87 In this section, the main
investigations on ternary oxides as photocathodes for water splitting are reviewed
according with their crystalline structure.
1.6.1. Spinels
Spinels are compounds with the general formula AB2O4, where A and B are metals with
oxidation states +2 and +3, respectively. Normal spinel crystalline structures are usually
cubic with a close-packed lattice, although the unit cell can derive in that of the
48
Introduction
orthorhombic group. The structure of the mineral spinel MgAl2O4 is represented in
Figure 1.25 in which B3+ cations occupy octahedral sites while A2+ occupy tetrahedral
ones. Inverse spinels, such as NiFe2O4, have the same crystalline structure but the B
cations occupy the tetrahedral sites while the octahedral sites are occupied half by A and
half by B.
Figure 1.25. Cubic structure of a normal spinel with a chemical formula AB 2O4. The unit cell
parameters correspond to MgAl2O4, concretely to the pattern 101-0129 in the COD database
with a space group Fd-3m (cubic).
Ferrite type (B = Fe) spinels constitute the most studied group in photocatalysis,88
especially for water remediation. They are good candidates as PEC materials because
they can exhibit both n-type and p-type behavior together with a relatively narrow band
gap, which enables them to absorb visible light. In fact, their photocatalytic ability for
H2 production has already been studied.88–91 Very recently, a research update has been
published summarizing literature experimental results on the photocatalytic and
photoelectrochemical behavior of ferrites.92 Although they have been presented as
promising non-toxic and Earth-abundant alternative p-type photoactive materials, up to
now only a relatively low number of works are based on ferrite photocathodes in spite
of the large diversity present in nature, such as those based on Mg, Zn, Ni, etc.88
As far as we know, calcium ferrite (CaFe2O4) is the most studied ternary oxide
photocathode for water splitting. Having a relatively narrow band gap (1.9 eV), it is able
49
Chapter 1
to absorb visible light, and it has suitable band edge positions for water reduction.
Matsumoto et al.93 were the first to describe the use of CaFe 2O4 as a photocathode for
the HER using a pellet synthesized at high temperatures (1000 ºC). They determined
that Fermi level pinning occurs at the semiconductor-electrolyte interface and
consequently, the rate determining step for the photocathodic HER is the
electrochemical process at the electrode surface. The electrode presented low efficiency
ascribed to a low electron concentration at the surface. Later, the same authors revealed
that FLP arises from the presence of surface states linked to the redox level Fe3+/Fe2+.94
Additionally, they managed to enhance the CaFe2O4 photoresponse by improving the
semiconductor/metal contact by using Au and Pt-Pd alloy between the oxide and the Ag
paste. Afterward, Ida et al.95 constructed a tandem cell for water splitting using an
(hk0)-oriented calcium ferrite electrode coated on a Pt substrate as a photocathode, an n-
TiO2 electrode as a photoanode and 0.1 M NaOH as the electrolytic solution. Although
H2 and O2 gases were detected, the main drawback of CaFe2O4 is the high temperature
required for electrode preparation and the slow decomposition of the material because
of photocorrosion reactions. Importantly, Cao and co-workers96 were able to coat a
transparent substrate (FTO) with a 100 nm-thick CaFe2O4 film by using pulsed laser
deposition (PLD) at relatively low temperatures for the first time. This is an important
question as deposition of the photoactive spinel on a transparent conducting substrate is
highly advisable for the final device architecture.
Recently, other non-ferrite spinels have attracted some interest. For instance, CuBi2O4
has been also widely investigated since Arai et al. discovered its potential as a new
visible-light-responsive p-type semiconductor in 2007.97 After that, Hahn et al.98
achieved the synthesis of pure CuBi2O4 films composed by interconnected particles by
electrodeposition studying different synthetic parameters. The material possessed a band
gap of 1.8 eV and a suitable onset of photocurrent around 1.05 V vs RHE. Although the
photocurrents were modest (25-33 A cm 2), they were stable after an initial decrease of
photocurrent over one hour of illumination at pH 10.8, but unstable in acidic pHs. The
low photoconversion efficiency of these films was in part ascribed to hindered hole
transfer because of both the deficient particle attachment to the FTO (poor adhesion of
the films) and the existence of interparticle boundaries. Years later, Nakabayashi et al.99
further studied the electrodeposition procedure in order to improve the mechanical
stability and the contact with the FTO. Later, a more profound analysis of CuBi2O4
50
Introduction
electrodes prepared by drop-casting was performed by Berglund et al.100 at pH 6.65.
Their work revealed two different hole lifetimes of 32 ns and 819 ns, which implies that
some of the photogenerated carriers recombine quickly while other are relatively long-
lived. Furthermore the carrier diffusion length was much shorter (10 nm/52 nm) than the
light penetration depth (244 nm), most of the photogenerated carriers likely
recombining before reaching the semiconductor/electrolyte interface. As a result, they
concluded that the main limiting factor for the photoelectrochemical performance of this
material was the poor hole transport.
On the other hand, the photoelectrochemical properties of ZnRh2O4 have been
uncovered very recently.101 A porous photoelectrode was fabricated by electrophoresis
using ZnRh2O4 powder previously obtained through a solid state reaction. Having a
band gap of 2.0 eV and a very positive onset of potential (1.2 V vs RHE) at pH 6,
IPCEs ca. 8% were reached at 400 nm. Likewise, H2 gas was detected by irradiating at
> 420 nm. The authors established that there were no signs of photocorrosion based on
XPS and analysis of the metal cations in the electrolyte before and after the PEC
reaction, although the measured faradaic efficiency was estimated to be of 77%.
1.6.2. Delafossites
Delafossite type oxides are compounds with the formula ABO2, where A and B are
metals with oxidation states +1 and +3. They have a layered structure with a sheet of
linearly coordinated A cations and BO6 octahedral layers sharing edges. The structure is
shown in Figure 1.26 for the mineral delafossite, CuFeO2.
Concretely, Cu(I) based delafossites have received considerable attention in the fields of
TCOs (Transparent Conductive Oxides)102–104 in optoelectronic devices because most of
them are wide band gap semiconductors, such as CuYO 2, CuAlO2, CuGaO2 or CuInO2,
and exhibit high transparency and carrier mobility. Only a limited number of Cu(I)
delafossites have a band gap narrow enough for PEC applications, being among them
CuFeO2 and CuRhO2. The good stability in aqueous media is a central advantage of
Cu(I) delafossites, since this is one of the current challenges for achieving a practical
device.
51
Chapter 1
Figure 1.26. Delafossite structure with the chemical formula ABO2. The represented unit cell
corresponds to CuFeO2, concretely with the pattern 153-3117 in the COD database with a space
group R-3m (trigonal).
CuFeO2 is an attractive candidate for water photosplitting due to its narrow band gap
(about 1.6 eV) and its ability to absorb visible light, compared with other p-type
delafossite oxides. In 2012, Read et al.105 reported the first study related with H2
evolution over CuFeO2 as a photoelectrode. In this work a new electrodeposition route
was developed to prepare thin compact films over FTO substrates with a band gap of
1.55 eV. The authors highlighted the thermodynamic feasibility to producing H2 with
this material, which absorbs the entire range of the visible spectrum, reaching IPCEs
values up to 2% (at 350 nm). Afterward, Prévot et al.106 reported on a sol-gel citrate-
nitrate based technique to prepare thin CuFeO2 films that presents several advantages
compared with electrodes prepared via the electrodeposition route, such as
straightforward solution processing of the films or the ability to tune the layer thickness.
Using an electron scavenger such as O2, the photocurrents were reasonably large and
stable for days. Conversely, the photoresponse for H2O reduction was low, which was
attributed to both poor charge separation and transport properties and sluggish charge
52
Introduction
transfer at the semiconductor/electrolyte interface. In a very recent work,107 the same
authors reported a high mobility (0.2 cm2 V 1 s 1) and a relatively long lifetime (200 ns)
of the photogenerated carriers. Despite these favorable properties, they concluded that
the limitation of the photoactivity was related to the presence of a high density of
surface states (1014 cm 2) acting as electron traps supporting charge recombination and
causing Fermi level pinning at the surface.
p-type CuRhO2 was also investigated as a photocathode under visible irradiation having
a band gap of 1.9 eV.108 Surprisingly, H2 was produced with 1 M NaOH solutions
purged with either air or Ar. Generally, in the presence of O2, which behaves as an
electron scavenger, electrons are invested in reducing oxygen rather than in the
reduction of water. However, the authors concluded that water photoreduction was
preferred over oxygen photoreduction for this material. In addition, the photoelectrode
tended to corrode under illumination in an Ar-purged electrolyte with the formation of
Cu(0) (detected by XPS) while in the presence of O2, Cu(0) was not detected. In the
latter case, the measured faradaic efficiency for hydrogen generation was ~80%. It is
important to mention that despite the suitable properties as a photocathodes of Rh
oxides (see also ZnRh2O4, section 1.6.1), the practical use of materials containing
platinum-group metals is compromised due to their scarcity and, consequently, their
high cost.
Other wide-band gap delafossite materials have also been investigated in PEC for water
splitting. Recently, Díaz-García et al.109 examined the photoelectrochemical properties
of the CuCrO2 delafossite. Thin electrodes were prepared by a sol-gel route using FTO
as a substrate. Although the material band gap is too large to absorb the visible range
(3.15 eV) of the solar spectrum, the IPCE reaches values near 6% at 350 nm, which is
remarkably high compared with other p-type oxide materials. It is also noteworthy the
excellent stability of the photoresponse in both acidic and alkaline media. CuGaO2 is
another wide-band gap delafossite whose PEC properties have been investigated.110 In
fact, hydrogen gas was detected over this material using white light. Despite the
excellent electrical properties (high acceptor density and electrical conductivity) and
reaching IPCEs of 10% (at 300 nm), the material has too high band gap for a practical
PEC application, as it only absorbs light with wavelengths below 350 nm.
53
Chapter 1
1.6.3. Perovskites
Perovskite type oxides are compounds with the formula ABO 3, where A and B are
metals with different combinations of oxidation states (A+1 and B+5, A+2 and B+4, A+3
and B+3). The ideal structure is a cubic lattice (Figure 1.27) consisting of small B
cations in an oxygen octahedra and larger A cations coordinationed by 12 O atoms. It is
common that these materials present a distortion of the lattice leading to other variants
with lower symmetry as orthorhombic or hexagonal structures. 111,112 It is worth noting
that ilmenite oxides have the same stoichiometry, (i.e. ABO3). Nonetheless, ABO3
compounds adopt this structure when A and B cations have similar sizes, while the
perovskite structure prevails when one of the cations has a similar size to O2 and the
other is much smaller.113
Figure 1.27. Perovskite (chemical formula ABO3) cubic structure. The represented unit cell
corresponds to the unit cell of LaFeO3, concretely with the pattern 154-2144 in the COD
database with a space group Pm-3m (cubic).
To the best of our knowledge, the first works regarding perovskite oxides as
photocathodes were released in 1980 by Jarret et al.114,115 They constructed an
electrolytic cell that produced hydrogen without applying external potential using p-type
LuRhO3 as a photocathode and TiO2 as a photoanode. In this work, the properties of the
as-synthesized p-type LuRhO3 and n-type LuRhO3 (containing Th4+) were studied. It
was revealed that the Fermi level at the surface is pinned at the same potential for both
n- and p-type LuRhO3 electrodes. To the best of our knowledge, despite the long time
54
Introduction
elapsed from the publication of this paper, other relevant works on LuRhO3 as a
photocathode have not appeared. Years later in 1987, a preliminary study of the
photoelectrochemical behavior of LaCoO3 was also published.116
More recently, the group of Maggard studied for the first time the
117
photoelectrochemistry and electronic structure of CuNbO3. Photoelectrodes were
prepared by the doctor blade technique from presynthesized pure CuNbO3 powders. In
Ar-purged Na2SO4 solution (pH 6.3), IPCEs reached rather high values, exceeding 5%.
The absorption extended to the visible range (optical band gap 1.9 eV), although the
onset of photocurrent was not too positive (~0.7 V vs RHE). It is important to mention
the excellent stability of the photocurrent for several hours when applying a significant
bias potential.
Since Celorrio et al.118 studied the properties of a nanostructured LaFeO3 electrode as a
photocathode, this material has become the most studied perovskite oxide for PEC in
recent years. In this work high purity powders were prepared by a novel ionic-liquid-
based method and then screen printed over an FTO substrate resulting in thick LaFeO3
nanoparticulate films. Despite cathodic photocurrents start appearing at 1.1 V vs RHE,
the low mobility of the holes and the competition of oxygen evolution reaction with
hole collection were found to be limiting factors for obtaining a high efficiency for
hydrogen evolution. Based on this work, Yu and co-workers119 constructed a photocell
of p-LaFeO3/n-Fe2O3 displaying an effective and stable photoresponse (faradaic
efficiency of 80.2 %), at least for 120 h using visible light (AM 1.5G radiation) but with
the application of a significant bias. The photocathode was fabricated by loading a thin
100 nm layer of LaFeO3 by PLD over a conductive substrate. In this context, it is
important to mention that several authors have reported anodic photocurrents for
LaFeO3 photoelectrodes, behaving thus as photoanodes.120–123 The electrical properties
of the material seem to critically depend on the synthesis procedure (including the type
of electrode architecture and the presence of impurities in the final material).
1.6.4. Other ternary oxides
There are other ternary oxide materials that cannot be classified into one of the above
structural groups. One of them is Sr7Fe10O22, which was presented by Matsumoto et al.
55
Chapter 1
in 198793 together with CaFe2O4, as a new photocathodic material for hydrogen
evolution. Quantum efficiencies up to 10% were reached with visible light, but authors
established that the photocurrents were not very stable.
One group of materials studied recently as photocathodes are the Cu(I) tantalates. They
have a general formula of CuxTa3n+1O8n+3 with a structure consisting of layers of TaO7
pentagonal bipyramids alternating with layers of linearly coordinated Cu(I) and TaO6
octahedra.87 Moreover, p-type films of semiconducting Cu5Ta11O30 and Cu3Ta7O19 have
been studied by the group of Maggard.124,125 They exhibit large initial photocurrents,
which dramatically drop over time. In the context of looking for narrow band gap
photocathode materials, the same research group also tested Cu(I) niobates126,127
(including the previously described CuNbO3)117 and Cu(I) vanadate.128 The latter,
synthesized with a high purity solid-state method exhibit a remarkably small optical
band gap size of ~1.2 eV. CuNb3O8 was also prepared by a solid state reaction and the
powder was coated over an FTO substrate.126 The photoelectrochemical properties were
measured at pH 6.3 using a visible source (>430 nm), achieving IPCE values around 6%
at 400 nm and faradaic efficiencies of 62%. The material exhibits a high hole mobility
(145 cm2 V 1
s 1), but a low acceptor density (~7·1015 cm 3). These materials will be
further discussed in section 1.7.3.
A combinatorial approach has recently been adopted for discovering new complex
oxide materials in the framework of the SHArK (Solar Hydrogen Activity Research Kit)
project by the group of Parkinson.129–139 An inkjet printing of metal oxide precursors is
used to fabricate multi-metal oxides with different compositions on the same FTO plate.
Then, the film is immersed into an aqueous electrolyte and a laser scanner interrogates
locally the multi-element metal oxide thin film to detect photoelectrochemical activity
under visible light. Fe0.84Cr1.0Al0.16O3 was identified by a combinatorial optimization as
the stoichiometry containing Fe, Cr and Al with the highest PEC activity for the
hydrogen evolution reaction.134 The material featured a band gap of 1.8 eV but the IPCE
5
was very low (5.5 · 10 % at 530 nm). Most recently, synthesis routes to prepare large
scale photocathode morphologies of quaternary FeCrAl oxides were investigated.138 A
mesoporous Fe0.84Cr1.0Al0.16O3 film reached an IPCE of 5% at 350 nm with an onset
potential located at 1.1 V vs RHE. The employment of an alternative synthesis route
56
Introduction
using a template leads to a macroporous inverse-opal structure reaching IPCEs of 28%,
while retaining the positive onset potential.138
Table 1.2. Review to the main parameters for photoelectrochemical water splitting of the
studied ternary oxide photocathodes: band gap, photocurrent onset potential and IPCE (Dir. and
Ind. stands for the band gap values derived from Tauc plots for direct and indirect transitions,
respectively).
Band gap / Eonset (vs IPCE at 400
Material Ref. Electrolyte
eV RHE) / V nm / %
93 1.9 (Ind.) 0.7-0.8 -- 0.25 M K2SO4 (pH 6)
1.9
94 0.8 -- 0.1 M H2SO4
(ref.93)
CaFe2O4
0.1 M NaOH, 0.1 M
95 1.9a 1.24b ca. 3d
Na2SO4, 0.1 M H2SO4
96 1.89 (Dir.) 1.0b -- 0.1 M Na2SO4
1.4-1.5 (Dir.) 0.1 M Na2SO4
98 a 1.05 ca. 1.1d
1.8 adjusted pH 10.8
ca.3 (at 450
99 1.8 (Dir.) 0.55b 0.1 M Na2SO4 (pH 6)
nm)d
CuBi2O4
0.3 M K2SO4 and 0.2
100 1.8a >1.0 ca. 2d M phosphate buffer
(pH 6.65)
140 1.83 (Dir.) 1.1c -- 0.1 M NaOH (pH 12.8)
d
CuCrO2 109 3.15 (Dir.) 1.06 ca. 1 0.1 M HClO4
105 ca. 1.55a 0.98 ca. 1.5 1 M NaOH (pH 13.5)
CuFeO2 3.10 (Dir.)
106 0.9 -- 1 M NaOH (pH 13.6)
1.47 (Ind.)
2.7 / 3.7 (Dir. 0.1 M KH2PO4 (pH
CuGaO2 110 0.9b ca. 0d
or Ind.) 4.8)
CuNbO3 117 ca. 2a 0.7b ca. 3.5d 0.5 M Na2SO4 (pH 6.3)
1.47 (Dir.)
CuNb3O8 126 0.7b ca. 5.5d 0.5 M Na2SO4 (pH 6.3)
1.26 (Ind.)
CuRhO2 108 ca. 1.9a 1.0b ca. 18d 1 M NaOH
Cu5Ta11O30 2.59 (Ref.141)a >0.7b --
124 141 a b
0.5 M Na2SO4 (pH 6.3)
Cu3Ta7O19 2.47 (Ref. ) >0.7 ca. 3.5d
1.17 (Dir.) /
Cu3VO4 128 0.7b,c -- 0.5 M Na2SO4 (pH 5.8)
1.14 (Ind.)
57
Chapter 1
Band gap / Eonset (vs IPCE at 400
Material Ref. Electrolyte
eV RHE) / V nm / %
5.5·10 5 (at H2(g) saturated-0.1 M
134 1.8e (Ind.) 1.0c
530 nm) HClO4
ca. 3
(mesoporou
Fe0.84CrAl0.16O3
1.8 s)d
138 1.1 0.1 M HClO4
(Ref.) ca. 28
(macroporou
s)
2.1
La2CuO4 142 0.9b ca. 0.2d KOH in water, pH 13
(Ref.143,144)
0.1 M Na2SO4
118 2.56 (Dir.) 1.1 ca. 0.35d
LaFeO3 adjusted pH to 12
119 -- 1.0b,c ca. 1.8d 1 M NaOH
114, Aqueous, adjusted to
LuRhO3 2.2 1.3b --
115 pH 7
Sr7Fe10O22 93 1.8 (Ind.) 0.7-0.8 -- 0.25 M K2SO4 (pH 6)
ZnRh2O4 101 2.0 (Ind.) 1.2 ca. 8d 0.1 M Na2SO4 (pH 6)
a
value extrapolated from the absorption edge
b
calculated as E(V vs RHE) = 0.059pH+ref, where ref is the potential of the reference electrode with respect to
SHE. When the measured pH value is not specified, the theoretically estimated value from the electrolyte
composition is considered (except for 0.1 M Na2SO4 for which pH = 6 is considered). Eref(SCE) = 0.244 V vs SHE
and Eref(Ag/AgCl(sat.)) = 0.199 V vs SHE.145
c
obtained visually from a linear voltammogram under (chopped) illumination
d
interpolated from an IPCE vs plot
e
extrapolated from an IPCE vs photon energy plot
1.7. Strategies for improving photoelectrochemical performance
The overall PEC water splitting reaction involves three main steps: (i) light absorption
by the semiconductor to generate electron-hole pairs, (ii) charge separation and
migration/diffusion of the minority charge carriers to the semiconductor surface (and
the majority charge carriers to the back contact), and (iii) reactions at the
semiconductor/electrolyte interface for water reduction or oxidation. Engineering the
electrode is the primary focus of the research aiming to improve the materials
performance in order to obtain an efficient solar water splitting device (as described in
section 1.5). In a general vein, common strategies73,146,147 focus on:
58
Introduction
I. Increasing light absorption by engineering the electronic structure of the
semiconductor by doping.
II. Enhancing majority charge carrier collection in the back contact by creating
p-n junctions, or increasing the conductivity by doping.
III. Diminishing surface recombination by loading passivation layers.
IV. Increasing the rate of charge carrier transfer to the electrolyte by
nanostructuring the material, attaching a surface cocatalyst or creating
heterojunctions.
V. Improving the stability (reducing photocorrosion) by loading protective
layers.
The organization of this section will be based on the effects generated by a particular
strategy. For each strategy, the mechanism by which the photoactivity increases is first
described. Second, the works dealing with it are reviewed, emphasizing those linked to
ternary oxides as photocathodes. A specific section for the enhancement of stability of
the materials is not included as few papers are devoted to this question in ternary oxide
materials. However, this effect is concomitant to other strategies, and as such they will
appear in this section.
Together with those mentioned above, it is important to mention that nanostructuring is
another widely used strategy in photoelectrochemistry (see section 1.4). Concretely, the
fabrication of nanostructured electrodes is one of the most employed approaches in
ternary oxide materials. Among the advantages of nanostructuring are the shorter carrier
collection pathway and the improvement of light distribution induced by light
scattering, which is notable for relatively large particles (microstructuring). On the
contrary, with nanostructured electrodes there is an increase of surface recombination
(because of the larger surface area), a reduced space charge layer thickness (the charge
carriers move by diffusion rather than by drift) and slow interparticle charge transport.
One of the challenges that need to be faced for optimal performance of ternary oxides
seems to be the slow mobility of carriers; thus the formation of 1D structures is thought
to be a promising route to enhance the PEC performance, as the minority carrier
pathway to reach the SEI is shortened. Apart from random ordered nanostructured
59
Chapter 1
electrodes, there is a lack of procedures for the formation of ordered nanostructures in
ternary oxide photocathodes. In this respect, it is important to mention a few works
addressing this issue.10,148
1.7.1. Improving light absorption by doping
One important challenge in finding photoelectrodes that meet the requirements specified
in section 1.5.2, is that the materials demonstrating high (photo)electrochemical stability
usually have too large band gap values, as in the case of TiO2. Thus, they absorb weakly
the visible range of the solar spectrum. The intentional introduction of dopants produces
extra energy levels within the forbidden gap that may enhance visible light absorption.
Additionally, these states can also act as recombination centers or bring new
recombination routes. Most of the studies on the enhancement of light absorption were
targeted on TiO2. It is reported that the introduction of C, N, and S as dopants in TiO 2
provide additional electronic states above the valence band edge of pure TiO 2, which
produces a red-shift absorption.149 Unfortunately, the improvement in the conversion
efficiency for TiO2 achieved following this strategy has not been satisfactory enough up
to now.150,151 In the context of ternary oxide materials, there is a lack of studies focusing
on the implementation of this approach. A few works show an enhancement of light
absorption by the introduction of a third metal in the structure (doping). The most
remarkable one is the Rh-doping of SrTiO3152,153 which not only switches the n-type
character of SrTiO3 to p-type (further detailed in next section) but also increases the
absorption coefficient in the visible region. Iwashina et al.152 established that an
absorption band appearing at ~400 nm is related with electronic transitions from
electron-donor levels consisting of Rh3+ and the other absorption band at ~600 nm was
linked to transitions involving acceptor levels consisting of Rh4+. Au and CuO doping in
CaFe2O4 has been shown to induce a red-shift of photoabsorption.154 On the other hand,
the introduction of Ta in CuNbO3 photoelectrodes yields a small blue shift of the band
gap size,155 although the effect does not produce a significant impact on the
photocurrent values.
60
Introduction
1.7.2. Improving separation of charge carriers by heterojunctions
In this methodology, a single-semiconductor photoelectrode is brought in contact with
an additional semiconductor forming a heterojunction, which is the junction between
two semiconductors having different band gaps. They are classified into several types,7
but the most employed for improving the PEC performance is the so-called type II p-n
junction. In this type of contact, electrons diffuse across the junction to the p-type
region, and holes diffuse to the n-type region until reaching equilibrium. This process
produces a lack of charge neutrality near the contact between both materials, creating a
positively charged region on the n-type side and negatively charged region in the p-type
side, generating a depletion region. Rectification occurs at the interface between both
semiconductors, and ideally electrons can flow only in one direction (from the p-type to
the n-type semiconductor). This strategy makes more efficient the extraction of the
photogenerated charge carriers, using an appropriate design. Figure 1.28 illustrates the
formation of a p-n junction for improving the performance both of a photoanode and of
a photocathode.
61
Chapter 1
e-
A.
Back contact
e-
Eredox
h+
h+
n-type p-type
B. e-
e-
Back contact
Eredox
h+
h+
p-type n-type
Figure 1.28. Band energy diagrams under photoexcitation showing the electron and hole flows
in a p-n junction for a photoanode (A) and a photocathode (B).
It is important to mention that heterojunctions are extensively employed in the field of
photocatalysis, including ternary oxides. However, in the present section, we will focus
mainly on the implementation for PEC reactions. Among them, apart from the p-n
heterojunction, p-p heterojunctions with adequate band alignments between both
materials are also employed to improve the photoresponse. Concerning ternary oxides,
there is a small number of reports dealing with heterojunctions for PEC H2 evolution.
One of them is the coating of ZnO on p-type ZnRh2O4 following the sol-gel method.101
The ZnO thin overlayer forms a space charge depletion region at the ZnO/ZnRh 2O4
interface (p-n heterojunction), which results in a better separation of the charge carriers
(besides improving electrical contact between ZnRh 2O4 particles). This approach results
in an IPCE value at 400 nm of ca. 13%, being ca. 8% in the untreated electrode. The
additional material for the heterojunction can also be formed by reaction of the ternary
oxide. For instance, the application of an oxidative heat treatment to Cu3VO4128 that
62
Introduction
generate CuO at the surface together with an oxygen excess in the structure (metal
deficiency) has been studied. After the treatment, an increase of the photoresponse was
observed. Later, Sullivan et al.125 identified the oxide/CuO junction in Cu5-xTa11O30
ternary oxide as the main cause for the enhancement of the photocurrent. This strategy
will be further described in section 1.7.3.
Regarding other oxide photocathodes, Cu2O and CuO are the most studied (especially
the former) and a substantial effort has been devoted to improve both their
photoresponse and their photostability. This has motivated the implementation of
heterojunctions with materials such as TiO2,156,157 -FeOOH158, ZnO159 and NiO78 for
solar cell applications and photocatalysis. The most relevant reports dealing with the use
of heterojunctions using Cu2O as an electrode material have been performed by the
group of Graetzel, which has developed a novel method for both improving the
photoresponse and protecting Cu2O photoelectrodes from photocorrosion.80,85,86,160 In an
earliest work, Paracchino et al.80 studied the deposition of TiO2 and ZnO/Al2O3 layers
by PLD over Cu2O for hydrogen evolution in 0.5 M Na2SO4 aqueous electrolyte, using
Pt as a cocatalyst. A Cu2O/21 nm ZnO/11 nm TiO2/Pt electrode exhibits photocurrents
three times larger than those of Cu2O/Pt. The ZnO buffer probably provides a more
uniform hydroxylated surface for TiO2 growth, in addition to forming a rectifying
junction with Cu2O. The introduction of ZnO/Al2O3 multilayers greatly enhanced the
stability of the photoresponse with time. The local electrostatic field formed at the
Cu2O/ZnO p-n junction may assist in extracting photoelectrogenerated electrons from
Cu2O. Later, following the method of Tilley et al. in Cu 2O electrodes,86 Prévot and
coworkers improved the photoresponse of the ternary oxide CuFeO2, with the drawback
of shifting the onset potential toward more negative values.106 Likewise, the most
important effect of the implemented junction was in the presence of Pt as cocatalyst. In
fact, direct loading of the catalyst at the CuFeO2 surface led to inconsistent results.
Another approach for increasing the extraction of holes was employed by Prévot et al.161
It consists in introducing a mesoporous p-CuAlO2 (scaffold layer) between the FTO and
the light absorber, p-CuFeO2. In this work, the optimization of the scaffold thickness
results in a 2.4-fold increase in photocurrent in the presence of an electron scavenger
(O2), while the photocurrent onset remained unaltered. The study suggests that a
63
Chapter 1
selective extraction of photogenerated holes toward the substrate at the CuFeO2/CuAlO2
interface is likely due to the favorable band alignment between both materials.
Ternary oxides have also been employed to improve both anodic and cathodic
photoresponses by the formation of p-n junctions. An example of such improvement in
p-type photoresponses is the electrode configuration FTO/CuO/CuBi2O4. It doubles the
photoactivity compared with the homo-layered electrode,162 which is explained based
on the enhanced electron-hole separation because of the valence band offset existing at
the solid-solid interface. In fact, when the heterojunction is inverted, the photocurrent
decreases because of a promotion of electron-hole recombination. On the other hand,
CaFe2O4 has mainly been employed in p-n junctions with the aim of enhacing
photoanodic responses by the combination with n-type materials, such as BiVO4,163,164
-Fe2O3165 or TaON.166 In the latter case, when CaFe2O4 was deposited over TaON by
electrophoresis, IPCE values increased six times (from 5 to 30 %) at 400 nm. Calcium
ferrite has also been involved in p-n heterojunctions in the field of photocatalysis, i.e.
CaFe2O4/WO3167 or CaFe2O4/MgFe2O4.168 Furthermore, CuFe2O4 was deposited on an
n-type WO3 photoelectrode,169 resulting in a 2.68 fold increase in the photoanodic
current compared to pristine WO3. The combination of CuFe2O4 with -Fe2O3 is also
effective for improved photoanodic performance.170
1.7.3. Improving charge collection by doping
Introducing a third metal (impurity) in the structure is expected to produce a doping
effect. There exists a considerable number of studies on doping ternary oxide
photocathodes with metals. Doping can enhance the p-type character via improving
majority carrier conductivity by the increase in the extrinsic charge carrier density
and/or the mobility of the charge carriers. Several effects derived from this strategy are
described in the literature, which include even switching the conductivity type (p- or n-
). Typically, a third metal with similar radius to one of the metal constituting the
intrinsic ternary oxide and having different oxidation state is introduced in the oxide
lattice. This induces the formation of impurities that can act as either donor or acceptor
levels, increasing the density of majority carriers, and consequently, increasing the
conductivity of the bulk material. Concomitantly, the introduction of such impurities
64
Introduction
can produce structural changes in the lattice that alter the mobility of carriers. In the
case of p-type oxide semiconductors, the third metal must have a lower oxidation state
of the substituted metal, inducing the creation of electron acceptor impurities. In this
section, the studies about doping of ternary oxides deriving in an increase of the p-type
conductivity (i.e. increase of the p-type character) will be reviewed.
Matsumoto et al. were the first to improve the photocurrent of a ternary oxide by doping
with metals. They introduced Mg and Na as dopants in a CaFe2O4/Pt-Pd pelletized
electrode.171 The low conductivity of the bare CaFe2O4 and the non-ohmic contact
created in the CaFe2O4/metal interface motivated doping with Na+ and Mg2+ near this
interface. They reported that Fe4+ was created by the substitution of Fe3+ by Na+ and
Mg2+, respectively. Fe4+ acts as an acceptor to create a hole in the valence band
according to Equation (1.63):
e4 e h (1.63)
The increment in hole density increases the material conductivity. The chemical formula
of the doped oxide can be written as a function of the dopant concentrations of Na+(x)
and Mg2+(y) as: a a e y e4y gy 4. It is important to mention that doping
the oxide in the CaFe2O4/metal interface produce the formation of an ohmic contact
(instead of the former non-ohmic), which also contributes to the improvement of the
photoresponse.
A similar doping effect was observed by Van´t Spijker et al.,142 who studied the
photoelectrochemical response of La2CuO4. They observed that doping with Ca, Ba and
Sr results in an increase of the photocurrents. An optimum PEC performance was found
for 4% of Ca, in which the IPCE reached 2% at 400 nm. More recently, Kang et al.140
introduced Ag in CuBi2O4. Ag+ replaced Bi3+ ions in the structure and increased hole
concentration. In addition to the positive effect on the photoresponse in the presence of
O2, photocorrosion of CuBi2O4 was also suppressed.
The introduction of acceptor impurities can change the n- to p-type character as stated
above. A representative example for this effect is the doping of SrTiO3 with Rh (already
mentioned in the above section) studied by Iwashina et al. for the first time.152
Electrodes with different values of x (ratio Rh/Ti) were prepared by a squeegee method.
65
Chapter 1
The cathodic photocurrents increased with larger percentages of Rh/(Ti+Rh) up to 7%,
and then decreased. Hydrogen evolution was detected in a SrTiO3:Rh(7%) by
determining a faradaic efficiency close to 100%. Later, Kawasaki et al.153 also studied
Rh-doped SrTiO3 forming epitaxial thin films fabricated by PLD. The oxygen pressure
and temperature selected during film growth were found to affect the oxidation state of
the Rh dopant (either Rh4+ or Rh3+). In the material containing Rh4+, the photocurrent
onset was shifted 0.6 V toward negative potentials with respect to the material
containing Rh3+.
Apart from the effect on the acceptor impurities, the conductivity can also be improved
by attaining higher carrier mobility, as stated above. Sekizawa et al.154 reported that Ag
doping in CaFe2O4 triggered an improvement in the symmetry around the Fe atom,
which induces high mobility (higher diffusion length of the charge carriers), which was
considered the main reason for the 23-fold increase in the cathodic photocurrent.
Another approach to improve the p-type character in ternary oxides is the application of
an oxidative treatment consisting in heating the material in an O2-rich atmosphere,
which is typically air. This leads to a metal deficiency (i.e. oxygen excess) in the
structure, generally linked to an increase in charge carrier density. This approach was
already mentioned in section 1.7.2. The Maggard group has extensively applied this
method to improve photoactivity in many ternary oxide materials. For instance, the
photoelectrochemical performance of CuNbO3 electrodes discussed in section 1.5.3 was
measured after a mild oxidation of the material at 250ºC.117 Before the heat treatment
the onset was located at lower potentials and the dark current associated to cathodic
corrosion was larger. Subsequent work155 revealed that after heating at 250 ºC in air, Cu
deficiency was generated owing to the formation of CuO at the surface, leading to a
stoichiometry Cu0.984NbO3. Similarly, an oxidation treatment at 250-550 °C in air of
Cu5Ta11O30 and Cu3Ta7O19 materials124 also produced a notable increase in the cathodic
photocurrents. In the Cu5Ta11O30 case, the enhancement was primarily attributed to
acceptor densities 4 orders of magnitude larger in the case of the oxidized samples (due
to the introduction of additional oxygen atoms as dopants). Later, a more detailed study
of the temperature and duration of the oxidation treatment was performed for
Cu5Ta11O30 electrodes.125 In this case, the largest increase of the photocurrents was
observed for 350 ºC, which was correlated with the largest changes in the amount of Cu
vacancies and also with the formation of CuO surface islands, which improved electron-
66
Introduction
hole separation (as commented in the above section). Nevertheless, the stability of these
materials under illumination in aqueous media was limited and a drop of photocurrent
with time was observed. Cu-deficiency was detected in p-type Cu3VO4 after heating
between 300 and 350 ºC.128 Again, the photocurrent enhancement was attributed to both
a higher degree of metal deficiency and the formation of a junction between the bare
material and CuO (Cu3V2O8 nanorods are also formed in the case of Cu3VO4). More
recently, a CuFeO2 photocathode was also submitted to an oxidation treatment at 300
°C106 and again, an increase in carrier density (detected by a Mott-Schottky analysis)
was linked to the improved photoactivity.
1.7.4. Decreasing surface recombination by passivation layers
Passivation layers denote the coating of additional layers that increase the PEC
performance with respect to the bare material by suppressing other competitive
processes consuming photogenerated carriers.147 This reaction can be a corrosion
reaction of the material or the recombination of photogenerated electron-holes at the
surface. In this section we will mainly focus on the latter effect. Very thin passivation
layers (frequently 1-2 nm) are usually used to suppress charge trapping in surface states
in order to avoid surface recombination. In this context, ALD is a useful technique, as it
allows for the deposition of thin and conformal coatings.
To the best of our knowledge, this approach has yet not been described for ternary oxide
materials. In the context of the improvement of PEC performance by passivation layers,
hematite received the most attention. 147 Le Formal and coworkers evinced the
passivation of surface states in hematite electrodes by coating a very thin layer (0.1 2
nm) of Al2O3 by ALD. The main effect of the modification was the decrease of the
overpotential for water photooxidation. Al2O3 thin layers were also deposited at
hematite surfaces by Cibrev et al.172 achieving a large increase of the photocurrents
values. The authors attributed such an effect to a change in the Fe-O bond properties,
although they did not discard a passivation layer effect.
Regarding p-type materials, only a few studies have been published until now related to
passivation effects (considering only very thin passivation layers). The main materials
used for this goal were Al2O3173,174 and TiO2175,176 for semiconductors such as Si and
67
Chapter 1
InP. In most cases, a decrease of the overpotential for PEC HER is achieved by this
approach, evinced by a shift of the photocurrent onset toward positive potentials due to
the passivation of the trapping centers.173 It is important to mention that, depending on
the passivation film thickness, usually another effect derived from the junction between
the absorber semiconductor and the coating may be present. Sometimes it is difficult to
differentiate between both effects.
1.7.5. Improving interfacial reaction kinetics by cocatalysts
Incorporating a cocatalyst on the absorber is an almost indispensable step in designing
highly efficient photoelectrodes. Even if the photogenerated electrons or holes possess
potentials appropriate for driving the water splitting reactions, they may recombine if a
reduced number of reaction sites exists on the surface of the semiconductor. Then, a
surface loaded cocatalyst is needed to facilitate the transfer of either electrons or holes
through the semiconductor/electrolyte interface as it reduces the energy barrier for the
reaction. Transition metals, especially noble metals, are traditionally used as effective
catalysts for H2 evolution. Metals not only act as electron sinks, but also provide
effective proton reduction sites. In principle, the larger the work function (that is, the
lower the Fermi level), the more facile trapping of electrons. Figure 1.29A displays the
contact between a metal cocatalyst loaded at the interface between the semiconductor
and the electrolyte. The Fermi level of the semiconductor (F), that of the cocatalyst
(F,M) and that of the redox couple in the electrolyte (redox) end up having the same
value at equilibrium. Under illumination, the photogenerated holes can be transferred to
the metal and subsequently to the electrolyte (Figure 1.29B). It is assumed that the
pseudo-Fermi level of electrons in the semiconductor is equilibrated with the metal
(cocatalyst) Fermi level.
Among transition metals, Pt is an extremely good electrocatalyst for HER, and
possesses a high work function. However, its low abundance on the Earth’s crust and its
high price restrict its use for scaling up devices for water splitting. This motivates the
search of abundant and cheap material catalysts compounds as an alternative to Pt.
68
Introduction
A. B. e-
F,M
*
F EF,n
F,M redox redox
F
h+
photoanode electrolyte photoanode electrolyte
Figure 1.29. Band energy diagrams in an n-type material in contact with a metal cocatalyst at
equilibrium (A) and under illumination (B).
Platinum metal has been demonstrated to be effective with the most often studied
photocathode materials, such as Cu2O,85,177,178 and Si.179,180 In the case of ternary oxides,
Pt was deposited electrochemically on CaFe2O4 and Sr7Fe10O22 surfaces, leading to
larger photocurrents due to the acceleration of the kinetics of H2 evolution and the
inhibition of surface recombination.93 Pt was also coated on CuBi2O4
photocathodes.100,140,162 In the work of Berglund et al.,100 IPCE values rose from about
3% for the bare material to 20% at 300 nm when Pt is loaded on CuBi2O4. Furthermore,
Pt was also deposited on the AZO (Al:ZnO) overlayer of CuFeO2, as mentioned in
section 1.7.2.106 To the best of our knowledge, this metal is the only cocatalyst studied
for the ternary oxide photoelectrodes sustaining HER.
The question that arises is why Pt is the only cocatalyst employed for ternary oxides. It
may be because most studies on PEC properties of these materials are fundamental in
nature still relatively far away from a practical device (in which a cheaper cocatalyst
would be required). Herein, the most employed cocatalyst for HER in photoelectrodes
are described. These materials can be studied in the future (instead of Pt) in ternary
oxide photocathodes.
Ni, NiO and NiOH have been described as effective low-cost cocatalysts for H2
generation.181–187 Although there are no precedents in the case of ternary oxide
photocathodes, its effectiveness has been shown in the case of p-Cu2O electrodes.
Somasundaram et al.188 reported a simple photodeposition method of Ni on Cu2O that
69
Chapter 1
increases the photocurrents in a PEC cell using Pt as a counter electrode (without the
application of a bias potential), and in the presence of a sacrificial hole acceptor (in the
absence of a hole acceptor photocurrents were very low because the photogenerated
holes are not sufficiently energetic as to oxidize water to O2). Another work based on a
novel nanocomposite Cu2O/NiOx, presented by Lin et al.189 allows for enhanced
catalysis and photostability of Cu2O. More recently, applied bias photon-to-current
efficiency (ABPE) of 2.7% have been reached for a Cu2O/CuO electrode decorated with
Ni.190
Ni-Mo cocatalysts have also demonstrated their catalytic activity for H2 generation as an
inexpensive alternative to noble metals. McKone et al.184 investigated the activity of the
Earth abundant cocatalysts Ni and Ni-Mo on Si electrodes being the energy conversion
efficiency for Ni and Ni-Mo catalysts quite similar. Lin et al.179 obtained similar
photocathodic currents using Ni-Mo catalyst compared to Pt in amorphous Si. In
reference to metal oxide photocathodes, Morales-Guio et al.191 reported the modification
of Cu2O with Ni-Mo cocatalyst, reaching 6.3 mA/cm2 at 0 V vs RHE.
The catalytic activity of RuO2 for the HER falls between those of platinum and nickel.
In fact, only an additional overvoltage of 30 mV is required in comparison with Pt.192
The group of Graetzel tested the RuOx catalyst86 with the previously developed strategy
of protecting Cu2O by means of AZO and TiO2 layers (as described in section 1.7.2).
Then, the RuOx catalyst was compared with Pt for hydrogen production and a great
improvement of the photocurrent with time was found while reaching faradaic
efficiencies for H2 evolution of 100%.
Amorphous molybdenum sulfide films are also promising cocatalysts for ternary oxides
photocathodes as they are composed by Earth abundant elements and they are active
hydrogen evolution catalysts.193,194 It has been reported that MoS2 is catalytically active
at pH solutions from 0 to 13, with relatively low overvoltage requirement for HER (ca.
100-200 mV).195,196 Tran et al.197 reported a photoassisted electrochemical method to
deposit MoS2 catalysts on a silicon nanowire array. The photocurrents at 0.2 V vs RHE
were 50 times higher and the onset shifted 200 mV toward positive potentials compared
to the untreated Si nanowires. Regarding oxide photocathodes, Morales-Guio and
coworkers integrated amorphous MoS2+x cocatalyst on surface-protected Cu2O,
70
Introduction
obtaining photocurrents similar to that obtained with the Ni-Mo catalyst mentioned
previously.191
Table 1.3. Review of the main strategies employed in ternary oxide photocathodes and the
effects produced in the photoelectrochemical performance.
Material Ref. Strategy Effects Observations
Ag doping leads to the
- Improving light best photoresponse.
Metal doping with Ag, absorption IPCE ca. 0.3% for
CaFe2O4 154
Au and CuO - Improving charge undoped and 2% for
mobility with Ag doping doped (at 420 nm in O2-
purged 0.2 M K2SO4).
- Improving charge Increase of the
CaFe2O4 and Loading a cocatalyst, transfer and inhibition photocurrents for both
93
Sr7Fe10O22 Pt of surface materials after Pt
recombination deposition.
- Increase of carrier The photocurrent
concentration and increase from ca. 0.35 to
Doping with Na and 2
improving conductivity 1.4 mA cm (in 0.1 M
CaFe2O4 171 Mg
- Generation of an ohmic H2SO4) after the
(Ca1 xNaxFe1 yMgyO4)
contact at the introduction of Na and
metal/oxide interface Mg.
- Increase reaction rate at IPCE increases from ca.
Loading a cocatalyst, the surface and/or 2% to ca. 9% after Pt
CuBi2O4 100
Pt decrease the surface loading (in pH 6.65 at
recombination 400 nm).
Increase of the
photocurrent values.
- Increase of hole density
Doping with Ag: IPCE of ca. 4.5% for Ar-
- Protection from
CuBi2O4 CuBi2O4/Ag- purged and ca. 8.5% in
photocorrosion (in
140 doped CuBi2O4 O2-purged 0.1 M NaOH
presence of O2)
(at 400 nm) for the
doped electrodes.
Loading a cocatalyst, Increase of the
CuBi2O4 Improving charge transfer
Pt photocurrents.
CuO/CuBi2O4/Pt doubles
- Heterojunction the photocurrent for the
heterojunction + improves charge homo-layered
CuBi2O4 162 cocatalyst, Pt: separation electrodes (also with Pt).
CuO/CuBi2O4/Pt - Cocatalyst increase IPCE reaches more than
reaction rate for HER 30% (in pH 6.8 at 400
nm).
71
Chapter 1
Material Ref. Strategy Effects Observations
- Improving light
absorption (due to the Photocurrents 2.5-fold
CuAlO2 layer as
different architecture) higher with the
CuFeO2 161 scaffold (p-p
- Improving hole composite (in O2-
junction)
extraction toward the purged 1 M NaOH).
substrate
Large increase of the
photocurrents up to 1
mA cm 2 (Ar-purged
Heterojunction + Understanding the
buffer electrolyte at pH
CuFeO2 106 cocatalyst, Pt: interface is subjected
6.1), albeit with onset
CuFeO2/AZO/TiO2/Pt to a future work
shift toward negative
values, from 0.9 to
0.4 V vs RHE.
Increase of the
acceptor density of two
- Increase of the times and the
acceptor density photocurrents increase
Oxidative heat
CuFeO2 106 - Better charge up to 100% in 2-layer
treatment
separation in the electrodes after the
depletion region oxidation by heat
treatment at 300 °C for
1 h.
Increase of the
- CuO formation photocurrents after the
Oxidative heat
CuNbO3 155 - Inducing metal oxidation heat
treatment
deficiency treatment of
250-350 °C for 3 h.
- Small increase of the
band gap with Ta.
Increase of light - Small differences in
CuNbO3 155 Doping with Ta
absorption the photocurrent
values with the
introduction of Ta.
Improving acceptor
Increase of the
density and
Oxidative heat photocurrent after the
Cu3Ta7O19 124 conductivity by
treatment heat treatment at 350
inducing metal
and 550 °C for 3 h.
deficiency
Improving acceptor The photocurrents
density and increase dramatically
Oxidative heat
Cu5Ta11O30 124 conductivity by after the heat
treatment
inducing metal treatment at
deficiency 250-350 °C for 3 h.
- Improving charge
carrier separation by
CuO formation Increase of the
(heterojunction) photocurrent after the
Oxidative heat
Cu5Ta11O30 125 heat treatment at
treatment - CuO also acts a
350- 550 °C for 15-60
cocatalyst
min.
- Inducing metal
deficiency
72
Introduction
Material Ref. Strategy Effects Observations
- Improving charge
carrier separation by
Decrease of dark
formation of CuO
currents and increase
Oxidative heat islands and Cu3V2O8
Cu3VO4 128 of photocurrent after
treatment rods at the surface
an oxidative treatment
(heterojunction)
(300-350 °C 15 min).
- Inducing metal
deficiency
IPCE increase from ca.
0.2% for undoped to
Doping with metals: Presumably increase in
ca. 2.5% for 4% Ca-
Ca, Ba and Sr the p-type character
doped (in pH 13 at 400
La2CuO4 142 nm).
- Improving hydrogen STH for Ca-doped
Loading a cocatalyst: desorption kinetics La2CuO4(4% ) are ca.
Pt and NiO - Protect from 0.1% using NiO and ca.
photocorrosion 1% using Pt.
Cathodic photocurrents
increase with the Rh
- Improving light content, up to an
absorption optimum for
SrTiO3 152 Doping with Rh
- Change n- to p-type Rh/(Rh+Ti) = 0.07.
character Faradaic efficiency of
100%.
- Improving light Photocurrent onset
Doping with Rh: Rh4+ absorption 0.6 V more positive for
SrTiO3 153
Rh3+:SrTiO3 than for
and Rh3+ - Change n- toe p-type
character Rh4+:SrTiO3.
IPCE rise from ca. 8%
ZnRh2O4 to ca. 13% for
p-n junction: Improving charge
ZnRh2O4 101 ZnO/ZnRh2O4 (at 400
ZnO/ZnRh2O4 separation
nm) in Ar-purged 0.1 M
Na2SO4 at 0 V vs RHE.
1.8. Bibliography
(1) Barber, J.; Tran, P. D. From natural to artificial photosynthesis. J. R. Soc. Interface 2013, 10,
20120984.
(2) Likvern, R. The Powers of Fossil Fuels; https://fractionalflow.com/2014/10/10/the-powers-of-
fossil-fuels/#more-869.
(3) Arrhenius, S. On the Influence of Carbonic Acid in the Air upon the Temperature of the Ground.
Philos. Mag. J. Sci. 1896, 41, 237–279.
73
Chapter 1
(4) Intergovernmental Panel on Climate Change (IPCC); http://www.ipcc.ch/.
(5) European commission; https://ec.europa.eu/energy/en/topics/nuclear-energy.
(6) Nuclear Energy Institute; https://www.nei.org/Knowledge-Center/Nuclear-Statistics/US-Nuclear-
Power-Plants/US-Nuclear-Generating-Statistics.
(7) Roel, van de K.; Grätzel, M. Chapter 1. Introduction. In Photoelectrochemical Hydrogen
Production; Roel, van de K., Grätzel, M., Eds.; Springer Science+Business Media, LLC: New
York (USA), 2012.
(8) Lewis, N. S.; Crabtree, G. Basic Research Needs for Solar Energy Utilization. In Report on Basic
Energy Sciences Workshop on Solar Energy Utilization, U.S. Department of Energy; Washington,
DC.
(9) ndez u iz, . . uervo arc a, R. Energía Solar Fotovoltaica undaci n onfemetal:
Madrid (Spain), 2007.
(10) Osterloh, F. E. Inorganic nanostructures for photoelectrochemical and photocatalytic water
splitting. Chem. Soc. Rev. 2013, 42, 2294–2320.
(11) Lewis, N. S. Research opportunities to advance solar energy utilization. Science 2016, 351,
aad1920.
(12) renberth, K. . asullo, . . rac ing arth’s nergy: rom l i o to lobal arming. Surv.
Geophys. 2012, 33, 413–426.
(13) Peter, L. M. Photoelectrochemical Water Splitting. A Status Assessment. Electroanalysis 2015,
27, 864–871.
(14) Office of Energy Efficiency and Renewable Energy;
https://energy.gov/eere/fuelcells/photoelectrochemical-working-group.
(15) Strategic Energy Technology (SET) Plan, European Commission. Annex I: Research and
innovation acction. 2017, p 154.
(16) Pleskov, Y. V.; Gurevich, Y. Y. Semiconductor Photoelectrochemistry; Consultants Bureau: New
York, 1986.
(17) Sato, N. Electrochemistry at Metal and Semiconductor Electrodes; Elsevier Science B. V.:
Amsterdam (The Netherland), 1998.
(18) Memming, R. Semiconductor Electrochemistry; WILEY-VCH Verlag GmbH: Weinheim
(Germany), 2001.
(19) Sze, S. M.; Ng, K. K. Physics of Semiconductor Devices; John Wiley & Sons, Inc.: Hoboken,
New Jersey (USA), 2007.
(20) Morrison, S. R. Electrochemistry at Semiconductor and Oxidized Metal Electrodes; Plenum
Press: New York (USA), 1980.
74
Introduction
(21) Shockley, W. On the surface states associated with a periodic potential. Phys. Rev. 1939, 56, 317–
323.
(22) Mertens, K. Textbook Photovoltaics; John Wiley & Sons Ltd., 2014.
(23) Rajeshwar, K. Fundamentals of Semiconductor Electrochemistry and Photoelectrochemistry;
Wiley-VCH: Weinheim (Germany), 2002.
(24) Jankulovska, M. Study of the electrochemical properties of nanostructured TiO 2 electrodes, PhD
thesis, Universitat d’ lacant, licante ain , .
(25) Peter, L. M. Chapter 1. Semiconductor Electrochemistry. In Photoelectrochemical Solar Fuel
Production: From Basic Principles to Advanced Devices; Giménez, S., Bisquert, J., Eds.;
Springer International Publishing: Switzerland, 2016.
(26) Trasatti, S. The Absolute Electrode Potential: an Explanatory Note. Pure Appl. Chem. 1986, 58,
955–966.
(27) Marcus, R. A. Reorganization free energy for electron transfers at liquid-liquid and dielectric
semiconductor-liquid interfaces. J. Phys. Chem. 1990, 94, 1050–1055.
(28) Gerischer, H. Heterogeneus electrochemical system for solar energy conversion. Pure. Appl.
Chem. 1980, 52, 2694–2667.
(29) Schmickler, W. Interfacial Electrochemistry; Oxford University Press, Inc.: New York (USA),
1996.
(30) Peter, L. M. Chapter 1. Photoelectrochemistry: From Basic Principles to Photocatalysis. In
Photocatalysis: Fundamentals and Perspectives; Schneider, J., Bahnemann, D., Ye, J., Puma, G.
L., Dionysiou, D. D., Eds.; The Royal Society of Chemistry: Cambridge (UK), 2016.
(31) Gärtner, W. W. Depletion-layer photoeffects in semiconductors. Phys. Rev. 1959, 116, 84–87.
(32) Peter, L. M. Dynamic aspects of semiconductor photoelectrochemistry. Chem. Rev. 1990, 90,
753–769.
(33) Berger, T.; Monllor-Satoca, D.; Jankulovska, M.; Lana-Villarreal, T.; Gómez, R. The
electrochemistry of nanostructured titanium dioxide electrodes. ChemPhysChem 2012, 13, 2824–
2875.
(34) Grätzel, M. Dye-sensitized solar cells. J. Photochem. Photobiol. C Photochem. Rev. 2003, 4,
145–153.
(35) Hagfeldt, A.; Boschloo, G.; Sun, L.; Kloo, L.; Pettersson, H. Dye-sensitized solar cells. Chem.
Rev. 2010, 110, 6595–6663.
(36) Nozik, A. J. Quantum dot solar cells. Physica E 2002, 14, 115–120.
(37) Kamat, P. V. Quantum dot solar cells. Semiconductor nanocrystals as light harvesters. J. Phys.
Chem. C 2008, 112, 18737–18753.
75
Chapter 1
(38) Bisquert, J.; Zaban, A.; Salvador, P. Analysis of the Mechanisms of Electron Recombination in
Nanoporous TiO2 Dye-Sensitized Solar Cells. Nonequilibrium Steady-State Statistics and
Interfacial Electron Transfer via Surface States. J. Phys. Chem. B 2002, 106, 8774–8782.
(39) Berger, T.; Lana-Villarreal, T.; Monllor-Satoca, D.; Gómez, R. An electrochemical study on the
nature of trap states in nanocrystalline rutile thin films. J. Phys. Chem. C 2007, 111, 9936–9942.
(40) Soedergren, S.; Hagfeldt, A.; Olsson, J.; Lindquist, S.-E. Theoretical Models for the Action
Spectrum and the Current-Voltage Characteristics of Microporous Semiconductor Films in
Photoelectrochemical Cells. J. Phys. Chem. 1994, 98, 5552–5556.
(41) Becquerel, E. Recherches sur les effets de la radiation chimique de la lumière solaire, au moyen
des courants électriques. C. R. Acad. Sci. 1839, 9, 45–149.
(42) 1954: Bell Labs demonstrates the first practical silicon solar cell. APS News. Am. Phys. Soc. April
2009, 18.
(43) Fujishima, A.; Honda, K. Electrochemical photolysis of water at a semiconductor electrode.
Nature 1972, 238, 37–38.
(44) Chen, Z.; Dinh, H. N.; Miller, E. Photoelectrochemical Water Splitting. Standars, Experimental
Methods, and Protocols; Springer Science & Business Media, 2013.
(45) Grätzel, M. Photoelectrochemical cells. Nature 2001, 414, 338–344.
(46) Sivula, K.; Grätzel, M. Chapter 4. Tandem Photoelectrochemical Cells for Water Splitting. In
Photoelectrochemical Water Splitting. Materials, Processes and Architectures; Lewerenz, H.,
Peter, L., Eds.; The Royal Society of Chemistry: Cambridge (UK), 2013.
(47) 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.
(48) Hanna, M. C.; Nozic, A. J. Solar conversion efficiency of photovoltaic and photoelectrolysis cells
carrier multiplication absorbers. J. Appl. Phys. 2006, 100, 074510.
(49) Pinaud, B. A.; Benck, J. D.; Seitz, L. C.; Forman, A. J.; Chen, Z.; Deutsch, T. G.; James, B. D.;
Baum, K. N.; Baum, G. N.; Ardo, S.; Wang, H.; Miller, E.; Jaramillo, T. F. Technical and
economic feasibility of centralized facilities for solar hydrogen production via photocatalysis and
photoelectrochemistry. Energy Environ. Sci. 2013, 6, 1983–2002.
(50) Döscher, H.; Geisz, J. F.; Deutsch, T. G.; Turner, J. A. Sunlight absorption in water – efficiency
and design implications for photoelectrochemical devices. Energy Environ. Sci. 2014, 7, 2951–
2956.
(51) Ni, M.; Leung, M. K. H.; Leung, D. Y. C.; Sumathy, K. A review and recent developments in
photocatalytic water-splitting using TiO2 for hydrogen production. Renew. Sustain. Energy Rev.
2007, 11, 401–425.
76
Introduction
(52) Lin, Y.; Yuan, G.; Liu, R.; Zhou, S.; Sheehan, S. W.; Wang, D. Semiconductor nanostructure-
based photoelectrochemical water splitting: A brief review. Chem. Phys. Lett. 2011, 507, 209–
215.
(53) ivula, K. e ormal, . rätzel, . olar water s litting: Progress using hematite -Fe2O3)
photoelectrodes. ChemSusChem 2011, 4, 432–449.
(54) Qiu, Y.; Leung, S.-F.; Zhang, Q.; Hua, B.; Lin, Q.; Wei, Z.; Tsui, K.; Zhang, Y.; Yang, S.; Fan,
Z. Efficient Photoelectrochemical Water Splitting with Ultrathin films of Hematite on Three-
Dimensional Nanophotonic Structures. Nano Lett. 2014, 14, 2123–2129.
(55) Tang, P.; Xie, H.; Ros, C.; Han, L.; Biset-Peiró, M.; He, Y.; Kramer, W.; Pérez Rodríguez, A.;
Saucedo, E.; Galán-Mascarós, J. R.; Andreu, T.; Morante, J. R.; Arbiol, J. Enhanced
photoelectrochemical water splitting of hematite multilayer nanowire photoanodes by tuning the
surface state via bottom-up interfacial engineering. Energy Environ. Sci. 2017, 10, 2124–2136.
(56) Carraro, G.; Maccato, C.; Gasparotto, A.; Warwick, M. E. A.; Sada, C.; Turner, S.; Bazzo, A.;
ndreu, . Plie hova, . Korte, . avrenčič Štangar, U. an endeloo, . orante, . .
Barreca, D. Hematite-based nanocomposites for light-activated applications: Synergistic role of
TiO2 and Au introduction. Sol. Energy Mater. Sol. Cells 2017, 159, 456–466.
(57) Fujimoto, I.; Wang, N.; Saito, R.; Miseki, Y.; Gunji, T.; Sayama, K. WO 3/BiVO4 composite
photoelectrode prepared by improved auto-combustion method for highly efficient water splitting.
Int. J. Hydrogen Energy 2014, 39, 2454–2461.
(58) Zhang, L.; Lin, C. Y.; Valev, V. K.; Reisner, E.; Steiner, U.; Baumberg, J. J. Plasmonic
enhancement in BiVO4 photonic crystals for efficient water splitting. Small 2014, 10, 3970–3978.
(59) Quiñonero, J.; Lana–Villarreal, T.; Gómez, R. Improving the photoactivity of bismuth vanadate
thin film photoanodes through doping and surface modification strategies. Appl. Catal. B
Environ. 2016, 194, 141–149.
(60) Murcia-López, S.; Fàbrega, C.; Monllor-Satoca, D.; Hernández-Alonso, M. D.; Penelas-Pérez, G.;
Morata, A.; Morante, J. R.; Andreu, T. Tailoring Multilayered BiVO 4 Photoanodes by Pulsed
Laser Deposition for Water Splitting. ACS Appl. Mater. Interfaces 2016, 8, 4076–4085.
(61) Bocarsly, A. B.; Bookbinder, D. C.; Dominey, R. N.; Lewis, N. S.; Wrighton, M. S.
Photoreduction at Illuminated p-Type Semiconducting Silicon Photoelectrodes. Evidence for
Fermi Level Pinning. J. Am. Chem. Soc. 1980, 102, 3683–3688.
(62) Sim, U.; Jeong, H.-Y.; Yang, T.-Y.; Nam, K. T. Nanostructural dependence of hydrogen
production in silicon photocathodes. J. Mater. Chem. A 2013, 1, 5414–5422.
(63) Coridan, R. H.; Shaner, M.; Wiggenhorn, C.; Brunschwig, B. S.; Lewis, N. S. Electrical and
Photoelectrochemical Properties of WO3/Si Tandem Photoelectrodes. J. Phys. Chem. C 2013,
117, 6949–6957.
77
Chapter 1
(64) Huang, S.; Zhang, H.; Wu, Z.; Kong, D.; Lin, D.; Fan, Y.; Yang, X.; Zhong, Z.; Huang, S.; Jiang,
Z.; Cheng, C. Large-area ordered p-type Si nanowire arrays as photocathode for highly efficient
photoelectrochemical hydrogen generation. ACS Appl. Mater. Interfaces 2014, 6, 12111–12118.
(65) Ros, C.; Andreu, T.; Hernández-Alonso, M. D.; Penelas-Pérez, G.; Arbiol, J.; Morante, J. R.
Charge Transfer Characterization of ALD-Grown TiO2 Protective Layers in Silicon
Photocathodes. ACS Appl. Mater. Interfaces 2017, 9, 17932–17941.
(66) Ma, Q.-B.; Kaiser, B.; Ziegler, J.; Fertig, D.; Jaegermann, W. XPS characterization and
photoelectrochemical behaviour of p-type 3C-SiC films on p-Si substrates for solar water
splitting. J. Phys. D. Appl. Phys. 2012, 45, 325101.
(67) Ma, Q. B.; Ziegler, J.; Kaiser, B.; Fertig, D.; Calvet, W.; Murugasen, E.; Jaegermann, W. Solar
water splitting with p-SiC film on p-Si: Photoelectrochemical behavior and XPS characterization.
Int. J. Hydrogen Energy 2014, 39, 1623–1629.
(68) Kato, M.; Yasuda, T.; Miyake, K.; Ichimura, M.; Hatayama, T. Epitaxial p-type SiC as a self-
driven photocathode for water splitting. Int. J. Hydrogen Energy 2014, 39, 4845–4849.
(69) Gao, L.; Cui, Y.; Wang, J.; Cavalli, A.; Standing, A.; Vu, T. T. T.; Verheijen, M. A.; Haverkort,
J. E. M.; Bakkers, E. P. A. M.; Notten, P. H. L. Photoelectrochemical Hydrogen Production on
InP Nanowire Arrays with Molybdenum Sulfide Electrocatalysts. Nano Lett. 2014, 3715–3719.
(70) Lin, Y.; Kapadia, R.; Yang, J.; Zheng, M.; Chen, K.; Hettick, M.; Yin, X.; Battaglia, C.; Sharp, I.
D.; Ager, J. W.; Javey, A. Role of TiO2 surface passivation on improving the performance of p-
InP photocathodes. J. Phys. Chem. C 2015, 119, 2308–2313.
(71) Gao, L.; Cui, Y.; Vervuurt, R. H. J.; Van Dam, D.; Van Veldhoven, R. P. J.; Hofmann, J. P.; Bol,
A. A.; Haverkort, J. E. M.; Notten, P. H. L.; Bakkers, E. P. A. M.; Hensen, E. J. M. High-
efficiency InP-based photocathode for hydrogen production by interface energetics design and
photon management. Adv. Funct. Mater. 2016, 26, 679–686.
(72) Awad, N. K.; Ashour, E. A.; Allam, N. K. Recent advances in the use of metal oxide-based
photocathodes for solar fuel production. J. Renew. Sustain. Energy 2014, 6, 022702.
(73) Chen, S.; Thind, S. S.; Chen, A. Nanostructured materials for water splitting - State of the art and
future needs: A mini-review. Electrochem. Commun. 2016, 63, 10–17.
(74) Huang, Q.; Ye, Z.; Xiao, X. Recent Progress in Photocathodes for Hydrogen Evolution. J. Mater.
Chem. A 2015, 3, 15824–15837.
(75) Kang, D.; Kim, T. W.; Kubota, S. R.; Cardiel, A. C.; Cha, H. G.; Choi, K. S. Electrochemical
Synthesis of Photoelectrodes and Catalysts for Use in Solar Water Splitting. Chem. Rev. 2015,
115, 12839–12887.
(76) Lianos, P. Review of recent trends in photoelectrocatalytic conversion of solar energy to
electricity and hydrogen. Appl. Catal. B Environ. 2017, 210, 235–254.
78
Introduction
(77) Jongh, P. E. De; Vanmaekelbergh, D.; Kelly, J. J. Photoelectrochemistry of Electrodeposited
Cu2O. J. Electrochem. Soc. 2000, 147, 486–489.
(78) Fernando, C. A. N.; De Silva, L. A. A.; Mehra, R. M.; Takahashi, K. Junction effects of p-Cu2O
photocathode with layers of hole transfer sites (Au) and electron transfer sites (NiO) at the
electrolyte interface. Semicond. Sci. Technol. 2001, 16, 433–438.
(79) Brown, K. E. R.; Choi, K. Electrochemical synthesis and characterization of transparent
nanocrystalline Cu2O films and their conversion to CuO films. Chem. Commun. 2006, 3311–
3313.
(80) Paracchino, A.; Laporte, V.; Sivula, K.; Grätzel, M.; Thimsen, E. Highly active oxide
photocathode for photoelectrochemical water reduction. Nat. Mater. 2011, 10, 456–461.
(81) Zhang, Z.; Wang, P. Highly stable copper oxide composite as an effective photocathode for water
splitting via a facile electrochemical synthesis strategy. J. Mater. Chem. 2012, 22, 2456–2464.
(82) Hsu, Y.-K.; Yu, C.-H.; Lin, H.-H.; Chen, Y.-C.; Lin, Y.-G. Template synthesis of copper oxide
nanowires for photoelectrochemical hydrogen generation. J. Electroanal. Chem. 2013, 704, 19–
23.
(83) Bornoz, P.; Abdi, F. F.; Tilley, S. D.; Dam, B.; Van De Krol, R.; Graetzel, M.; Sivula, K. A
bismuth vanadate-cuprous oxide tandem cell for overall solar water splitting. J. Phys. Chem. C
2014, 118, 16959–16966.
(84) Yang, Y.; Xu, D.; Wu, Q.; Diao, P. Cu2O/CuO Bilayered Composite as a High-Efficiency
Photocathode for Photoelectrochemical Hydrogen Evolution Reaction. Sci. Rep. 2016, 6, 35158.
(85) Paracchino, A.; Mathews, N.; Hisatomi, T.; Stefik, M.; Tilley, S. D.; Grätzel, M. Ultrathin films
on copper(I) oxide water splitting photocathodes: a study on performance and stability. Energy
Environ. Sci. 2012, 5, 8673–8681.
(86) illey, . . chreier, . zevedo, . tefi , . rätzel, M. Ruthenium Oxide Hydrogen
Evolution Catalysis on Composite Cuprous Oxide Water Splitting Photocathodes. Adv. Funct.
Mater. 2014, 24, 303–311.
(87) Sullivan, I.; Zoellner, B.; Maggard, P. A. Copper(I)-Based p-Type Oxides for
Photoelectrochemical and Photovoltaic Solar Energy Conversion. Chem. Mater. 2016, 28, 5999–
6016.
(88) Casbeer, E.; Sharma, V. K.; Li, X. Z. Synthesis and photocatalytic activity of ferrites under
visible light: A review. Sep. Purif. Technol. 2012, 87, 1–14.
(89) Saadi, S.; Bouguelia, A.; Trari, M. Photoassisted hydrogen evolution over spinel CuM 2O4 (M=Al,
Cr, Mn, Fe and Co). Renew. Energy 2006, 31, 2245–2256.
(90) Yang, H.; Yan, J.; Lu, Z.; Cheng, X.; Tang, Y. Photocatalytic activity evaluation of tetragonal
CuFe2O4 nanoparticles for the H2 evolution under visible light irradiation. J. Alloys Compd. 2009,
476, 715–719.
79
Chapter 1
(91) Rekhila, G.; Bessekhouad, Y.; Trari, M. Visible light hydrogen production on the novel ferrite
NiFe2O4. Int. J. Hydrogen Energy 2013, 38, 6335–6343.
(92) Dillert, R.; Taffa, D. H.; Wark, M.; Bredow, T.; Bahnemann, D. W. Research Update:
Photoelectrochemical water splitting and photocatalytic hydrogen production using ferrites
(MFe2O4) under visible light irradiation. APL Mater. 2015, 3, 104001.
(93) Matsumoto, Y.; Omae, M.; Sugiyama, K.; Sato, E. New Photocathode Materlals for Hydrogen
Evolution: CaFe2O4 and Sr7Fe10O22. J. Phys. Chem. 1987, 91, 577–581.
(94) Matsumoto, Y.; Sugiyama, K.; Sato, E. Photocathodic Hydrogen Evolution Reactions at p-Type
CaFe2O4 Electrodes with Fermi Level Pinning. J. Electrochem. Soc. 1988, 135, 98–104.
(95) Ida, S.; Yamada, K.; Matsunaga, T.; Hagiwara, H.; Matsumoto, Y.; Ishihara, T. Preparation of p-
type CaFe2O4 photocathodes for producing hydrogen from water. J. Am. Chem. Soc. 2010, 132,
17343–17345.
(96) Cao, J.; Kako, T.; Li, P.; Ouyang, S.; Ye, J. Fabrication of p-type CaFe2O4 nanofilms for
photoelectrochemical hydrogen generation. Electrochem. Commun. 2011, 13, 275–278.
(97) Arai, T.; Konishi, Y.; Iwasaki, Y.; Sugihara, H.; Sayama, K. High-Throughput Screening Using
Porous Photoelectrode for the Development of Visible-Light-Responsive Semiconductors. J.
Comb. Chem 2007, 9, 574–581.
(98) Hahn, N. T.; Holmberg, V. C.; Korgel, B. A.; Mullins, C. B. Electrochemical Synthesis and
Characterization of p-CuBi2O4 Thin Film Photocathodes. J. Phys. Chem. C 2012, 116, 6459–
6466.
(99) Nakabayashi, Y.; Nishikawa, M.; Nosaka, Y. Fabrication of CuBi2O4 photocathode through novel
anodic electrodeposition for solar hydrogen production. Electrochim. Acta 2014, 125, 191–198.
(100) Berglund, S. P.; Abdi, F. F.; Bogdanoff, P.; Chemseddine, A.; Friedrich, D.; van de Krol, R.
Comprehensive Evaluation of CuBi2O4 as a Photocathode Material for Photoelectrochemical
Water Splitting. Chem. Mater. 2016, 28, 4231–4242.
(101) Kamimura, S.; Higashi, M.; Abe, R.; Ohno, T. Fabrication of a porous ZnRh2O4 photocathode for
photoelectrochemical water splitting under visible light irradiation and a significant effect of
surface modification by ZnO necking treatment. J. Mater. Chem. A 2016, 4, 6116–6123.
(102) Kawazoe, H.; Yanagi, H.; Ueda, K.; Hosono, H. Transparent p-Type Conducting Oxides: Design
and Fabrication of p-n Heterojunctions. MRS Bull. 2000, 25, 28–36.
(103) Banerjee, A. N.; Chattopadhyay, K. K. Recent developments in the emerging field of crystalline
p-type transparent conducting oxide thin films. Prog. Cryst. Growth Charact. Mater. 2005, 50,
52–105.
(104) Vanaja, K. A. Growth and characterisation of p-type delafossite transparent conducting thin films
for heterojunction applications, PhD thesis, Cochin University of Science and Technology,
Cochin, Kerala (India), 2011.
80
Introduction
(105) Read, C. G.; Park, Y.; Choi, K.-S. Electrochemical Synthesis of p-Type CuFeO2 Electrodes for
Use in a Photoelectrochemical Cell. J. Phys. Chem. Lett. 2012, 3, 1872–1876.
(106) Prévot, M. S.; Guijarro, N.; Sivula, K. Enhancing the Performance of a Robust Sol-Gel-Processed
p-Type Delafossite CuFeO2 Photocathode for Solar Water Reduction. ChemSusChem 2015, 8,
1359–1367.
(107) Prévot, M. S.; Jeanbourquin, X. A.; Bourée, W. S.; Abdi, F.; Friedrich, D.; van de Krol, R.;
Guijarro, N.; Le Formal, F.; Sivula, K. Evaluating Charge Carrier Transport and Surface States in
CuFeO2 Photocathodes. Chem. Mater. 2017, 29, 4952–4962.
(108) Gu, J.; Yan, Y.; Krizan, J. W.; Gibson, Q. D.; Detweiler, Z. M.; Cava, R. J.; Bocarsly, A. B. p-
type CuRhO2 as a self-healing photoelectrode for water reduction under visible light. J. Am.
Chem. Soc. 2014, 136, 830–833.
(109) Díaz-García, A. K.; Lana-Villarreal, T.; Gómez, R. Sol–gel copper chromium delafossite thin
films as stable oxide photocathodes for water splitting. J. Mater. Chem. A 2015, 3, 19683–19687.
(110) Lee, M.; Kim, D.; Yoon, Y. T.; Kim, Y. Il. Photoelectrochemical Water Splitting on a Delafossite
CuGaO2 Semiconductor Electrode. Bull. Korean Chem. Soc. 2014, 35, 3261–3266.
(111) Mark Levy. Crystal Structure and Defect Properties in Ceramic Materials, PhD thesis, Imperial
College London, London (UK), 2005.
(112) Ishihara, T. Chapter 1. Structure and Properties of Perovskite Oxides. In Perovskite Oxide for
Solid Oxide Fuel Cells; Ishihara, T., Ed.; Springer Science+Business Media, LLC: New York
(USA), 2009.
(113) Cotton, F. A.; Wilkinson, G. Química Inorgánica Avanzada; Editorial Limusa S. A.: Limusa
(Mexico), 1993.
(114) Company, P. Photoelectrochemical Properties of LuRhO 3. Surf. Sci. 1980, 101, 205–213.
(115) Jarrett, H. S.; Sleight, A. W.; Kung, H. H.; Gillson, J. L. Photoelectrochemical and solid-state
properties of LuRhO3. J. Appl. Phys. 1980, 51, 3916–3925.
(116) Joiret, S.; Campet, G.; Claverie, J. Photoelectrochemical behaviour of LaCoO 3 ceramic and single
crystal. Matterial Lett. 1987, 5, 468–474.
(117) Joshi, U. A.; Palasyuk, A. M.; Maggard, P. A. Photoelectrochemical Investigation and Electronic
Structure of a p-Type CuNbO3 Photocathode. J. Phys. Chem. C 2011, 115, 13534–13539.
(118) Celorrio, V.; Bradley, K.; Weber, O. J.; Hall, S. R.; Fermín, D. J. Photoelectrochemical Properties
of LaFeO3 Nanoparticles. ChemElectroChem 2014, 1, 1667–1671.
(119) Yu, Q.; Meng, X.; Wang, T.; Li, P.; Liu, L.; Chang, K.; Liu, G.; Ye, J. Highly Durable p-
LaFeO3/n-Fe2O3 Photocell for Effective Water Splitting under Visible Light. Chem. Commun.
2015, 51, 3630–3633.
81
Chapter 1
(120) Sora, I. N.; Fontana, F.; Passalacqua, R.; Ampelli, C.; Perathoner, S.; Centi, G.; Parrino, F.;
Palmisano, L. Photoelectrochemical properties of doped lanthanum orthoferrites. Electrochim.
Acta 2013, 109, 710–715.
(121) Peng, Q.; Shan, B.; Wen, Y.; Chen, R. Enhanced charge transport of LaFeO 3 via transition metal
(Mn, Co, Cu) doping for visible light photoelectrochemical water oxidation. Int. J. Hydrogen
Energy 2015, 40, 15423–15431.
(122) Peng, Q.; Wen, Y.; Shan, B.; Chen, R. Enhanced Oxygen Evolution Reaction Activity of LaFeO 3
Coated with Co-Pi. ECS Trans. 2015, 64, 27–35.
(123) Peng, Q.; Wang, J.; Wen, Y. W.; Shan, B.; Chen, R. Surface modification of LaFeO 3 by Co-Pi
electrochemical deposition as an efficient photoanode under visible light. RSC Adv. 2016, 6,
26192–26198.
(124) Fuoco, L.; Joshi, U. A.; Maggard, P. A. Preparation and Photoelectrochemical Properties of p-
type Cu5Ta11O30 and Cu3Ta7O19 Semiconducting Polycrystalline Films. J. Phys. Chem. C 2012,
116, 10490–10497.
(125) Sullivan, I.; Sahoo, P. P.; Fuoco, L.; Hewitt, A. S.; Stuart, S.; Dougherty, D.; Maggard, P. A. Cu-
Deficiency in the p-Type Semiconductor Cu − Ta11O30: Impact on Its Crystalline Structure,
Surfaces, and Photoelectrochemical Properties. Chem. Mater. 2014, 26, 6711–7621.
(126) Joshi, U. A.; Maggard, P. A. CuNb3O8: A p-type semiconducting metal oxide photoelectrode. J.
Phys. Chem. Lett. 2012, 3, 1577–1581.
(127) Kamimura, S.; Murakami, N.; Tsubota, T.; Ohno, T. Fabrication and characterization of a p-type
Cu3Nb2O8 photocathode toward photoelectrochemical reduction of carbon dioxide. Appl. Catal. B
Environ. 2015, 174–175, 471–476.
(128) Sahoo, P. P.; Zoellner, B.; Maggard, P. A. Optical, electronic, and photoelectrochemical
properties of the p-type Cu − VO4 semiconductor. J. Mater. Chem. A 2015, 3, 4501–4509.
(129) Woodhouse, M.; Parkinson, B. A. Combinatorial discovery and optimization of a complex oxide
with water photoelectrolysis activity. Chem. Mater. 2008, 20, 2495–2502.
(130) Woodhouse, M.; Parkinson, B. A. Combinatorial approaches for the identification and
optimization of oxide semiconductors for efficient solar photoelectrolysis. Chem Soc Rev 2009,
38, 197–210.
(131) He, J.; Parkinson, B. A. Combinatorial investigation of the effects of the incorporation of Ti, Si,
and l on the erformance of -Fe2O3 photoanodes. ACS Comb. Sci. 2011, 13, 399–404.
(132) Seley, D.; Ayers, K.; Parkinson, B. A. Combinatorial search for improved metal oxide oxygen
evolution electrocatalysts in acidic electrolytes. ACS Comb. Sci. 2013, 15, 82–89.
(133) Jiang, C.; Wang, R.; Parkinson, B. A. Combinatorial Approach to Improve Photoelectrodes Based
on BiVO4. ACS Comb. Sci. 2013, 15, 639–645.
82
Introduction
(134) Rowley, J. G.; Do, T. D.; Cleary, D. A.; Parkinson, B. A. Combinatorial discovery through a
distributed outreach program: Investigation of the photoelectrolysis activity of p-type Fe, Cr, Al
oxides. ACS Appl. Mater. Interfaces 2014, 6, 9046–9052.
(135) Newhouse, P. F.; Parkinson, B. A. Combinatorial optimization of spinel Co 3-xMxO4 M = (Al, Ga,
In) alloyed thin films prepared by ink jet printing: photoelectrochemical, optical, and structural
properties. J. Mater. Chem. A 2015, 3, 5901–5907.
(136) Skorupska, K.; Maggard, P. A.; Eichberger, R.; Schwarzburg, K.; Shahbazi, P.; Zoellner, B.;
Parkinson, B. A. Combinatorial Investigations of High Temperature CuNb Oxide Phases for
Photoelectrochemical Water Splitting. ACS Comb. Sci. 2015, 17, 742–751.
(137) Sliozberg, K.; Stein, H. S.; Khare, C.; Parkinson, B. A.; Ludwig, A.; Schuhmann, W. Fe-Cr-Al
containing oxide semiconductors as potential solar water-splitting materials. ACS Appl. Mater.
Interfaces 2015, 7, 4883–4889.
(138) Kondofers y, I. üller, . unn, . K. Ivanova, . Štefanić, . hrens erger, . cheu, C.;
Parkinson, B. A.; Fattakhova-Rohlfing, D.; Bein, T. Nanostructured Ternary FeCrAl Oxide
Photocathodes for Water Photoelectrolysis. J. Am. Chem. Soc. 2016, 138, 1860–1867.
(139) Skorupska, K.; Parkinson, B. A. Chapter 10. Combinatorial Synthesis and Screening of Oxide
Materials for Photoelectrochemical Energy Conversion. In Photoelectrochemical Solar Fuel
Production: From Basic Principles to Advanced Devices; Jiménez, S., Bisquert, J., Eds.; Springer
International Publishing: Switzerland, 2016.
(140) Kang, D.; Hill, J. C.; Park, Y.; Choi, K.-S. Photoelectrochemical Properties and Photostabilities
of High Surface Area CuBi2O4 and Ag-Doped CuBi2O4 Photocathodes. Chem. Mater. 2016, 28,
4331–4340.
(141) Palasyuk, O.; Palasyuk, A.; Maggard, P. A. Syntheses, optical properties and electronic structures
of copper(I) tantalates: Cu5Ta11O30 and Cu3Ta7O19. J. Solid State Chem. 2010, 183, 814–822.
(142) van’t i er, . imon, . oms, . Photocatalytic water s litting by means of undo ed and
doped La2CuO4 photocathodes. Int. J. Hydrogen Energy 2008, 33, 6414–6419.
(143) Falck, J. P.; Levy, A.; Kastner, M. A.; Birgeneau, R. J. Charge-Transfer Spectrum and Its
Temperature Dependence in La2CuO4. Phys. Rev. Lett. 1992, 69, 1109–1112.
(144) Quijada, M. A.; Tanner, D. B.; Chou, F. C.; Johnston, D. C.; Cheong, S.-W. Optical Properties of
single-crystal La2CuO4 δ. Phys. Rev. B 1995, 52, 15485–15503.
(145) Sawyer, D. T.; Sobkowiak, A.; Roberts, J. L. Electrochemistry for Chemists; John Wiley & Sons:
New York (USA), 1995.
(146) Zhao, Y.; Zhang, Y.; Yang, Z.; Yan, Y.; Sun, K. Synthesis of MoS 2 and MoO2 for their
applications in H2 generation and lithium ion batteries: a review. Sci. Technol. Adv. Mater. 2013,
14, 043501.
83
Chapter 1
(147) Liu, R.; Zheng, Z.; Spurgeon, J.; Yang, X. Enhanced photoelectrochemical water-splitting
performance of semiconductors by surface passivation layers. Energy Environ. Sci. 2014, 7,
2504–2517.
(148) Mao, Y.; Park, T.-J.; Wong, S. S. Synthesis of classes of ternary metal oxide nanostructures.
Chem. Commun. 2005, 5721–5735.
(149) Chen, X.; Burda, C. The Electronic Origin of the Visible-Light Absorption Properties of C- , N-
and S-Doped TiO2 Nanomaterials. J. Am. Chem. Soc. 2008, 130, 5018–5019.
(150) Prévot, M. S.; Sivula, K. Photoelectrochemical Tandem Cells for Solar Water Splitting. J. Phys.
Chem. C 2013, 117, 17879–17893.
(151) Liu, H.; Imanishi, A.; Nakato, Y. Mechanisms for Photooxidation Reactions of Water and
Organic Compounds on Carbon-Doped Titanium Dioxide. J. Phys. Chem. C 2007, 111, 8603–
8610.
(152) Iwashina, K.; Kudo, A. Rh-Doped SrTiO3 Photocatalyst Electrode Showing Cathodic
Photocurrent for Water Splitting under Visible-Light Irradiation. J. Am. Chem. Soc. 2011, 133,
13272–13275.
(153) Kawasaki, S.; Nakatsuji, K.; Yoshinobu, J.; Komori, F.; Takahashi, R.; Lippmaa, M.; Mase, K.;
Kudo, A. Epitaxial Rh-doped SrTiO3 thin film photocathode for water splitting under visible light
irradiation. Appl. Phys. Lett. 2012, 101, 033910.
(154) Sekizawa, K.; Nonaka, T.; Arai, T.; Morikawa, T. Structural improvement of CaFe 2O4 by metal
doping toward enhanced cathodic photocurrent. ACS Appl. Mater. Interfaces 2014, 6, 10969–
10973.
(155) Zoellner, B.; Stuart, S.; Chung, C.-C.; Dougherty, D. B.; Jones, J. L.; Maggard, P. A.
CuNb − TaxO3 ≤ . solid solutions: im act of a substitution and u I deficiency on
their structure, photocatalytic, and photoelectrochemical properties. J. Mater. Chem. A 2016, 4,
3115–3126.
(156) Siripala, W.; Ivanovskaya, A.; Jaramillo, T. F.; Baeck, S. H.; McFarland, E. W. A Cu 2O/TiO2
heterojunction thin film cathode for photoelectrocatalysis. Sol. Energy Mater. Sol. Cells 2003, 77,
229–237.
(157) Tian, X.; Li, S.; Cao, Y.; Xu, Y.; Zhang, G. Preparation, optical property, and photocatalytic
activity of cubic Cu2O/amorphous TiO2 and spheric CuO/TiO2 core-shell nanocomposites. Mater.
Lett. 2014, 131, 86–89.
(158) Vequizo, J. J. M.; Ichimura, M. Fabrication of Cu 2 γ-FeOOH heterojunction solar cells using
electrodeposition. Appl. Phys. Express 2014, 7, 045501.
(159) Kargar, A.; Jing, Y.; Kim, S. J.; Riley, C. T.; Pan, X.; Wang, D. ZnO/CuO heterojunction
branched nanowires for photoelectrochemical hydrogen generation. ACS Nano 2013, 7, 11112–
11120.
84
Introduction
(160) Azevedo, J.; Steier, L.; Dias, P.; Stefik, M.; Sousa, C. T.; Araújo, J. P.; Mendes, A.; Graetzel, M.;
Tilley, S. D. On the stability enhancement of cuprous oxide water splitting photocathodes by low
temperature steam annealing. Energy Environ. Sci. 2014, 7, 4044–4052.
(161) Prévot, M. S.; Li, Y.; Guijarro, N.; Sivula, K. Improving charge collection with delafossite
photocathodes: a host–guest CuAlO2/CuFeO2 approach. J. Mater. Chem. A 2016, 4, 3018–3026.
(162) Park, H. S.; Lee, C.-Y.; Reisner, E. Photoelectrochemical reduction of aqueous protons with a
CuO|CuBi2O4 heterojunction under visible light irradiation. Phys. Chem. Chem. Phys. 2014, 16,
22462–22465.
(163) Kim, E. S.; Kang, H. J.; Magesh, G.; Kim, J. Y.; Jang, J.; Lee, J. S. Improved
Photoelectrochemical Activity of CaFe2O4/BiVO4 Heterojunction Photoanode by Reduced
Surface Recombination in Solar Water Oxidation. ACS Appl. Mater. Interfaces 2014, 6, 17762–
17769.
(164) Srinivasan, N.; Sakai, E.; Miyauchi, M. Balanced Excitation between Two Semiconductors in
Bulk Heterojunction Z-Scheme System for Overall Water Splitting. ACS Catal. 2016, 6, 2197–
2200.
(165) Ahmed, M. G.; Kandiel, T. a.; Ahmed, A. Y.; Kretschmer, I.; Rashwan, F. A.; Bahnemann, D. W.
Enhanced Photoelectrochemical Water Oxidation on Hematite Photoanode via p-CaFe2O4/n-
Fe2O3 Heterojunction Formation. J. Phys. Chem. C 2015, 119, 5864–5871.
(166) Kim, E. S.; Nishimura, N.; Magesh, G.; Kim, J. Y.; Jang, J.; Jun, H.; Kubota, J.; Domen, K.; Lee,
J. S. Fabrication of CaFe2O4/TaON Heterojunction Photoanode for Photoelectrochemical Water
Oxidation. J. Am. Chem. Soc. 2013, 135, 5375–5383.
(167) Liu, Z.; Zhao, Z. G.; Miyauchi, M. Efficient visible light active CaFe 2O4/WO3 based composite
photocatalysts: Effect of interfacial modification. J. Phys. Chem. C 2009, 113, 17132–17137.
(168) Kim, H. G.; Borse, P. H.; Jang, J. S.; Jeong, E. D.; Jung, O.-S.; Suh, Y. J.; Lee, J. S. Fabrication
of CaFe2O4/MgFe2O4 bulk heterojunction for enhanced visible light photocatalysis. Chem.
Commun. 2009, 5889–5891.
(169) Liu, Y.; He, H.; Li, J.; Li, W.; Yang, Y. H.; Li, Y.; Chen, Q. Preparation and enhanced
photoelectrochemical performance of p-n heterojunction CuFe2O4/WO3 nanocomposite film. RSC
Adv. 2015, 5, 99378–99384.
(170) Hussain, S.; Hussain, S.; Waleed, A.; Tavakoli, M. M.; Wang, Z.; Yang, S.; Fan, Z.; Nadeem, M.
A. Fabrication of CuFe2O4 -Fe2O3 Composite Thin Films on FTO Coated Glass and 3-D
Nanospike Structures for Efficient Photoelectrochemical Water Splitting. ACS Appl. Mater.
Interfaces 2016, 8, 35315–35322.
(171) Matsumoto, Y.; Sugiyama, K.; Sato, E.-I. Improvement of CaFe2O4 photocathode by doping with
Na and Mg. J. Solid State Chem. 1988, 74, 117–125.
85
Chapter 1
(172) Tallarida, M.; Das, C.; Cibrev, D.; Kukli, K.; Tamm, A.; Ritala, M.; Lana-Villarreal, T.; Gómez,
R.; Leskelä, M.; Schmeisser, D. Modification of Hematite Electronic Properties with Trimethyl
Aluminum to Enhance the Efficiency of Photoelectrodes. J. Phys. Chem. Lett. 2014, 5, 3582–
3587.
(173) Choi, M. J.; Jung, J.-Y.; Park, M.-J.; Song, J.-W.; Lee, J.-H.; Bang, J. H. Long-term durable
silicon photocathode protected by a thin Al2O3/SiOx layer for photoelectrochemical hydrogen
evolution. J. Mater. Chem. A 2014, 2, 2928–2933.
(174) Park, M.-J.; Jung, J.-Y.; Nam, Y.-H.; Song, J.-W.; Jeong, C.; Lee, J.-H. Improved
photoelectrochemical hydrogen evolution using a defect-passivated Al2O3 thin film on p-Si. Thin
Solid Films 2016, 616, 550–554.
(175) Lee, M. H.; Takei, K.; Zhang, J.; Kapadia, R.; Zheng, M.; Chen, Y.-Z.; Nah, J.; Matthews, T. S.;
Chueh, Y.-L.; Ager, J. W.; Javey, A. p-Type InP Nanopillar Photocathodes for Efficient Solar-
Driven Hydrogen Production. Angew. Chemie Int. Ed. 2012, 51, 10760–10764.
(176) Seger, B.; Pedersen, T.; Laursen, A. B.; Vesborg, P. C. K.; Hansen, O.; Chorkendorff, I. Using
TiO2 as a Conductive Protective Layer for Photocathodic H2 Evolution. J. Am. Chem. Soc. 2013,
135, 1057–1064.
(177) Chatchai, P.; Nosaka, A. Y.; Nosaka, Y. The Effect of Platinum Deposition on the Water Photo-
Reduction at p-Cu2O Semiconductor Electrodes with Visible Light Irradiation. Electrochemistry
2011, 79, 821–825.
(178) Jin, Z.; Hu, Z.; Yu, J. C.; Wang, J. Room temperature synthesis of a highly active Cu/Cu 2O
photocathode for photoelectrochemical water splitting. J. Mater. Chem. A 2016, 4, 13736–13741.
(179) Lin, Y.; Battaglia, C.; Boccard, M.; Hettick, M.; Yu, Z.; Ballif, C.; Ager, J. W.; Javey, A.
Amorphous Si thin film based photocathodes with high photovoltage for efficient hydrogen
production. Nano Lett. 2013, 13, 5615–5618.
(180) Dasgupta, N. P.; Liu, C.; Andrews, S.; Prinz, F. B.; Yang, P. Atomic layer deposition of platinum
catalysts on nanowire surfaces for photoelectrochemical water reduction. J. Am. Chem. Soc. 2013,
135, 12932–12935.
(181) Sreethawong, T.; Suzuki, Y.; Yoshikawa, S. Photocatalytic evolution of hydrogen over
mesoporous TiO2 supported NiO photocatalyst prepared by single-step sol-gel process with
surfactant template. Int. J. Hydrogen Energy 2005, 30, 1053–1062.
(182) Jang, J. S.; Choi, S. H.; Kim, D. H.; Jang, J. W.; Lee, K. S.; Lee, J. S. Enhanced photocatalytic
hydrogen production from water-methanol solution by nickel intercalated into titanate nanotube.
J. Phys. Chem. C 2009, 113, 8990–8996.
(183) Ran, J.; Yu, J.; Jaroniec, M. Ni(OH) 2 modified CdS nanorods for highly efficient visible-light-
driven photocatalytic H2 generation. Green Chem. 2011, 13, 2708–2713.
86
Introduction
(184) McKone, J. R.; Warren, E. L.; Bierman, M. J.; Boettcher, S. W.; Brunschwig, B. S.; Lewis, N. S.;
Gray, H. B. Evaluation of Pt, Ni, and Ni–Mo electrocatalysts for hydrogen evolution on
crystalline Si electrodes. Energy Environ. Sci. 2011, 4, 3573–3583.
(185) Tran, P. D.; Xi, L.; Batabyal, S. K.; Wong, L. H.; Barber, J.; Loo, J. S. C. Enhancing the
photocatalytic efficiency of TiO2 nanopowders for H2 production by using non-noble transition
metal co-catalysts. Phys. Chem. Chem. Phys. 2012, 14, 11596–11599.
(186) Sun, K.; Madsen, K.; Andersen, P.; Bao, W.; Sun, Z.; Wang, D. Metal on metal oxide nanowire
Co-catalyzed Si photocathode for solar water splitting. Nanotechnology 2012, 23, 194013.
(187) Ding, C.; Qin, W.; Wang, N.; Liu, G.; Wang, Z.; Yan, P.; Shi, J.; Li, C. Solar-to-hydrogen
efficiency exceeding 2.5% achieved for overall water splitting with an all earth-abundant dual-
photoelectrode. Phys. Chem. Chem. Phys. 2014, 16, 15608–15614.
(188) Somasundaram, S.; Raman Nair Chenthamarakshan, C.; de Tacconi, N. R.; Rajeshwar, K.
Photocatalytic production of hydrogen from electrodeposited p-Cu2O film and sacrificial electron
donors. Int. J. Hydrogen Energy 2007, 32, 4661–4669.
(189) Lin, C.-Y.; Lai, Y.-H.; Mersch, D.; Reisner, E. Cu2O/NiOx nanocomposite as an inexpensive
photocathode in photoelectrochemical water splitting. Chem. Sci. 2012, 3, 3482–3487.
(190) Dubale, A. A.; Pan, C.-J.; Tamirat, A. G.; Chen, H.-M.; Su, W.-N.; Chen, C.-H.; Rick, J.; Ayele,
D. W.; Aragaw, B. A.; Lee, J.-F.; Yang, Y.-W.; Hwang, B.-J. Heterostructured Cu2O/CuO
decorated with nickel as a highly efficient photocathode for photoelectrochemical water
reduction. J. Mater. Chem. A 2015, 3, 12482–12499.
(191) Morales-Guio, C. G.; Liardet, L.; Mayer, M. T.; Tilley, S. D.; Grätzel, M.; Hu, X.
Photoelectrochemical Hydrogen Production in Alkaline Solutions Using Cu 2O Coated with Earth-
Abundant Hydrogen Evolution Catalysts. Angew. Chemie Int. Ed. 2015, 54, 664–667.
(192) Kötz, E. R.; Stucki, S. Ruthenium dioxide as a hydrogen-evolving cathode. J. Appl. Electrochem.
1987, 17, 1190–1197.
(193) Merki, D.; Fierro, S.; Vrubel, H.; Hu, X. Amorphous molybdenum sulfide films as catalysts for
electrochemical hydrogen production in water. Chem. Sci. 2011, 2, 1262–1267.
(194) Morales-Guio, C. G.; Hu, X. Amorphous molybdenum sulfides as hydrogen evolution catalysts.
Acc. Chem. Res. 2014, 47, 2671–2681.
(195) Jaramillo, T. F.; Jørgensen, K. P.; Bonde, J.; Nielsen, J. H.; Horch, S.; Chorkendorff, I.
Identification of active edge sites for electrochemical H2 evolution from MoS2 nanocatalysts.
Science 2007, 317, 100–102.
(196) Merki, D.; Hu, X. Recent developments of molybdenum and tungsten sulfides as hydrogen
evolution catalysts. Energy Environ. Sci. 2011, 4, 3878–3888.
87
Chapter 1
(197) Tran, P. D.; Pramana, S. S.; Kale, V. S.; Nguyen, M.; Chiam, S. Y.; Batabyal, S. K.; Wong, L. H.;
Barber, J.; Loo, J. Novel assembly of an MoS 2 electrocatalyst onto a silicon nanowire array
electrode to construct a photocathode composed of elements abundant on the earth for hydrogen
generation. Chem. Eur. J. 2012, 18, 13994–13999.
88
Chapter 2
Experimental methods
2.1. Preparation of ternary oxide electrodes ................................................................ 91
2.1.1. Methods based on pre-synthesized particles.................................................. 93
2.1.2. Methods based on direct growth on the substrate ........................................ 98
2.2. Photoelectrochemical methods ............................................................................. 101
2.2.1. Linear and cyclic voltammetry ..................................................................... 104
2.2.2. Chronoamperometry ..................................................................................... 107
2.2.3. Chronopotentiometry and photopotential ................................................... 109
2.2.4. (Photo)electrochemical impedance spectroscopy (PEIS) ............................ 111
2.2.4.1. Constant frequency (variable potential), Mott-Schottky representation ......114
2.2.4.2. Constant potential (variable frequency) Nyquist representation ..................116
2.2.5. Electrochemical quartz-crystal microbalance (EQCM) ............................. 118
2.3. Spectroscopic methods .......................................................................................... 119
2.3.1. UV-visible spectroscopy ................................................................................. 119
2.3.2. X-ray Diffraction (XRD) ................................................................................ 123
2.3.3. X-ray photoelectron spectroscopy (XPS) ..................................................... 126
2.4. Microscopic methods ............................................................................................. 129
2.4.1. Transmission electron microscopy (TEM) ................................................... 131
2.4.2. Scanning electron microscopy (SEM) .......................................................... 133
2.5. Bibliography ........................................................................................................... 136
Experimental methods
2.1. Preparation of ternary oxide electrodes
The photoelectrochemical activity of the ternary oxide electrodes depends on their
physical properties such as crystallinity, morphology, crystal structure and surface area.
Such factors critically depend on the employed synthetic method. To fabricate the
electrode, the semiconductor must be attached to a conductive substrate that collects the
charge carriers at the back contact of the electrode material. In the literature, different
techniques can be found for preparing p-type ternary oxide electrodes. In this thesis,
some of these approaches are adopted to prepare photoelectrodes. The different methods
can be divided into two types: one is based on electrode fabrication from a pre-
synthesized powder and the other on the direct growth of the oxide on a substrate.
Producing high-quality ternary oxide photoelectrodes is a challenge, mainly because
they are thermodynamically less stable than simple binary compounds and binary oxide
impurities are easily formed. Additionally, when the ternary oxide electrode is
fabricated starting from a pre-synthesized powder, a good contact among the particles
and between the particles and the conducting substrate is often not achieved. This
generates poor charge transport and collection of charge carriers. Hence, electrode
preparation is a critical issue for a suitable photoelectrochemical performance.
91
Chapter 2
Electrodeposition
Methods based Spin Coating
on direct Dip coating
growth Drop casting
Hydrothermal
METHODS FOR
ELECTRODE PREPARATION
Nanoparticle preparation
Doctor Blade Methods based Commercial NPs
Electrophoretic deposition Solid State
on pre-synthesized
Screen printing Sol-Gel
Spin coating particles
Hydrothermal
Dip coating
Figure 2.1. Outline of the main preparation methods for p-type oxide electrodes.
The use of transparent substrates for supporting the semiconductor oxide is highly
advisable for photoelectrochemical applications. In fact, the most employed substrates
are composed by a glass sheet covered with a thin film of F:SnO2 (FTO) or In2O3:SnO2
(ITO). They are chosen because of their good properties: high transparency, high
conductivity and weak absorption in the visible. FTO presents several advantages over
ITO: it is chemically more stable, it has a broader electrochemical window and it also
withstands higher temperatures. Thus, FTO glass is more convenient for
photoelectrochemical experiments.
Generally, ternary oxides require synthesis temperatures higher than binary oxide
materials for the obtainment of highly pure phases. Thus, the threshold temperature of
the substrate is an imperative restriction for preparing these materials. In the most
common type of FTO, glass melting occurs at lower temperatures than F:SnO2
degradation, and consequently, the glass type used in the preparation of the optically
transparent electrode is crucial in this respect. FTO fabricated using borosilicate glass
92
Experimental methods
can withstand higher temperatures, but it is more expensive. Additionally, migration of
tin atoms from FTO into the absorber material can occur before glass melts. In fact,
doping with Sn coming from the substrate in the case of hematite photoelectrodes has
been reported,1,2 although in these studies, it has a positive effect in the photoresponse.
Apart from glass, other substrates can be employed with the shortcoming that the
electrode can only be illuminated from the front side (EE). They are usually metals and
alloys such as stainless steel, titanium, platinum, copper, etc. Despite their good
conductivity and the formation of favorable junctions with the SC, they have some
limitations. In some metals, in addition to the glass melting point threshold, heating at
high temperatures leads to the formation of a non-conductive oxide layer on the metal
surface (insulating or semiconductor layer). For instance, when Ti is heated at mild
temperatures in air, a TiO2 layer is easily formed over metallic Ti. This restricts the use
of many of these substrates because of the high temperatures needed for this application.
Platinum does not present these constrains and it has been employed as a substrate for
photoelectrodes being submitted to temperatures over 1000 ºC.3 However, its high cost
severely limits its applicability.
The above-mentioned shortcomings have motivated the development of synthetic
methods based on pre-synthesized pure-phase nanoparticles. In such cases, the
nanoparticles are deposited on the substrate and only a mild heat treatment is required
for sinterization. This is the most extended synthesis route for preparing p-type ternary
photoelectrodes. Unfortunately, it can result in poor electric contact among particles
and/or between particles and substrate, as stated above. On the other hand, direct growth
of the oxide on the substrate can be advisable when the synthesis conditions allow for it.
Based on the above approaches, the different methods employed for preparing ternary
oxide photoelectrodes in the framework of this thesis are described in the following.
2.1.1. Methods based on pre-synthesized particles
This section focuses on methods to prepare ternary oxide materials based on pre-
synthesized particles (Figure 2.1). First, we will deal with pre-synthesized particles
deposition methods and, then, with the different procedures for preparing nano- or
micro-particles.
93
Chapter 2
A well-known technique for preparing thin film nanostructured electrodes is the doctor
blade method developed in the 1940s for constructing thin sheets of piezoelectric
materials and capacitors.4 In this method, a well-mixed slurry consisting of a suspension
of particles and containing additives (such as binders or dispersants) is deposited on the
substrate. Previously, the substrate has been fixed to the bench and the deposit area has
been delimited by pieces of adhesive tape as illustrated in Figure 2.2. Then, the paste is
spread over the substrate using a glass rod. The thickness of the tape will determine the
amount of deposited slurry, and therefore, the final thickness of the film (together with
the slurry properties such as particle concentration, viscosity, density, etc.). The
employed additives are critical for the morphology, thickness and porosity of the final
electrode. In order to obtain the dry and pure material, an annealing treatment is
required with the purpose of evaporating the solvent and burning up the organic
additives. The temperature must be high enough to sinterize the particles among them
and with the substrate in order to achieve a good electric contact. This temperature is
typically in the range of 300-500 °C.
Adhesive tapes
A.
FTO
Glass rod
Delimited area
B.
C. slurry
FTO
FTO
slurry
Figure 2.2. Scheme illustrating the doctor blade procedure. (A) shows an FTO plate while the
slurry is spread over the area delimited by the adhesive tapes and (B) and (C) show the same set
after deposition, being (B) the upper view and (C) the lateral view. The thickness of the tape is
exaggerated in order to properly visualize the procedure.
94
Experimental methods
Electrophoretic deposition (EPD)5–8 is an electrochemical technique performed in two
steps. First, two electrodes are immersed in a liquid containing the suspended particles.
One electrode is the substrate (working electrode) and the other is a counter-electrode.
The particles are forced to move toward the electrode (substrate) by applying an electric
field (application of a high potential bias between both electrodes), which is the
electrophoresis step. In the second step, the particles are attracted to the electrode
forming a consistent deposit constituted by agglomerated particles on the substrate. An
additional step (which usually is a heat treatment) is required in order to sinter the
deposited particles, leading to a mechanically stable electrode. The particles in the
liquid can be charged by different mechanisms that include ion adsorption of solution
species on the solid particle or dissociation of species from the solid phase into the
liquid.8 Depending on the charge sign of the particles, the substrate will be the cathode
or the anode in the electrophoretic cell. The resulting film thickness depends on factors
such as applied potential, time, distance between electrodes or suspension
concentration.
Screen printing9,10 is also a useful technique to prepare porous films. It is similar to
doctor blade but having the advantage that it can be scaled up more easily. A paste is
formed from a powder and it is placed on top of a screen. Afterward, the paste is
extended using a squeegee that is drawn across the screen by applying pressure thereby
forcing the paste to pass through the open areas of the screen. The printing quality
depends on the composition of the paste, pressure and speed of the squeegee, among
other parameters.
Spin coating and dip coating are also alternative routes for preparing electrodes from a
suspension/dispersion of the pre-synthesized powder or by direct growth on the
substrate. These methods will be described in the next section.
The particle powder used for the above-mentioned methods can be either commercially
purchased or synthesized. The latter is the most employed option as not many ternary
oxide powders are commercially available. The traditional approach adopted with this
purpose is the solid state synthesis from binary oxide precursors, but it has the
drawback of the high temperatures and annealing times required for obtaining the pure
phase. It is important to mention that in several works a pellet of the ternary oxide is
fabricated by pressing the pre-synthesized powder to obtain the photoelectrode.11–14
95
Chapter 2
Afterward, the pellet is sintered at high temperatures and electrically contacted with a
metal wire. Generally, this method leads to low photoresponses, partly by the large
thickness of the electrode, which favors bulk recombination of the photogenerated
carriers.
Sol-gel15–18 is a wet chemical method. Together with the solid-state method is one of the
most extensively used for preparing nano- and micro-powders of the ternary metal
oxides. In a general vein, it consists in the formation of a sol from metal inorganic salts
or metal alkoxide solutions by hydrolysis reactions. A sol is a stable dispersion of
colloidal particles (or polymer molecules) of 1-100 nm in a solvent. By a dehydration
and/or pH control, a gel is generated by condensation and polymerization reactions. A
gel is an interconnected, rigid network with pores of submicrometer dimensions and
polymeric chains. The evaporation of the solvent results in a xerogel and a heat
treatment is required to obtain the pure-phase material. Figure 2.3 illustrates different
routes in the sol-gel method. The above-described procedure generates dense particles.
If the sol is coated on a substrate, dense films can be obtained (that will be mentioned in
the next section). The precipitation of the sol leads to uniform particles and the solvent
extraction of the gel results in an aerogel.
96
Experimental methods
Solution
Metal
solution
Xerogel
Heat Dense
Sol film
Uniform
particles
Gelation /
evaporation
Gel Aerogel
Xerogel
Heat Dense
particles
Figure 2.3. Scheme showing the main steps for the sol-gel method. Depending on the synthesis
route, uniform particles, aerogels, dense particles or dense films can be obtained.
Sol-gel processing allows for controlling the structure of the material on a nanometer
scale from the earliest stages of processing. The main advantage is that high-purity
materials are obtained if the precursors are pure, using lower temperatures than in the
solid state synthesis. Importantly, not only rigorous control of the stoichiometry of the
products can be achieved, but also the production of complex oxides and new hybrid
organic-inorganic materials is possible.19
Hydrothermal methods20,21 are also widely employed to prepare nanoparticles with a
variety of morphologies. In these methods, the material is generated by heating a
precursor aqueous solution to high temperatures under autogenous pressures.
Commonly the reactor is a sealed vessel, known as an autoclave. When an organic
solvent is used instead of water, it is called “solvothermal” method. The formation of
the nanoparticles occurs in two steps: (i) nucleation and (ii) growth. The particle size
97
Chapter 2
and morphology can be tuned by controlling temperature, pH, precursor concentration
or through the introduction of additives. It is important to mention that, in the literature,
the term “hydrothermal method” is also employed for processes at ambient temperature
involving water as a solvent.
Although most of the methods described above have been employed under the scope of
the investigations leading to this thesis, many of them are not reflected in the results
appearing herein. A commercial powder has been employed to prepare CuFe 2O4
electrodes in Chapter 3. Moreover, a sol-gel method has been employed in the synthesis
of CaFe2O4 particles (Chapter 4) and NiTiO3 nanorods (Chapter 7). The preparation of
YFeO3 nanoparticles (in Chapter 6) has been performed by using a novel ionic-liquid
method that was recently described.22 The pre-synthesized nanoparticles are deposited
on the substrate by a doctor blade method for the preparation of CuFe2O4, YFeO3 and
NiTiO3 electrodes. This method has been widely employed for TiO2 electrodes23,24 and
the specific procedure is usually adapted to other oxide materials. The suspensions have
been prepared using additives such as Triton 100X,25 which acts as a surfactant and
acetylacetone,26 which favors the disaggregation of the particles. CaFe2O4 powder was
deposited by direct loading of an alcoholic suspension of the nanoparticles over a Pt
substrate.
2.1.2. Methods based on direct growth on the substrate
The growth of a nanostructure or a compact architecture of the oxide directly on the
substrate can be achieved as long as the temperature required to obtain the pure-phase
oxide does not conflict with the heating substrate limitations commented above.
Generally, these methods provide good electrical continuity, which was the main
challenge pointed out previously for the methods based on pre-synthesized particles.
Electrochemical deposition or electrodeposition20,27–29 is a powerful technique for
growing high-quality films based on the occurrence of redox reactions at the
electrode/solution interface. The formation of nanostructures is also possible. Metal
electrodeposition is a widely employed technique in which metal is reduced on the
electrode surface.30,31 Deposition occurs in two steps: (a) nucleation followed by (b)
growth of the formed nuclei.
98
Experimental methods
Electrodeposition may not only generate a conformal deposit of the material, but also
allow for the easy control of the synthesis variables for tuning the resulting material
properties and morphologies (pH, additives, solvent, temperature, etc.). Both, the
applied current and potential are critical parameters that determine not only the amount
deposited, but also the stoichiometry and composition of the film. For these reasons, the
technique has been extensively adopted for the synthesis of metal oxides32–37 by
different mechanisms such as (a) cathodic deposition (b) anodic deposition) and (c) pH
changes that induce precipitation. In turn, the synthesis of the oxide by cathodic
deposition can be carried out in different ways. Namely, Cu2O has been deposited by
direct growth of the crystalline oxide.34,37 In other cases, the amorphous and/or hydrated
form of the oxide is deposited and a final heat treatment is required as in the
electrodeposition of WO3.29,38 Another strategy is based on electrodepositing a metal
layer that is subsequently thermally treated to produce the oxide. Fe2O3 and CuBi2O4
have been prepared in this way.39,40 Also, metal oxides can be prepared by anodic
deposition, which has been employed for the synthesis of hematite. 41 The
electrodeposition of Fe2+ leads to FeOOH and a subsequent heat treatment generates
Fe2O3. Finally, an electrochemical reaction can produce pH changes, inducing the
precipitation of the metal oxide. For instance, TiO2 has been prepared from a solution of
Ti(OH)442 in which the local alkalinization caused by H2O reduction decreases the
solubility of the Ti(OH)4, inducing TiO2 formation. The electrodeposition of complex
oxides has the drawback that metals must be deposited with a particular stoichiometry.
This limits the application of this approach for ternary oxides. Regarding photocathodes
and, as far as we know, exclusively CuFeO2 and CuBi2O4 have been prepared by this
route.29,40
Spin coating allows for the preparation of uniform thin films (even a few nanometers in
thickness) on flat substrates. A volume of a precursor suspension or colloidal solution is
deposited in the center of the substrate, and then it is rotated at a high speed. Centrifugal
forces cause that part of the liquid spreads, leaving a thin film on the substrate. The
solvent employed is usually volatile, and rapidly evaporates. The thickness of the film
critically depends on the nature of the suspension or solution (viscosity, density,
concentration, surface tension, vapor pressure of the solvent, etc.), the rotating speed
and time, and the nature of the substrate (i.e. roughness). The higher the applied rotation
speed, the thinner the obtained film. Finally, the films are heated to high temperatures to
99
Chapter 2
decompose the organics and crystallize the material. The above-described process is
schematized in Figure 2.4. The thickness can be increased by the deposition of
multilayers.
Suspension
or solution
w
Figure 2.4. Schematic representation of the spin coating procedure.
The dip-coating technique is used to prepare uniform thin films by dipping the substrate
vertically in a suspension or (colloidal) solution of the precursors. Once immersed, it is
pulled up at a constant rate and concomitantly a film is deposited. The withdrawal speed
is determinant for the film thickness. During this process the solution naturally and
homogeneously spreads over the surface of the substrate as a result of the competition
of several forces in the film deposition region such as surface tension, capillarity or
viscous drag.43 Then, the solvent evaporates from the liquid, forming the thin layer.
Similarly to the spin coating technique, a final heat treatment is required. Thicker films
can be obtained by repeating the above procedure.
Spin coating and dip coating techniques are usually combined with sol-gel processing
for preparing a wide variety of thin film materials, including p-type electrodes.44,45 In
spin coating, the precursor solution is commonly the sol and the generated film is
composed by a xerogel, which is a monolith resulted from the evaporation of the solvent
contained in the deposited sol. It contains the hydrous metal oxide15 and the anions
present in the metallic precursor salts. Usually acetates, nitrates and chlorides are
selected as to facilitate their removal upon heat treatment.
In the drop casting technique, a small volume of a precursor solution is dropped on a
delimited area of a substrate. The remnant solvent is evaporated under ambient
conditions or with mild heating. Then, it is heated at a particular temperature to form the
100
Experimental methods
desired material. This method allows for a direct control of the deposited metal/s by
setting the concentration and the deposited volume of the precursor solution. It has the
advantage of being a simple and practical method and there is a minimum waste of
reagents in the synthesis. It is important to mention that the ternary oxide CuBi2O4 has
been prepared by drop casting.46–48 Drop casting is frequently used for loading a
submonolayer coverage on a pre-synthesized material, for instance, the deposition of a
cocatalyst over the semiconductor.49,50
Among the results shown in this thesis, spin coating combined with the sol-gel method
has been employed for the fabrication of compact crystalline LaFeO 3 films (Chapter 5)
and amorphous YFeO3 films (Chapter 6).
2.2. Photoelectrochemical methods
(Photo)Electrochemical techniques are the basic tools to evaluate the
photoelectrochemical performance of among others, ternary oxide materials. The
methods employed in this thesis are detailed in the present section.
The setup for the experiments is a three-electrode system that can work under either
dark or illumination conditions. A scheme of a three-electrode cell is outlined in Figure
2.5A. It incorporates the working electrode (WE), a reference electrode (RE) and a
counter-electrode (or auxiliary electrode, CE) immersed in the electrolytic solution. The
RE has a well-known and stable potential used as a reference to measure the working
electrode potential. A bias is applied between the RE and the WE, and enables to
control the so called WE potential. The CE allows for closing the circuit and has the
role of carrying out the reaction coupled (oxidation or reduction) to that taking place on
the WE (reduction or oxidation). The applied potential is controlled by a
potentiostat,51,52 as schematized in Figure 2.5B. It adjusts the current flowing between
WE and CE, in order to set the potential bias between the WE and RE to the input value.
101
Chapter 2
POTENTIOSTAT
A. B.
Function
A generator
V
RE CE
RE
Potentiostat
WE CE E controlled WE
i(t) measured
Figure 2.5. (A) Scheme of a three-electrode cell. WE = working electrode, RE = Reference
electrode and CE = counter-electrode. (B) Basic arrangement for controlled-potential
experiments with a potentiostat.
The PEC cell and its components are shown in Figure 2.6. The cell is typically made of
glass and is outfitted with a threaded opening to adapt a fused quartz disc constituting a
window through which the electrode is illuminated. Electrode illumination through the
quartz window allows for UV and visible light to get into the cell. The cell is provided
with an additional compartment for placing the reference electrode, which is joined to
the main body of the cell through a porous glass plate. It also incorporates gas inlet and
outlet attachments. The gas inlet is equipped with a valve that regulates the gas flow
either through the solution or above the solution. This system allows to purge the
electrolyte by letting the gas flow through it and then switching to gas flow through the
head space of the cell during the measurements, thus avoiding the entry of ambient air
in the cell.
102
Experimental methods
Working electrode
Counter electrode Gas inlet
(Pt wire)
Reference electrode
(Ag/AgCl/KCl(sat))
Gas outlet
Electrolyte
Quarz
window
Figure 2.6. Three-electrode glass cell employed in PEC measurements with the different
components indicated with labels.
The RE employed in the measurements is a commercial Ag/AgCl/KCl(sat) electrode,
which is the most popular in electrochemistry together with the saturated calomel
electrode (SCE). The use of the latter has diminished in recent years because of the
health and environmental problems associated with mercury. Throughout this thesis, it
will be often required to refer the measured potential to the reversible hydrogen scales.
It will be considered that the reversible potential of Ag/AgCl,KCl(sat) is +0.197 V
against SHE. The potentials measured vs Ag/AgCl can be converted to the SHE and
reversible hydrogen electrode (RHE) scales as:
vs vs l . (2.1)
vs vs l . . p (2.2)
In PEC water splitting, electrolytes containing water are obviously required. The
working solutions in this thesis are aqueous electrolytes at different pHs. They can be 1
M or 0.1 M NaOH for basic, HClO4 for acid and Na2SO4 for neutral pHs. For the
103
Chapter 2
photocathodes, the presence of O2 is critical for the measured photocurrents, as oxygen
is an electron scavenger. Thus, the solution is purged with an inert gas (N2 or Ar) for
evaluating water reduction, and purged with O2 when oxygen reduction is the reaction
of interest. These experimental settings are detailed in each chapter.
The main illumination sources employed in this thesis are (a) an 1000 W Xe(Hg) arc
lamp (Newport) equipped with a water filter to suppress the IR radiation, whose
spectrum is displayed in Figure 2.7, and (b) a solar simulator SUN 2000 (ABET
Technologies), which has an emission spectrum corresponding to the AM1.5G (shown
in Figure 1.8). It is important to mention that the emitted irradiance is quite different in
both cases, while the Xe(Hg) arc lamp has a strong emission in the UV range, in the
solar simulator the irradiance is larger in the visible range.
Figure 2.7. Spectral irradiance of the 1000 W Xe(Hg) arc lamp (circled in yellow).53
2.2.1. Linear and cyclic voltammetry
Potential sweep techniques are widely used because they provide preliminary
information in a fast way about the electrodic processes, e.g. the potential at which an
104
Experimental methods
electrochemical process occurs. The simplest technique is the linear sweep voltammetry
(LSV) that consists on sweeping the electrode potential between two potential limits (E1
to E2) at a certain scan rate, . As mentioned above, the potentiostat controls the
potential difference between WE and RE and the current flowing between WE and CE
is recorded. The concept of cyclic voltammetry (CV) is essentially the same, but when
potential reaches E2, the sweep is inverted, as illustrated in Figure 2.8A. The common
representation is a current vs potential plot, as presented in Figure 2.8B, in which the
current is usually normalized per geometric electrode area (current density, j).
A. E B.
t = tf /2
E2 t = tf /2
-2
j/Am
t= 0
t = tf
E1
E1 E2
t=0 t = tf time E/V
Figure 2.8. (A) Variation of potential with time and (B) variation of current density as a
function of potential in a cyclic voltammogram.
The recorded current results from both faradaic and non-faradaic contributions. In a
faradaic process, there is an electron transfer through the electrode-electrolyte interface
associated with a redox reaction O + ne R. Such reactions are governed by
Faraday´s law:
d dn
ne ne v (2.3)
dt dt
where jF is the faradaic current density, Q is the charge involved in the Faradaic process,
ne is the number of electrons exchanged per mole of reactant, n is the number of moles
that have reacted and v is the reaction rate per unit geometric area. For pure faradaic
processes, the current can be directly related with the reaction rate of the redox process.
105
Chapter 2
In cyclic voltammetry, when the current is limited by the diffusion of the species, it can
be described by the Randles–Sevcik equation,54 which established that the current is
proportional to the square root of the scan rate.
On the contrary, in a non-faradaic process there is no charge transfer across the
interface. It is typically a reversible capacitive process in which the current is
proportional to the scan rate, :
d
(2.4)
dt
where jC is the capacitive current density and C is the capacitance per unit geometric
area.
Among the non-faradaic processes, two cases can be distinguished: (a) pure capacitive
processes, dominated by electrostatic charge accumulation at the interface and
associated exclusively with the charge/discharge of the double layer and (b) pseudo-
capacitive processes in which changes in the oxidation state of the material
simultaneously occur. Processes such as electrosorption or intercalation are pseudo-
capacitive although Equation (2.4) does not always apply. In semiconductors for water
splitting, the faradaic processes are usually related with the oxygen and hydrogen
evolution reactions and/or reduction or oxidation of the electrode, and the non-faradaic
ones are capacitive processes related to the charge/discharge of the interfacial double
layer.
In the context of semiconductor electrochemistry, dark voltammetry is mainly employed
to analyze charge accumulation in the so-called charge accumulation region (Chapter 1,
section 1.3.3). This measurement provides information about the approximate value of
the conduction band edge for an n-type and of the valence band edge for a p-type
material, and about the density of states within the band gap near the band edges. In
addition to the electrostatic charge/discharge of the double layer, electrosorption and
intercalation processes (i.e. pseudo-capacitive processes) also take place in the charge
accumulation region, these processes are particularly important for nanostructured
electrodes for which the real interfacial area is huge.
106
Experimental methods
An extensively used technique to evaluate photocurrent as a function of potential is the
LSV under transient illumination. In such measurements, LSV is performed at low scan
rates (2-5 mV s−1) and upon intermittent light transients (that last a few seconds). An
example of LSV with transient illumination for a photocathode is presented in Figure
2.9A in the next section. This representation provides insights about the location of the
onset of photocurrent, Eonset, defined as the potential from which photocurrent starts to
appear, and about the magnitude of the photocurrent as a function of potential. For very
slow scan rates, the photocurrents approach the photostationary values obtained at
constant potential (see next section, 2.2.2). Similar insights can be obtained from a
comparison of cyclic voltammograms in the dark with those obtained under
illumination. However, LSV under illumination allows for discriminating smaller
photocurrents, and also displays the shape of the photocurrent transients (non-stationary
photocurrent).
2.2.2. Chronoamperometry
In chronoamperometry the potential applied between WE and RE is set and the current
flowing between CE and WE is recorded versus time. Potential steps can be applied to
analyze the current response upon an instantaneous potential change.
A very useful technique in photoelectrochemistry is chronoamperometry under transient
illumination. First, a constant potential (positive to the onset for an n-type electrode or
negative to the onset for a p-type electrode) is applied while the current is recorded
under the dark for a few seconds and then light is turned on. The absolute value of the
current will increase upon illumination for photoactive electrodes. After a certain time,
light is interrupted. The difference between the dark current (or current density, j dark),
and the current under illumination (or current density, jlight) is called “photocurrent”, (or
photocurrent density, jph):
ph li ht dar (2.5)
Figure 2.9B displays the photocurrent transients measured by chronoamperometry at
two different potentials marked in the LSV shown in Figure 2.9A. Just when light is
turned on, the photocurrent is non-stationary. After a period of time, it will tend to
107
Chapter 2
stabilize reaching finally a stationary value. If a photocorrosion reaction is taking place,
the current will not stabilize. Then, in the absence of photocorrosion, photostationary
currents recorded by this method can be considered purely faradaic (dE/dt=0), and their
magnitudes can provide information about the kinetics of the photocatalytic reactions.
As mentioned in the above section, LSV under illumination at very slow scan rate can
be a good approximation to the steady state photocurrents for a wide range of potentials.
A. Eonset
jdark
-2
j/Am
jph
jlight
Eb E/V Ea
B. light
OFF
light light light
ON ON OFF
j / A m-2
j / A m-2
E = Ea E = Eb
time time
Figure 2.9. (A) Linear voltammogram under transient illumination for a photocathode
illustrating the photocurrent onset and the components of the photocurrent (jdark and jlight) (B)
photocurrent transients measured at a constant potential by chronoamperometry. The transient
of Ea shows surface recombination (presence of spikes upon illumination), which are absent at
higher potential bias (as exemplified for E = Eb).
The intensity of the spikes under illumination and light interruption (non-stationary
currents) provides information about the surface recombination processes. For instance,
108
Experimental methods
the initial spike is usually associated with a fast trapping of the minority carriers at
surface states (or reaction intermediates), and the subsequent recombination with
majority carriers, decreasing the current until a steady state is reached (as in the
transient at Ea, Figure 2.9B). When the photocurrent transient has a quasi-square shape
it is usually indicative of the absence of surface recombination (case of the transient at
Eb in Figure 2.9B). Evidence of surface recombination (i.e. photocurrent spikes) is
usually observed near the photocurrent onset, and then it decreases as higher potential
bias is applied, which is tantamount to a higher driving force.
Chronoamperometries together with LSVs under transient illumination and CVs in the
dark are routinely carried out in all the photoelectrochemical studies undertaken in this
thesis.
2.2.3. Chronopotentiometry and photopotential
In chronopotentiometry, the potential is recorded as a function of time under a constant
value of current between WE and CE. Open circuit potential (OCP) can be measured by
chronopotentiometry at zero current.
A common measurement in photoelectrochemistry is the OCP under transient
illumination, as shown in Figure 2.10. It is employed to evaluate the “photopotential,
Eph” defined as the difference between the open circuit potential in the dark (Edark) and
under illumination (Elight).
ph li ht dar (2.6)
109
Chapter 2
E(OCP) light
OFF
light
ON
Eph
time
Figure 2.10. Open circuit potential under transient illumination showing the magnitude of the
photopotential (Eph).
The OCP is linked to the Fermi level of the substrate. When the electrode is immersed
in the electrolyte, the Fermi level of the semiconductor is governed by the equilibrium
with the species present in the electrolyte as long as they form a redox couple (Figure
2.11A). Once the electrode is irradiated, the minority carrier density increases and quasi-
Fermi levels for electrons and holes are established. As stated in Chapter 1, section
1.3.4, the photopotential, Eph (Figure 2.11C), is the difference between the quasi-Fermi
level under illumination and the Fermi level in the dark. The magnitude of Eph and the
shape of the OCP decay upon light interruption, can provide information about the
location of the band edges, the lifetime of the charge carriers, the concentration of
accumulated charges and about their reactivity toward the electrolyte species.9
110
Experimental methods
B. e-
A. e-
C.
e-
EF,n = EF,p Eredox EF,s
E*F,n
EF,s
Eph EF,s
h+ E*F,p
h+ h+
FTO FTO FTO
Figure 2.11. A p-type semiconductor in contact with an FTO substrate in equilibrium with a
redox couple in the electrolyte (A), collection of holes in the back contact when it is illuminated
(B) and quasi-equilibrium under illumination (C). The photopotential (Eph) is given by the
difference between the quasi-Fermi level of the substrate upon illumination and the Fermi level
in the dark.
In this thesis, chronopotentiometry at constant current has been carried out only to
determine the faradaic efficiency for H2 evolution (Chapter 3). In such an experiment it
was imperative maintaining identical currents in two separate experiments, in which a
cathodic current was applied for a certain time, to either a Pt electrode in the dark or to
the photoelectrode under illumination. OCP experiments under illumination have been
carried out in the study of the electrochemically pretreated CuFe2O4 electrode.
2.2.4. (Photo)electrochemical impedance spectroscopy (PEIS)
Techniques in the frequency domain may provide information on the electrode
processes that cannot be obtained with the methods presented above. In this context,
Electrochemical Impedance Spectroscopy (EIS)55,56 is a powerful technique that allows
for obtaining valuable information not only about the structure of the interface and the
kinetics of the interfacial processes, but also about the charge carrier density and an
approach to the band edge locations. In this technique, a periodic ac potential
perturbation with an angular frequency w = 2f is applied to the electrode. It can be
expressed as:
111
Chapter 2
t sin t (2.7)
where E(t) is the potential perturbation, t is the time and E 0 is the potential
perturbation amplitude (see Figure 2.12). As a consequence, an ac current perturbation
is generated having the same frequency, w but different amplitude and a phase shift :
t sin t (2.8)
where I(t) is the resulting current perturbation and I0 the current perturbation
amplitude. In ac circuits, impedance is equivalent to resistance in dc circuits, both
governed by the Ohm law:
t t t (2.9)
t sin t sin t
t (2.10)
t sin t sin t
Then, Z(t) can be described as a complex number:
cos i sin i (2.11)
Where i = . Z' is the real part of the impedance and Z'' is the imaginary part.
112
Experimental methods
I /A
A. B. E(t)
I = f(E)
I(t) I0
I(t)
time
E0
E/ V
E (t)
Figure 2.12. (A) I vs E dc curve for a certain electrochemical system and a scheme showing a
superimposed sinusoidal potential signal E(t) with a small amplitude that produces a
sinusoidal current signal I(t) with the same frequency and with an amplitude I0. (B) sinusoidal
waves for E(t) and I(t) illustrating the phase shift of the current with respect to the voltage
perturbation.
The Frequency Response Analyzer (FRA) is the standard piece of equipment for
performing impedance measurements. The WE is subjected to a small sinusoidal
potential perturbation of a set frequency and constant amplitude of 5-15 mV that is
superimposed to a dc constant bias potential. Then, the resulting sinusoidal ac current is
measured and analyzed. When the measurements are taking place under illumination the
technique is called Photoelectrochemical Impedance Spectroscopy (PEIS).
In the following, the most employed EIS and PEIS methodologies in
photoelectrochemistry are detailed. One consists in the obtainment of the capacitance of
the electrode from impedance data, at a constant frequency and varying the potential.
The so-called Mott-Schottky equation is then used with the aim of obtaining
information about the charge carrier density and flat band potential (the conditions of its
applicability will be established below). This information is usually given under dark
conditions (EIS). The other is the Nyquist plot representation that is the most employed
to determine both the nature of the double layer phenomena and the kinetics of the
113
Chapter 2
photoelectrochemical reactions. For the latter, illumination of the electrode is usually
applied (PEIS).
2.2.4.1. Constant frequency (variable potential), Mott-Schottky representation
In Chapter 1, the Mott-Schottky equation was deduced for both n-type and p-type
photoelectrodes (see Equations (1.33) and (1.34)). The applicability of these equations
is restricted to the condition CSC << CH, and consequently >> . This
implies that the electrode capacitance coincides with the space charge layer capacitance.
Then, the applied potential drop fully occurs in the space charge layer, while the
potential drop in the Helmholtz layer would remain constant (BEP condition). Then,
sc fb , where Efb is the flat band potential and E is the applied potential and
Equation 1.33 converts into:
fb (2.12)
e e
The above form of the equation is extensively used in Electrochemistry for the
determination of flat band potentials and majority charge carrier densities. In the case of
n-type materials, the vs E plot gives a straight line with a positive slope, in which
the intercept with the x-axis roughly corresponds with Efb, and ND can be calculated
from the slope:
d (2.13)
e d
-
In this case, the vs E representation would give a straight line of negative slope.
The Mott-Schottky equation has been developed under some assumptions that must be
considered for its practical application, namely:57,58
(i) The resistance of substrate and electrolyte are very low and the measured
capacitance is not distorted by any leakage current across the interface.
(ii) Only one type of donor/acceptor impurities is present and they are both fully
ionized and homogeneously distributed within the space charge region.
114
Experimental methods
(iii) The measured capacitance does not include contributions from surface states,
adsorption or the Helmholtz layer.
(iv) The dielectric constant of the material () is frequency-independent.
(v) The interface is completely planar and two-dimensionally infinite.
Throughout this thesis, capacitance values (C) are expressed in units of farads per unit
geometric area. Equations (2.12) to (2.18) are valid when the electrodes present low
values of the roughness factor (close to 1). When the deviation from the flatness is
large, the physical (real) area must be considered instead of the geometric area unless
the characteristic dimension of the surface features is smaller than the space charge
region width, in which case the MS equation is not applicable.58,59 Likewise, the real
area can be expressed as a function of the roughness factor:
eom real r (2.14)
where Areal is the real surface area, Ageom is the geometric (projected) surface area and r
is the roughness factor. This treatment should be applied for nanostructured electrodes,
in which the real area differs from the geometric one. Considering an n-type
nanostructured photoelectrode, equations (2.12) and (2.13) can be expressed as a
function of the roughness factor as:
fb (2.15)
er e
d (2.16)
er
d
An expression analogous to Equation (2.15) can be written for p-type electrodes:
fb (2.17)
r e e
d (2.18)
er d
115
Chapter 2
A method for obtaining CSC consists in measuring the impedance by EIS. From the
complex impedance data, the capacitance of the electrode can be calculated. The
impedance of a pure capacitor, ZC, is:
(2.19)
i
This means that a pure capacitor is characterized by an imaginary impedance, being the
real part zero. Taking into account that Z’’ = −ZC, the capacitance can be calculated as a
function of the measured imaginary part of the impedance:
(2.20)
More often than not the capacitance of surface states contributes to the total capacitance
values. To minimize these effects, the EIS measurement is performed at high frequency
values, as the relaxation times of the surface states are usually long. The typical values
of frequency experiments are usually within the range of 0.1 to 10 kHz. It is important
to mention that Fabregat-Santiago et al. have proposed corrections of the Mott-Schottky
equation based on different covering of the substrate with the semiconductor and a
certain degree of band edge unpinning.60
In this thesis, Mott-Schottky analysis has been employed in nanostructured electrodes
by considering the roughness factor through Equations (2.15) and (2.16). On the other
hand, Mott-Schottky plots from equations (2.17) and (2.18) have been applied for
compact (low roughness) LaFeO3 electrodes in order to compare among pristine and
doped samples the acceptor density and flat band potential values.
2.2.4.2. Constant potential (variable frequency) Nyquist representation
A Nyquist plot consists in the representation of minus the imaginary part of the
impedance (Z’’) versus its real part (Z’) measured for discrete values of frequency. This
is the common representation derived from PEIS measurements and it is usually
characterized by a single semicircle or a sequence of them, as exemplified in Figure
2.13. Straight lines can also appear, for instance, related with diffusion controlled
processes.
116
Experimental methods
Z’’
1 kHz
10 kHz
0.1 kHz
Z’
Figure 2.13. An example of Nyquist plot representation displaying two semicircles. Each
experimental point corresponds to a different value of frequency.
In the literature, two main approaches have been adopted with the aim of interpreting
the EIS measurements and thus unveiling the mechanism of the photoelectrochemical
reactions. The most common approach is based on the modeling of the SEI by the so-
called equivalent circuits. They consist of an electrical circuit composed of different
elements such as capacitors, resistors and inductors in a particular arrangement that
represents the behavior of the SEI. Although they are advantageous because of their
simplicity, the drawback is the presence of ambiguities in the interpretation (or physical
meaning) of some of the elements of the circuits. Alternatively, a kinetic model can be
built from the elementary steps taking place at the illuminated SEI. This approach
provides specific values for the charge transfer and recombination constants together
with the electrochemical capacitance values at the SEI, but the analysis presents a
greater complexity. In both cases, a mathematical model is fitted to the experimental
EIS data to obtain the values of a number of parameters.
A large amount of work with PEIS has dealt with unveiling the mechanism involved in
water splitting under illumination in photoelectrodes such as TiO 2,61,62 hematite63,64 or
p-InP.65 On the other hand, it is important to mention that Intensity Modulated
Photocurrent Spectroscopy (IMPS) was also intensively employed with the same aim.
In this case, an ac perturbation of the light intensity is superimposed to the stationary
illumination of the interface. The periodic photocurrent response is measured as a
117
Chapter 2
function of the frequency of the perturbation. In Intensity Modulated photoVoltage
Spectroscopy (IMVS) the light intensity is modulated generating an ac voltage. IMPS
and IMVS are mainly used in Dye Sensitized Solar Cell research.
In this thesis, PEIS impedance and Nyquist plots are shown in Chapter 4. A kinetic
model is developed for the photoassited water reduction reaction on a photocathode.
The model is fitted to the experimental PEIS data to obtain valuable information about
the capacitances and kinetic constants (for charge transfer and recombination) at the
SEI. A qualitative analysis of the Nyquist plots is also performed in the case of doped
and undoped LaFeO3 electrodes in Chapter 5.
2.2.5. Electrochemical quartz-crystal microbalance (EQCM)
The use of the quartz-crystal microbalance (QCM)51,66,67 is based on the fact that the
resonance frequency of a quartz crystal is sensitive to mass changes. The QCM
measurement is based on the piezoelectric properties of the quartz crystal. When an
electric field is applied, the piezoelectric material deforms. If a sinusoidal electrical
signal is applied, the piezoelectric crystal oscillates at a resonant frequency.
Importantly, changes in mass of the crystal per unit area (m) result in changes in the
resonant frequency (f).
The basic idea behind the electrochemical quartz crystal microbalance (EQCM) is
coupling the QCM to an electrochemical measurement, allowing to record changes in
the mass in parallel to the electrochemical measurements. Commercial QCs for this
application have a metallic thin film onto which an oxide thin film can in turn be
deposited. Their typical resonant frequency is 5 MHz. A linear relationship there exists
between f and m, following the Sauerbrey equation:
f m (2.21)
Kf is the sensitivity factor of the crystal that combine several constants such as the
density and shear modulus of quartz and the resonant frequency of the quartz plate. It
has a value of 56.6 Hz cm2 g−1 for a 5 MHz crystal in air. Theoretically, the frequency
can be measured with an accuracy of 1 Hz, and therefore mass changes down to 20 ng
118
Experimental methods
cm−2 can be detected. EQCM is used to investigate processes involving mass changes,
such as the electrochemical deposition of materials or adsorption/desorption and
intercalation processes, etc.
The EQCM has been employed in Chapter 3 to analyze the adsorption/intercalation and
desorption/de-intercalation processes associated with the accumulation region of the
photoelectrode in the course of cyclic voltammograms.
2.3. Spectroscopic methods
Spectroscopy studies the interaction of the electromagnetic radiation with matter.
Spectroscopic methods are useful for evaluating, among others, the electronic structure
and composition of the sample. In the following, the spectroscopic techniques employed
in the framework of this thesis are briefly described.
2.3.1. UV-visible spectroscopy
This technique studies the interaction of the UV and visible ranges of the light spectrum
with matter.27,68,69 The equipment used for this aim is the spectrophotometer, whereby
the absorbed, transmitted or reflected light can be measured as a function of the
wavelength (i.e. the photon energy). One of the most important factors in
photoelectrochemistry is the light harvesting by the semiconductor, being the UV-
visible spectroscopy the most important tool to evaluate it. Absorption spectra provide
information about the electronic structure of the semiconductor. As stated in Chapter 1,
the band gap can be estimated through Equation (1.18), which requires determination of
the variation of the absorption coefficient with wavelength (obtainment of the
absorbance spectrum).
In a spectrophotometer, a collimated beam from a lamp passes through a
monochromator, and a slit allows selecting light of certain wavelengths. When a light
119
Chapter 2
beam of intensity, i, interacts with the solid, it can be split into different components,
as schematized in Figure 2.14. This can be expressed as:
i d (2.22)
Where IA is the light absorbed, I is the transmitted light, RSP is the specular reflected
light, Rd is the diffuse reflected light and S accounts for the refracted light redirected
off-axis. The latter component will be considered negligible in the subsequent
discussion. In the following, two different modes to estimate the variation of the
absorption coefficient with the wavelength will be described.
Transmittance measurements
The main elements needed for the transmittance mode are sketched in Figure 2.14. In a
general vein, the incident beam of a particular wavelength, I0, interacts with the sample
and the transmitted beam (I) reaches the detector and it is compared with a reference
baseline of 100% of transmission. The reference measurement must consider the light
absorption by the supporting materials, such as cuvette walls (in liquids) or the glass
substrate in the case of solids supported on glass.
IA
Rd S
White
light I0 I Detector
Rs
Light source Monochromator
Slit Sample
Figure 2.14. Scheme of the transmission configuration of a spectrophotometer.
This method allows for measuring the transmittance as a function of the wavelength,
which is defined as:
120
Experimental methods
(2.23)
i
Transmittance can also be expressed as a percentage (%T = 100T).
For estimating the absorption coefficient, it is assumed that light can only be either
transmitted or absorbed by the material (reflections and scattering are not considered).
Then Equation (2.22) simplifies to:
i (2.24)
The absorbed light is represented by the absorbance (A), which is calculated from
transmittance as:
lo lo (2.25)
i
The Lambert-Beer law, which is applicable under these conditions, relates absorbance
with absorption coefficient as:
i
lo (2.26)
where A is the absorbance, and L is the path of the light beam within the material.
For compact materials with high transparency, most of the incident light is either
transmitted or absorbed. In such cases, the transmittance configuration is most
appropriate for estimating both absorbance and material band gap. When FTO is being
used as a substrate, its contribution to the absorbance can be neglected above 350 nm. In
this thesis, the transmittance configuration has been employed to measure the absorption
spectra of doped and undoped LaFeO3 electrodes (Chapter 5).
Diffuse reflectance measurements
In the diffuse reflectance configuration, the diffusely reflected light is measured (R d).
This mode is indicated for estimating light absorption in thick and rough electrodes,
which typically are opaque and matt. In such a case, when the incident beam interacts
with the sample, light is reflected in two ways (see Equation (2.22)). One is the specular
121
Chapter 2
reflection (RSP) that occurs when the incident beam is reflected at an angle that is equal
to the incident angle. The second is the diffuse reflectance (Rd), in which the incident
beam is reflected consecutively at a rough surface, resulting in the emission of light in
all directions from the sample.
The measurement of the diffusely reflected light is achieved by the use of an integrating
sphere, which is illustrated in Figure 2.15. This component is internally covered with a
highly refractive material that allows for the total reflection of light. In such a way, the
diffusely reflected light is collected and directed to the detector, and it is compared with
a reference of 100% diffuse reflectance, which is generally a coating of BaSO4 powder.
In practice, specular reflections decrease the accuracy of the measurement. This is the
reason why integrating spheres incorporate a specular reflectance plug that reduces this
contribution.
Specular
I0 ref lectance
plug
Rs
Rd
To detector
IA
Transmitted
Figure 2.15. Scheme of the diffuse reflectance configuration, involving an integrating sphere
equipped with a specular reflectance plug. The components resulting from I0 are indicated in the
picture.
The Kubelka-Munk model70,71 is the most popular approach to estimate the variation of
the absorption coefficient with photon energy from diffuse reflectance values. In this
model, the incident light is supposed to be either scattered or absorbed. Moreover, only
diffuse reflectance is considered (non-specular reflections). The Kubelka-Munk
function, F(R) is defined as72:
122
Experimental methods
(2.27)
s
where is the diffuse reflectance of an infinitely thick layer, is the absorption
coefficient and s is the scattering coefficient. In practice, R is the measured diffuse
reflectance. For thick and highly rough electrodes, the light is mostly scattered and
absorbed, rather than being transmitted. In this regard, the Kubelka-Munk equation has
been widely employed to estimate the absorption coefficient and thus, to determine the
band gap of a large number of microstructured and nanostructured semiconductor
materials.
In this thesis, diffuse reflectance has been measured for nanostructured CuFe2O4,
YFeO3 and NiTiO3 electrodes by using an integrating sphere and BaSO4 as a reference
(Chapters 3, 6 and 7) with the aim of determining the optical band gap of these
materials.
2.3.2. X-ray Diffraction (XRD)
X-Ray Diffraction28,73,74 is a powerful technique to identify crystal structure, including
the lattice parameters (a, b, c, ) and the spacing between lattice planes (hkl Miller
indices). The technique is based on the X-ray diffraction phenomenon: when an incident
X-ray beam interacts with an ordered crystal, constructive and non-constructive
interferences between the scattered rays are generated, as the distance between the
scattering centers are of the same order of magnitude as the wavelength of the incident
radiation. This generates a diffraction pattern that is characteristic of every crystalline
compound. The Bragg law establishes the angles for constructive interferences:
n d sin (2.28)
where d is the interplanar space of the crystal, n is a positive integer number and is the
incident angle of the X-rays. A sketch of the diffraction of X-rays in a crystalline
arrangement is depicted in Figure 2.16. For any not satisfying the Bragg law, the
interference will be destructive. In other words, for a given X-ray wave (with a
wavelength, impinging with an angle , a constructive interference must be produced
123
Chapter 2
when there exists such a d-spacing that the wave can cross the distance 2d sen (in
Figure 2.16) n times.
Incident
X-rays
d
d sen
Figure 2.16. Scheme showing the X-rays scattered by two different atoms in a crystalline
arrangement spaced a distance “d”. The distance (d sen is depicted in the triangle in the right
side inset.
A typical XRD analysis consists in the measurement of the intensity of the diffracted X-
rays from a beam of a known wavelength as a function of the scattering angle
(diffraction angle). The instrument for performing the analysis is called diffractometer;
a Bruker D8-Advance diffractometer is shown in Figure 2.17A. The apparatus consist
basically of a sample holder, an X-ray source, and an X-ray detector, as schematized in
Figure 2.17B. The angle between the plane of the sample and the direction of the X-ray
beam () is maintained equal to that between the detector (the diffracted beam) and the
plane of the sample. Then, the diffracted beam is analyzed by scanning the diffraction
angle between two limit values. The resulting plot presents different peaks at certain 2
angle values, corresponding to constructive interferences of the X-rays with the sample,
and it is called “diffraction pattern”. Each peak represents a given set of planes (hkl
indice) of the crystalline sample, characterized by a particular d-spacing. The higher the
2 angle value, the lower the d-spacing. The reference diffraction patterns for the
different phase-pure crystalline materials are gathered in different databases, such as the
International Center of Diffraction Data (ICDD) or the Crystallography Open Database
(COD). For the identification of crystalline compounds, the experimental pattern is
compared with the reference patterns. This is usually performed by using a specific
software.
124
Experimental methods
Depending on the synthesis procedure the materials can exhibit preferred orientations,
especially when the oxide is synthesized by direct growth on the substrate. In those
cases, only peaks ascribed to the above mentioned orientations appears in the diffraction
pattern. The preferred orientations in the crystal will predetermine the physical
properties and photoelectrochemical activity of the electrode.
A. C. Inlet
water
Outlet
water
Ground
Anode Berilium window
(metallic
target)
X-Rays
Cathode
(tungsten
filament) Vacuum
chamber
B.
- Circuit to heat
the filament
Sample
holder
Figure 2.17. (A) A Bruker D8-Advance diffractometer photograph; (B) scheme of the basic
arrangement for the diffraction analysis, indicating the different components in the instrument
of (A); (C) scheme of the X-ray tube, indicating the different elements integrated into the
equipment.
The X-rays are produced in an X-ray tube, whose components are shown in Figure
2.17C. It consists of a high vacuum chamber, in which a high voltage (several kV) is
applied between two electrodes. At such a high voltage, the electrons are drawn out
from the cathode (which is a tungsten filament) at high speed, i.e. with high kinetic
energy, and bump into the anode (metal target). The collision slows down the electrons
and a fraction of the kinetic energy carried by them is converted into X-rays. These X-
rays are called continuous X-rays. Some of the electrons from the cathode have enough
125
Chapter 2
kinetic energy to eject an electron from an inner shell (for example, a K shell electron,
see below, Figure 2.24) and to leave a hole in this electron shell. Then, such a hole can
be filled by an outer shell electron, this process involving the production of X-rays with
a particular wavelength corresponding to an energy equal to the difference between the
electronic energy levels. These are called characteristic X-rays, and they are emitted
with high intensity. The production of X-rays in this way has a near 1% of efficiency
and the rest of the energy is converted into heat. A monochromator filters the radiation
to a specific transition line of the generated characteristic X-rays. The most widely
employed monochromatic X-rays are those corresponding to the Cu K line.
In this thesis, XRD has been an essential technique not only for identifying the
crystalline phases forming the semiconductor photoelectrodes, but also for estimating
the particle size of the material by the Scherrer equation, as described in Chapter 3.
Therefore, XRD analysis is present in all the following chapters of this thesis.
2.3.3. X-ray photoelectron spectroscopy (XPS)
X-ray photoelectron spectroscopy (XPS)51,73 was introduced by K. Siegbahn, work by
which he received the Nobel prize in 1981. This technique provides information not
only on the atomic composition, but also on the chemical environment and the oxidation
state of the elements in the sample at a surface level. It is based on the determination of
the kinetic energy of the emitted electrons when monochromatic X-rays interact with
the sample. This process is schematized in Figure 2.18. The E1, E2 and E3 lines
represent the electron energies corresponding to the inner shells of an atom, while the
upper lines represent the electron energy in the outer shells. When a photon from the
monochromatic beam impinges on the sample, it provides energy to extract the electron
from one of the levels:
h e (2.29)
where A is an atom or ion and is the excited state after electron extraction. In
Figure 2.18, Ek is the kinetic energy of the emitted electron. The binding energy is the
126
Experimental methods
energy required to remove an electron in the atom from a certain level. It can be
calculated as:
indin ner y h (2.30)
where is the work function of the spectrometer, which has a very small value that
needs to be considered for precise measurements. The produced photoelectrons are not
able to travel through more than 1-5 nm of solid; therefore XPS is a useful technique to
obtain information about the surface.73
Ek = h – E1
E6
E5
Binding energy
E4
h
E3
E2
E1
Figure 2.18. Scheme showing the emission of an electron because of the interaction of the atom
with an X-ray photon.
The XPS equipment is an electron spectrometer, which is schematized in Figure 2.19.
Basically, it consists of an X-ray source, a sample holder, an analyzer (function
analogous to that of a monochromator), a detector and a signal processor. The X-rays
produced in the X-ray source interact with the sample. Then, an electronic lens collect
the emitted electrons and the analyzer disperses the electrons according to their kinetic
energies. Finally, the count rate of electrons is determined by means of a detector as a
function of the kinetic energy. The most employed X-ray sources are X-ray tubes with
magnesium or aluminum blanks. The K lines of these elements have narrow
bandwidths. They provide monochromatic radiation of 1486.6 eV for the Al K and
1253.6 eV for Mg K.
127
Chapter 2
Figure 2.19. Schematic diagram of an XPS spectrometer consisting of an X-ray source, a
sample and electron analysis components.
The typical plot in XPS analysis is the count rate of electrons vs binding energy.
Binding energies appear as peaks in the XPS spectrum and they have well-defined
values that characterize the different atomic levels. They provide valuable information
as they do not only depend on the oxidation state of the atom, but also on the type of
bonds with neighboring atoms. Thus, it allows for determining the different elements
present within a few nanometers from the material surface, as well as the oxidation
states of these elements. The atomic percent of the elements present in the sample can
also be determined by this technique from the peak areas. The equipment can also
incorporate an Ar+ sputtering system that allows for removing part of the material and
thus performing analysis as a function of depth.
In this thesis, XPS analysis has been performed for CuFe2O4 electrodes in Chapter 3 to
analyze the changes in surface composition taking place upon immersion of the
electrode in a basic electrolyte, after an electrochemical treatment or upon illumination
of the electrode. In Chapter 5, doped and undoped LaFeO3 electrodes were examined
with XPS to determine dopant composition. In this case, Ar+ sputtering was used to
obtain a compositional depth profile. In addition, XPS has been studied in YFeO3
electrodes, allowing us to reach the conclusion of existence of Y-enrichment of the
surface.
128
Experimental methods
Auger electron spectroscopy (AES) is often incorporated in the XPS instrument. It is
also based on measuring the kinetic energy of the emitted electrons produced by the
interaction of either X-rays or an electron beam with the sample. The incident X-rays
(or electrons) cause the emission of a photoelectron, leaving a hole. This hole is filled
by an electron from a higher energy level. The energy of this transition is invested in the
emission of a second electron, which is the Auger electron. These electrons possess
energies between a few hundred eV to a few keV. Auger photoelectron peaks are
present in the XPS spectra, however, their kinetic energy is independent of the energy of
the radiation source.
In the context of this thesis, Auger spectroscopy has been employed for the analysis of
the oxidation states of Cu in CuFe2O4 electrodes (Chapter 3), as XPS analysis does not
allow for discriminating between the oxidations states zero and +1 of copper.
2.4. Microscopic methods
Microscopic methods allow for unveiling the morphology (and composition) of the
samples. They have a major importance in the context of nanostructured materials. They
are able to determine the morphology (shape and size of the nano- or microstructure).
They also provide crystallographic information (such as the arrangement of the atoms in
the material). Elemental composition can also be obtained simultaneously with the
adequate equipment.
129
Chapter 2
Incident
electron
EDX beam
X-Rays
Backscattered electrons
SEM
Visible light
Auger electrons
Secondary electrons SEM
sample
Absorbed electrons
Elastically scattered electrons Inelastically scattered electrons
Transmitted electrons
TEM
Figure 2.20. Different signals coming from the interaction of electrons with matter.
Before describing the basic microscopic techniques employed in this thesis, we will first
make an overview of the interactions of electrons (ionizing radiation) with matter.
When an electron beam impinges on a solid, different signals are emitted. Figure 2.20
depicts some of these processes. Photons in the visible spectrum can be emitted by a
phenomenon known as cathodoluminescence. Also, as explained in section 2.3.2, X-
rays can be generated by the deceleration of electrons (which is known as
Bremsstrahlung radiation) or by electron promotions between energetic levels
(characteristic X-rays). In a similar way, Auger electrons can also be emitted, process
which was described in section 2.3.3.
On the other hand, the electrons can be deflected from their path by electrostatic forces
within the atoms (with the nucleus or the electrons); this is called scattering. Depending
on the scattering angle, forward and back scattering can be distinguished. Back-
scattered electrons (and secondary electrons) are detected in SEM, while transmitted
electrons and forward scattered electrons are detected in TEM.
130
Experimental methods
2.4.1. Transmission electron microscopy (TEM)
Transmission electron microscopy (TEM) is used to visualize samples at micro and
nanoscopic levels with an excellent resolution.28,73,75,76 A TEM microscope operates
with the transmitted electrons depicted in Figure 2.20. An electron beam of uniform
intensity passes through the sample and the transmitted electrons are collected. Part of
the beam that straddles the sample has the same direction as the incident beam and
another part is scattered by the atoms in the sample. Scattering changes both the spatial
and angular distribution of electrons, yielding a non-uniformity that provides structural
and chemical information of the sample. Scattering is elastic when there is no energy
loss or inelastic when it involves energy loss. Different analyses can be performed
depending on the examined electrons: (i) Bright field images are obtained when
electrons from the direct beam are those contributing to the formation of the image. It is
used for typical morphological characterization. (ii) Dark field images are produced by
scattered electrons. This mode is often employed to study crystal defects. (iii)
Diffraction image (pattern) results from Bragg scattering. The electrons are diffracted
by the atomic planes constituting the sample, yielding a pattern resulting from
constructive and destructive signals.
A photograph of a JEM-2010 (JEOL) microscope is shown in Figure 2.21, and a scheme
showing the different components is displayed on the left hand side. The electron gun is
located in the upper part. It consists of a filament (that is made of W or LaB6) that acts
as a cathode, and an anode at earth potential. A small emission current is applied to the
filament to achieve the release of electrons (thermal emission). Additionally, a high
voltage is applied between the anode and the cathode, accelerating the electrons. The
Wehnelt cylinder serves as a convergent electrostatic lens that focuses and controls the
electron beam. When a small negative bias is applied to the Wehnelt, the electrons from
the cathode converge at a point called “crossover”, which is located between the
Wehnelt and the anode. The generated beam passes through a condenser lens that
focuses it on the sample, which transmits some of the electrons. Then, the objective lens
forms the image with these transmitted electrons. This image is magnified by the
projector lenses and projected on a fluorescent screen or in a digital device.
131
Chapter 2
The typical voltages in TEM are in the range 40-400 kV. The higher the voltages, the
higher the brightness, the better the resolution of the image and the lower the heating of
the sample.
Filament current
(Cathode) supply
Wehnelt
cylinder - High voltage
+ supply
(40-400 kV)
Condenser
Anode lens aperture
electrons Objective
aperture
Condenser lens Area selection
aperture
Sample
Objective lens
Projector lens
Fluorescent
screen
Figure 2.21. A JEM-2010 (JEOL) TEM microscope photograph and a scheme indicating the
different components integrated into the instrument.
The samples analyzed by TEM should have a thickness below 100 nm and they are
supported on a metal grid (usually made of copper). In this thesis, TEM images have
been obtained in order to determine the morphology and size distribution of the particles
in CuFe2O4 nanostructured electrodes in Chapter 3 and YFeO3 samples in Chapter 6.
The NiTiO3 nanorods have also been analyzed by TEM.
132
Experimental methods
2.4.2. Scanning electron microscopy (SEM)
The scanning electron microscopy technique51,73,75 operates with secondary and
backscattered electrons, which were defined in Figure 2.20. While secondary electron
detection allows for the inspection of the morphology and topography of the sample,
backscattered electrons are adequate for analyzing differences in composition. After
interacting with the atoms of the sample, these electrons emerge with energies between
zero and the energy of the incident electrons. Secondary electrons are those emitted with
energies below 50 eV and arise from the surface region of the sample, (in the range of
5-50 nm in depth).73 Backscattered electrons have energies above 50 eV and they
interact with deeper atoms in the sample. They are sensitive to the atomic weight of the
elements in the sample; the higher they are, the brighter the image appears.
Energy of the
Secondary Backscattered incident electrons
electrons electrons
Number of emitted e
50 eV
Energy of the emitted e
Figure 2.22. The number of emitted secondary or backscattered electrons as a function of their
energy (between zero and the energy of the incident electrons).
A photograph of a JSM-840 (JEOL) microscope is shown in Figure 2.23, including a
diagram depicting the different components. The mechanism of the electron gun is the
same as in TEM. The voltage supply is ranging between 2 and 50 kV. The magnetic
lens generates a rotational and symmetric magnetic field that focuses the electron beam.
Together with a condenser lens and an objective lens, they determine the final size of
133
Chapter 2
the beam on the sample. The scanning coils move the beam allowing for scanning the
sample in the XY plane. Finally, the microscope is equipped with detectors for both
backscattered and secondary electrons.
For the analysis, the sample is supported on a conductive substrate. Additionally, the
sample must be conductive. When the material presents low conductivity, a thin layer of
a metal is deposited by sputtering as to obtain a conformal thin coating. A typical metal
used for such a coating is Au.
Filament current
(Cathode) supply
Wehnelt
cylinder High voltage
-
+ supply
(2-50 kV)
Anode
electrons
Magnetic lens
Scanning coils
Backscattered
electron
detector Secondary
electron
detector
Figure 2.23. A JSM-840 (JEOL) SEM microscope photograph and a scheme indicating the
different components integrated into the equipment.
In this thesis, SEM characterization has been performed for the NiTiO3 electrodes, in
Chapter 7.
The field emission scanning electron microscope (FE-SEM) is a modern SEM
equipment in which images are achieved at a higher magnification than in traditional
SEM microscopes, allowing for a higher resolution. As an electron gun, it utilizes a
field emission source instead of a thermionic filament. The field emission gun provides
narrower probing beams at both low and high electron energies, which improves the
134
Experimental methods
spatial resolution at lower acceleration potentials. This also minimizes the charging
effects on the samples, avoiding their damage and improving the image quality (then,
there is no need for placing conducting coatings when the samples present low
conductivity). FE-SEM images are shown in Chapter 5 illustrating the LaFeO3 film
morphology and in Chapter 6 to compare YFeO3 photoelectrodes of different nature.
The energy dispersive X-ray technique (EDX or EDS)73,76,77 is usually incorporated into
the TEM and SEM equipment, and it is employed to determine the elemental
composition of the sample. It explores the characteristic X-rays emitted by the sample
as described in Figure 2.20 when a high-energy electron beam interacts with the sample.
Such interaction can lead to the extraction of an electron from an inner shell of the
atom. The minimum energy required to extract the atom from a given energy level is
called critical energy, Ec. The hole left after extraction is filled by an electron from an
outer shell, generating an X-ray photon of a characteristic energy that is equal to the
difference between the energetic levels involved in the transition. In Figure 2.24, the
nomenclature employed for the different transitions is displayed. In the case of the K
transition, the energy of the emitted X-rays will be EL−EK, being EL and EK the energies
of the L and K energy levels, respectively.
Energy
n=4
M
n=3
L L
n=2
K K K
n=1
Figure 2.24. Nomenclature employed for the different transitions leading to the characteristic
X-rays.
The acceleration voltage of the electrons must be selected higher than the Ec value for
the elements to be analyzed. Optimally, the beam energy must be from 1.5 to 3 times
higher than Ec. The higher the atomic number of an atom, the higher the E c value for a
given energy level. It is important to mention that wavelength dispersive X-ray
135
Chapter 2
spectroscopy (WDS) is a similar technique that analyzes the wavelength of the
generated X-rays instead of the energy of the photons.
X-ray microanalysis mapping is used to determine the spatial distribution of elements in
a sample. EDX or WDS techniques can be used to generate a two-dimensional
elemental map. The image is defined by rastering the electron beam through the surface,
resulting in a point-by-point compositional spectrum, which is evaluated to determine
the content of the chemical elements being analyzed. Thus, an elemental map is
constructed that can be delivered as individual element images (one map for each
element) or as multi-color images in which each color represents one element. It is
important to mention that the resolution of these maps is significant lower than that of
the typical SEM images.
In this thesis, X-ray microanalysis mapping has been acquired for YFeO3
photoelectrodes, reflecting the composition over the sample surface for both
nanostructured (undoped and doped) and compact films.
2.5. Bibliography
(1) Sivula, K.; Zboril, R.; Formal, F. Le; Robert, R.; Weidenkaff, A.; Tucek, J.; Frydrych, J.; Grätzel,
M. Photoelectrochemical Water Splitting with Mesoporous Hematite Prepared by a Solution-
Based Colloidal Approach. J. Am. Chem. Soc. 2010, 132, 7436–7444.
(2) Ling, Y.; Wang, G.; Wheeler, D. A.; Zhang, J. Z.; Li, Y. Sn-Doped Nanostructures for
photoelectrochemical water splitting. Nano Lett. 2011, 11, 2119–2125.
(3) Ida, S.; Yamada, K.; Matsunaga, T.; Hagiwara, H.; Matsumoto, Y.; Ishihara, T. Preparation of p-
type CaFe2O4 photocathodes for producing hydrogen from water. J. Am. Chem. Soc. 2010, 132,
17343–17345.
(4) Berni, A.; Mennig, M.; Schmidt, H. 2.2.8. Doctor Blade. In Sol-Gel Technologies for Glass
Producers and Users; Aegerter, M. A., Mennig, M., Eds.; Springer US, 2004; pp 89–92.
(5) Sakar, P.; Nicholson, P. S. Electrophoretic Deposition (EPD): Mechanism, Kinetics, and
Application to Ceramics. J. Am. Ceram. Soc. 1996, 79, 1987–2002.
(6) Boccaccini, A. R.; Zhitomirsky, I. Application of electrophoretic and electrolytic deposition
techniques in ceramics processing. Curr. Opin. Solid State Mater. Sci. 2002, 6, 251–260.
136
Experimental methods
(7) Corni, I.; Ryan, M. P.; Boccaccini, A. R. Electrophoretic deposition: From traditional ceramics to
nanotechnology. J. Eur. Ceram. Soc. 2008, 28, 1353–1367.
(8) Van der Biest, O. O.; Vandeperre, L. J. Electrophoretic Deposition of Materials. Annu. Rev.
Mater. Res. 1999, 29, 327–352.
(9) Berger, T.; Monllor-Satoca, D.; Jankulovska, M.; Lana-Villarreal, T.; Gómez, R. The
electrochemistry of nanostructured titanium dioxide electrodes. ChemPhysChem 2012, 13, 2824–
2875.
(10) Banks, C. E.; Foster, C. W.; Kadara, R. O. Chapter 2. Screen-Printing Electrochemical
Architectures. In Screen-Printing Electrochemical Architectures; Springer, 2016.
(11) Matsumoto, Y.; Omae, M.; Sugiyama, K.; Sato, E. New Photocathode Materlals for Hydrogen
Evolution: CaFe2O4 and Sr7Fe10O22. J. Phys. Chem. 1987, 91, 577–581.
(12) Matsumoto, Y.; Sugiyama, K.; Sato, E.-I. Improvement of CaFe2O4 photocathode by doping with
Na and Mg. J. Solid State Chem. 1988, 74, 117–125.
(13) Matsumoto, Y.; Sugiyama, K.; Sato, E. Photocathodic Hydrogen Evolution Reactions at p-Type
CaFe2O4 Electrodes with Fermi Level Pinning. J. Electrochem. Soc. 1988, 135, 98–104.
(14) Gu, J.; Yan, Y.; Krizan, J. W.; Gibson, Q. D.; Detweiler, Z. M.; Cava, R. J.; Bocarsly, A. B. p-
type CuRhO2 as a self-healing photoelectrode for water reduction under visible light. J. Am.
Chem. Soc. 2014, 136, 830–833.
(15) Livage, J.; Henry, M.; Sanchez, C. Sol-gel chemistry of transition metal oxides. Prog. Solid State
Chem. 1988, 18, 259–341.
(16) Hench, L. L.; West, J. K. The sol-gel process. Chem. Rev. 1990, 90, 33–72.
(17) Brinker, C. J.; Scherer, G. W. Sol-Gel Science. The Physics and Chemistry of Sol-Gel Processing;
Academic Press, Inc.: San Diego (USA), 1990.
(18) Niederberger, M.; Pinna, N. Chapter 2. Aqueous and Nonaqueous Sol-Gel. In Metal Oxide
Nanoparticles in Organic Solvents; Springer-Verlag: London (UK), 2009.
(19) Pierre, A. C. Introduction to Sol-Gel Processing; Springer Science+Business Media: New York
(USA), 1998.
(20) P. Jeevanandam. Chapter 7. Nanorods. In Nanoscale materials in chemistry; John Wiley & Sons,
Inc.: Hoboken (New Jersey), 2009.
(21) Li, J.; Wu, Q.; Wu, J. Chapter 12. Synthesis of Nanoparticles via Solvothermal and Hydrothermal
Methods. In Handbook of nanoparticles; Aliofkhazraei, M., Ed.; Springer International
Publishing: Switzerland, 2016.
(22) Green, D. C.; Glatzel, S.; Collins, A. M.; Patil, A. J.; Hall, S. R. A new general synthetic strategy
for phase-pure complex functional materials. Adv. Mater. 2012, 24, 5767–5772.
137
Chapter 2
(23) Arabatzis, I. M.; Antonaraki, S.; Stergiopoulos, T.; Hiskia, A.; Papaconstantinou, E.; Bernard, M.
C.; Falaras, P. Preparation, characterization and photocatalytic activity of nanocrystalline thin
film TiO2 catalysts towards 3,5-dichlorophenol degradation. J. Photochem. Photobiol. A Chem.
2002, 149, 237–245.
(24) Andronic, L.; Duta, A. Titanium dioxide thin film for photodegradation of methyl orange. Adv.
Mater. Res. 2007, 23, 325–328.
(25) Andronic, L.; Andrasi, D.; Enesca, A.; Visa, M.; Duta, A. The influence of titanium dioxide phase
composition on dyes photocatalysis. J. Sol-Gel Sci. Technol. 2011, 58, 201–208.
(26) Syrrokostas, G.; Giannouli, M.; Yianoulis, P. Effects of paste storage on the properties of
nanostructured thin films for the development of dye-sensitized solar cells. Renew. Energy 2009,
34, 1759–1764.
(27) Zoski, C. G. Handbook of Electrochemistry; Elsevier B.V.: Oxford (England), 2007.
(28) Jankulovska, M. Study of the electrochemical properties of nanostructured TiO 2 electrodes, PhD
thesis, Universitat d’ lacant, licante (Spain), 2015.
(29) Kang, D.; Kim, T. W.; Kubota, S. R.; Cardiel, A. C.; Cha, H. G.; Choi, K. S. Electrochemical
Synthesis of Photoelectrodes and Catalysts for Use in Solar Water Splitting. Chem. Rev. 2015,
115, 12839–12887.
(30) Budevski, E.; Staikov, G.; Lorenz, W. J. Electrocrystallization: Nucleation and growth
phenomena. Electrochim. Acta 2000, 45, 2559–2574.
(31) icelli, L. P.; ozzini, .; Mele, .; ’Urzo, L. review of nanostructural aspects of metal
electrodeposition. Int. J. Electrochem. Sci. 2008, 3, 356–408.
(32) Izaki, M.; Omi, T. Transparent zinc oxide films prepared by electrochemical reaction. Appl. Phys.
Lett. 1996, 68, 2439–2440.
(33) Golden, T. D.; Shumsky, M. G.; Zhou, Y.; Vanderwerf, R. A.; Leeuwen, R. A. Van; Switzer, J.
A. Electrochemical Deposition of Copper (I) Oxide Films. Chem. Mater. 1996, 8, 2499–2504.
(34) de Jongh, P. E.; Vanmaekelbergh, D.; Kelly, J. J. Cu 2O: Electrodeposition and Characterization.
Chem. Mater. 1999, 11, 3512–3517.
(35) Pauport , T.; Lincot, D. Heteroepitaxial electrodeposition of zinc oxide films on gallium nitride.
Appl. Phys. Lett. 1999, 75, 3817.
(36) Karuppuchamy, S.; Nonomura, K.; Yoshida, T.; Sugiura, T.; Minoura, H. Cathodic
electrodeposition of oxide semiconductor thin films and their application to dye-sensitized solar
cells. Solid State Ionics 2002, 151, 19–27.
(37) Paracchino, A.; Laporte, V.; Sivula, K.; Grätzel, M.; Thimsen, E. Highly active oxide
photocathode for photoelectrochemical water reduction. Nat. Mater. 2011, 10, 456–461.
138
Experimental methods
(38) Meulenkamp, E. A. Mechanism of WO3 electrodeposition from peroxy-tungstate solution. J.
Electrochem. Soc. 1997, 144, 1664 -1671.
(39) Mcdonald, K. J.; Choi, K.-S. Photodeposition of Co-Based Oxygen Evolution Catalysts on -
Fe2O3 Photoanodes. Chem. Mater 2011, 23, 1686–1693.
(40) Hahn, N. T.; Holmberg, V. C.; Korgel, B. A.; Mullins, C. B. Electrochemical Synthesis and
Characterization of p-CuBi2O4 Thin Film Photocathodes. J. Phys. Chem. C 2012, 116, 6459–
6466.
(41) Spray, R. L.; Choi, K. S. Photoactivity of transparent nanocrystalline Fe 2O3 electrodes prepared
via anodic electrodeposition. Chem. Mater. 2009, 21, 3701–3709.
(42) Lokhande, C. D.; Min, S. K.; Jung, K. D.; Joo, O. H. S. Cathodic electrodeposition of amorphous
titanium oxide films from an alkaline solution bath. J. Mater. Sci. 2004, 39, 6607–6610.
(43) Jeffrey Brinker, C. Chapter 10. Dip coating. In Chemical Solution Deposition of Functional Oxide
Thin Films; Schneller, T., Waser, R., Kosec, M., Payne, D., Eds.; Wien (Austria), 2013.
(44) Díaz-García, A. K.; Lana-Villarreal, T.; Gómez, R. Sol–gel copper chromium delafossite thin
films as stable oxide photocathodes for water splitting. J. Mater. Chem. A 2015, 3, 19683–19687.
(45) Prévot, M. S.; Guijarro, N.; Sivula, K. Enhancing the Performance of a Robust Sol-Gel-Processed
p-Type Delafossite CuFeO2 Photocathode for Solar Water Reduction. ChemSusChem 2015, 8,
1359–1367.
(46) Berglund, S. P.; Lee, H. C.; Núñez, P. D.; Bard, A. J.; Mullins, C. B. Screening of transition and
post-transition metals to incorporate into copper oxide and copper bismuth oxide for
photoelectrochemical hydrogen evolution. Phys. Chem. Chem. Phys. 2013, 15, 4554–4565.
(47) Park, H. S.; Lee, C.-Y.; Reisner, E. Photoelectrochemical reduction of aqueous protons with a
CuO|CuBi2O4 heterojunction under visible light irradiation. Phys. Chem. Chem. Phys. 2014, 16,
22462–22465.
(48) Berglund, S. P.; Abdi, F. F.; Bogdanoff, P.; Chemseddine, A.; Friedrich, D.; van de Krol, R.
Comprehensive Evaluation of CuBi2O4 as a Photocathode Material for Photoelectrochemical
Water Splitting. Chem. Mater. 2016, 28, 4231–4242.
(49) Iwase, A.; Ito, H.; Jia, Q.; Kudo, A. Solar-driven BiVO4 Photoanodes Prepared by a Facile Screen
Printing Method. Chem. Lett. 2016, 45, 152–154.
(50) Bazri, B.; Lin, Y.-C.; Lu, T.-H.; Chen, C.-J.; Kowsari, E.; Hu, S.-F.; Liu, R.-S. A heteroelectrode
structure for solar water splitting: integrated cobalt ditelluride across a TiO 2-passivated silicon
microwire array. Catal. Sci. Technol. 2017, 7, 1488–1496.
(51) Bard, A. J.; Faulkner, L. R. Electrochemical Methods. Fundamentals and Applications; John
Wiley & Sons, Inc.: New York (USA), 2001.
(52) Metrohm Autolab.B.V. Basic overview of the working principle of a potentiostat/galvanostat
(PGSTAT) – Electrochemical cell setup. Autolab Application Note EC08. pp 1–3.
139
Chapter 2
(53) Spectral irradiance (Oriel product training) http://assets.newport.com/webdocuments-
en/images/light_sources.pdf.
(54) Pletcher, D.; Greff, R.; Peat, R.; Peter, L. M.; Robinson, J. Instrumental Methods in
Electrochemistry; Ellis Horwood Limited: Chichester (England), 1985.
(55) Park, S.-M.; Yoo, J.-S. Electrochemical impedance spectroscopy for better electrochemical
measurements. Anal. Chem. 2003, 75, 455A–461A.
(56) Lasia, A. Electrochemical Impedance Spectroscopy and Its Applications; Springer
Science+Business Media: New York, 2014.
(57) Pleskov, Y. V.; Gurevich, Y. Y. Semiconductor Photoelectrochemistry; Consultants Bureau: New
York, 1986.
(58) Cardon, F.; Gomes, W. P. On the determination of the flat-band potential of a semiconductor in
contact with a metal or an electrolyte from the Mott-Schottky plot. J. Phys. D. Appl. Phys. 1978,
11, L63–L67.
(59) Cots, A.; Cibrev, D.; Bonete, P.; Gómez, R. Hematite Nanorod Electrodes Modified with
Molybdenum: Photoelectrochemical Studies. ChemElectroChem 2017, 4, 585–593.
(60) Fabregat-Santiago, F.; Garcia-Belmonte, G.; Bisquert, J.; Bogdanoff, P.; Zaban, A. Mott-Schottky
Analysis of Nanoporous Semiconductor Electrodes in Dielectric State Deposited on SnO 2(F)
Conducting Substrates. J. Electrochem. Soc. 2003, 150, E293–E298.
(61) Liu, H.; Li, X. Z.; Leng, Y. J.; Li, W. Z. An Alternative Approach to Ascertain the Rate-
Determining Steps of TiO2 Photoelectrocatalytic Reaction by Electrochemical Impedance
Spectroscopy. J. Phys. Chem. B 2003, 107, 8988–8996.
(62) Leng, W. H.; Zhang, Z.; Zhang, J. Q.; Cao, C. N. Investigation of the Kinetics of a TiO 2
Photoelectrocatalytic Reaction Involving Charge Transfer and Recombination through Surface
States by Electrochemical Impedance Spectroscopy. J. Phys. Chem. B 2005, 109, 15008–15023.
(63) 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.
(64) 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.
(65) Ponomarev, E. A.; Peter, L. M. A comparison of intensity modulated photocurrent spectroscopy
and photoelectrochemical impedance spectroscopy in a study of photoelectrochemical hydrogen
evolution at p-InP. J. Electroanal. Chem. 1995, 397, 45–52.
(66) Izutsu, K. Electrochemistry in Nonaqueous Solutions; WILEY-VCH Verlag GmbH & Co. KGaA:
Weinheim (Germany), 2002.
(67) Bagotsky, V. S. Fundamentals of Electrochemistry; John Wiley & Sons, Inc.: Hoboken (New
Jersey), 2006.
140
Experimental methods
(68) Memming, R. Semiconductor Electrochemistry; WILEY-VCH Verlag GmbH: Weinheim
(Germany), 2001.
(69) Chen, Z.; Dinh, H. N.; Miller, E. Photoelectrochemical Water Splitting. Standars, Experimental
Methods, and Protocols; Springer Science & Business Media, 2013.
(70) Kubelka, P. New Contributions to the Optics of Intensely Light-Scattering Materials. Part I. J.
Opt. Soc. Am. 1948, 38, 448–457.
(71) Murphy, A. B. Band-gap determination from diffuse reflectance measurements of semiconductor
films, and application to photoelectrochemical water-splitting. Sol. Energy Mater. Sol. Cells
2007, 91, 1326–1337.
(72) Nobbs, J. H. Kubelka- Munk Theory and the Prediction of Reflectance. Rev. Prog. Color. 1985,
15, 66–75.
(73) Skoog, D. A.; Holler, F. J.; Crouch, S. R. Principios de Análisis Instrumental; Cengage Learning:
Mexico, 2008.
(74) Waseda, Y.; Matsubara, E.; Shinoda, K. X-Ray Diffraction Crystallography. Introduction,
Examples and Solved Problems; Springer-Verlag: Berlin Heidelberg (Germany), 2011.
(75) Oura, K.; Lifshits, V. G.; Saranin, A. A.; Zotov, A. V.; Katayama, M. Surface Science. An
Introduction; Springer-Verlag: Berlin Heidelberg (Germany), 2003.
(76) David B. Williams; Carter, C. B. Transmission Electron Microscopy. Part 1: Basics; Springer
Science+Business Media, LLC, 2009.
(77) Introduction to Energy Dispersive X-ray Spectrometry (EDS). Cent. Facil. Adv. Microsc.
Microanal. (University California, Riverside); https://cfamm.ucr.edu/documents/eds-intro.pdf.
141
Chapter 3
CuFe2O4 photocathodes: improving
photoactivity by electrochemical
pretreatment
3.1. Introduction ........................................................................................................... 145
3.2. Experimental section ............................................................................................. 146
3.2.1. CuFe2O4 electrode preparation ..................................................................... 146
3.2.2. Electrode characterization ............................................................................ 146
3.3. Nanostructured CuFe2O4 films from commercial nanoparticles ...................... 148
3.3.1. Physical and optical characterization ........................................................... 148
3.3.2. Photoelectrochemical response of the as-prepared electrodes ................... 151
3.3.3. Electrochemical pretreatment ....................................................................... 153
3.3.4. Mechanism of the electrochemical pretreatment ........................................ 161
3.3.5. Photoelectrochemical characterization of pretreated p-type CuFe2O4
electrodes ....................................................................................................................... 166
3.4. Conclusions ............................................................................................................ 172
3.5. Bibliography ........................................................................................................... 173
CuFe2O4 photocathodes: improving its photoactivity by electrochemical pretreatment
3.1. Introduction
As it was discussed in the Introduction (section 1.5.1.), most spinel ferrites are good
candidates as PEC materials. Among them, CuFe2O4 is an attractive option because it is
a non-toxic, Earth abundant material. Previous works also reveal that it possesses a
relatively narrow band gap, which enables them to absorb visible light, and it exhibits p-
type behavior.1 Copper ferrite oxide has been reported not only as an anode in Li-ion
batteries2–4 and gas sensing,5–7 but also as a photocatalyst, particularly when forming
heterojunctions with CdS,8 SnO2,9 Bi4Ti3O12,10 C3N4,11 including H2 production in the
heterostructure CuFe2O4/TiO2.1 Recently, CuFe2O4/TiO2 catalysts were also found to
reduce CO2 into methanol using visible light.12 As mentioned in chapter 1 (section
1.6.3), coupling CuFe2O4 with WO313 and -Fe2O314 photoelectrodes led to an important
increase of the anodic photocurrents with respect to the bare photoanodic materials.
However, there is a lack of knowledge about the fundamental photoelectrochemical
behavior of this material as thin film electrodes. Only a fundamental study has been
described for the tetragonal phase of CuFe2O4 in the form of pellets.1
In this chapter, a photocathode made of CuFe2O4, a non-toxic, eco-friendly narrow band
gap spinel able to absorb visible light is described. This novel photocathode is
composed of nanoparticles deposited on a transparent conductive glass substrate.
Although the as-prepared electrodes are basically inactive, they exhibit p-type
photoactivity after a simple and fast electrochemical treatment based on a controlled
anodic polarization. This pretreatment leads to the development of cathodic
photocurrents upon illumination, and to the appearance of a charge accumulation region
at positive potentials near the oxygen evolution region. Electrochemical quartz crystal
microbalance measurements have been used to elucidate the nature of the processes that
lead to the enhancement of the p-type photoelectrochemical response. Even though this
work only deals with CuFe2O4, we believe that the electrochemical treatment described
herein could also be applied to other p-type materials.
145
Chapter 3
3.2. Experimental section
3.2.1. CuFe2O4 electrode preparation
Nanoporous films were fabricated by doctor blading commercial CuFe2O4 nanoparticles
(Sigma-Aldrich; 98.5 %). An aqueous slurry was prepared by grinding 100 mg
CuFe2O4, 10 L acetylacetone (Sigma-Aldrich; +99%), 10 L Triton 100X (Sigma-
Aldrich; laboratory grade) with 0.350 mL of ultrapure water. The slurry was spread over
F:SnO2 – coated glass (FTO, Pilkington TEC15), dried at ambient temperature, and
annealed at 450 ºC for 1 h in air to sinter the nanoparticles among them and with the
substrate. The thickness of the resulting films was measured with an Alpha Step D-100
profilometer.
3.2.2. Electrode characterization
A morphological examination of the films was carried out by transmission electron
microscopy (TEM) with a JEM-2010 (JEOL) microscope. Field emission scanning
electron microscopy (FE-SEM) images were obtained by a Merlin VP Compact (Zeiss)
microscope. The crystal structure was studied by X-ray diffraction (XRD) with a Seifert
JSO-Debyeflex 2002 diffractometer using the Cu Kα line (= 1.5406 Å). High
resolution XPS spectra were measured with a VG-Microtech Mutilab 3000
spectrometer. A Shimadzu UV-2401PC spectrophotometer equipped with an integrating
sphere coated with BaSO4 was used to measure UV-visible diffuse reflectance spectra.
(Photo)electrochemical measurements were performed at room temperature in a three-
electrode cell equipped with a fused silica window. All the potentials were measured
against and are referred to an Ag/AgCl/KCl(sat) reference electrode. A platinum wire
was used as a counter electrode and an N2-saturated 0.1 M NaOH (Panreac, p.a.)
solution as a working electrolyte. Measurements were carried out with a computer-
controlled Autolab PGSTAT30 potentiostat. A 1000 W Hg(Xe) arc lamp, equipped with
a water filter and a cutoff filter (cutoff λ=350 nm, Newport model FSR-KG3), was used
as a UV-visible light source unless otherwise stated. The applied light irradiance was
500 mW·cm2 measured with a power meter (Thorlabs PM100D) equipped with a
thermal power sensor (Thorlabs S310C). The electrodes were illuminated from the
146
CuFe2O4 photocathodes: improving its photoactivity by electrochemical pretreatment
substrate-electrode side (SE illumination). The chemical composition of the N2-purged
0.1 M NaOH working solution was measured by Inductively Coupled Plasma Mass
Spectrometry (ICP-MS).
Experiments for H2 detection were carried out in a closed and sealed glass cell with two
compartments separated by a NafionTM membrane NM-117. The working electrode
(photocathode) was an electrochemically pretreated FTO/CuFe2O4 electrode (2 cm2 of
geometric area). In the counter-electrode compartment, a platinum wire was placed.
Both compartments were filled with N2-purged 0.1 M NaOH, and the cell was closed by
sealing the ground glass joints with silicone. Samples of 300 L were extracted from the
headspace of the working electrode compartment and analyzed with a Hewlett Packard
5890 gas chromatograph equipped with a thermal conductivity detector (TCD). Prior to
the experiments, the headspace of the cell was purged using a vacuum pump and filled
with nitrogen gas several times. Before the (photo)electrolysis, an initial sample was
analyzed to check that the headspace was free of oxygen. Then, a constant cathodic
current of 25 A was applied during 2 h under illumination in the CuFe2O4
photoelectrode and at the end, the photogenerated H2 was measured. A reference
experiment was performed using as a working electrode a platinum sheet (of about 2
cm2 of geometric area) in the dark under the same conditions, assuming that the faradaic
efficiency is 100 % in this case.
A QCM200 quartz crystal microbalance (QCM) was used to record the quartz crystal +
electrode resonant frequency evolution during the electrochemical measurements. A
CuFe2O4 thin film of about 2 μm in thickness was deposited from a dilute slurry on an
AT-cut quartz crystal (Stanford Research Systems QCM crystal Ti/Au, 5 MHz) by
doctor blading. The film was sintered at 350 ºC in air for 1 h. The quartz crystal was
placed in a holder, connected to a QCM25 crystal oscillator and introduced in an
electrochemical glass cell.
147
Chapter 3
3.3. Nanostructured CuFe2O4 films from commercial nanoparticles
3.3.1. Physical and optical characterization
The dark brown FTO/CuFe2O4 electrodes were characterized by XRD and TEM after
the thermal annealing treatment. Figure 3.1 shows the X-ray diffraction pattern for the
commercial powder (A), the FTO/CuFe2O4 electrode (B) and the FTO substrate (C).
The commercial powder pattern exhibit well-defined peaks, indicating a high degree of
crystallinity for the CuFe2O4 nanoparticles. An analysis of the peak positions and
intensities confirms that the nanopowder is composed of cubic CuFe2O4 (PDF 01-077-
0010 card). In the case of the FTO/CuFe2O4 electrode (apart from the peaks
corresponding to the fluorine doped tin oxide) the pattern reveals the same phase as in
the commercial powder. It is worth noting that most of the previous studies on the
photocatalytic properties of this material were done with the tetragonal phase.1,9,11,15
A.
Intensity / a.u.
B.
C.
10 20 30 40 50 60
2 / degrees
Figure 3.1. X-ray diffraction patterns for (A) CuFe2O4 commercial nanoparticles (B)
FTO/CuFe2O4 sample after annealing at 450 ºC and (C) an FTO substrate. The reflections of the
cubic CuFe2O4 pattern (PDF 01-077-0010 card) are indicated at the bottom of the diffractogram
figures.
The TEM image of Figure 3.2A shows the morphology of the CuFe2O4 commercial
nanoparticles. Nanoparticles obtained after mechanically detaching them from an
electrode sintered at 450 ºC present a similar morphology (Figure 3.2B). No apparent
148
CuFe2O4 photocathodes: improving its photoactivity by electrochemical pretreatment
changes are observed in the nanoparticle morphology after heating at 450 ºC. The
nanoparticles have a quasi-spherical shape, with a relatively broad particle size
distribution, as displayed in the histogram in the inset of Figure 3.2B. The calculated
average size of the nanoparticles is 30 nm with a standard deviation of 20 nm.
A. B. C.
Frequency 20 40 60 80100
Particle size (nm)
20 nm 20 nm 20 nm
Figure 3.2. TEM image of (A) CuFe2O4 commercial nanoparticles, (B) CuFe2O4 nanoparticles
after detaching them from the FTO substrate of an FTO/CuFe2O4 electrode and (C) after
immersing the electrode for 30 min in 0.1 M NaOH. Inset: Particle size distribution determined
from TEM images (over 200 nanoparticles were counted).
Upon immersing the CuFe2O4 electrodes in 0.1 M NaOH for 30 minutes, a clear
modification of the nanoparticle morphology takes place. As can be seen in Figure 3.2C,
smaller nanoparticles are generated. The morphology change suggests that a dissolution
process of the material occurs in this medium. It is important to mention that the XRD
pattern remains unaltered after immersion and therefore a change in the bulk material is
discarded. In order to further investigate this issue, XPS prior and after immersion in
N2-purged 0.1 M NaOH for 30 minutes were measured. The Cu 2p, Fe 2p, and O 1s
regions in both cases are shown in Figure 3.3. One of the major changes occurring in the
sample after its exposure to this solution is the shift of the Cu 2p3/2 signal. The peak
initially located at 934.6 eV shifts to 933.2 eV after immersion. The initial location of
the Cu 2p3/2 agrees well with that reported for Cu(OH)2,16 while the location after
immersion is in agreement with values reported for CuFe2O4.16,17 Since Cu(OH)2 is
mildly amphoteric, this outer layer can be dissolved in the NaOH solution forming
[Cu(OH)4]2. These results points to the fact that Cu(OH)2 is present in the outer layer
of the commercial CuFe2O4 nanoparticles and upon immersion in alkaline media, the
149
Chapter 3
nanoparticle surface becomes free of Cu(OH)2. Longer periods of time in 0.1 M NaOH
do not trigger any additional morphological change.
Cu 2p Fe 2p
Counts / a.u.
Counts / a.u.
930 940 950 960 700 710 720 730 740
Binding energy / eV Binding energy / eV
O 1s
Counts / a.u.
Air
After immersion in 0.1 M NaOH
for 30 min
525 530 535 540 545
Binding energy / eV
Figure 3.3. High resolution XPS spectra for a CuFe2O4 electrode prior and after immersion in
0.1 M NaOH for 30 minutes.
The UV-visible diffuse reflectance spectra for a CuFe2O4 thin film is shown in Figure
3.4A. It exhibits a rather well-defined absorption edge at about 550 nm. To determine
more precisely the band gap, (Ah)2 (being A the absorbance and h the photon energy)
was plotted versus the photon energy (Figure 3.4B). A well-defined linear region can be
observed, according to the behavior expected for a direct transition. The intercept with
the x-axis provides the direct optical band gap, which has a value of 2.1 eV. The
determined band gap proves the ability of this material to absorb light in the visible
range. An indirect optical band gap of 1.8 eV can also be calculated (Figure 3.4C). The
band gap values obtained in this study are higher than those reported by Helaïli et al.18
for the cubic phase of CuFe2O4.
150
CuFe2O4 photocathodes: improving its photoactivity by electrochemical pretreatment
A. B. 12 C.
Reflectance / % 2
40
0.5
2
(hA) / eV
/ eV
8
1
0.5
20
(hA)
4
2.1 eV 1.8 eV
0 0 0
400 600 800 1.5 2.0 2.5 3.0 3.5 1.5 2.0 2.5 3.0 3.5
/ nm Photon energy, h / eV Photon energy, h / eV
Figure 3.4. (A) UV-visible diffuse reflectance spectrum and Tauc plots for the determination of
the (B) direct and (C) indirect optical band gap for a CuFe2O4 thin film (thickness: 2 μm).
3.3.2. Photoelectrochemical response of the as-prepared electrodes
Figure 3.5 shows the characteristic cyclic voltammogram in the dark for an
FTO/CuFe2O4 electrode. The main feature is the presence of a cathodic current
appearing at potentials more negative than 0.2 V, concretely at 0.32 V, together with
the corresponding oxidation process at more positive potentials (anodic peak located at
0.03 V, with a small shoulder at about 0.11 V). In previous works, a couple of peaks
at similar potentials have been described for copper oxides in alkaline media.19,20 The
cathodic process has been ascribed to the reduction of Cu(II) to Cu(I), and the anodic
processes to the back oxidation of Cu(I) to Cu(II). However, the participation of Cu(0)
in such redox processes cannot be completely excluded. Taking into account the
Pourbaix diagram for iron,21 the reduction of iron (III) in this potential range is unlikely,
although a minor reduction of Fe(III) to Fe(II) at potentials lower than 0.6 V cannot be
discarded, which would lead to the formation of iron (II) species, such as HFeO2. We
can thus assign the redox peaks observed in Figure 3.5 mainly to the reduction and
subsequent oxidation of Cu(II) ions present in the CuFe2O4 structure. It is worth noting
that the charge under the anodic peak is much smaller than that under the cathodic
wave. This suggests that the reduction product is soluble, diffusing to the solution bulk.
In fact, in strongly alkaline solutions the soluble species Cu(OH)2 has been described.22
151
Chapter 3
0.15
0.10
0.05
-2
j / mA cm
FTO
0.00
15
-0.05
10
-2
j / A cm
-0.10 5
0
-0.15
-0.2 0.0 0.2 0.4
E (vs Ag/AgCl) / V
-0.20
-0.6 -0.4 -0.2 0.0 0.2 0.4 0.6
E (vs Ag/AgCl) / V
Figure 3.5. Cyclic voltammograms in the dark for an FTO/CuFe2O4 electrode (thickness 14 μm)
and for an FTO substrate in N2-purged 0.1 M NaOH. Scan rate: 20 mV s1. Inset: Linear sweep
voltammogram (negative-going scan) under transient 500 mW cm2 illumination. Scan rate:
5 mV s1.
An additional relevant feature of the cyclic voltammogram is the faradaic oxidation
current that starts at about 0.5 V. This current can be ascribed to the oxygen evolution
reaction, as in our electrolyte the equilibrium potential for the O2/H2O redox couple is
located at 0.26 V. For the sake of comparison, the cyclic voltammogram for an FTO
electrode in N2-purged 0.1 M NaOH has also been included in Figure 3.5. With the
current density scale employed, the contribution of the FTO substrate can be considered
as negligible. This clearly suggests that O2 evolution takes place in the dark on the
CuFe2O4 nanoparticles.
The photoelectrochemical behavior of the native CuFe2O4 films was studied by linear
scan voltammetry under transient UV-visible illumination. In order to avoid Cu(II)
reduction, the measurements were performed in a restricted potential range (see inset of
Figure 3.5). Upon illumination, small anodic photocurrents were detected at potentials
more positive than 0.2 V. This behavior agrees well with that expected for an n-type
semiconductor with a very low photoactivity and contrasts with the expected p-type
character of CuFe2O4.1,8 The difference with previous studies could be related to the fact
that the present study has been done with the cubic instead of the most common
152
CuFe2O4 photocathodes: improving its photoactivity by electrochemical pretreatment
tetragonal phase. Due to the nanostructured nature of the electrodes and the presumably
low doping level, the photoresponse is dominated by the kinetics of the charge carrier
transfer at the interface. We can tentatively ascribe the observed behavior to an impeded
electron transfer to the electrolyte or to a hindered hole transport in the nanoporous
matrix, favoring a high recombination rate.
3.3.3. Electrochemical pretreatment
It has been reported for n-type TiO2 nanoporous electrodes that a reductive doping takes
place when a negative enough potential is applied. Such an electrochemical treatment
induces a photoactivity enhancement.23 On the other hand, an oxidative stripping
treatment has been shown to turn nanoparticulate zinc copper indium sulfide electrodes
from n-type into p-type.24 An analogous behavior can be expected for nanostructured
CuFe2O4 electrodes, i.e. an electrochemical treatment at positive enough potentials
could induce an oxidative doping, promoting p-type photoactivity. In this respect, for
the as-prepared CuFe2O4 nanoporous electrodes, the application of positive potentials
near 0.6 V provokes the development of cathodic photocurrents at 0.2 V. Additionally,
successive voltammetric cycles, involving a positive enough potential limit, also
produce the same effect. Therefore, it is possible to tune the photoelectrochemical
behavior of copper ferrite both potentiodynamically and potentiostatically. Both
procedures have been optimized to maximize the photoactivity of the resulting CuFe2O4
electrodes. The effect of the potentiostatic procedure is illustrated in Figure 3.6. To
evaluate more quantitatively the improvement in photoactivity, photocurrent transients
at 0.2 V were measured. This potential value was selected as a compromise to avoid
the region where the reduction of copper (II) occurs and to be negative enough with
respect to the photocurrent onset. The potentiostatic method consists in applying a
potential of 0.6 V to the electrode for different times, and then, measure the
photocurrent at 0.2 V (inset of Figure 3.6). The photocurrent increases with the
polarization time until a maximum is reached. Then, the photocurrent remains constant
with time, at least, up to the maximum polarizing time measured, that is 12500 s.
153
Chapter 3
15
10
-2
jph / A cm
0.01 LIGHT
LIGHT
ON OFF
0.00
-2
j A cm
-0.01
5
-0.02
-0.03
0 100 200 300 400
0 Time / s
0 2500 5000 7500 10000 12500
Time / s
Figure 3.6. Photocurrent as a function of the polarization time at 0.6 V for an FTO/CuFe2O4
electrode (13 μm in thickness), measured after 300 s of illumination (transients at 0.2 V vs
time for 0, 120, 360, 960, 2160, 3360, 5760 and 12960 s pretreatments are shown in the inset).
The potentiodynamic electrochemical pretreatment consists in applying consecutive
cycles between 0.2 V and a variable positive potential limit comprised between 0.5
and 0.7 V, using different scan rates. The enhancement of the p-type character after the
electrochemical pretreatment is observed independently of the scan rate. However, the
faster the cycling, the sooner the electrode develops and enhances a p-type behavior.
When the potentiostatic and the potentiodynamic procedures were compared (using a
potential of 0.6 V as the positive potential limit in the potentiodynamic method and as
the polarization potential in the potentiostatic procedure), it was noticeable that the
potentiostatic electrochemical treatment leads to smaller photocurrent values and, in
addition, the polarization time required to reach the same enhancement in the
photocurrent is significantly longer than that needed in the potentiodynamic method.
Therefore, subsequent work focused on the potentiodynamic procedure.
154
CuFe2O4 photocathodes: improving its photoactivity by electrochemical pretreatment
LIGHT LIGHT
ON OFF
A. 0
0
50
-20 150 400 1000
700
-2
j / A cm
-40 30
0.65 V
0.70 V
-2
jph / A cm
-60 20
0.60 V
10
-80 0.55 V
0 0.50 V
0 400 800
-100 Number of cycles
0 100 200 300 400
Time / s
B.
0.6 0 700
400
-2
j / A cm
-20
0.4
-2
j / mA cm
-40 150
-60 1000
0.2 0 1000 2000 3000 4000
50
Time / s
0
0.0
-0.2
-0.2 0.0 0.2 0.4 0.6
E (vs Ag/AgCl) / V
Figure 3.7. (A) Photocurrent transients at 0.2 V for an FTO/CuFe2O4 electrode (13 μm in
thickness) after 0, 50, 150, 400, 700, and 1000 cycles between 0.2 and 0.65 V at 500 mV s1.
Inset: Stationary photocurrent versus the number of voltammetric cycles for different positive
potential limits. (B) Cyclic voltammograms in the dark after 0, 50, 150, 400, 700, and 1000
cycles between 0.2 and 0.65 V at 500 mV s1 (scan rate: 20 mV s1). Inset: Current vs time plot
for an electrochemically pretreated FTO/CuFe2O4 electrode at 0.2 V under chopped light
coming from a 300 W Xe arc lamp with a >400 nm cut-off filter (500 mW cm2) (green arrows
= light on, black arrows = light off).
Figure 3.7A shows the photocurrent transients after a different number of voltammetric
cycles between 0.2 and 0.65 V (scan rate 500 mV s1) for a 13 m-thick electrode. A
dramatic increase in the photocurrent initially occurs, reaching an optimum after 400
cycles. Afterward, the value of the resulting photocurrent slowly decreases with the
155
Chapter 3
number of cycles. For other positive potential limits, the observed behavior is similar,
but the enhancement of the photocurrent is smaller (inset of Figure 3.7A).
The electrochemical pretreatment also triggers the development of capacitive currents at
sufficiently positive potentials. As shown in Figure 3.7B, the capacitive currents
increase with the number of cycles. These currents, starting at about 0.1 V, seem to
define the typical charge accumulation region of an n-type nanostructured electrode, but
inverted. By analogy, we can associate this region with the reversible accumulation of
holes either in the VB or more likely in surface states located just above it. The
emergence of the capacitive behavior with cycling is probably due to the fact that the
pretreatment induces p-type doping and thus an increase in the conductivity of the
sample. Prior to doping, the behavior of the electrode is similar to that of an insulator
and only when a sufficient degree of electrochemically induced doping is achieved, the
typical response of a semiconductor electrode is finally revealed. It is important to
mention that the development of a well-defined accumulation region is facilitated by the
nanoporous nature of the electrode.25 In order to discard the influence of a change in the
electrode morphology produced by the electrochemical pretreatment, TEM images were
obtained for the electrochemical pretreated electrode. As shown in Figure 3.8A,
nanoparticles exhibit identical shape as in the Figure 3.2C (before the electrochemical
pretreatment). To get more detail about the electrode morphology, Figure 3.8B shows
the FE-SEM micrographs of the FTO/CuFe2O4 electrode after the pretreatment. The
particles exhibit quasi-spherical shape, as expected, despite the modification
experienced after dipping in the 0.1 M NaOH electrolyte (discussed previously).
20 nm A. B.
100 nm
Figure 3.8. (A) TEM images for nanoparticles of an electrochemically pretreated FTO/CuFe2O4
after detaching them from the FTO substrate and (B) FE-SEM image of an electrochemically
pretreated electrode.
156
CuFe2O4 photocathodes: improving its photoactivity by electrochemical pretreatment
The question that arises is whether the changes in the electrochemical behavior
triggered by the pretreatment are permanent. To answer this question, an electrode was
electrochemically pretreated, and immersed in the electrolyte for 20 h. In the subsequent
photoelectrochemical characterization, the magnitude of the electrode capacitive
currents remained unaltered, while only a small diminution of the photocurrent was
recorded. In this respect, it is important to mention that, even if a decay in the
photocurrent is initially observed in the first transient, the subsequent transients are
stable (inset of Figure 3.7B). These results suggest that the electrochemical pretreatment
induces irreversible changes in the electrode.
Another critical issue to investigate is whether the photocurrents measured in N2-purged
0.1 M NaOH can be ascribed to the HER or to other electrochemical processes such as
CuFe2O4 photocorrosion (photoreduction) or the reduction of oxidized species present
in the pores of the nanostructure. In the photocorrosion case, a priori, one would expect
that Cu(II) is the species more susceptible to be reduced under illumination, given that
copper in the CuFe2O4 structure can be reduced electrochemically at more positive
potentials than iron, as previously discussed. Figure 3.9 shows the cyclic
voltammogram (red line) recorded immediately after a photocurrent transient at 0.2 V
for 550 s (Figure 3.9B). For comparison, the cyclic voltammogram after polarizing the
electrode at 0.2 V for 550 s in the dark is also shown (black lines). Both cyclic
voltammograms are virtually identical.
157
Chapter 3
(A) 1.0
A.
-2
j / mA·cm
0.5
0.0
-0.5
-0.2 0.0 0.2 0.4 0.6
E (vs Ag/AgCl) / V
B.
(B) 50
-2
j / A·cm 0.1V -0.2V
0 dark
light
-50
-100
0 200 400 600
Time / s
Figure 3.9. (A) Cyclic voltammogram at 20 mV s1 in the dark for an electrochemically
pretreated FTO/CuFe2O4 electrode recorded immediately after measuring the photocurrent
transient for 550 s at 0.2 V (B) (red line). As a control experiment, the cyclic voltammogram
(A) after performing the chronoamperometric measurement (B) without illuminating the
electrode is also shown (black lines).
Note that when the working potential window is opened down to potentials more
negative than 0.32 V, the reduction of Cu(II) mainly to Cu(I) occurs as for the non-
pretreated samples (Figure 3.5). The corresponding back oxidation is observed in the
positive-going scan at potentials more positive than 0.2 V. However, the charge
involved in the back oxidation is just a fraction of that flowing during the reduction as
pointed out above. Therefore, the fact that the cyclic voltammogram after polarization
under illumination in Figure 3.9 does not show an anodic peak (as in Figure 3.5),
indicates either that the possible products resulting from the photoreduction of the
electrode material do not remain at the electrode surface or that they cannot be oxidized
in the studied potential window. In this respect, it is important to mention the existence
of the Cu(OH)2 species at pH > 12.9 at 25 ºC.22 To determine the possible formation of
soluble products during photoelectrolysis, electrodes polarized at 0.2 V were
illuminated for 4 h and chemical analysis of the working solution was performed by
158
CuFe2O4 photocathodes: improving its photoactivity by electrochemical pretreatment
ICP-MS (Sample I in Table 3.1). For the sake of comparison, another FTO/CuFe2O4
electrode that was only immersed in the electrolyte, which was also analyzed (Sample D
in Table 3.1). The analytical results indicate that upon illumination there is an increase
in the amount of copper present in the electrolyte, which strongly suggests that a
substantial fraction of the charge measured in the chronoamperometric experiment is
employed in the reduction of Cu(II) to Cu(OH)2 (40% of faradaic efficiency). The
chemical analysis also demonstrates the presence of a minor amount of Fe in solution
that could result from the reduction of Fe(III) to HFeO2 (around 5% of faradaic
efficiency). Part of the photogenerated Fe(II) would remain on the surface (see below).
Table 3.1. Amount of Cu and Fe detected by ICP-MS in N2-purged 0.1 M NaOH working
solution resulting from the following experiments using FTO/CuFe2O4 electrodes. Sample I
corresponds to a solution in contact with an electrode electrochemically pretreated and then
illuminated for 4 h under polarization at 0.2 V (the overall time in contact with solution was 6
h). Sample D corresponds to an electrolyte in contact with the FTO/CuFe2O4 sample for 6 h.
Metal Amount detected / mol
Sample D Sample I
Cu 0.2806 0.6817
Fe 0.0320 0.0834
Finally, hydrogen gas detection for photoelectrochemical water reduction in the
CuFe2O4 photoelectrode was performed by gas chromatography. In Figure 3.10 the
potential evolution during 2 h at an applied cathodic current of 25 A is shown for a Pt
sheet (in the dark) and the FTO/CuFe2O4 electrode (under illumination). The
equilibrium potential for hydrogen evolution in this electrolyte is 0.97 V, which is
close to the value measured during the electrolysis for the Pt sheet. In the CuFe2O4
photoelectrode under illumination, a continuous decrease in potential is observed after
30 min, which is attributed to the H2 evolution process coupled with the photoreduction
of the CuFe2O4 material. In fact, compared with the amount of H2 detected in the
experiment with the Pt sheet, the faradaic efficiency for H2 evolution for the CuFe2O4
photoelectrode can be estimated to be about 15-20 %. Therefore, in addition to the H2
generation and photocorrosion, other reduction processes such as those of oxygen or the
159
Chapter 3
species resulting from the pretreatment and remaining in the nanopores of the electrode
should also take place.
0.0
FTO/CuFe2O4
E (vs Ag/AgCl) / V
-0.2
-0.4
-0.6
-0.8
Pt sheet
-1.0
0 40 80 120
Time / min
Figure 3.10. Electrode potential vs time curves during electrolysis (applied cathodic current of
25 A) for either an FTO/CuFe2O4 illuminated electrode (500 mW cm2) or a Pt sheet in the
dark, both in N2-purged 0.1 M NaOH electrolyte.
XPS measurements were performed before and after the electrochemical pretreatment,
and also after illumination for 2 h under reverse bias potential as shown in Figure 3.11.
XPS spectra are virtually identical before and after the electrochemical pretreatment. In
the Cu 2p region, the main peak is detected at 933.2 eV for all the samples, indicating
that Cu(II) is present at the electrode surface without contributions from Cu(I) or Cu(0).
In fact, the CuLM2 Auger spectrum clearly shows the absence of metallic copper,
which would be noticeable at 568 eV.26 On the other hand, an important feature can be
observed in the Fe 2p XPS spectrum. In the non-illuminated electrodes, the peaks for Fe
2p3/2 and Fe 2p1/2 are located at 710.8 eV and 724.3 eV, respectively, which is in
agreement with the existence of Fe(III) species.27,28 Nonetheless, after illumination
under reverse bias, the peaks shift toward lower energy values and for the Fe 2p3/2
contribution a binding energy of 710.3 eV is attained. As Fe(II) peak appears around
709.5 eV,27 the observed shift seems to be indicative that Fe(II) is generated on the
material surface, in agreement with the previous discussion.
160
CuFe2O4 photocathodes: improving its photoactivity by electrochemical pretreatment
Cu 2p Fe 2p
Counts / a.u.
Counts / a.u.
930 940 950 960 700 710 720 730 740 CuFe2O4 after inmersion in
Binding energy / eV Binding energy / eV 0.1 M NaOH after 20 min
CuFe2O4 after electrochemical
pretreatment
O 1s CuFe2O4 after illumination at
Counts / a.u.
CuLM2
Counts / a.u.
reverse bias potential for 2 h
525 530 535 540 545 560 565 570 575 580
Binding energy / eV Binding energy / eV
Figure 3.11. High resolution XPS spectra for FTO/CuFe2O4 electrodes with and without
electrochemical pretreatment, and after illuminating for 2 h a pretreated electrode with an
applied cathodic current of 25 A in N2-purged 0.1 M NaOH.
Typically, when an electrode suffers photocorrosion, the photocurrent decays all over
the experiment. However, in this case, the photocurrent is constant for 30 min and then,
it begins to decrease (or the potential decreases at constant current as in Figure 3.10). It
is worth noting that the electrochemical pretreatment in neutral media does not produce
an improvement of the photoresponse as in basic media. However, the electrode exhibits
a p-type behavior in neutral media once pretreated in basic media, but showing lower
photocurrents (see section 3.3.5.). This may be due to the suppression of
photocorrosion. In this way, the photocurrent would be due mainly to the
photogeneration of hydrogen.
3.3.4. Mechanism of the electrochemical pretreatment
To get some insights into the mechanism that triggers the p-type photoactivity
enhancement, QCM measurements were carried out. The change in the resonant
161
Chapter 3
frequency of the QC can be directly related with a change in mass through the
Sauerbrey’s equation, as mentioned in Chapter 2, section 2.2.5.
In this case, m is the theoretical change in mass of the film deposited on the quartz
crystal (Au/CuFe2O4). However, it is important to mention that the resonant frequency
can also be affected by other changes in the film and/or in the surrounding medium, not
directly related to a change in mass. For our particular sample, a nanoporous thin film
deposited on a gold-coated quartz crystal, the measured frequency shift during the
electrochemical measurements (Δf) corresponds to:
Δf = ΔfCuFe2 O4 + Δfpore + Δfsubs (3.1)
where ΔfCuFe2 O4 reflects the changes in mass in the CuFe2O4 material due to
electrochemically induced processes, such as electrosorption, insertion of electrolyte
species, etc; Δfpore represents the change in mass ascribed to porosity variations due to
possible alterations of the CuFe2O4 nanoparticle morphology and mainly to changes in
the electrolyte composition within the pores, and Δfsubs corresponds to the frequency
contribution from the substrate. To estimate the latter contribution, cyclic voltammetric
experiments were performed with a bare Au-coated quartz substrate. The typical
electrochemical response of gold, corresponding to the quasi-reversible
oxidation/reduction of the Au surface, was displayed in alkaline media in Figure 3.12.
Importantly, when the CuFe2O4 nanoparticles were deposited, this behavior was not
observed, being the electrochemical behavior analogous to that obtained for
FTO/CuFe2O4 electrodes. This indicates that the gold substrate is well covered by the
CuFe2O4 nanoparticles, preventing a significant direct contact with the electrolyte.
Therefore, the contribution of Δfsubs has been neglected in the discussion. On the other
hand, during the electrochemical pretreatment, the contribution of Δfpore is considered
to be minor as no changes in the nanoparticle morphology were found by TEM (not
shown). In the same way, no important changes in the composition of the electrolyte in
the pores are expected because of the absence of significant (faradaic) currents. Thus,
the changes in the resonant frequency (f) are mainly attributed to changes in the mass
associated with the electrochemically induced processes in the CuFe2O4 film
(ΔfCuFe2 O4 ).
162
CuFe2O4 photocathodes: improving its photoactivity by electrochemical pretreatment
0.01 A.
0.00
-2
j / mA·cm
-0.01
-0.02
0 B.
F / Hz
-5
-10
0.1 0.2 0.3 0.4 0.5 0.6 0.7
E (vs Ag/AgCl) / V
Figure 3.12. Cyclic voltammogram for a gold-covered quartz crystal in N2-purged 0.1 M NaOH
(A) and the corresponding resonance frequency evolution during the cyclic voltammetry
experiment (B). Scan rate: 5 mV s1.
Time-resolved electrochemical QCM measurements for a gold-covered quartz crystal
supporting a deposit of CuFe2O4 nanoparticles are shown in Figure 3.13. Concretely,
Figure 3.13A shows the resonant frequency evolution during the electrochemical
pretreatment. It exhibits two main features. The general trend reveals an irreversible
decrease in Δf with time, together with a cyclic quasi-reversible behavior directly
related to the cyclic voltammetric perturbation.
163
Chapter 3
B. 10
A.
-2
-20
0
j / A cm
-22 5
F / Hz
-24
-26
-10 700 750 800
Time / s 0
F / Hz
-20 C. 1.0
0.5
f / Hz
-30
0.0
0 200 400 600 800 1000
-0.5
Time / s 0.1 0.2 0.3 0.4 0.5 0.6
E (vs Ag/AgCl) / V
Figure 3.13. (A) Resonant frequency variation for a CuFe2O4 nanoporous film deposited on a
gold-coated crystal during successive voltammetric cycles between 0.1 and 0.6 V at 100 mV s1.
The inset shows a zoom of the curve. (B) Detail of a single voltammetric scan at 5 mV s1 and
(C) the corresponding resonant frequency shift. The experiments were performed in N 2-purged
0.1 M NaOH.
The clarification of the electrochemical pretreatment mechanism seems to be a
challenging task, as no apparent changes occur in nanoparticle morphology. In any case,
it is reasonable to consider that the effect of an electrochemical pretreatment based on a
certain number of voltammetric cycles should be a consequence of the electrochemical
processes occurring in each cycle. In Figure 3.13B and Figure 3.13C, a single cyclic
voltammogram measured at a slow scan rate, once the electrode was pretreated, together
with the corresponding frequency shift are plotted. The positive potential was limited to
0.55 V to avoid water oxidation. The capacitive currents are lower than in the case of
the FTO/CuFe2O4 electrode in Figure 3.7, probably because of the smaller thickness of
the electrode. Additional studies (not shown) revealed that a lower development of the
capacitive currents also occurs in the case of very thin films supported on FTO.
As shown in Figure 3.13B and C, the positive going scan leads to an increase in
frequency, while during the inverse scan there is an approximate recovery of the initial
frequency, leading to a quasi-reversible behavior. For a p-type semiconductor, the
164
CuFe2O4 photocathodes: improving its photoactivity by electrochemical pretreatment
accumulation of holes in the VB and/or in surface states/traps above the VB is expected
to occur at positive enough potentials. Chemically, this charge accumulation should
correspond to an oxidation process. In the case of CuFe2O4, we could tentatively assign
this process to the reaction:
CuII Fe2 O4 (OH )ads + OH + h+ ↔ CuII Fe2 O4 (O )ads + H2 O (3.2)
leading to a net mass loss (an increase in the resonant frequency) of one proton per
accumulated hole. Considering that the existence of Cu(III) has been widely described
in the literature,20,29,30 we can tentatively assume that CuFe2O4 may sustain the transfer
of an electron from the oxygen to the metal cation, leading to:
CuII + (O )ads ↔ CuIII + O2 (3.3)
During the back scan, the inverse processes should occur: electrons would be
transferred to CuFe2O4, while a proton would be adsorbed leading to a mass increase
(i.e. a decrease in the resonant frequency). It is important to mention that the H+
adsorption process allows for electroneutrality. This behavior is analogous to that
reported for some n-type semiconductors, such as TiO2, but in an inverse manner.25 As
the material becomes modified in a permanent way, we need to assume that reactions
(3.2) and (3.3) are slightly irreversible, which allows for the accumulation of a small
concentration of Cu(III) ions. This would also explain the increase in mass observed
during the electrochemical pretreatment, which could be ascribed to the conversion of
CuFe2O4 into CuFe2O4+x on the basis of the mechanism delineated above. The Cu(III)
ions would behave as electron acceptors, leading to a p-type character. In this context, it
is noteworthy that some authors have established that O species can be inserted into the
oxide network, doping the host material. This process has been studied for the
superconductor La2CuO4+δ.19,20 A limited amount of O temporarily bound to the oxide
surface can diffuse into the oxide network and it can react receiving an electron from
the metal cation. In fact, this procedure (electrochemical oxygen intercalation) is
applied in order to prepare superconducting materials. For instance, the superconductor
La2CuO4+δ has been prepared by means of this technique. A value of δ equal to 0.07 was
enough to transform the material from an insulator into a superconductor.31 From the
results of this section, we can tentatively ascribe the enhancement of the p-type
character mainly to CuFe2O4 doping, which would favor hole transport through the
165
Chapter 3
nanoporous structure. The fact that the effects of the cycling associated with the
electrochemical pretreatment saturate after a few hundreds of cycles suggest that at this
point doping would already extend to the whole film. The nanoporosity of the film and
the relatively small particle size are critical for attaining a homogeneous doping
throughout the film. In any case, we cannot completely discard the possibility of a
surface modification (for example, hydroxylation) that could enhance electron transfer
to species in solution, increasing the measured cathodic photocurrents.
3.3.5. Photoelectrochemical characterization of pretreated p-type CuFe2O4
electrodes
Figure 3.14 shows a linear scan voltammogram under transient illumination, together
with the Mott-Schottky plot for an electrochemically pretreated FTO/CuFe2O4
electrode. As observed, a cathodic photocurrent is developed already at potentials more
negative than 0.2 V. Near this value, the photocurrent increases very quickly while it
grows more slowly at more negative potentials. This behavior slightly differs from that
typical of a nanoporous electrode and suggests the existence of a space-charge layer.
The shape of the photocurrent transients clearly points to a high recombination rate, at
least at potentials close to the onset. In fact, remarkable spikes appear upon both
illumination and light interruption. This behavior has been attributed to the
accumulation of photogenerated electrons at surface states/traps (negative spike)
followed by recombination with holes when illumination is interrupted (positive spike).
The fact that electrons become trapped at the interface reveals kinetic limitations for the
electron transfer to water molecules. Interestingly, at more negative potentials, the
spikes tend to disappear, which indicates that electron surface trapping under
photostationary conditions diminishes as a result of a higher driving force for hole
transport toward the substrate. Admittedly, the photocurrents are very low, which
indicates that recombination is very fast, even at potentials substantially more negative
than the onset.
166
CuFe2O4 photocathodes: improving its photoactivity by electrochemical pretreatment
20
-2
jph / A cm
-20
4
2
C / F cm
-2
-40 1
-2
-10
0
10
-60 -0.2 0.0 0.2 0.4 0.6
E (vs Ag/AgCl) / V
-0.2 0.0 0.2 0.4
E (vs Ag/AgCl) / V
Figure 3.14. Linear scan voltammogram under transient illumination (green arrows = light on,
black arrows = light off) for an FTO/CuFe2O4 electrode (14 m in thickness) after the
electrochemical pretreatment in N2-purged 0.1 M NaOH. Scan rate: 5 mV/s. Inset: Mott-
Schottky plot for a 1 kHz frequency and ΔV=10 mV.
Mott-Schottky plots in the dark have been obtained from impedance measurements (see
Chapter 1, section 1.3.2). As shown in the inset of Figure 3.14, a linear region with a
negative slope is present, which is typical for p-type materials, although the
applicability of the model is doubtful for a nanostructured electrode. If the apparent flat
band potential were to be calculated from the Mott-Schottky plot, a value of 0.12 V
would be obtained, which is close to the photocurrent onset. This value could also be
affected by a deviation from the band edge pinning condition. In such cases, the
intercept with the x-axis shifts negatively from the real flat band potential,32,33 which
would be in agreement with the somewhat lower value of the intercept being slightly
less positive (0.12 V) than the photocurrent onset (0.2 V).
On the other hand, in Figure 3.15 an open circuit potential transient shows a shift
toward more positive values under illumination, and, when light is interrupted, a
relatively fast decay. The photopotential decay method has been commonly employed
for n-type nanoporous electrodes to evaluate the electron lifetime.25,34–36 By analogy, the
fast decay observed in this case can be ascribed to a relatively high recombination rate.
167
Chapter 3
OFF
0.05 ON
E (vs Ag/AgCl) / V
0.00
-0.05
-0.10
-0.15
0 100 200 300 400 500
Time / s
Figure 3.15. Photopotential transient for an electrochemically pretreated FTO/CuFe2O4
electrode in N2-purged 0.1 M NaOH.
In Figure 3.16, linear sweep voltammograms for the electrochemically pretreated
FTO/CuFe2O4 are shown either in N2 and O2 purged electrolyte. In the presence of an
efficient electron scavenger such as O2, the hole lifetime should increase.37 Figure 3.16
shows an increase of the cathodic photocurrent in the presence of O2 and a lower decay
of the photocurrent is observed upon illumination.
20
0
-2
-20
j / A cm
-40
N2 O2
-60
-80
-0.2 -0.1 0.0 0.1 0.2 0.3 0.4
E (vs Ag/AgCl) / V
Figure 3.16. Linear sweep voltammograms under transient illumination for an
electrochemically pretreated FTO/CuFe2O4 electrode (14 μm in thickness) at 5 mV s1 in the
presence (green line) or in the absence (red line) of O2. The current in the dark is also shown
(black line).
168
CuFe2O4 photocathodes: improving its photoactivity by electrochemical pretreatment
It is worthwhile emphasizing the very positive value of the photocurrent onset potential,
which is a critical parameter for evaluating the suitability of a photocathode response in
a water splitting device. The value is as high as 1.15 V vs RHE. As far as we know, this
is one of the highest values reported for an oxide absorbing in the visible. The measured
onset of the photocurrent should be close to the VB edge location (although there exists
a degree of uncertainty on how close it really is). Taking this into account, together with
the band gap energy of CuFe2O4, obtained from the UV-visible reflectance
measurements, the location of the CuFe2O4 CB edge can be tentatively estimated. The
corresponding energy band diagram is shown in Figure 3.17, together with the
potentials for the O2/H2O and H2O/H2 redox couples. As observed, CuFe2O4 possesses a
favorable energy band location for the photoreduction of water, with the CB lying more
negative than the potential for the reduction of water and the VB slightly negative to the
potential for water oxidation. Under illumination, part of the electrons photogenerated
in CuFe2O4 is transferred to water with the generation of hydrogen. In principle, the
photogenerated holes are not sufficiently oxidant as to photo-oxidize water, however, if
the bands are unpinned, as it was deduced previously, O2 evolution could take place
under illumination upon the application of a high enough potential.
Figure 3.17. Energy band diagram for a CuFe2O4 nanoporous electrode in contact with
0.1 M NaOH (pH 12.8). The redox levels for water reduction/oxidation are also
indicated in the figure.
169
Chapter 3
Cyclic voltammetry in the dark was also measured in neutral media (0.1 M Na2SO4
adjusted to pH 7 with NaOH). As the electrochemical pretreatment cannot be induced in
the FTO/CuFe2O4 electrode in neutral media (as stated above), the electrode was
previously pretreated in 0.1 M NaOH and the voltammetric behavior was compared in
subsequent measurements in neutral media. Both current-voltage responses are shown
in Figure 3.18 referred to the reversible hydrogen electrode. In both cases, an
accumulation region is present at positive potentials. The beginning of the capacitive
current shifts with the pH about 47 mV per pH unit, a lower value than the theoretical of
59 mV per pH unit. In addition, in basic media an irreversible current at 1.5 V vs RHE
attributed to water oxidation (as commented in section 3.2.2.) appears, while it is absent
up to 1.6 V vs RHE in neutral media. This suggests that the overpotential for O2
evolution in neutral media is larger than in basic media.
A linear sweep voltammogram in neutral media for the electrochemically pretreated
electrode in basic media is depicted in the inset of Figure 3.18. Generally, lower
photocurrent values have been observed in neutral media when the same electrode is
compared with basic media. Nonetheless, direct comparison with Figure 3.16 is not
conceivable, because different illumination sources are employed. The photocurrent
onset in neutral media is located at around 0.4-0.5 V (1.0-1.1 V vs RHE), indicating that
the onset also shifts with the pH, as expected.
170
CuFe2O4 photocathodes: improving its photoactivity by electrochemical pretreatment
pH 12.8
pH 7.0
0.4
-2
j / mA cm
0.2 2 A cm
-2
0.0 0.2 0.4 0.6
E (vs RHE) / V
0.0
-0.2
0.6 0.8 1.0 1.2 1.4 1.6
E (vs RHE) / V
Figure 3.18. Cyclic voltammogram in 0.1 M Na2SO4 adjusted to pH 7 with NaOH (black line)
and N2-purged 0.1 M NaOH (pH 12.8, red line) electrolytes in the dark for an electrochemically
pretreated FTO/CuFe2O4 electrode; at a scan rate of 20 mV s1. Inset: Linear sweep
voltammogram with transient illumination with a 300 W Xe lamp using a light intensity of
80 mW cm2 and a 1.5G filter in 0.1 M Na2SO4 at 5 mV s1.
The deposition of a cocatalyst on the electrode surface is a common strategy to enhance
photoactivity in semiconductor materials, as commented in Chapter 1 (section 1.6.4.).
Based on previous works on metal oxides, an attempt to enhance the photoresponse by
the loading of cocatalysts on CuFe2O4 has been made. (i) Ni/NiOx (by photodeposition38
and spin coating of a Ni2+ alcoholic solution39), (ii) MoSx40 (using a photo-assisted
electrochemical method) and known metal catalysts for HER such as (iii) Au (by
photoelectrodeposition) and (iv) Pt (by photodeposition) have been deposited over a
CuFe2O4 nanostructured electrode under different conditions. Neither of these
cocatalysts resulted in an increase of the photocurrent values. It is important to highlight
that intriguingly only a few reports have been published dealing with the deposition of
cocatalysts on ternary oxide materials for photoelectrochemical HER (see section
1.6.4.). Deposition of Pt over the ternary oxide was only reported for CuBi2O4
electrodes. Attaching directly Pt over CuFeO2 led to inconsistent results,41 and an
intermediate extracting layer between the absorbing material and the cocatalyst was
required to improve the photoresponse. The same strategy was employed previously to
favor the photostability and photoresponse of Cu2O electrodes.42 As the copper ferrite
171
Chapter 3
electrode suffers from photocorrosion coupled with HER, as demonstrated above, we
believe that the deposition of protective and/or extracting layers, similar to those
implemented for CuFeO2 and Cu2O electrodes would be an adequate strategy to
increase the photoelectrochemical hydrogen production on CuFe2O4 photoelectrodes.
It is worth noting that fueled by the growing interest in producing hydrocarbons and
other C1 and C2 compounds from CO2,43,44 CO2 photoelectrochemical reduction was
checked in the electrochemical pretreated FTO/CuFe2O4 electrode. The experiments
were performed in neutral media instead of basic media to avoid the presence of
carbonated species that can compete with CO2 reduction, and bubbling CO2 in a
solution previously purged with N2. The pH of the solution was adjusted in such a way
that similar values were measured before and after bubble CO2. The photocurrent values
and the shape of the illumination transients were almost identical. These results suggest
that CO2 reduction does not taking place on the FTO/CuFe2O4 electrode.
3.4. Conclusions
In summary, we have shown that copper ferrite (CuFe2O4) has the potential to be
implemented as a photocathode material in water splitting devices. Copper ferrite has
important advantages that derive from its low cost, its environmental benignity, and the
fact that it is composed of earth-abundant materials. Our results show that CuFe2O4
electrodes prepared by doctor blading a paste made of commercial powder on
conducting glass exhibit a well-defined p-type behavior in contact with 0.1 M NaOH or
0.1 M Na2SO4 solutions after being submitted to an electrochemical annealing
procedure in a basic electrolyte. Such an electrochemical pretreatment consists in
performing multiple voltammetric cycles up to a sufficiently positive potential (1.6 V vs
RHE), which leads to a change from the initial n-type behavior of the as-prepared
electrodes to a clear p-type behavior as revealed by the appearance of a voltammetric
accumulation region at positive potentials, and by the development of cathodic
photocurrents. We believe that this behavior is a result of the electrochemical doping of
the material, probably consisting in the generation of Cu(III) ions that can act as
172
CuFe2O4 photocathodes: improving its photoactivity by electrochemical pretreatment
acceptor states. Importantly, such a procedure could be extended to other oxides to
develop and enhance their p-type behavior.
The photoelectrochemical behavior of the pretreated copper ferrite electrodes is
characterized by some favorable features. On the one hand, the photocurrent onset is
located at a very positive value. In fact, as far as we know, this material exhibits one of
the highest onset potential values (1.1 V vs RHE) among all the oxides absorbing in the
visible region. This is critical as it ensures that it can be coupled with several oxide
photoanodes for carrying out water splitting in bias-free devices (tandem cells). For
instance, hematite could be a candidate as a photoanode. Even though the
semiconductor suffers photocorrosion, it shows more stability than other p-type oxide
materials (binary copper oxides). Admittedly, the photocurrent is still low, but we
believe that there are several strategies that can be applied in order to significantly
enhance CuFe2O4 photoactivity and photostability. Among them, we believe that the
implementation of thin films for electron extraction combined with an adequate
cocatalysts for hydrogen evolution is imperative. In addition, preparing thin films with
different microstructure/nanostructure (including compact films) or introducing into the
structure a third metal for chemical doping would improve the water photoreduction rate
over this material.
3.5. Bibliography
(1) Kezzim, A.; Nasrallah, N.; Abdi, A.; Trari, M. Visible light induced hydrogen on the novel
hetero-system CuFe2O4/TiO2. Energy Convers. Manag. 2011, 52, 2800–2806.
(2) Kalai Selvan, R.; Kalaiselvi, N.; Augustin, C. O.; Doh, C. H.; Sanjeeviraja, C. CuFe 2O4/SnO2
nanocomposites as anodes for Li-ion batteries. J. Power Sources 2006, 157, 522–527.
(3) Bomio, M.; Lavela, P.; Tirado, J. L. Electrochemical evaluation of CuFe 2O4 samples obtained by
sol–gel methods used as anodes in lithium batteries. J. Solid State Electrochem. 2008, 12, 729–
737.
(4) Luo, L.; Cui, R.; Qiao, H.; Chen, K.; Fei, Y.; Li, D.; Pang, Z.; Liu, K.; Wei, Q. High lithium
electroactivity of electrospun CuFe2O4 nanofibers as anode material for lithium-ion batteries.
Electrochim. Acta 2014, 144, 85–91.
173
Chapter 3
(5) Tao, S.; Gao, F.; Liu, X.; Sørensen, O. T. Preparation and gas-sensing properties of CuFe2O4 at
reduced temperature. Mater. Sci. Eng. B 2000, 77, 172–176.
(6) Sun, Z.; Liu, L.; Jia, D. zeng; Pan, W. Simple synthesis of CuFe 2O4 nanoparticles as gas-sensing
materials. Sensor. Actuat. B-Chem. 2007, 125, 144–148.
(7) Singh, A.; Singh, A.; Singh, S.; Tandon, P. Fabrication of copper ferrite porous hierarchical
nanostructures for an efficient liquefied petroleum gas sensor. Sensor. Actuat. B-Chem. 2017,
244, 806–814.
(8) Nasrallah, N.; Kebir, M.; Koudri, Z.; Trari, M. Photocatalytic reduction of Cr(VI) on the novel
hetero-system CuFe2O4/CdS. J. Hazard. Mater. 2011, 185, 1398–1404.
(9) Karunakaran, C.; SakthiRaadha, S.; Gomathisankar, P.; Vinayagamoorthy, P. Nanostructures and
optical, electrical, magnetic, and photocatalytic properties of hydrothermally and sonochemically
prepared CuFe2O4/SnO2. RSC Adv. 2013, 3, 16728–16738.
(10) Zhao, W.; Jin, Y.; Gao, C. H.; Gu, W.; Jin, Z. M.; Lei, Y. L.; Liao, L. S. A simple method for
fabricating p–n junction photocatalyst CuFe2O4/Bi4Ti3O12 and its photocatalytic activity. Mater.
Chem. Phys. 2014, 143, 952–962.
(11) Yao, Y.; Lu, F.; Zhu, Y.; Wei, F.; Liu, X.; Lian, C.; Wang, S. Magnetic core-shell
CuFe2O4@C3N4 hybrids for visible light photocatalysis of Orange II. J. Hazard. Mater. 2015,
297, 224–233.
(12) Uddin, M. R.; Khan, M. R.; Rahman, M. W.; Yousuf, A.; Cheng, C. K. Photocatalytic reduction
of CO2 into methanol over CuFe2O4/TiO2 under visible light irradiation. React. Kinet. Mech.
Catal. 2015, 116, 589–604.
(13) Liu, Y.; He, H.; Li, J.; Li, W.; Yang, Y. H.; Li, Y.; Chen, Q. Preparation and enhanced
photoelectrochemical performance of p-n heterojunction CuFe2O4/WO3 nanocomposite film. RSC
Adv. 2015, 5, 99378–99384.
(14) Hussain, S.; Hussain, S.; Waleed, A.; Tavakoli, M. M.; Wang, Z.; Yang, S.; Fan, Z.; Nadeem, M.
A. Fabrication of CuFe2O4/α-Fe2O3 Composite Thin Films on FTO Coated Glass and 3-D
Nanospike Structures for Efficient Photoelectrochemical Water Splitting. ACS Appl. Mater.
Interfaces 2016, 8, 35315–35322.
(15) Yang, H.; Yan, J.; Lu, Z.; Cheng, X.; Tang, Y. Photocatalytic activity evaluation of tetragonal
CuFe2O4 nanoparticles for the H2 evolution under visible light irradiation. J. Alloys Compd. 2009,
476, 715–719.
(16) Faungnawakij, K.; Shimoda, N.; Fukunaga, T.; Kikuchi, R.; Eguchi, K. Crystal structure and
surface species of CuFe2O4 spinel catalysts in steam reforming of dimethyl ether. Appl. Catal. B
Environ. 2009, 92, 341–350.
174
CuFe2O4 photocathodes: improving its photoactivity by electrochemical pretreatment
(17) Zhao, Y.; He, G.; Dai, W.; Chen, H. High Catalytic Activity in the Phenol Hydroxylation of
Magnetically Separable CuFe2O4 –Reduced Graphene Oxide. Ind. Eng. Chem. Res. 2014, 53,
12566–12574.
(18) Helaïli, N.; Bessekhouad, Y.; Bachari, K.; Trari, M. Synthesis and physical properties of the
CuFe2−xMnxO4 (0 ≤ x ≤ 2) solid solution. Mater. Chem. Phys. 2014, 148, 734–743.
(19) Sander, U.; Strehblow, H.-H.; Dohrmann, J. K. In Situ Photoacoustic Spectroscopy of Thin Oxide
Layers on Metal Electrodes. Copper in Alkaline Solution. J. Phys. Chem. 1981, 85, 447–450.
(20) Dong, S.; Xie, Y.; Cheng, G. Cyclic Voltammetric and Spectroelectrochemical Studies of Copper
in Alkaline Solution. Electrochim. Acta 1992, 37, 17–22.
(21) Pourbaix, M. Atlas of Electrochemical Equilibria in Aqueous Solutions, Second edi.; NACE
International: Houston (USA), 1974.
(22) Beverskog, B.; Puigdomenech, I. Revised Pourbaix Diagrams for Copper at 5-150 oC (SKI
Report 95:73); Swedish Nuclear Power Inspectorate: Stockholm (Sweden), 1995.
(23) Berger, T.; Lana-Villarreal, T.; Monllor-Satoca, D.; Gómez, R. Charge transfer reductive doping
of nanostructured TiO2 thin films as a way to improve their photoelectrocatalytic performance.
Electrochem. Commun. 2006, 8, 1713–1718.
(24) Guijarro, N.; Lana-Villarreal, T.; Gómez, R. Modulating the n- and p-type photoelectrochemical
behavior of zinc copper indium sulfide quantum dots by an electrochemical treatment. Chem.
Commun. 2012, 48, 7681–7683.
(25) Berger, T.; Monllor-Satoca, D.; Jankulovska, M.; Lana-Villarreal, T.; Gómez, R. The
electrochemistry of nanostructured titanium dioxide electrodes. ChemPhysChem 2012, 13, 2824–
2875.
(26) Park, J.-Y.; Jung, Y. S.; Cho, J.; Choi, W.-K. Chemical reaction of sputtered Cu film with PI
modified by low energy reactive atomic beam. Appl. Surf. Sci. 2006, 252, 5877–5891.
(27) McIntyre, N. S.; Zetaruk, D. G. X-ray photoelectron spectroscopic studies of iron oxides. Anal.
Chem. 1977, 49, 1521–1529.
(28) Singh, S.; Yadav, B. C.; Prakash, R.; Bajaj, B.; Lee, J. R. Synthesis of nanorods and mixed
shaped copper ferrite and their applications as liquefied petroleum gas sensor. Appl. Surf. Sci.
2011, 257, 10763–10770.
(29) Grenier, J.-C.; Pouchard, M.; Wattiaux, A. Electrochemical synthesis: Oxygen intercalation.
Curr. Opin. Solid State Mater. Sci. 1996, 1, 233–240.
(30) Magnone, E.; Cerisola, G.; Ferretti, M.; Barbucci, A. Electrochemical investigation of oxygen
intercalation into La2CuO4+δ phases. J. Solid State Chem. 1999, 144, 8–15.
175
Chapter 3
(31) Grenier, J.-C.; Wattiaux, A.; Lagueyte, N.; Park, J. C.; Marquestaut, E.; Etourneau, J.; Pouchard,
M. A new superconductor obtained by electrochemical oxidation of La 2CuO4. Physica C 1991,
173, 139–144.
(32) Fabregat-Santiago, F.; Garcia-Belmonte, G.; Bisquert, J.; Bogdanoff, P.; Zaban, A. Mott-Schottky
Analysis of Nanoporous Semiconductor Electrodes in Dielectric State Deposited on SnO2(F)
Conducting Substrates. J. Electrochem. Soc. 2003, 150, E293–E298.
(33) Wang, G.; Wang, Q.; Lu, W.; Li, J. Photoelectrochemical study on charge transfer properties of
TiO2-B nanowires with an application as humidity sensors. J Phys Chem B 2006, 110, 22029–
22034.
(34) Zaban, A.; Greenshtein, M.; Bisquert, J. Determination of the electron lifetime in nanocrystalline
dye solar cells by open-circuit voltage decay measurements. ChemPhysChem 2003, 4, 859–864.
(35) Monllor-Satoca, D.; Gómez, R. Electrochemical method for studying the kinetics of electron
recombination and transfer reactions in heterogeneous photocatalysis: The effect of fluorination
on TiO2 nanoporous layers. J. Phys. Chem. C 2008, 112, 139–147.
(36) Guijarro, N.; Lana-Villarreal, T.; Gómez, R. Electron lifetime in quantum-dot-sensitized
photoanodes by open-circuit-potential measurements. ChemPhysChem 2012, 13, 3589–3594.
(37) Lana-Villarreal, T.; Gómez, R. Tuning the photoelectrochemistry of nanoporous anatase
electrodes by modification with gold nanoparticles: Development of cathodic photocurrents.
Chem. Phys. Lett. 2005, 414, 489–494.
(38) Somasundaram, S.; Raman Nair Chenthamarakshan, C.; de Tacconi, N. R.; Rajeshwar, K.
Photocatalytic production of hydrogen from electrodeposited p-Cu2O film and sacrificial electron
donors. Int. J. Hydrogen Energy 2007, 32, 4661–4669.
(39) Lin, C.-Y.; Lai, Y.-H.; Mersch, D.; Reisner, E. Cu2O/NiOx nanocomposite as an inexpensive
photocathode in photoelectrochemical water splitting. Chem. Sci. 2012, 3, 3482–3487.
(40) Tran, P. D.; Pramana, S. S.; Kale, V. S.; Nguyen, M.; Chiam, S. Y.; Batabyal, S. K.; Wong, L. H.;
Barber, J.; Loo, J. Novel assembly of an MoS2 electrocatalyst onto a silicon nanowire array
electrode to construct a photocathode composed of elements abundant on the earth for hydrogen
generation. Chem. Eur. J. 2012, 18, 13994–13999.
(41) Prévot, M. S.; Guijarro, N.; Sivula, K. Enhancing the Performance of a Robust Sol-Gel-Processed
p-Type Delafossite CuFeO2 Photocathode for Solar Water Reduction. ChemSusChem 2015, 8,
1359–1367.
(42) Paracchino, A.; Laporte, V.; Sivula, K.; Grätzel, M.; Thimsen, E. Highly active oxide
photocathode for photoelectrochemical water reduction. Nat. Mater. 2011, 10, 456–461.
(43) Khraisheh, M.; Khazndar, A.; Al-Ghouti, M. A. Visible light-driven metal-oxide photocatalytic
CO2 conversion. Int. J. Energy Res. 2015, 39, 1142–1152.
176
CuFe2O4 photocathodes: improving its photoactivity by electrochemical pretreatment
(44) Kibria, M. G.; Mi, Z. Artificial photosynthesis using metal/nonmetal-nitride semiconductors:
current status, prospects, and challenges. J. Mater. Chem. A 2016, 4, 2801–2820.
177
Chapter 4
Investigating water splitting with CaFe2O4
photocathodes by electrochemical
impedance spectroscopy
4.1. Introduction ......................................................................................................181
4.2. Experimental section ........................................................................................182
4.2.1. CaFe2O4 photoelectrode preparation ..................................................... 182
4.2.2. Electrode characterization ...................................................................... 182
4.3. PEIS mathematical model for hydrogen reduction .......................................183
4.3.1. Mechanism and theoretical approach .................................................... 183
4.3.1. Simulation of the impedance function .................................................... 193
4.4. CaFe2O4 compact film electrodes ....................................................................198
4.4.1. Physical and chemical characterization ................................................. 199
4.4.2. Investigation on the mechanism of photoelectrochemical water
reduction .................................................................................................................. 200
4.5. Conclusions........................................................................................................211
4.6. Bibliography ......................................................................................................212
Investigating water splitting with CaFe2O4 photocathodes by electrochemical impedance spectroscopy
4.1. Introduction
As far as we know, CaFe2O4 is among the most studied ternary oxide photocathodes for
water splitting. Having a relatively narrow band gap (1.9 eV), it is able to absorb visible
light, and it also has suitable band edge positions for water reduction. In addition, it can
be prepared from cheap reagents. Several works dealing with the photoelectrochemistry
of this material for hydrogen evolution have been published, as noted in sections 1.6.1
and 1.7.3 of Chapter 1.1–6 The low resulting efficiency and the high temperature needed
for electrode fabrication are the main drawbacks for its practical use. Nowadays, finding
a cost-effective method for CaFe2O4 electrode preparation is still a challenge.
The knowledge of the different processes and reactions that take place in the
photoactive material is critical for the improvement of the photoelectrochemical
performance of the electrodes, which can approach us to the goal of a practical device.
Not only the semiconductor/electrolyte interface (SEI) plays a key role, but also the
characteristics of the material bulk, where recombination processes may take place.
Non-stationary photoelectrochemical impedance spectroscopy is a powerful in situ
technique for the understanding of the processes at illuminated SEIs. It can provide
useful information about the kinetics of the different steps and the nature of the
electrochemical capacitance. For instance, it could help identify the rate determining
step of the charge transfer reaction. The most common approach to analyze the
electrochemical impedance spectroscopy (EIS) results is based upon an equivalent
circuit of capacitors and resistances representing the SEI.7–12 This strategy leads to a
good fit of the experimental values, and provides information about resistive and
capacitive elements related with the phenomena occurring at the SEI, but it has the
disadvantage that some of the elements of the circuits have an unclear physical
meaning, especially those related with the kinetics and reaction mechanisms involving
the charge carriers.
In this chapter, not only the electrochemical capacitance in the SEI, but also the kinetics
of the hydrogen evolution photoelectrochemical reaction for a p-type CaFe2O4 electrode
is analyzed by means of EIS. A model is developed on the basis of the work of Leng et
al.,13 which focused on the kinetics of a photooxidation reaction on TiO2 electrodes, but
adapted to H2 evolution over p-type photoelectrodes. The importance of both defining
the characteristics of the interface employed to build the mathematical model, and
181
Chapter 4
establishing convenient approximations in view of extracting electrochemical
information from the experimental analysis are highlighted.
4.2. Experimental section
4.2.1. CaFe2O4 photoelectrode preparation
CaFe2O4 powder was prepared using a sol gel method described in the literature.4,14
First, 0.2 M of calcium (II) and iron (III) hydrated nitrates were dissolved in 25 mL of
water, and then, 5 mL of 5% aqueous polyethylene glycol (PEG 6000) solution was
added. The solvent was evaporated under stirring at 120 °C. The obtained gel was
calcined at 450 °C for 2 h and then at 1050 °C for 2 h, originating a brown CaFe 2O4
powder.
Electrode fabrication was performed using a platinum sheet as a substrate. 4 50 mg of the
powder was suspended in 100 L of ethanol and was kept in an ultrasonic bath for 15
min. The suspension was dropped over the Pt sheet and the solvent was evaporated at
room temperature. The deposited amount of CaFe 2O4 was about 25 mg. Finally, the
electrode was calcined at 1200 °C for 2 h and a dark brown film of CaFe 2O4 was
obtained.
4.2.2. Electrode characterization
X-ray diffraction characterization was performed on a Seifert JSO-Debyeflex 2002
diffractometer using the Cu Kα line ( = 1.5406 Å). A Shimadzu UV-2401PC
spectrophotometer equipped with an integrating sphere coated with BaSO 4 was used to
measure UV-visible diffuse reflectance spectra.
Photoelectrochemical measurements were performed using a three-electrode cell
equipped with a fused silica window. In order to avoid the platinum response in the
electrochemical analysis, the bare substrate and part of the deposit were isolated by
covering them with epoxy resin (PoxiPol ®) and then with Teflon tape over the resin.
182
Investigating water splitting with CaFe2O4 photocathodes by electrochemical impedance spectroscopy
The active working electrode area was of 0.2 cm2. A platinum wire was used as a
counter electrode and all the potentials are measured and referred to an
Ag/AgCl/KCl(sat) reference electrode. All the (photo)electrochemical measurements
were carried out by using a NOVA potenciostat coupled with a frequency response
analyzer. The impedance measurements were performed using a mV amplitude
perturbation in the range of frequencies from 10 kHz to 3 mHz. A solar simulator SUN
2000 (Abet Technologies) was employed as a light source, using ca. 1 sun of power
irradiance and incorporating neutral density filters to set the irradiance as required. The
actual irradiance ranged from 5 to 93 mW cm2.
A. B.
Teflon /
Pt Epoxi resin
substrate
Pt
Figure 4.1. Pt/CaFe2O4 electrode (A) before and (B) after isolating the bare substrate by
covering it with epoxy resin and then with Teflon tape.
4.3. PEIS mathematical model for hydrogen reduction
4.3.1. Mechanism and theoretical approach
The photoelectrochemical interpretation of EIS experiments using kinetic models
instead of equivalent circuits has been mainly developed for n-type
semiconductors.13,15,16 Generally, the resulting equations are also valid for p-type
electrodes with the appropriate current sign reversal. Ponomarev and Peter17
investigated recombination, charge transfer rate constants, and space charge capacitance
for a p-type InP electrode. However, to the best of our knowledge, the present work is
the first in investigating the water reduction process in a p-type metal oxide electrode on
183
Chapter 4
the basis of a kinetic model for PEIS. The general description of the proposed model
derives from the analysis performed by Leng et al.,13 considering not only the transfer of
electrons to the electrolyte from the conduction band, but also from surface states, as
well as the recombination of electron-hole pairs in the bulk and/or at surface states.
Figure 4.2. Scheme showing the different processes of charge transfer at the p-type/electrolyte
interface under illumination.
As shown in Figure 4.2, the mechanism considers surface states situated at a certain
energy value above the Fermi level of the semiconductor. The mechanism that governs
the water reduction process can be represented by the following kinetic reactions:
C h h eC Process 0
(4.1)
Process 1
eC ads
n (4.2)
Process 2
eC
184
Investigating water splitting with CaFe2O4 photocathodes by electrochemical impedance spectroscopy
n n (4.3)
Process 3
ads ads
n n (4.4)
h Process 4
s
n n (4.5)
Process 5
ads ads
(4.6)
The generated electron-hole pairs under illumination (given by process 0) can
recombine in the bulk or move to the surface. It is supposed that the flux of electrons
that reach the surface (I0) is directly proportional to the total incident photon flux (0),
being the proportionality factor, K<1. K depends on different factors such as the nature
and concentration of defects in the bulk of the semiconductor, the bending of the bands
or the diffusion length of the electrons in the bulk.
K (4.7)
Conduction band electrons can be transferred to solution (process 1) or to surface states
(process 2). It is assumed that (under stationary conditions) only a minor fraction of the
surface states are occupied by trapped electrons and the electrons trapped in surface
states can react with adsorbed water (process 3) to produce adsorbed hydrogen.
Photogenerated holes in the valence band can recombine with trapped electrons (process
4). Note that process (4.2) would correspond to the Volmer reaction step in the classical
hydrogen evolution reaction (HER) mechanism. It is assumed that the second step of the
HER is the Tafel reaction (process 5). However, the Heyrovsky electrodesorption step
could also occur. In fact, both Tafel and Heyrovsky steps have been reported for the
HER on metal oxide electrodes.18-20
185
Chapter 4
In the following, we hypothesize that process 3 is unlikely, and that the main pathway
for water reduction is electron transfer from the conduction band. Moreover, it is
assumed that the value of is high and constant. Process 5 is assumed to be fast and
thus process 1 is considered to be the rate determining step of the HER. The rate of
reactions 0-5 (vi) are expressed in cm2 s1 and the rate constants k1, k2, k3 and k4 are
given in s1.
The time-dependent concentrations of free minority carriers at the surface (n 1) and of
minority carriers trapped at surface states (n 2) are defined by kinetic differential
equations:
dn (4.8)
n n
dt
dn (4.9)
n n
dt
This approach has been used in previous works13,15,17 for either n-type or p-type
semiconductor electrodes.
Under steady state conditions we may write down dn 1/dt = 0, dn2/dt = 0, and Equations
(4.8) and (4.9) can be written as (4.10) and (4.11), respectively:
dn
n
dt
n (4.10)
dn
n n
dt
n
n (4.11)
where and are the steady state concentrations for free electrons at the surface
(cm2) and for electrons trapped at surface states, respectively.
186
Investigating water splitting with CaFe2O4 photocathodes by electrochemical impedance spectroscopy
The steady-state photocurrent density (jst) results from the transfer of electrons from the
conduction band to the electrolyte:
st
n e e (4.12)
On the other hand, three different capacitances in the SEI can be distinguished: the
Helmholtz double layer capacitance (CH), the space charge region capacitance (CSC) and
the capacitance of the surface states (CSS). The importance of surface states in the
photoresponse of semiconductor electrodes was extensively studied by Li and Peter;21
they determined that the presence of surface states induces a shift of the photocurrent
onset toward positive potentials in the case of an n-type semiconductor. In the present
work, the existence of surface states is considered to be described by a surface state
capacitance charged through a resistor (RSS), with a characteristic time constant for the
trapping of minority carriers that can be expressed as C .
The drop of potential across the SEI is a function of these capacitances, as described in
the classic theory of semiconductor electrochemistry:22,23
C
(4.13)
where jSS is a capacitive current density related with the charge and discharge of surface
states.
On the other hand, the total electron flux across the interface can be expressed as the
sum of a faradaic com onent (su erscri t “F”) and a ca aciti e component.
Furthermore, the net flux of electrons (current) in the semiconductor side ( electrode ) of
the interface must be equal to that in the electrolyte side ( electrol te
). Similar theoretical
treatments can be found in the literature.13,24,25
electrol te electrode (4.14)
187
Chapter 4
where:
F d
electrol te
C (4.15)
dt
F d C
d
electrode C
C C C (4.16)
dt dt
F
e n (4.17)
F
C
e n (4.18)
Some assumptions or approximations are considered in the following discussion. First,
the flux of electrons toward the surface (0) may be assumed as independent of the
applied potential as long as the penetration depth of the light is smaller than the width of
the depletion region13,15 or if the depletion region is not significantly affected by the
applied potential.
Traditionally, the electric double layer in the SEI has been modeled as a combination of
ideal capacitors. However, in real systems different non-idealities (porosity, surface
roughness, chemical inhomogeneities, etc.) at the surface can produce frequency
dispersion because of a non uniform distribution of the current density. The most
frequently employed approximation in EIS to represent the deviation from purely
capacitive behavior at both the semiconductor side and the electrolyte side of the SEI is
the introduction of a constant phase element (CPE).26–29 In this work, the electrical
Helmholtz double layer is modeled as a CPE30:
cos i sen (4.19)
i
instead of using the expression for a pure capacitance typical of an ideally polarizable
electrode ZH = 1/(CH i). In Equation (4.19), RH is a constant and 0 < < 1 (where =
1 corresponds to an ideal capacitor). As a result, the corresponding admittance can be
expressed as:
188
Investigating water splitting with CaFe2O4 photocathodes by electrochemical impedance spectroscopy
i (4.20)
where Q = cos(/2) and P = sen(/2).
The potential modulation is supposed to affect the band bending and, consequently, the
concentration of majority carriers at the surface ( ), which influences, in turn, the rate
constant k4 as defined by the following equation:
e C (4.21)
s
e
is the rate constant when C
= 0 (when the bands are flat) and is an ideality
factor.
The modulation of the potential across de Helmholtz layer will modify the activation
energy and thus affect the rate constant k1. Kühne et al.31 proposed the Tafel equation to
describe the relationship between the interfacial charge transfer constant and the
potential drop across the Helmholtz layer for p-InP. Analogously, here k1 can be
expressed as:
αc e
e (4.22)
Finally, we consider that the photoelectrochemical water reduction is not limited by
mass transfer, and that the rate determining step is a charge transfer step.
The electrochemical impedance spectroscopy technique consists in the application of an
ac perturbation to the potential dc signal producing, in turn, a perturbation in the
variables involved in the SEI. Frequently, a sinusoidal potential perturbation is applied
to the electrode with a frequency f = /2
e i t (4.23)
Where E0 is the amplitude of the perturbation and its angular frequency in rad s−1. It
is important to mention that corresponds to the potential perturbation E, defined in
Chapter 2, section 2.2.4.
189
Chapter 4
The ac component of the pertinent variables will be marked with a tilde ( etc ). The
admittance, Y(), is defined as the ratio of the modulated current density and the
modulated applied potential. Thus, the electrochemical impedance, Z(), the inverse of
the admittance, will be given by:
(4.24)
The sinusoidal perturbation of the rate constants can be obtained by derivation of
Equations (4.21) and (4.22):
αC e αC e αC e
e (4.25)
e e C e
e C C C (4.26)
e αC e
where A = and B =
Assuming that I0 is constant in the range of the potential perturbation, the time
dependent density of surface electrons, n , can be worked out from the time-dependent
perturbation form of Equation (4.8):
dn
i n n n n (4.27)
dt
n
(4.28)
i i
Analogous analysis can be applied to Equation (4.9):
dn
i n n n n (4.29)
dt
Substitution of Equations (4.26) and (4.28) into (4.29) yields:
190
Investigating water splitting with CaFe2O4 photocathodes by electrochemical impedance spectroscopy
n
n n n
i C (4.30)
n
i i
The derivation of Equation (4.13) results in:
C
(4.31)
Following the procedure of Leng et al.13 the time perturbation of the capacitive current,
, is related to the charge and discharge of surface states; i C is
substituted into (4.31), and an expression for is obtained:
(4.32)
i
On the other hand, the expressions for the time perturbation of the photocurrent
densities from Equations (4.15) and (4.16) are:
electrol te
n e ne i C
electrode
n e n e i C C C i C (4.33)
The introduction of the CPE element in the model, by means of Equations (4.19) and
(4.20), yields Equation (4.34) for electrol te
:
electrol te
n e ne i (4.34)
After the substitution of Equations (4.25), (4.26), (4.28), (4.30) and (4.32) into
Equations (4.33) and (4.34), and after dividing by the following expressions for the
admittance are obtained:
n e
ne i (4.35)
i
191
Chapter 4
n e n e
ne
i i i
(4.36)
i C
i C C
i
The combination of Equations (4.35) and (4.36) together with the substitution of n and
n according to Equations (4.10) and (4.11) results in a complex expression for the ratio
of the time-dependent variation of the Helmholtz layer drop and time-dependent
variation of the applied potential (take into account that ), :
C
C C
i i
(4.37)
i C
C C i
i i i i
The substitution of Equation (4.37) into either (4.35) or (4.36) yields the following
expression for Y():
i i Css
i i
i Csc
i
i ss (4.38)
i i i Css
i
i Csc
i ss
i
i i
Finally, the expression of Z() as function of the applied frequency is obtained as the
inverse of Y().
i i i Css
i Csc i
i i i i ss (4.39)
s
i i Css
i i Csc
i i i ss
where:
e (4.40)
192
Investigating water splitting with CaFe2O4 photocathodes by electrochemical impedance spectroscopy
αc e (4.41)
The electrolyte resistance, Rs, is also considered and introduced in the final equation of
the impedance (4.39).
4.3.1. Simulation of the impedance function
Theoretical simulations of the electrochemical impedance can be performed on the basis
of Equation (4.39). Taking into account the large number of parameters considered in
the model, many possibilities for illustrating the impedance response can be conceived.
Consequently, EIS plots may display a wide variety of shapes, depending on the relative
values of the different parameters. A complete analysis of the different cases is beyond
the scope of this work. The analysis will focus on the conditions under which the
general expression of Z() results in two semicircles, one at high frequencies (first
semicircle) and the other at low frequencies (second semicircle). At high frequencies,
the value of the rate constants can be neglected with respect to i in Equation (4.39).
The resulting expression for the first semicircle would be:
Css
i Csc i
i ss
first s
Css
(4.42)
i i Csc
i ss
An expression for the intercept of the semicircle with the x-axis of the Nyquist plot
(Zcross) can be obtained from the low frequency limit (0) of Equation (4.42):
(4.43)
cross s s s
e K e
The frequency at which the imaginary part of the impedance is maximum (max) can
also be determined by setting equal to zero the derivative with respect to of the
imaginary part of Equation (4.42) (assuming that 1 >> iSS). By neglecting the CPE
contribution, the following equation is obtained:
193
Chapter 4
K (4.44)
ma e e
C C C C C C C C C
where e( ) and αc e ( ). In the region of the second semicircle, the terms
related with the charge/discharge of both the surface states and the space charge layer
(capacitive contributions, except that related with the CPE) can be neglected:
i i i
i
i i i (4.45)
second s
i i
i
i i
The EIS experimental results for an FTO/CaFe 2O4 electrode will be analyzed in section
4.4.2. This section will illustrate changes in the theoretical electrochemical impedance
response as a result of variations of either E or I0.
-5
2
Z'' / k cm
-10 2.5 Hz
0
-2
1.5 Hz
eI0 / mA cm
-15 -2
0.89 Hz
0.01
0.02
-4 0.04
0.32 Hz
-20 0 2 4 6
0.08
0 5 10 15 20
2
Z' / k cm
Figure 4.3. Nyquist plots calculated according to the theoretical model (Equation (4.39) with
RH = 2·104 F cm2 s, = 0.5, CSC = 1.5·105 F cm2, CSS = 6·105 F cm2, SS = 104 s, k1 =
1.7·102 s1, k2 = 1.7·102 s1, k4 = 3·102 s1, RS = 65 C = 0.5, = 1 for eI0 values ranging
from 0.01 to 0.08 mA cm2. Inset shows a detail of the plots.
194
Investigating water splitting with CaFe2O4 photocathodes by electrochemical impedance spectroscopy
First, the effect of I0 on the EIS response is analyzed on the basis of Figure 4.3. Two
semicircles are consistently observed in the Nyquist representation for the values chosen
for the different parameters. As expected, the real and imaginary parts of the impedance
increase with decreasing I0. An analysis of the first semicircle on the basis of Equations
(4.43) and (4.44) reveals that, as long as the rate constants and K are not affected by the
light intensity, Zcross should be proportional to 1/0 and max should be proportional to 0.
0
A.
2
-2
Z'' / k cm
-0.5
-2
CSC / F cm
1.5 Hz 150
-4 -1.0 75
15
1.5 1.5
Hz 0.89 Hz 0.53 Hz
-1.5
1 2 3
0
B.
2
Z'' / k cm
-2
-0.5
-2
CSS / F cm
-4 -1.0
600
300
60
4.2 Hz 15
1.5 0.31 0.19 Hz
Hz Hz
-1.5
1 2 3
-6
0 2 4 6
2
Z' / k cm
Figure 4.4. Nyquist plots built according to the theoretical model (Equation (4.39) with eI0 =
0.04 mA cm2, RH = 2·104 F cm2 s, = 0.5, SS = 104 s, k1 = 1.7·102 s1, k2 = 1.7·102 s1,
k4 = 3·102 s1, RS = 65 C = 0.5, = 1 (A) effect of CSC ranging from 1.5·104 to 1.5·106 F
cm2 with CSS = 6·105 F cm2 and (B) effect of CSS ranging from 6·104 to 1.5·105 F cm2 with
CSC = 1.5·105 F cm-2. Inset shows a detail of the plots.
The influence of the magnitude of the depletion layer capacitance, C SC, and the surface
state capacitance, CSS, on the impedance spectra is shown in Figure 4.4. The semicircle
195
Chapter 4
at low frequencies is practically unaltered by changes of the capacitance values as it is
predicted by Equation (4.45). As the capacitance values change, the first semicircle is
deformed in such manner that Zcross remains constant, while max increases as either CSC
or CSS decreases. This is in agreement with model Equations (4.43) and (4.44). For the
lowest values of CSC in Figure 4.4A, the impedance response is virtually the same. This
fact is explained on the basis of Equation (4.39), in which CSC and CSS always appear
combined as CSC + CSS/(1+iSS). If the condition CSC << CSS/(1+iSS) is fulfilled, CSC
becomes negligible and the EIS response becomes independent of the depletion layer
capacitance. In this case, Band Edge Pinning (BEP) would be precluded and thus, the
capacitance of the surface states together with the Helmholtz layer capacitance would
govern the SEI.
In Figure 4.5, the effect of the recombination constant (k4) and charge transfer constant
(k1) values is illustrated. Figure 4.5A displays the effect of k4 on the impedance spectra,
the first semicircle is unaltered by this parameter, in agreement with Equation (4.42).
However, the effect in the second semicircle is obvious, producing an increase of
impedance with k4. Figure 4.5B illustrates that an increase in k 1 can either produce an
increase or a decrease in impedance depending on frequency. The effect of this variable
on both the first and second semicircles is significant. The value of Zcross decreases with
k1 according to Equation (4.43). Analysis of this equation unveils that Zcross increases
with k1 when k1>k2 and decreases with k1 when k2>k1. Importantly, the first condition
implies the transfer of electrons from the conduction band to the electrolyte rather than
their trapping at SS. The second condition represents the opposite situation. On the
other hand, max decreases when k1 increases. When k1<<k2, max is constant, according
to Equation (4.44). However, the effects on Zcross and max when k2>>k1 cannot be
appreciated because when k1 is sufficiently small (k1 = 3.4·103 s1 in Figure 4.5), the
two semicircles converge to a curve without a well-defined shape.
196
Investigating water splitting with CaFe2O4 photocathodes by electrochemical impedance spectroscopy
0
A.
1.5 Hz
2
-5
Z'' / k cm
-2 -1
-10 k4 / cm s
0.3
0.15
0.03
0.006
0
B.
2
-5
Z'' / k cm
0
1.5 Hz
-1 k1 / 10 cm s
-2 -2 -1
-10 -2
8.5
3.4
0.89 Hz 1.7
0.53 Hz
-3 0.34
0 1 2 3 4
-15
0 5 10 15
2
Z' / k cm
Figure 4.5. Nyquist plots calculated according to the theoretical model (Equation (4.39) with eI0
= 0.04 mA cm2, RH = 2·104 F cm-2 s, = 0.5, CSC = 1.5·105 F cm2, CSS = 6·105 F cm2,
SS = 104 s, k2 = 1.7·102 s1, RS = 65 C = 0.5, = 1 (A) effect of k1 ranging from 8.5·102
to 3.4·103 s1 with k4 = 3·102 s1 and (B) effect of k4 ranging from 0.3 to 6·103 s1 with k1 =
1.7·102 s1. Inset in panel B shows a detail of the plots.
The variation of k1 also affects the potential drop in the Helmholtz layer up to 10 Hz, as
shown in the vs frequency plot in Figure 4.6. adopts values close
to unity in the high frequency range. The frequency dispersion of the current in the
double layer (represented by the CPE element) produces an enhancement of the
potential drop over this region at high frequencies. At low frequencies, is
dominated by the concentration of free electrons at the surface and the concentration of
trapped electrons at surface states. The effect of k1 on is especially strong at
the lowest frequencies.
197
Chapter 4
1.0
0.8
0.6
∆ϕH ∆ϕT
-2 -1
k1 / 10 s
0.4 8.5
3.4
1.7
0.2 0.34
0.0 -3 -1 1 3
10 10 10 10
f / Hz
Figure 4.6. Effect of k1 on vs frequency plot with eI0=0.04 mA cm2, RH=2·104
F cm2s, =0.5, CSC=1.5·105 F cm2, CSS=6·105 F cm-2, SS=104 s, k2=1.7·102 s1,
k4=3·102 s1, C=0.5 and =1.
In brief, in this section, it has been shown that the semicircle at low frequencies is
mainly influenced by the kinetics of charge transfer and recombination processes, while
the semicircle at high frequencies is not only strongly linked to the capacitances in the
SEI, but also to the values of the kinetic constants k1 and k4.
4.4. CaFe2O4 compact film electrodes
In this section, the validity of the developed model with the approximations established
in section 4.1.1 will be checked for a CaFe2O4 photoelectrode. First, a brief discussion
on the physicochemical characterization of the material is given. Second, the
electrochemical responses in the dark and under illumination are analyzed. Then,
photoelectrochemical impedance spectroscopy is presented and discussed under
different conditions. Finally, the experimental EIS data will be fitted with the developed
model. In such a way, the model will be able to provide valuable information on the
parameters determining the behavior of the illuminated SEI.
198
Investigating water splitting with CaFe2O4 photocathodes by electrochemical impedance spectroscopy
4.4.1. Physical and chemical characterization
The characterization of the powder calcined at 1050 °C was performed by means of
XRD and UV/vis spectroscopy. The XRD pattern (Figure 4.7A) shows well defined
peaks assigned to orthorhombic CaFe2O4 (JCPDS card 00-032-0168). The UV/vis
absorption spectrum (Figure 4.7B) shows a defined absorption edge starting at 650 nm,
demonstrating that the material is able to absorb light in a rather wide range of the
visible spectrum. The Tauc plot for a direct optical transition (inset in Figure 4.7B)
reveals a band gap of 1.85 eV in agreement with the value given in a previous report. 14
A.
(320)
(201)
Intensity / a.u.
(040)
(170)
(331) (241)
(311)
(410)
(131)
(360)
(420)
(600)
(261)
(200)
(141)
(411)
(260)
(051)
(031)
(111)
(011)
(220)
(020)
20 30 40 50 60 70
2 / degrees
B. 0.5 1.5
2
0.4
(F(R)h)
1.0
(1-R) /2R
0.5
0.3
2
0.0
1.5 2.0 2.5 3.0 3.5
0.2 h / eV
0.1
0.0
400 600 800
/ nm
Figure 4.7. (A) X-ray diffraction pattern for the CaFe2O4 powder before being attached to a
platinum substrate (B) UV/vis absorption spectrum calculated by means of the Kubelka-Munk
function: F(R) = (1-R)2/(2R) from diffuse reflectance (R) measurements of a CaFe 2O4 film on
an FTO substrate and the corresponding Tauc plot for a direct optical transition (inset).
199
Chapter 4
4.4.2. Investigation on the mechanism of photoelectrochemical water reduction
The method described for the electrode synthesis requires high temperatures. At this
point, it is important to mention that a more versatile technique for electrode fabrication
was attempted. In this respect, FTO was used as a cheaper and transparent substrate
(more adequate for photoelectrochemical applications). The CaFe2O4 powder was
deposited on the FTO by the doctor blade method followed by a heat treatment at 640
ºC. Unfortunately, the photoresponse of the electrodes was extremely low, which can be
related with the observed low adherence of the resulting CaFe2O4 films. Then, in an
attempt to improve the contact among particles and with the substrate, three strategies
were employed: (i) ball milling of the primary CaFe2O4 powder in order to decrease the
particle size (ii) increase of the contact area with the FTO substrate by depositing ultra-
thin films of In2O3:SnO2 or SnO2:Sb nanoparticles (conductive materials) between the
CaFe2O4 and the FTO layers (iii) treatment with CaCl 2 or FeCl3 based on the work of
Kim et al.32 in which FTO/TaON samples were treated with TaCl5. None of these
strategies produced a substantial enhancement of the photocathodic currents. Hence, the
following results are referred to the fabricated Pt/CaFe2O4 electrode described in section
4.2.1.
The current-potential response in the dark for the Pt/CaFe2O4 electrode in N2-purged 0.1
M Na2SO4 is shown in Figure 4.8A. In the positive potential region, a reversible
capacitive process appears, as expected for a p-type semiconductor electrode
(accumulation region). At potentials positive to 0.5 V, irreversible faradaic currents
attributed to the reaction of water oxidation on the CaFe 2O4 surface are observed. The
pair of shoulders located at 0.3 V (oxidation peak) and 0.2 V (reduction peak) could be
assigned to the Pt-O formation and reduction, because similar peaks are present (with a
much higher intensity) in the voltammogram when the Pt substrate of the CaFe 2O4
electrode is not isolated from solution (as in Figure 4.1A). The residual presence of the
oxidation/reduction process of Pt might be due to the existence of pin holes or
microcracking in the compact CaFe 2O4 layer, because the entire bare platinum surface
surrounding the deposit was carefully isolated. The small charge under the Pt residual
signal indicates that the portion of platinum that would be exposed to the electrolyte is
rather small.
200
Investigating water splitting with CaFe2O4 photocathodes by electrochemical impedance spectroscopy
0.08
4
/ F cm
0.06 40
-2
-2
j / mA cm
20
-12
0.04
C · 10
0
-0.2 0.0 0.2 0.4 0.6 0.8
-2
0.02 E (vs Ag/AgCl) / V
0.00
-0.02 A.
0.02
-2
j / mA cm
0.00
-0.02
-0.04 B.
-0.2 0.0 0.2 0.4 0.6 0.8
E (vs Ag/AgCl) / V
Figure 4.8. (A) Cyclic voltammogram in the dark in N 2-purged 0.1 M Na2SO4 (pH = 6) at 20
1
mV s and Mott-Schottky representation calculated from the imaginary part of the impedance
data at 1 kHz (inset) (B) Linear sweep voltammogram with transient solar simulator
illumination of 93 mW cm 2.
The current-potential curve under chopped illumination in Figure 4.8B reveals cathodic
photocurrents in a broad range of potentials, as expected by considering the p-type
nature of the material. The onset of the photocurrent is located at about 0.60 V vs
Ag/AgCl (1.15 V vs RHE) at pH 6 (although more significant photocurrents are
observed only below 0.4 V). This value is more positive than that reported by Cao et al.5
for FTO/CaFe2O4 electrodes prepared by pulsed laser deposition. The fact that the
photocurrent onset is located within the accumulation region appearing in the high
potential range suggests that this region can be attributed to the filling/emptying of
surface states rather than to hole accumulation in the valence band.
The capacitance of the electrode was measured at different potentials and the Mott-
Schottky plot is presented in the inset of Figure 4.8A. The capacitance is almost
constant in the investigated potential window. When the electrode interface is mainly
governed by BEP, the Mott-Schottky plot exhibits a linear region33–36 (with a negative
slope in the case of a p-type semiconductor). As such linear region is not clearly defined
201
Chapter 4
in the present case, it is deduced that a modification of the applied potential is not
mainly reflected in a change in the potential drop through the space charge region. The
independence of the capacitance with the potential in CaFe 2O4 pellets was already
observed by Matsumoto et al.1 They suggested that FLP occurs in the SEI due to the
presence of surface states in the CaFe2O4 electrode.
The photocurrent response was also recorded at pH 13 as shown in Figure 4.9. Higher
photocurrents were found in alkaline media, being the observed photocurrent onset in
agreement with that given by Ida et al.4 In addition, the shift in the photocurrent onset
with pH follows a Nernstian behavior of 0.06 V per pH unit at least from pH 6 to pH 13.
E (vs Ag/AgCl) / V
-0.6-0.4 -0.2 0.0 0.2
0.05 pH = 13
-2
j / mA cm
0.00
-0.05
-0.10
0.4 0.6 0.8 1.0 1.2
-0.2 0.0 0.2 0.4 0.6
0.05 pH = 6
-2
j / mA cm
0.00
-0.05
-0.10
0.4 0.6 0.8 1.0 1.2
E vs RHE / V
Figure 4.9. Linear scan voltammograms with transient illumination in 0.1 M NaOH (pH=13)
and 0.1 M Na2SO4 (pH=6) N2-purged electrolytes under ca. 1 sun illumination (93 mW cm 2).
A determination of the band gap and the flat band potential enables to locate the
energetic position of the conduction and valence band edges. Unfortunately, the
available data does not allow the determination of the flat band potential accurately;
however, the onset of photocurrent (1.15 V vs RHE) is usually close to the flat band
202
Investigating water splitting with CaFe2O4 photocathodes by electrochemical impedance spectroscopy
potential value. Following this approach, the conduction band edge is roughly estimated
to be at 0.7 V vs RHE. This means a maximum overvoltage of 0.7 V for water
reduction. Electrons trapped in surface states have lower energies and consequently,
their transfer to the reactive H2O species would be more difficult or even impeded if the
required overpotential for hydrogen evolution is higher in absolute value than the
difference between the potential of the surface states and the redox potential for the
hydrogen couple in the electrolytic medium. In this respect, it is envisagable that the
transfer of electrons from the CaFe2O4 photoelectrode to water molecules is more likely
occurring from the conduction band than from surface states.
The experimental EIS data shown in Figure 4.10 for the CaFe2O4 electrode was
obtained by setting the light intensity and recording the impedance spectra at different
potentials from 0.4 V to 0.3 V. The photoelectrochemical response in the dark and
after transient illumination was recorded before and after the EIS experiments at every
light intensity. Cyclic voltammograms in the dark remained unchanged after the EIS
experiments, although the magnitude of the photocurrent was somewhat altered,
especially at the highest light intensities (an 8% decrease in the photocurrent was
recorded at 0.3 V using 93 mW cm2 of illumination). This behavior may be attributed
to an incipient irreversible reduction of the material during long illumination times
under bias potential.
203
Chapter 4
0 0 0
2 2 2
-10 100 mW/cm -10 50 mW/cm -10 20 mW/cm
2
Z'' / k cm
-20 -20 -20
-30 -30 -30
-40 8.4 mHz -40 -40
5.0 mHz 5.0 mHz
-50 -50 -50
0 10 20 30 40 50 0 10 20 30 40 50 0 10 20 30 40 50
2 2 2
Z' / k cm Z' / k cm Z' / k cm
0 0
-10 -10
2
Z'' / k cm
E / V vs Ag/AgCl
-20 -20 0.4
0.2
-30 -30 0
2 -0.2
-40 10 mW/cm -40 -0.3
2
-50 5.0 mHz -50 5 mW/cm
0 10 20 30 40 50 0 10 20 30 40 50
2 2
Z' / k cm Z' / k cm
Figure 4.10. Nyquist plots recorded at different light intensities (from 93 to 5 mW cm2) in N2-
purged 0.1 M Na2SO4 at different potentials (0.4, 0.2, 0, 0.2 and 0.3 V vs Ag/AgCl) with a
perturbation amplitude of E= mV. A frequency was varied from 10 kHz to 3 mHz.
Photoelectrochemical impedance spectra for an applied potential of 0 V at different
applied power irradiance values are shown in Figure 4.11. The applied power irradiance
ranges from 5 to 93 mW cm2. Based on the absorption spectrum of the material, and
the illumination source employed (solar simulator), for a quantum yield equal to one,
the values of the power irradiance correspond to equivalent currents (e0) ranging from
0.8 to 17 mA cm2 (as estimated by considering a step function absorbance at the
wavelength corresponding to the band gap).37 Visibly, a semicircle at high frequencies
can be identified in the Nyquist plots. In addition, a second region at lower frequencies
is present. The assignment of this region to a transport process related with diffusion of
the products and/or the charge carriers in the semiconductor 38–43 was discarded because
diffusion processes emerge in EIS as straight lines with a slope of 45º (usually they are
represented as Warburg elements in equivalent circuits). In this case, the angle is
204
Investigating water splitting with CaFe2O4 photocathodes by electrochemical impedance spectroscopy
comprised between 40-60º, and it changes with the applied potential. These facts
suggest that the lower frequency part corresponds to a second semicircle of large radius.
2
Z'' / k cm
-25 e 0 / mA cm
-2
0
0.8
2
4
9 -10
17
0 10
-50
0 25 50
2
Z' / k cm
Figure 4.11. Experimental Nyquist plots in N2-purged 0.1 M Na2SO4 at a constant applied
potential of 0 V vs Ag/AgCl and different e0 values.
The impedance decreases when e0 increases for all the studied potentials as was
illustrated in Figure 4.10. A direct proportionality between the intercept of the
semicircle with the x-axis (Zcross) and the inverse of the irradiated light intensity is
observed in Figure 4.12A for applied potentials of 0 V, 0.2 and 0.3 V. The
experimental results clearly match the tendency predicted by the model (Figure 4.3). On
the other hand, the slope of the Zcross vs 01 plot is similar for applied potentials from 0
to 0.3 V. In fact, according to Equation (4.43), the proportionality factor in Equation
(4.7) is independent of irradiance. In addition, it can be deduced that k1 and k2 are
almost independent of both the applied irradiance (Zcross is independent of k4) and the
applied potential from 0 to 0.3 V. Although, a priori, k1 could be affected by the light
intensity in an indirect way by an alteration of the potential drop across the Helmholtz
layer, this effect is not present in this case, probably because of the low values of I 0 (see
below). In this respect, Ponomarev and Peter17 found by IMPS that the recombination
and transfer rate constants were affected by light intensity under strong illumination.
205
Chapter 4
40 12
E = -0.3 V E = -0.3 V
E = -0.2 V E = -0.2 V
E=0V E=0V
30
2
8
-1
Zcross / k·cm
max / rad s
20
4
10
A. B.
0 0
0 2 4 6 8 10 0 2 4 6 8 10 12
-1 -17 2 16 -2 -1
0 / 10 cm s 0 / 10 cm s
Figure 4.12. (A) Zcross vs and (B) max vs for the first semicircle of the Nyquist plots at
0, 0.2 and 0.3 V shown in Figure 4.10.
Similar conclusions can be drawn from the plot of max vs 0 shown in Figure 4.12B
analyzed according to Equation (4.44). However, a linear correlation of this parameter
with 0 is not so obvious probably because of the high error in the determination of
max.
As mentioned earlier, working with a kinetic model instead of with an equivalent circuit
is advantageous because all the parameters in the simulation have a well-defined
physical meaning. In Figure 4.13 the most relevant results on the effect of the potential
on the impedance at a selected light intensity of 93 mW cm2 are shown. It is important
to highlight that the theoretical points fit very well the experimental data for every
frequency in the whole range studied, which is clearly demonstrated in the Bode plot for
0.3 V shown in the inset of Figure 4.13. As observed, there is a very good agreement
between the calculated and the experimental Bode plots.
206
Investigating water splitting with CaFe2O4 photocathodes by electrochemical impedance spectroscopy
0 0
E = -0.3 V
1.5 Hz
2
1.5 Hz
-2 -2
Z'' / k cm
0.0
phase / rad
-0.4
-4 -4
-0.8
-3 -1 1 3
10 10 10 10 E = -0.2 V
f / Hz
-6 -6
0 2 4 6 0 2 4 6
0 0
1.5 Hz -10
2
-2
Z'' / k cm
-20
-4 -30
E=0 V -40 E = 0.2 V
-6
0 2 4 6 0 10 20 30 40
2 2
Z' / k cm Z' / k cm
Figure 4.13. Nyquist plots using a light intensity of 93 mW cm 2 in N2-purged 0.1 M Na2SO4
including experimental points (black solid symbols) and the simulated curves from the
theoretical model (red open symbols) at different potentials vs Ag/AgCl. The points
corresponding to max are marked in green. Inset represents experimental and simulated Bode
plots at -0.3 V. Simulation parameters: RH = 2.3·104 F cm2 s, = 0.5, CSC = 1.5·105 F
cm2, SS = 1.7·103 s, RS = 68 , C = 0.5, = 1, and I0, CSS, k1, k2 and k4 are given in Figure
4.14.
Admittedly, for the highest potentials the fit is worse. The reason may be the
contribution of the dark component to the overall current, because the model is
developed by considering only the minority carriers generated under illumination.
Consequently, its applicability is limited to the case when dark currents are negligible
compared with the photocurrents.
The impedance is virtually potential-independent in the higher frequency range. This
region is governed by parameters related with the capacitances of the double layer,
revealing that some of these parameters such as RH, , CSC, and SS, can be considered
as independent of the applied potential.
For all the studied light intensities, impedance adopts comparable values at high reverse
bias, and increase largely at potentials above 0 V, close to the photocurrent onset, as
207
Chapter 4
illustrated in Figure 4.10. In fact, in Figure 4.13, the impedance of the first semicircle
reached similar values for applied potentials of 0.3, 0.2 and 0 V. Concretely, Zcross
and max are very similar in the three cases (see also Figure 4.12). According to the
equations (4.43) and (4.44), the observed behavior suggests that I0, k1, k2, CSC and CSS
should change only slightly within this range of potential. On the other hand, Figure
4.14A shows that the magnitude of I0 does not change much upon varying the applied
potential, which supports the assumption that I0 only depends on 0, as previously
discussed.
6
-1
I0 / 10 cm s
A.
-2
4
14
0
0.6 B.
-2
CSS / mF cm
0.4
0.2
0.0
10
0
C. k4
-1
ki / s
-1
10
k2
-2
10
k1
-3
10
-0.3 -0.2 -0.1 0.0 0.1 0.2
E (vs Ag/AgCl) / V
Figure 4.14. Parameters obtained by fitting the experimental impedance data in Figure 4.13 to
the theoretical model as a function of potential: (A) I0 (B) CSS and (C) k1, k2 and k4.
As observed in Figure 4.14B, CSS increases with potential, especially for the highest
potential values. Previously, the capacitive currents (accumulation region) in the dark
starting at 0 V (Figure 4.8) were correlated with the filling/emptying of surface states
located at energies within the band gap (close to the valence band edge). The increase in
CSS can be tentatively correlated with such voltammetric region.
The fitting parameters obtained from the experimental data also reveal that C SS is larger
than CSC for all the studied potentials. Thus, the part of the applied potential that drops
208
Investigating water splitting with CaFe2O4 photocathodes by electrochemical impedance spectroscopy
in the semiconductor side of the interface, produce the charge/discharge of surface
states, rather than an alteration of the space charge region. This is in agreement with the
observed C2 vs E behavior (inset, Figure 4.8A).
Working with CPE elements does not allow for a direct obtainment of the CH value.
However, the “effecti e equi alent ca acitance C H,eff” of the elmholtz la er can be
estimated44 as ma
and it has a value of 7.5·105 F cm2. The value of max has
been taken from the first semicircle in the Nyquist plots. The fact that CH,eff reaches the
same order of magnitude than CSS (which is significantly larger than CSC) implies that
changes in the applied potential would also affect the Helmholtz layer potential drop.
Regarding the dependence of the kinetic constants with the applied potential, the
behavior observed in Figure 4.14C is the expected one: k4 increases with potential, k1
decreases with the applied potential and k 2 does not significantly depend on potential.
An increase in k4 of about two orders of magnitude in the whole range of potential (0.5
V) is deduced after fitting the experimental impedance data. However, according to
Equation (4.21), a change of eight orders of magnitude would be expected for pure BEP
behavior.
Figure 4.15 simulate vs frequency from Equation (4.37) using the fitting
parameters at each potential (according to Figure 4.13 and Figure 4.14). Two regions for
the values of can be distinguished, one at high frequencies (above 10 Hz) and
other at low frequencies (<10 Hz). At high frequencies, when capacitive contributions
dominate, a high portion of the drop of potential is inverted in the Helmholtz layer
(close to the FLP case, but considering the non-ideality of the surface by the substitution
of CH by a CPE). However, at low frequencies the potential drop is mainly governed by
the kinetics of the different processes illustrated in Figure 4.2, causing a potential drop
variation distributed between the Helmholtz and the space charge layer, because of the
accumulation of free and trapped electrons at the surface. This is in agreement with the
fact that CH,eff reaches the same order of magnitude as CSS (discussed above).
209
Chapter 4
1.0
0.8
0.6
ϕT
ϕH ∆ϕ
0.4
∆ϕ E/V
0.2
0
0.2 -0.2
-0.3
0.0 -3 -1 1 3
10 10 10 10
f / Hz
Figure 4.15. as a function of the frequency for different applied potentials. The
values of the different parameters are given in the captions for Figure 4.13 and Figure 4.14.
An average value of 0.7-0.8 can be roughly estimated based on Figure 4.15.
About a fourth of the applied potential change is reflected in the potential drop in the
space charge region, which is in agreement with an increase of k 4 of two orders of
magnitude instead of eight when increasing the applied potential by 0.5 V. Similarly,
the decrease of k1 with potential is smaller than that expected from Equation (4.22),
which reflects the fact that the electrode is not under FLP either.
The maximum flux of photogenerated electrons (0) can be roughly estimated as
1.1·1017 cm2 s1 (17 mA cm2).37 Therefore, a large difference is observed between 0
and the actual electron flux toward the surface (I0); the I0/0 ratio attains an average
value of 3·103. This means that only 0.3 % of the photogenerated electrons reach the
surface. As an important conclusion, this reveals that the main path for the
recombination of electrons and holes occurs in the bulk of the semiconductor. The
relatively high thickness of the electrode and especially the fact that the electrode does
not follow a BEP behavior favors recombination in the bulk of the semiconductor and
impede an effective collection of charge carriers at the surface. This suggests that the
concentration of electrons trapped in surface states is low, which is confirmed by the
absence of large spikes upon tuning the light on and off in the photocurrent transients of
210
Investigating water splitting with CaFe2O4 photocathodes by electrochemical impedance spectroscopy
Figure 4.8. In addition, this indicates that band bending cannot be greatly enhanced by
applying potentials negative to the flat band potential as in such a case a drastic increase
in I0 should be observed, which is not the case.
The ratio k1/(k1+k2) provides information about the fraction of electrons that are
transferred to solution and the fraction that recombines with holes at surface states. For
the highest potential value, the majority of electrons recombine (k1/(k1+k2)<<1), this is
reflected by small spikes (due to the low value of I0) when light is turned on in Figure
4.8 at 0.2 V. For lower potentials, the ratio is estimated to be 0.5-0.6, which would be
reflected in higher stationary photocurrent values given by Equation (4.12).
The excellent agreement at both qualitative and quantitative levels between experiments
and simulations clearly indicates that the assumptions upon which the model is based
are plausible. The most critical hypothesis is that direct CB electron transfer is the
predominant pathway for water reduction. This is in agreement with the fact that an
attempt to adjust the experimental EIS data to a model equivalent to that of Leng et al.,13
in which indirect electron transfer from surface states is considered, leads to
inconsistencies in the fitting parameters.
4.5. Conclusions
In this chapter, a novel photoelectrochemical impedance model based on a kinetic
mechanism has been developed and applied for the investigation of hydrogen generation
at the SEI of the Pt/CaFe2O4 p-type metal oxide photoelectrode. The mechanism
proposed in the model is fully consistent with the experimental data. The presence of
surface states is considered but only as electron trapping and recombination centers,
being the conduction band electrons those directly transferred to water molecules. The
quantitative agreement between experimental and calculated impedance data
demonstrated by a simultaneous fit of both Nyquist and Bode plots is remarkable. It is
worth nothing that the fit has been achieved by using only parameters with a clear
physical meaning. Interestingly, the fit of the experimental data to the model reveals
that most of the photogenerated charge carriers recombine in the bulk of the
211
Chapter 4
semiconductor, and that only a minor portion (0.3%) of the photogenerated electrons
reaches the electrode surface. The concentration of trapped electrons in surface states is
low, which is reflected in the chopped light voltammogram by the absence of spikes in
the light transients. The values of the different interfacial capacitances resulting from
the model reveal that the behavior of the CaFe 2O4 SEI does not follow either Fermi
level pinning or band edge pinning in agreement with the experimental Mott-Schottky
plot. Importantly, the model considers the variation of the rate constants with the
applied potential and it is found that all of them follow the expected trend, within the
studied potential ranges, which is a strong validation of the underlying hypotheses.
In a more general vein, this chapter illustrates how a physicochemical model can be
used to fit the EIS experimental behavior of a p-type electrode in the context of
photoelectrochemical water splitting, providing critical information on the water
photoreduction mechanism. The identification of the main losses in performance should
help develop strategies for the improvement of the photoelectrode.
4.6. Bibliography
(1) Matsumoto, Y.; Omae, M.; Sugiyama, K.; Sato, E. New Photocathode Materlals for Hydrogen
Evolution: CaFe2O4 and Sr7Fe10O22. J. Phys. Chem. 1987, 91, 577–581.
(2) Matsumoto, Y.; Sugiyama, K.; Sato, E. Photocathodic Hydrogen Evolution Reactions at p-Type
CaFe2O4 Electrodes with Fermi Level Pinning. J. Electrochem. Soc. 1988, 135, 98–104.
(3) Matsumoto, Y.; Sugiyama, K.; Sato, E.-I. Improvement of CaFe2O4 photocathode by doping with
Na and Mg. J. Solid State Chem. 1988, 74, 117–125.
(4) Ida, S.; Yamada, K.; Matsunaga, T.; Hagiwara, H.; Matsumoto, Y.; Ishihara, T. Preparation of p-
type CaFe2O4 photocathodes for producing hydrogen from water. J. Am. Chem. Soc. 2010, 132,
17343–17345.
(5) Cao, J.; Kako, T.; Li, P.; Ouyang, S.; Ye, J. Fabrication of p-type CaFe2O4 nanofilms for
photoelectrochemical hydrogen generation. Electrochem. Commun. 2011, 13, 275–278.
(6) Sekizawa, K.; Nonaka, T.; Arai, T.; Morikawa, T. Structural improvement of CaFe2O4 by metal
doping toward enhanced cathodic photocurrent. ACS Appl. Mater. Interfaces 2014, 6, 10969–
10973.
212
Investigating water splitting with CaFe2O4 photocathodes by electrochemical impedance spectroscopy
(7) Lopes, T.; Andrade, L.; Ribeiro, H. A.; Mendes, A. Characterization of photoelectrochemical
cells for water splitting by electrochemical impedance spectroscopy. Int. J. Hydrogen Energy
2010, 35, 11601–11608.
(8) Shimizu, K.; Lasia, A. Electrochemical Impedance Study of the Hematite/Water Interface.
Langmuir 2012, 28, 7914–7920.
(9) 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.
(10) 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.
(11) Gimenez, S.; Dunn, H. K.; Rodenas, P.; Fabregat-Santiago, F.; Miralles, S. G.; Barea, E. M.;
Trevisan, R.; Guerrero, A.; Bisquert, J. Carrier density and interfacial kinetics of mesoporous
TiO2 in aqueous electrolyte determined by impedance spectroscopy. J. Electroanal. Chem. 2012,
668, 119–125.
(12) Wang, Y.; Cui, X.; Zhang, Y.; Gao, X.; Sun, Y. Preparation of Cauliflower-like ZnO Films by
Chemical Bath Deposition: Photovoltaic Performance and Equivalent Circuit of Dye-sensitized
Solar Cells. J. Mater. Sci. Technol. 2013, 29, 123–127.
(13) Leng, W. H.; Zhang, Z.; Zhang, J. Q.; Cao, C. N. Investigation of the Kinetics of a TiO2
Photoelectrocatalytic Reaction Involving Charge Transfer and Recombination through Surface
States by Electrochemical Impedance Spectroscopy. J. Phys. Chem. B 2005, 109, 15008–15023.
(14) Liu, Z.; Zhao, Z. G.; Miyauchi, M. Efficient visible light active CaFe2O4/WO3 based composite
photocatalysts: Effect of interfacial modification. J. Phys. Chem. C 2009, 113, 17132–17137.
(15) Peter, L. M.; Ponomarev, E. A.; Fermín, D. J. Intensity-modulated photocurrent spectroscopy:
reconciliation of phenomenological analysis with multistep electron transfer mechanisms. J.
Electroanal. Chem. 1997, 427, 79–96.
(16) Liu, H.; Li, X. Z.; Leng, Y. J.; Li, W. Z. An Alternative Approach to Ascertain the Rate-
Determining Steps of TiO2 Photoelectrocatalytic Reaction by Electrochemical Impedance
Spectroscopy. J. Phys. Chem. B 2003, 107, 8988–8996.
(17) Ponomarev, E. A.; Peter, L. M. A comparison of intensity modulated photocurrent spectroscopy
and photoelectrochemical impedance spectroscopy in a study of photoelectrochemical hydrogen
evolution at p-InP. J. Electroanal. Chem. 1995, 397, 45–52.
(18) Losiewicz, B.; Budniok, A.; Rowinski, E.; Lagiewka, E.; Lasia, A. Effect of heat-treatment on the
mechanism and kinetics of the hydrogen evolution reaction on Ni-P + TiO2 + Ti electrodes. J.
Appl. Electrochem. 2004, 34, 507–516.
213
Chapter 4
(19) Lačn e ac U Č ; Jo ić M ; Jo ić D ; Krsta ić Determination of inetic arameters
for the hydrogen evolution reaction on the electrodeposited Ni–MoO2 composite coating in
alkaline solution. J. Electroanal. Chem. 2012, 677–680, 31–40.
(20) Aruna, K. K.; Manoharan, R. Electrochemical hydrogen evolution catalyzed by SrMoO 4 spindle
particles in acid water. Int. J. Hydrogen Energy 2013, 38, 12695–12703.
(21) Li, J.; Peter, L. Surface recombination at semiconductors electrodes. Part III. Steady-state and
intensity modulated photocurrent response. J. Electroanal. Chem. 1985, 193, 27–47.
(22) Pleskov, Y. V.; Gurevich, Y. Y. Semiconductor Photoelectrochemistry; Consultants Bureau: New
York, 1986.
(23) Bard, A. J.; Bocarsly, A. B.; Fan, F.-R. F.; Walton, E. G.; Wrighton, M. S. The concept of Fermi
level pinning at semiconductor/liqud junctions. Consequences for energy conversion efficiency
and selection of useful redox couples in solar devices. J. Am. Chem. Soc. 1980, 102, 3671–3677.
(24) Vanmaekelbergh, D. Direct and surface state mediated electron transfer at
semiconductor/electrolyte junctions—I. A comparison of steady-state results. Electrochim. Acta
1997, 42, 1121–1134.
(25) Hens, Z. The Electrochemical Impedance of One-Equivalent Electrode Processes at Dark
Semiconductor/Redox Electrodes Involving Charge Transfer through Surface States. 1. Theory. J.
Phys. Chem. B 1999, 103, 122–129.
(26) Pleskov, Y. V.; Elkin, V. V.; Abaturov, M. A.; Krotova, M. D.; Mishuk, V. Y.; Varnun, V. P.;
Teremetskaya, I. G. Synthetic semiconductor diamond electrodes: Elucidation of the equivalent
circuit for the case of frequency-dependent impedance. J. Electroanal. Chem. 1996, 413, 105–
110.
(27) Spagnol, V.; Sutter, E.; Debiemme-Chouvy, C.; Cachet, H.; Baroux, B. EIS study of photo-
induced modifications of nano-columnar TiO2 films. Electrochim. Acta 2009, 54, 1228–1232.
(28) Hirschorn, B.; Orazem, M. E.; Tribollet, B.; Vivier, V.; Frateur, I.; Musiani, M. Constant-Phase-
Element Behavior Caused by Resistivity Distributions in Films. II. Applications. J. Electrochem.
Soc. 2010, 157, C458–C463.
(29) Lasia, A. Electrochemical Impedance Spectroscopy and Its Applications; Springer
Science+Business Media: New York, 2014.
(30) Rammelt, U.; Reinhard, G. On the applicability of a constant phase element (CPE) to the
estimation of roughness of solid metal electrodes. Electrochim. Acta 1990, 35, 1045–1049.
(31) Kühne, H.-M.; Schefold, J. Tafel Plots from Illuminated Photoelectrodes. A New Insight Into
Charge Transfer Mechanism. J. Electrochem. Soc. 1990, 137, 568–575.
(32) Kim, E. S.; Nishimura, N.; Magesh, G.; Kim, J. Y.; Jang, J.; Jun, H.; Kubota, J.; Domen, K.; Lee,
J. S. Fabrication of CaFe2O4/TaON Heterojunction Photoanode for Photoelectrochemical Water
Oxidation. J. Am. Chem. Soc. 2013, 135, 5375–5383.
214
Investigating water splitting with CaFe2O4 photocathodes by electrochemical impedance spectroscopy
(33) Khan, S. U. M.; Akikusa, J. Photoelectrochemical Splitting of Water at Nanocrystalline n-Fe2O3
Thin-Film Electrodes. J. Phys. Chem. B, 1999, 103, 7184–7189.
(34) Paracchino, A.; Mathews, N.; Hisatomi, T.; Stefik, M.; Tilley, S. D.; Grätzel, M. Ultrathin films
on copper(I) oxide water splitting photocathodes: a study on performance and stability. Energy
Environ. Sci. 2012, 5, 8673–8681.
(35) Gurudayal, G.; Chiam, S. Y.; Kumar, M. H.; Bassi, P. S.; Seng, H. L.; Barber, J.; Wong, L. H.
Improving the efficiency of hematite nanorods for photoelectrochemical water splitting by doping
with manganese. ACS Appl. Mater. Interfaces 2014, 6, 5852–5859.
(36) Digdaya, I. A.; Han, L.; Buijs, T. W. F.; Zeman, M.; Dam, B.; Smets, A. H. M.; Smith, W. A.
Extracting large photovoltages from a-SiC photocathodes with an amorphous TiO2 front surface
field layer for solar hydrogen evolution. Energy Environ. Sci. 2015, 8, 1585–1593.
(37) Smestad, G. P.; Krebs, F. C.; Lampert, C. M.; Granqvist, C. G.; Chopra, K. L.; Mathew, X.;
Takakura, H. Reporting solar cell efficiencies in Solar Energy Materials and Solar Cells. Sol.
Energy Mater. Sol. Cells 2008, 92, 371–373.
(38) Hens, Z.; Gomes, W. P. On the Diffusion Impedance at Semiconductor Electrodes. J. Phys.
Chem. B 1997, 101, 5814–5821.
(39) Bisquert, J.; Garcia-Belmonte, G.; Bueno, P.; Longo, E.; Bulhões, L. O. S. Impedance of constant
phase element (CPE)-blocked diffusion in film electrodes. J. Electroanal. Chem. 1998, 452, 229–
234.
(40) Bisquert, J.; Garcia-Belmonte, G.; Fabregat-Santiago, F.; Bueno, P. R. Theoretical models for ac
impedance of diffusion layers exhibiting low frequency dispersion. J. Electroanal. Chem. 1999,
475, 152–163.
(41) Bisquert, J.; Compte, A. Theory of the Electrochemical Impedance of Anomalous Diffusion. J.
Electroanal. Chem. 2001, 499, 112–120.
(42) Arutyunyan, V. M.; Arakelyan, V. M.; Shakhnazaryan, G. E.; Stepanyan, G. M.; Turner, J. A.;
Khaselev, O. Investigation of ceramic Fe2O3(Ta) photoelectrodes for solar energy
photoelectrochemical converters. Int. J. Hydrogen Energy 2002, 27, 33–38.
(43) Aroutiounian, V. M.; Arakelyan, V. M.; Shahnazaryan, G. E.; Hovhannisyan, H. R.; Wang, H.;
Turner, J. A. Photoelectrochemistry of tin-doped iron oxide electrodes. Sol. Energy 2007, 81,
1369–1376.
(44) Lvovich, V. F. Impedance Spectroscopy: Applications to Electrochemical and Dielectric
Phenomena; John Wiley & Sons: Hoboken (New Jersey), 2012.
215
Chapter 5
LaFeO3 photocathodes: metal doping as
way to enhance the photoelectrochemical
behavior
5.1. Introduction ........................................................................................................... 219
5.2. Experimental section ............................................................................................. 220
5.2.1. LaFeO3 thin film preparation ....................................................................... 220
5.2.2. Electrode characterization ............................................................................ 221
5.3. Photoelectrochemical characterization of pristine and doped LaFeO 3 thin
film electrodes ................................................................................................................... 222
5.3.1. Physical characterization of the electrodes .................................................. 222
5.3.2. Effect of Mg and Zn doping in the PEC response ....................................... 226
5.3.3. Investigating the role of doping..................................................................... 235
5.4. Conclusions............................................................................................................. 240
5.5. Bibliography ........................................................................................................... 240
LaFeO3 photocathodes: metal doping as way to enhance the photoelectrochemical behavior
5.1. Introduction
Perovskite-type oxides may be seen as a good alternative as p-type semiconductors
because most of them exhibit p-type semiconducting nature and high stability under
illumination in aqueous solutions. In the introduction (Chapter 1) the main works in
photoelectrochemistry related with perovskite ternary oxides were reviewed. As we
mentioned in section 1.6.3, LaFeO3 appears as a promising candidate, and several
reports investigating this material as a photocathode in the context of
photoelectrochemistry have been recently released.1,2 It is worth noting that the
development of not only cathodic, but also anodic photocurrents has been reported for
this material depending on the particular conditions of the experiment. In fact, May and
co-workers3 have studied ultrathin LaFeO 3 films over Nb:SrTiO3 prepared by PLD. For
all the electrodes, anodic photocurrents are observed. In addition, cathodic
photocurrents appear in the presence of O2 for LaFeO3 films with thickness equal or
higher than 5 nm. Peng et al.4 also demonstrated that nanoparticulate LaFeO3 electrodes
can develop either cathodic or anodic photocurrents. On the other hand, several works
have focused on the properties of this material as a photoanode, reporting an
enhancement of photoelectrochemical water oxidation by doping with metals5 and by
incorporating co-catalysts.4,6 For instance, surface modification with Co-Pi as a
cocatalyst leads to an increase of the n-type character, suppressing the photocathodic
current and shifting the onset toward more negative values. 4
Doping may also be an appropriate strategy for increasing the p-type character of the
ternary oxides.7 It has been reported that doping with metals having a +2 oxidation state
and radii similar to that of Fe3+ (such as Mg2+ or Zn2+) triggers changes in the structural,
electrical and magnetic properties of the pristine material. 8–10 Concretely, using Mg and
Zn as dopants in the LaFeO3 perovskite has been demonstrated to be an effective
strategy for the development of gas sensors,11–13 oxidation catalysts9 and
photocatalysts.14
In this chapter, LaFeO3 thin film photoelectrodes prepared on transparent conductive
substrates following a simple and scalable sol-gel procedure are shown to exhibit
cathodic photocurrents in the studied range of potentials. Such a p-type character is
enhanced by the partial substitution of Fe by Mg or Zn in the perovskite structure. This
strategy produces a drastic increase of the O2 photoreduction reaction. Electrochemical
219
Chapter 5
ac methods (Mott-Schottky and Nyquist plots) help reveal important information about
the effect of the substitution of Fe by Mg and Zn on the photoresponse.
5.2. Experimental section
5.2.1. LaFeO3 thin film preparation
A citric sol-gel route is employed for the preparation of the different films using FTO
conducting glass as a substrate. To synthesize the pristine LaFeO 3 sample,
La(NO3)3·6H2O (0.3 M) and Fe(NO3)3·9H2O (0.3 M) were dissolved in water and the
solution was stirred for 1 h. Then, the appropriate amount of citric acid monohydrate
was added to reach a concentration of 0.6 M and the solution was stirred again for 20 h.
The LaFe1xMgxO3 and LaFe1xZnxO3 (with x=0.01, 0.025, 0.05 and 0.1) electrodes
were obtained using the above described procedure, but substituting the corresponding
amount of the Fe salt by Mg(NO3)2·6H2O or Zn(NO3)2·6H2O according to the final
desired stoichiometry.
The above mentioned solutions were mixed with acetylacetone (30 L/mL) and Triton
100X (30 L/mL) and the resultant liquid was spin-coated on an FTO substrate at a
rotating speed of 3000 rpm for 20 s and then the samples were calcined at 500 °C for 1
h. This procedure was repeated twice (two-layers) and a final heat treatment at 640 °C
for 2 h was applied, in order to obtain the crystalline perovskite films. The thickness of
the resulting electrodes was 80-100 nm (two-layer electrodes), as measured with an
Alpha Step D-100 profilometer. Thicker electrodes can be obtained by increasing the
number of spin-coating and calcination cycles. Unless otherwise stated, the results are
referred to two-layer electrodes.
220
LaFeO3 photocathodes: metal doping as way to enhance the photoelectrochemical behavior
5.2.2. Electrode characterization
A morphological examination of the films was carried out by Field Emission Scanning
Electron Microscopy (FE-SEM) with a ZEISS Merlin VP Compact microscope. The
crystal structure was studied by X-ray diffraction (XRD) with a Seifert JSO-Debyeflex
2002 diffractometer using the Cu Kα line (=1.5406 Å). The optical characterization
was performed by measuring their UV-visible transmittance spectra using an FTO plate
as a blank by means of a Shimadzu UV-2401PC spectrophotometer. XRD and UV/vis
analysis were performed using 10-layer thick LaFeO3, LaFe1-xMgxO3 or LaFe1-xZnxO3
electrodes (with x=0.05 and 0.1). Core level photoemission spectra were collected in
normal emission at room temperature with a (K-Alpha of Thermo-Scientific) X-ray
Photoelectron Spectrometer (XPS) using an Al Kα X-ray source and equipped with an
Ar ion sputter gun to perform depth profile analysis. Ta 2O5 was used as a reference to
estimate the depth.
(Photo)electrochemical measurements were performed at room temperature in a three-
electrode cell equipped with a fused silica window. All the potentials were measured
against and are referred to an Ag/AgCl/KCl(sat) reference electrode. A platinum wire
was used as a counter electrode and an N2- or O2-purged 0.1 M NaOH (Panreac, p.a.)
solution as a working electrolyte. The photoelectrochemical experiments were
controlled by an Autolab PGSTAT30 potentiostat coupled with a frequency response
analyzer. The electrodes were irradiated with 1 sun (100 mW cm −2) illumination from
the electrolyte-electrode side (EE illumination) using a solar simulator SUN 2000 (Abet
Technologies). All the impedance measurements were performed using a 10 mV
amplitude perturbation. Mott-Schottky plots were performed at 1 kHz and using a
potential step of 10 mV. Experiments at constant potential were measured in a
frequency range from 10 kHz to 50 mHz.
221
Chapter 5
5.3. Photoelectrochemical characterization of pristine and doped
LaFeO3 thin film electrodes
5.3.1. Physical characterization of the electrodes
The crystalline structure of the electrodes was characterized by XRD using 10-layer-
thick samples to improve the intensity of the diffraction peaks. The XRD patterns of
LaFeO3 and the Mg- and Zn-doped samples for x=0.05 and 0.10 are presented in Figure
5.1A. Apart from the signals associated with the fluorine-doped tin oxide (FTO)
substrate, all the diffraction peaks in the patterns of the pristine and doped samples are
fully compatible with those expected for the cubic form of LaFeO3 (01-075-0541 card
in the ICDD database), and no other phases are detected. This is in agreement with
reports that confirm the prevalence of the substitution of the Fe atoms in the perovskite
structure by Mg or Zn instead of the formation of new phases, at least for dopant
concentrations up to x=0.3.8,9,11,14 The FE-SEM images displayed in Figure 5.1B reveal
the formation of compact films with a degree of porosity and a similar surface
morphology for both pristine and doped samples.
A. LaFe0.90Zn0.10O3 B.
(110)
200 nm
(100)
(200)
(111)
(211)
(220)
LaFe0.95Zn0.05O3 LaFeO3
Intensity / a. u.
LaFe0.90Mg0.10O3
LaFe0.95Mg0.05O3
LaFe0.95 Mg0.05 O3
LaFeO3
LaFe0.95 Zn0.05 O3
20 30 40 50 60 70
2 / degree
Figure 5.1. (A) XRD patterns and (B) FE-SEM images of different pristine and doped LaFeO3
electrodes.
The high-resolution XPS spectra in the La 3d, Fe 2p, and O 1s regions for LaFeO3
presented in Figure 5.2 are very similar for the pristine and doped samples. There are no
222
LaFeO3 photocathodes: metal doping as way to enhance the photoelectrochemical behavior
shifts in the peak positions triggered by the introduction of Mg or Zn. The La 3d doublet
is located at a binding energy (BE) of 833.5 and 850.4 eV for La 3d5/2 and La 3d3/2,
respectively (Figure 5.2A). These peak positions15,16 suggest the +3 oxidation state for
La, further confirmed by a spin–orbit gap of approximately 16.8 eV between La 3d5/2
and La 3d3/2.17 In the Fe 2p region the peak positions for Fe 2p3/2 and Fe 2p1/2 are at
BE=709.8 and 723.7 eV, respectively (Figure 5.2B). The satellite at 718.0 eV is
indicative of the presence of Fe3+.18 However, the BEs are slightly lower than those
reported for Fe3+, and thus the presence of a minor amount of Fe2+ in the material cannot
be discounted.
A. La 3d5/2 La 3d B. Fe 2p3/2
Fe 2p C. Olatt
O 1s
Counts / a.u.
La 3d3/2 O-H
C=O
Fe 2p1/2 528 532
830 840 850 860 700 710 720 730 740 528 532 536
Binding energy / eV Binding energy / eV Binding energy / eV
D. Zn 2p E. Mg 1s F.
Zn 2p3/2 0.12
Mg/(La+Fe)
Counts / a.u.
0.08
Zn 2p1/2
0.04
0.00
0 10 20 30 40
1010 1020 1030 1040 1050 1300 1305 1310
Depth / nm
Binding energy / eV Binding energy / eV
Figure 5.2. High-resolution XPS spectra that show the (A) La 3d, (B) Fe 2p, and (C) O 1s
regions of the LaFeO3 (blue line), LaFe0.95Mg0.05O3 (black line), and LaFe0.95Zn0.05O3 (red line)
electrodes. The inset shows the three components that result from the deconvolution of the O1s
spectrum for the LaFeO3 sample. High-resolution XPS spectra in the (D) Zn 2p region for
LaFe0.95Zn0.05O3 and (E) Mg 1s region for LaFe0.95Mg0.05O3. (F) Atomic Mg/(La+Fe) ratio from
the XPS spectra as a function of the depth for a LaFe 0.95Mg0.05O3 electrode (the green line
indicates the expected value).
223
Chapter 5
The O1s XPS spectra (Figure 5.2C) can be deconvoluted into three chemically shifted
components in all cases. The component at BE= 528.9 eV is assigned to oxygen in the
perovskite lattice.15 The second peak at 530.8 eV is ascribed to OH species as the rare
Earth oxides are highly hygroscopic. It is known that La2O3 absorbs water rapidly with
a concomitant conversion of surface anions to give rise to a hydroxylated surface.15,19
Finally, the component at BE=532 eV can be associated with O atoms in adsorbed
carbonated species rather than to adsorbed water, the signal of which is located typically
at higher BE values. The carbonated species likely correspond to adventitious impurities
that come from the exposure of the electrode to ambient air before the measurements.
Table 5.1. Atomic ratio obtained by XPS analysis for three different electrodes. The figures in
red between brackets represent the corresponding theoretical values in the bulk.
Atomic ratio
La/Fe Mg/Fe Zn/Fe OH/Olatt
LaFeO3 1.8 (1.0) - - 0.41
LaFe0.95Mg0.05O3 2.0 (1.05) 0.34 (0.05) - 0.52
LaFe0.95Zn0.05O3 1.9 (1.05) - 0.4 (0.05) 0.51
More importantly, the XPS spectra shown in Figure 5.2D and E confirm the presence of
either Mg or Zn at the surface. BEs of 1304.0 eV for Mg 1s and 1021.7 eV for Zn 2p3/2
are in accordance with a +2 oxidation state in both cases, as expected. The surface
composition obtained from the peak areas in the photoemission spectra for the La 3d, Fe
2p, Mg 1s, and Zn 2p regions are presented in Table 5.1. The La/Fe ratio derived from
XPS in these samples is about twice the stoichiometric ratio, revealing a segregation of
La in the surface, which was also reported by Barbero et al. 20 Moreover, for the
LaFe0.95Mg0.05O3 and LaFe0.95Zn0.05O3 electrodes, the surface ratios of Mg/Fe and Zn/Fe
are 7–8 times the expected bulk concentration. The evident segregation of these ions
may be correlated with the observed increase in the hydroxyl surface concentration for
the doped electrodes (Table 5.1) because the oxygenated Mg and Zn species are more
hygroscopic than the Fe ones. In this respect, the values of the Mg/(La+Fe) atomic ratio
as a function of the depth are given in Figure 5.2F. Mg segregates to the surface,
although the Mg concentration is virtually constant and close to the expected value for
depths equal to or greater than 9 nm.
224
LaFeO3 photocathodes: metal doping as way to enhance the photoelectrochemical behavior
The UV/vis absorption spectra for 10-layer-thick pristine and Zn- or Mg-doped LaFeO3
electrodes are presented in Figure 5.3A. The absorbance values were calculated as
A=2−log(%T) (see Equation 2.25) from the percent transmittance (%T). Tauc plots are
represented in Figure 5.3B and C in order to obtain the indirect and direct optical band
gaps, respectively. For the pristine LaFeO3 thin-film electrode in Figure 5.3A, an
absorption edge at approximately =600 nm is present, which is in agreement with an
indirect optical band gap of 2.0 eV calculated by means of a Tauc plot (Figure 5.3B).
Such a band gap value is in line with other reported values.3,5 A direct transition band
gap of 2.7 eV can also be calculated (Figure 5.3C), which is slightly wider than that
found by other authors (2.47–2.56 eV) for the direct transition of LaFeO3
nanoparticles,1,21,22 and closer to the 2.65 eV value obtained for 25–55 nm thick films.23
Notably, the indirect and direct band gaps change slightly with the doping level.
However, irrespective of the presence of either Mg or Zn as dopants, the transmittance
vs curves are virtually identical for the same value of x.
A. 1.5
LaFeO3
Absorbance
LaFe0.95Mg0.05O3
1.0 LaFe0.90Mg0.10O3
LaFe0.95Zn0.05O3
LaFe0.90Zn0.10O3
0.5
0.0
400 500 600 700 800
/ nm
B. 3 C. 12
1/2
2
/ eV
(Ah) / eV
2 8
1/2
2
(Ah)
1 4
Eg = 2.0 eV Eg = 2.7 eV
0 0
2 3 4 2 3 4
h / eV h / eV
Figure 5.3. (A) UV/vis absorption spectra for different pristine and doped LaFeO3 electrodes
and the corresponding Tauc plots for (B) indirect and (C) direct optical transitions. The obtained
band gap value (Eg) is indicated by means of arrows.
225
Chapter 5
5.3.2. Effect of Mg and Zn doping in the PEC response
The photoelectrochemical response for pristine and doped LaFeO3 electrodes in the
form of linear scan voltammograms under chopped illumination recorded in either N2-
or O2- purged 0.1 M NaOH are shown in Figure 5.4. The current–potential response for
the LaFeO3 electrode in N2-purged electrolyte exhibits photocathodic currents at
potentials as high as 0.5 V (photocurrent onset). The photocurrent transients present
spikes upon both illumination and light interruption, which indicates a high
recombination rate of the charge carriers caused by an effective trapping of
photogenerated electrons at the surface.24 In any case, the excellent stability of the
electrodes under illumination, together with the reported ability of this material for H2
production in photocatalysis,25 points to the assignment of the observed photocurrents in
N2-purged electrolyte to the H2 evolution reaction.
0.01
0.00
-2
-0.01
j / mA cm
-0.02
N2
-0.03 O2
-0.04
-0.4 -0.2 0.0 0.2 0.4
E (vs Ag/AgCl) / V
Figure 5.4. Linear scan voltammograms (negative-going scan) under chopped 1 sun
illumination in a 0.1 M NaOH electrolyte purged with either N2 or O2 at a scan rate of 5 mV s−1
for a LaFeO3 electrode.
Figure 5.4 displays similar photocurrent onset in both N2- and O2-purged electrolytes.
Oxygen can be employed as an electron scavenger in photoelectrochemical studies.
When the electrolyte is purged with O2, a drastic increase in photocurrent below 0.2 V
is observed together with the suppression of the spikes in the illumination transients at
potentials lower than approximately 0 V. These facts indicate that surface electron
trapping is significantly hindered in the presence of O2.
226
LaFeO3 photocathodes: metal doping as way to enhance the photoelectrochemical behavior
A. LaFeO3
B. LaFe0.99Mg0.01O3 F. LaFe0.99Zn0.01O3
C. LaFe0.975Mg0.025O3 G. LaFe0.975Zn0.025O3
-2
j / mA cm
D. LaFe0.95Mg0.05O3 H. LaFe0.95Zn0.05O3
E. LaFe0.9Mg0.1O3 I. LaFe0.9Zn0.1O3
-2
25 A cm 25 A cm
-2
-0.4 -0.2 0.0 0.2 0.4 -0.4 -0.2 0.0 0.2 0.4
E (vs Ag/AgCl) / V E (vs Ag/AgCl) / V
Figure 5.5. Linear scan voltammograms (negative-going scan) under 1 sun chopped
illumination in a 0.1 M NaOH electrolyte purged with N2 at a scan rate of 5 mV s−1 for (A)
LaFeO3, (B–E) LaFe0.95Mg0.05O3, and (F–I) LaFe0.95Zn0.05O3 electrodes with x = 0.01, 0.025,
0.05, and 0.1.
Figure 5.5 shows the photoelectrochemical behavior in N2-purged electrolyte for
LaFeO3 electrodes either undoped or doped with Mg and Zn. Notably, higher current
spikes upon illumination and light interruption are developed for the doped electrodes
although the magnitude of the stationary photocurrents remains almost unaltered. Thus,
it seems that electron transfer from surface states to water remains kinetically hindered.
Linear sweep voltammograms upon light interruption were also studied in the presence
of O2 for the undoped and doped LaFeO3 electrodes (Figure 5.6). In this case, the
photocurrents dramatically increase upon doping with Mg and Zn in contrast with the
case of N2-purged electrolyte. The enhancement is relatively large even for the lowest
doping concentration. Again, the photocurrents are significant below ca. 0.2 V.
227
Chapter 5
A. LaFeO3
LaFe0.99Mg0.01O3 LaFe0.99Zn0.01O3
B. F.
LaFe0.975Mg0.025O3 LaFe0.975Zn0.025O3
C. G.
LaFe0.95Mg0.05O3 LaFe0.95Zn0.05O3
D. H.
LaFe0.9Mg0.1O3 LaFe0.9Zn0.1O3
E. I.
2 2
25 A cm 25 A cm
-0.4 -0.2 0.0 0.2 0.4 -0.4 -0.2 0.0 0.2 0.4
E (vs Ag/AgCl) / V E (vs Ag/AgCl) / V
Figure 5.6. Linear scan voltammograms (negative-going scan) under chopped 1 sun
illumination in a 0.1 M NaOH electrolyte purged with O 2 at a scan rate of 5 mV s−1 for (A)
LaFeO3, (B–E) LaFe0.95Mg0.05O3, and (F–I) LaFe0.95Zn0.05O3 electrodes with x=0.01, 0.025, 0.05,
and 0.1.
Cyclic voltammograms in the dark shown in Figure 5.7 reveal a capacitive region at
positive potentials that fits well with the accumulation region for a p-type material,26
which was also present for CuFe2O4 electrodes in Chapter 3. As the potential of the
photocurrent onset is close to the onset of the capacitive currents in the dark, such an
accumulation region can tentatively be related to the filling/emptying of valence band
states. However, the photocurrents are noteworthy only below 0.2 V then, the
contribution of surface states to the accumulation in 0.2-0.5 V potential region is likely.
In this respect, it is important to highlight that DFT calculations reveal that the upper
part of the valence band consists mainly of O 2p states with an important contribution of
the Fe 3d eg states.27,28 Consequently, free holes in LaFeO3 can be ascribed to both O
228
LaFeO3 photocathodes: metal doping as way to enhance the photoelectrochemical behavior
and Fe4+ species. Minor differences have been observed in the cyclic voltammograms in
the doped electrodes. In the Zn-doped ones, a slight reduction of the accumulation
region is observed, which can be linked to the formation of a thin ZnO layer at the
surface because of the Zn segregation detected by XPS (Table 5.1).
A.
LaFeO3
B. F.
LaFe0.99Mg0.01O3 LaFe0.99Zn0.01O3
C. G.
LaFe0.975Mg0.025O3 LaFe0.975Zn0.025O3
D. H.
LaFe0.95Mg0.05O3 LaFe0.95Zn0.05O3
-2 -2
20 A cm 20 A cm
E. I.
LaFe0.9Mg0.1O3 LaFe0.9Zn0.1O3
0.0 0.2 0.4 0.6 0.0 0.2 0.4 0.6
E (vs Ag/AgCl) / V E (vs Ag/AgCl) / V
Figure 5.7. Cyclic voltammograms in the dark in N2-purged 0.1 M NaOH electrolyte at 20 mV
s1 for (A) LaFeO3, (BE) LaFe1xMgxO3 and (FI) LaFe1xZnxO3 electrodes.
The shape of the cyclic voltammogram in the dark together with the appearance of
cathodic photocurrents evinces the p-type nature of the pristine LaFeO3 perovskite,
which is thought to be because of the presence of small amounts of Fe 4+.29 In this
respect, Porta et al.8 determined a Fe4+/Fetotal ratio of 0.024 by using temperature-
programmed reduction (TPR) and of 0.032 by using chemical titration for LaFeO 3
powder obtained by a sol–gel method using the citrate route. The presence of Fe 4+ is
compensated for by the corresponding level of oxygen excess to yield a stoichiometry
LaFeO3+.
229
Chapter 5
A. 0.00 LaFeO3
-0.02
-2
j / mA·cm
-0.04 LaFe0.95Zn0.05O3
-0.06 LaFe0.95Mg0.05O3
0 20 40 60 80 100
t/s
-2 B.
jph / mA·cm 0.06
0.04
0.02 Mg
Zn
0.00
0.00 0.04 0.08
x
Figure 5.8. (A) Photocurrent transients at 0.25 V for pristine and Mg- and Zn-doped LaFeO3
electrodes for x=0.05. (B) Stationary photocurrent at 0.25 V as a function of x for
LaFe1xMgxO3 and LaFe1xZnxO3 electrodes. Electrolyte: O2-purged 0.1 M NaOH.
Stationary photocurrents under illumination in the presence of O2 were also recorded for
the different electrodes at 0.25 V. The transients for pristine LaFeO3 and samples
doped with x=0.05 are shown in Figure 5.8A. Except for a small initial spike present in
the pristine material, the photocurrents are constant over the illumination time for all the
electrodes, which indicates again that electron surface trapping is minor. The stationary
photocurrent values measured after 60 s of illumination is plotted against the dopant
concentration in Figure 5.8B. The photocurrent increases with the dopant concentration
to reach an optimum at x=0.05 for both dopants, and then it drops. Importantly, doping
with equivalent quantities of either Mg or Zn leads to similar photocurrent values,
which suggests that the roles of Mg and Zn in the perovskite structure are similar.
230
LaFeO3 photocathodes: metal doping as way to enhance the photoelectrochemical behavior
A. LaFeO3 B. LaFe0.95Mg0.05O3
0.005
0.02
-2
0.00
j / mA·cm
0.000
-0.02
EE ilumination
-0.005 SE ilumination
-0.04
C. D.
0.00 0.00
-2
j / mA·cm
-0.04
-0.01
-0.08
-0.02
-0.4 -0.2 0.0 0.2 0.4 -0.4 -0.2 0.0 0.2 0.4
E (vs Ag/AgCl) / V E (vs Ag/AgCl) / V
Figure 5.9. Linear sweep voltammograms under 1-sun transient illumination for electrodes of
(A, C) LaFeO3 and (B, D) LaFe0.95Mg0.05O3 in (A, B) N2-purged and (C, D) O2-purged 0.1 M
NaOH electrolyte using EE illumination (black line) or SE illumination (red line). Scan rate: 5
mV s−1.
For the two-layer electrode, the illumination direction (electrolyte–electrode (EE) vs
substrate–electrode (SE) illumination) leads to similar values of photocurrent for both
pristine and doped LaFeO3 electrodes and in both N2- and O2-purged electrolytes
(Figure 5.9AD). This is a result of the low thickness of the electrode that (i) transmits a
significant fraction of the incident light and/or (ii) has a thickness smaller than the sum
of the hole diffusion length and the space charge width.
231
Chapter 5
A. B.
0.15 2-layer 0.15 2-layer
4-layer 4-layer
6-layer 6-layer
0.10 0.10
-2
j / mA·cm
0.05 0.05
0.00 0.00
-0.05 -0.05
-0.2 0.0 0.2 0.4 0.6 -0.2 0.0 0.2 0.4 0.6
E (vs Ag/AgCl) / V E (vs Ag/AgCl) / V
Figure 5.10. Cyclic voltammograms in the dark in N2-purged 0.1 M NaOH electrolyte at 20
mV s1 for 2, 4 and 6 layers (i.e., different thicknesses) for (A) LaFeO3, (B) LaFe0.95Mg0.05O3.
Electrodes consisting of four and six layers were also studied together with the two-
layer ones. In the dark cyclic voltammograms in Figure 5.10, higher pseudo-capacitive
currents are developed in the charge accumulation region as the thickness increases. In
nanoporous electrodes, the charge involved within the charge accumulation region is
considered proportional to the real electrode surface area. Therefore the accumulated
charge should scale with thickness, which indicates that LaFeO3 (doped and pristine)
electrodes are not ideally compact and flat, having a certain degree of porosity. In
addition, SEM micrographs in Figure 5.1B reveal the presence of pores in the material.
This electrode morphology can be advantageous as a compact nature allows for the
formation of a well-defined space charge accumulation region and the moderate
porosity provides larger surface areas.
232
LaFeO3 photocathodes: metal doping as way to enhance the photoelectrochemical behavior
A. LaFeO3 B. LaFe0.95Mg0.05O3
0.005 0.04
-2
j / mA·cm 0.000 0.00
2-layer
-0.005 -0.04 4-layer
6-layer
C. D.
0.00 0.00
-2
j / mA·cm
-0.04
-0.01
-0.08
-0.02
-0.12
-0.4 -0.2 0.0 0.2 0.4 -0.4 -0.2 0.0 0.2 0.4
E (vs Ag/AgCl) / V E (vs Ag/AgCl) / V
Figure 5.11. Linear sweep voltammograms under 1-sun transient illumination for (A, C)
LaFeO3 and (B, D) LaFe0.95Mg0.05O3 electrodes in (A, B) N2-purged and (C, D) O2-purged 0.1 M
NaOH electrolyte for 2, 4 and 6 layers (i.e., different thicknesses) using EE illumination at a
scan rate of 5 mV s−1.
Figure 5.11 displays the photocurrents under transient illumination for electrodes having
a different number of spin coated layers. In N2-purged electrolyte, the photocurrent does
not depend significantly on the electrode thickness, which could be attributed to the
small photoresponse found for H2 evolution. In contrast, the effect is significant in the
presence of 2, especially for the Mg-doped electrodes.
233
Chapter 5
A. LaFeO3 B. LaFe0.95Mg0.05O3
16 100
80
12
-2
j / A cm
60
8
40
4 EE 20 EE
SE SE
0 0
2 3 4 5 6 2 3 4 5 6
number of layers number of layers
Figure 5.12. Stationary photocurrents at 0.4 V with EE illumination for (A) LaFeO3 and (B)
LaFe1xMgxO3 electrodes in O2-purged 0.1 M NaOH electrolyte.
Figure 5.12 shows the stationary photocurrents at 0.4 V after 60 s of illumination in
O2-purged electrolyte (the shape of the transients is similar to those in Figure 5.8A). In
the pristine electrodes, an increase of the photocurrent is observed from 2 to 3 layers,
reaching an optimum for 3-4 layers and then, the photocurrents decrease. In the case of
the LaFe0.95Mg0.05O3 electrodes, the changes in the photocurrent with the number of
layers are more noticeable, defining clearly a maximum of approximately −0.09 mA
cm−2 for 4-layers LaFe0.95Mg0.05O3 electrodes. On the other hand, the slightly larger
values using EE illumination respect to SE illumination can be linked to the fact that the
material absorbs a significant part of the spectrum in the UV range and using SE
illumination, part of the UV light is absorbed by the FTO substrate before reaching the
oxide material. In any case, photocurrents under EE and SE illumination remain similar
for all the studied range of thickness. This could be indicative of large diffusion lengths
for charge carriers.
234
LaFeO3 photocathodes: metal doping as way to enhance the photoelectrochemical behavior
5.3.3. Investigating the role of doping
To gain some insights into the mechanism for photocurrent enhancement upon the
introduction of Mg and Zn in the structure, the Mott–Schottky approximation was
applied to investigate possible changes in carrier density and flat band potential. The
C−2 vs potential plots presented in Figure 5.13A and B show a straight line with a
negative slope in all cases, as expected for p-type semiconductor electrodes under band
edge pinning. In addition, at potentials that are substantially more negative than the flat
band potential, a close-to-linear region with a positive slope is observed. There are
several reasons that could explain this behavior such as a contribution linked to the
conducting glass substrate or a loss of the band edge pinning condition associated with a
contribution of surface states.30
A. C.
LaFe1-xMgxO3 LaFe1-xMgxO3 x=0
30 x = 0.01
4
4
x= 0.025
C / 10 F cm
x= 0.05
-Z'' / k cm
x = 0.1
-2
20 0.18 Hz
10
2 0.95 Hz
0.05 Hz
10
-2
B. 0 D. 0
LaFe1-xZnxO3 30 LaFe1-xZnxO3
4
4
C / 10 F cm
2
-Z'' / k cm
-2
20 0.18 Hz
10
2 0.95 Hz 0.05 Hz
10
-2
0 0
-0.4 -0.2 0.0 0.2 0.4 0.6 0 10 20 30
2
E (vs Ag/AgCl) / V Z' / k cm
Figure 5.13. (A, B) Mott–Schottky plots in the dark at a frequency of 1 kHz for pristine and
doped LaFeO3 electrodes in N2-purged 0.1 M NaOH electrolyte; (C,D) Nyquist plots under
illumination at 0.25 V in O2-purged 0.1 M NaOH electrolyte for pristine and doped LaFeO 3
electrodes.
In the Mg-doped samples, no shift in the flat band potential located near 0.5 V (as
calculated as the intercept of the straight line with the x-axis) is discernible. In contrast,
235
Chapter 5
a shift of 0.1 V toward positive potentials is observed for all the Zn-doped electrodes.
The position of the valence band edge for these materials can thus be estimated to be
located at 1.45–1.55 V vs the RHE, as the measured flat band potential is expected to be
very close to the valence band edge location. Remarkably, the flat band potential values
are located near the observed photocurrent onset in either N2- or O2-purged electrolytes.
If a band gap value of 2.0 eV is considered for LaFeO3, the location of the valence band
(VB) edge would translate into a potential for the conduction band (CB) edge of 0.5 V
vs RHE. The band edge potentials values reveal that the material should be able to
reduce both water and O2, although the kinetics of these processes is slow particularly in
the case of water reduction. The potentials of the CB and VB edges are very similar for
all the studied samples, and therefore, the enhancement of the photoresponse should not
be ascribed to an increase in the reducing power of the photogenerated electrons.
However, it is known that the slope of the straight line in the Mott–Schottky plot
(Figure 5.13A and B) is inversely proportional to the density of the majority charge
carriers (see Equation (2.18)). If we consider a dielectric constant of 75,3,29 the
calculated majority charge carrier density (NA) would be as high as 8.3·1018 cm−3 for the
LaFeO3 electrode. A generalized decrease in the slope values is observed as the dopant
concentration increases (Table 5.2), which can be attributed to an increase in hole
density. This effect is especially notable for the highest concentrations of Mg and Zn
whereas a minor change (if any) is observed for the lowest x values. As a caveat, the
quantitative information obtained from the Mott–Schottky analysis should be taken with
caution given the rather stringent conditions that need to be met for its full validity. The
changes of the dielectric constant with frequency and/or the contribution of surface
states and/or to the electrode surface being rough could yield to deviation in the
calculated values of the flat band potential and NA.30,31
236
LaFeO3 photocathodes: metal doping as way to enhance the photoelectrochemical behavior
Table 5.2. Absolute values of the slopes of the linear region in the Mott-Schottky plots shown
in Figure 5.13A and B for pristine and doped LaFeO3 electrodes.
10−10 Slope / F−2 cm4 V−1
x
Mg Zn
0 22.7
0.01 17.7 22.9
0.025 6.4 13.3
0.05 3.0 4.5
0.10 1.3 2.6
Nyquist plots under illumination at 0.25 V in the presence of O2 are shown in Figure
5.13C and D for the different electrodes studied. One semicircle is observed in all cases.
A generalized decrease in the impedance values (semicircle radius) is noticeable with
the increment of the Zn and Mg content. The decrease in impedance is correlated with
the increment in the stationary photocurrent with x in Figure 5.8B and it can also be
associated with a higher photogenerated electron flow toward the surface as advanced in
models developed to understand the electrochemical impedance spectroscopy (EIS)
response of photoelectrodes (see Chapter 4).32 Peng et al.5 also observed lower
impedance values for a LaFeO3 electrode after doping with Mn, Co, and Cu, which they
related to a decrease in either charge transport or charge transfer resistance.
From the experiments discussed above, we may also conclude that an increase in bulk
electrical conductivity is induced by the introduction Mg and Zn in the perovskite
structure. This is supported by the reported increase of the conductivity with the
introduction of Zn in LaFeO3.10 In addition, the resistivity has also been shown to
decrease with increasing values of x, for LaFe1xMgxO3 and LaFe1xZnxO3 in gas-
sensing applications.11,13 Studies about the doping effect in other semiconductor oxide
materials have led to the same conclusion for doping levels similar to those employed in
the present work.33,34
On the other hand, the electrical conductivity () depends on NA and the mobility (p)
of the majority carriers as = NA·p·e, in which e is the elementary charge. The
contribution of NA in the conductivity enhancement is evident, as deduced from the
237
Chapter 5
Mott–Schottky plots. Although precise values of NA cannot be calculated (the dielectric
constant is expected to change with the dopant concentration) and the roughness factor
is not accurately known, an increase in the charge carrier density of around one order of
magnitude for x=0.1 with respect to pristine LaFeO3 may be estimated. This is in
agreement with the increase of the Fe4+/Fetotal ratio upon Mg doping observed by Porta
et al.8 for LaFe1xMgxO3+. If the trivalent Fe ions are substituted by the divalent ions
Mg or Zn, charge compensation may be accomplished either by the oxidation of Fe 3+ to
Fe4+ or by a partial loss of the oxygen excess (that is, by decreasing ). The unusually
small change in the majority carrier density deduced for the lowest x values (being the
Mg or Zn content relatively high) may be caused by a loss of excess oxygen almost
tantamount to the density of divalent cations introduced rather than to an increase in
Fe4+ density. This observation could also be related to a significant surface segregation
of the dopant, as revealed by using XPS (see discussion above), which would lead to a
lower level of material bulk modification. However, the fact that Figure 5.2F showed a
bulk concentration similar to the stoichiometric value, suggests that this is not a main
factor.
From the previous discussion, the enhancement of the photoelectrochemical response
upon doping cannot be explained solely by an increase in NA. It seems that there is also
a contribution that comes from an increase in hole mobility. An increase of p with
dopant concentration was reported for In2O3 with several metal dopants (such as W,
Mo, and Ti)33–35 and for the perovskite SrTiO3 doped with La,36 which was linked to
lattice distortion triggered by doping. Several reports pointed out a variation of the unit
cell parameters of the LaFeO3 perovskite structure with the introduction of divalent Zn
and Mg ions.8,10,11,13,14 This variation has been associated with a combined effect of the
larger substituting Zn2+ (74.5 pm) or Mg2+ (72.0 pm) ions in the octahedral B site in
comparison with Fe3+ (64.5 pm) and to the oxidation of Fe 3+ to the smaller Fe4+ (58.5
pm). Importantly, a surface passivation effect in the doped samples cannot be discarded.
This would be caused by the segregation of Mg and Zn to the surface and its subsequent
further hydroxylation, confirmed by using XPS (Figure 5.2). The decrease in the
semicircle radii with the doping level could also be linked to an enhancement of the
electron transfer at the interface (a decrease in the charge transfer resistant) caused by
the Mg- and Zn-induced passivation of surface recombination centers, as discussed
238
LaFeO3 photocathodes: metal doping as way to enhance the photoelectrochemical behavior
previously. However, the photocurrents for water photoreduction remain almost
unaltered upon doping, which suggests that the latter effect is minor.
Aside from doping, other strategies have been implemented for improving the response
of the LaFeO3 electrodes. One of them was loading Pt on the electrodes by drop casting
an H2PtCl6 alcoholic solution with a subsequent heat treatment at 300 ºC. In N 2-purged
solutions, the electrodes exhibited an increase of the photocurrent after Pt deposition
(Figure 5.14) that was more significant at potentials near the onset. In any case, the
improvement of the photocurrent is very modest compared to that observed with other
materials when using Pt as a cocatalyst. The same Pt loading procedure was applied to
LaFe0.95Mg0.05O3 electrodes, leading to similar conclusions.
0.002
0.000
-2
j / mA·cm
-0.002
-0.004
-0.006 FTO/LaFeO3
FTO/LaFeO3/Pt
-0.008
-0.4 -0.2 0.0 0.2 0.4 0.6
E (vs Ag/AgCl) / V
Figure 5.14. Linear sweep voltammograms under 1-sun transient illumination for LaFeO3
electrodes in N2-purged 0.1 M NaOH electrolyte before and after Pt deposition, at a scan rate of
5 mV s−1.
Other implemented strategy consists in dipping the electrode during a certain time in a 1
M HClO4 solution (etching), which leads to small improvements of the photocurrents
and a notable shift of the onset of O2 reduction in the dark toward positive potentials.
These effects can be attributed to the removal of the La2O3 layer existing on the
electrode surface (see XPS analysis) together with an increase of the roughness of the
electrode because of the partial dissolution of the material. In fact, the cyclic
voltammograms after the etching showed larger accumulation regions.
239
Chapter 5
5.4. Conclusions
Sol–gel synthesized perovskite LaFeO3 electrodes can be used as stable photocathodes
for oxygen (and water) reduction. Their efficiency as photocathodes for oxygen
reduction can be improved drastically by doping with relatively large amounts of Mg2+
and Zn2+ (5% with respect to Fe atoms). This effect is attributed primarily to an increase
in the majority carrier concentration that results from a dopant-induced increase of the
Fe4+ density caused by charge compensation upon the substitution of Fe3+ by Mg2+ and
Zn2+. In addition, the distortions induced by the introduction of the dopants likely
increase hole mobility. Moreover, a contribution from Zn- or Mg-induced passivation
cannot be excluded as the X-ray photoelectron spectra point to their accumulation at the
surface of the material. The fact that similar photocurrents are observed with EE and SE
illumination (300-nm thickness electrodes suggests high diffusion lengths of the holes.
In principle, these electrodes have the potential to be applied as photocathodes in
tandem devices because of their very positive onset potential values and their relatively
narrow band gap. In addition, LaFeO3 is an attractive candidate because of its stability,
environmental friendliness, and the relatively high Earth abundance of the constituent
elements. Admittedly, the kinetics of the transfer of photogenerated electrons to water
needs to be enhanced as the electrons tend to get trapped at surface states thus favoring
recombination. The combination of extracting and/or passivating layers combined with
a cocatalyst can be a future route to improve both hole collection in the back contact and
electron transfer to the electrolyte for HER.
5.5. Bibliography
(1) Celorrio, V.; Bradley, K.; Weber, O. J.; Hall, S. R.; Fermín, D. J. Photoelectrochemical Properties
of LaFeO3 Nanoparticles. ChemElectroChem 2014, 1, 1667–1671.
(2) Yu, Q.; Meng, X.; Wang, T.; Li, P.; Liu, L.; Chang, K.; Liu, G.; Ye, J. Highly Durable p-
LaFeO3/n-Fe2O3 Photocell for Effective Water Splitting under Visible Light. Chem. Commun.
2015, 51, 3630–3633.
240
LaFeO3 photocathodes: metal doping as way to enhance the photoelectrochemical behavior
(3) May, K. J.; Fenning, D. P.; Ming, T.; Hong, W. T.; Lee, D.; Stoerzinger, K. A.; Biegalski, M. D.;
Kolpak, A. M.; Shao-Horn, Y. Thickness-dependent photoelectrochemical water splitting on
ultrathin LaFeO3 films grown on Nb:SrTiO3. J. Phys. Chem. Lett. 2015, 6, 977–985.
(4) Peng, Q.; Wang, J.; Wen, Y. W.; Shan, B.; Chen, R. Surface modification of LaFeO3 by Co-Pi
electrochemical deposition as an efficient photoanode under visible light. RSC Adv. 2016, 6,
26192–26198.
(5) Peng, Q.; Shan, B.; Wen, Y.; Chen, R. Enhanced charge transport of LaFeO 3 via transition metal
(Mn, Co, Cu) doping for visible light photoelectrochemical water oxidation. Int. J. Hydrogen
Energy 2015, 40, 15423–15431.
(6) Peng, Q.; Wen, Y.; Shan, B.; Chen, R. Enhanced Oxygen Evolution Reaction Activity of LaFeO 3
Coated with Co-Pi. ECS Trans. 2015, 64, 27–35.
(7) Perkins, J. D.; Paudel, T. R.; Zakutayev, A.; Ndione, P. F.; Parilla, P. A.; Young, D. L.; Lany, S.;
Ginley, D. S.; Zunger, A.; Perry, N. H.; Tang, Y.; Grayson, M.; Mason, T. O.; Bettinger, J. S.;
Shi, Y.; Toney, M. F. Inverse design approach to hole doping in ternary oxides: Enhancing p-type
conductivity in cobalt oxide spinels. Phys. Rev. B 2011, 84, 205207.
(8) Porta, P.; Cimino, S.; De Rossi, S.; Faticanti, M.; Minelli, G.; Pettiti, I. AFeO 3 (A = La, Nd, Sm)
and LaFe1−xMgxO3 perovskites: Structural and redox properties. Mater. Chem. Phys. 2001, 71,
165–173.
(9) Ciambelli, P.; Cimino, S.; De Rossi, S.; Lisi, L.; Minelli, G.; Porta, P.; Russo, G. AFeO 3 (A = La,
Nd, Sm) and LaFe1−xMgxO3 perovskites as methane combustion and CO oxidation catalysts:
Structural, redox and catalytic properties. Appl. Catal. B Environ. 2001, 29, 239–250.
(10) Bhat, I.; Husain, S.; Khan, W.; Patil, S. I. Effect of Zn doping on structural, magnetic and
dielectric properties of LaFeO3 synthesized through sol-gel auto-combustion process. Mater. Res.
Bull. 2013, 48, 4506–4512.
(11) Huang, S.; Qin, H.; Song, P.; Liu, X.; Li, L.; Zhang, R.; Hu, J.; Yan, H.; Jiang, M. The
formaldehyde sensitivity of LaFe1−xZnxO3-based gas sensor. J. Mater. Sci. 2007, 42, 9973–9977.
(12) Liu, X.; Cheng, B.; Hu, J.; Qin, H.; Jiang, M. Preparation, structure, resistance and methane-gas
sensing properties of nominal La1−xMgxFeO3. Sensor. Actuat. B-Chem. 2008, 133, 340–344.
(13) Liu, X.; Cheng, B.; Hu, J.; Qin, H.; Jiang, M. Semiconducting gas sensor for ethanol based on
LaMgxFe1−xO3 nanocrystals. Sensor. Actuat. B-Chem. 2008, 129, 53–58.
(14) Dong, S.; Xu, K.; Tian, G. Photocatalytic activities of LaFe 1−xZnxO3 nanocrystals prepared by sol-
gel auto-combustion method. J. Mater. Sci. 2009, 44, 2548–2552.
(15) Mickevičius, S.; Grebinskij, S.; Bondarenka, V.; Vengalis, B.; Šliužiene, K.; Orlowski, B. A.;
Osinniy, V.; Drube, W. Investigation of epitaxial LaNiO 3−x thin films by high-energy XPS. J.
Alloys Compd. 2006, 423, 107–111.
241
Chapter 5
(16) Cho, Y.-G.; Choi, K.-H.; Kim, Y.-R.; Jung, J.-S.; Lee, S.-H. Characterization and Catalytic
Properties of Surface La-rich LaFeO3 Perovskite. Bull. Korean Chem. Soc 2009, 30, 1368–1372.
(17) Dai, Z.; Lee, C.-S.; Kim, B.-Y.; Kwak, C.-H.; Yoon, J.-W.; Jeong, H.-M.; Lee, J.-H. Honeycomb-
Like Periodic Porous LaFeO3 Thin Film Chemiresistors Performances with Enhanced Gas
Sensing Performances. ACS Appl. Mater. Interfaces 2014, 6, 16217–16226.
(18) Yamashita, T.; Hayes, P. Analysis of XPS spectra of Fe2+ and Fe3+ ions in oxide materials. Appl.
Surf. Sci. 2008, 254, 2441–2449.
(19) De Asha, A. M.; Critchley, J. T. S.; Nix, R. M. Molecular adsorption characteristics of lanthanum
oxide surfaces: the interaction of water with oxide overlayers grown on Cu(111). Surf. Sci. 1998,
405, 201–214.
(20) Barbero, B. P.; Gamboa, J. A.; Cadús, L. E. Synthesis and characterisation of La 1−xCaxFeO3
perovskite-type oxide catalysts for total oxidation of volatile organic compounds. Appl. Catal. B
Environ. 2006, 65, 21–30.
(21) Sora, I. N.; Fontana, F.; Passalacqua, R.; Ampelli, C.; Perathoner, S.; Centi, G.; Parrino, F.;
Palmisano, L. Photoelectrochemical properties of doped lanthanum orthoferrites. Electrochim.
Acta 2013, 109, 710–715.
(22) Li, L.; Wang, X.; Zhang, Y. Enhanced visible light-responsive photocatalytic activity of LnFeO3
(Ln=La ,Sm) nanoparticles by synergistic catalysis. Mater. Res. Bull. 2014, 50, 18–22.
(23) Scafetta, M. D.; Xie, Y. J.; Torres, M.; Spanier, J. E.; May, S. J. Optical absorption in epitaxial
La1−xSrxFeO3 thin films. Appl. Phys. Lett. 2013, 102, 081904.
(24) Salvador, P. Kinetic Approach To the Photocurrent Transients in Water Photoelectrolysis at n-
TiO2 Electrodes. 1. Analysis of the Ratio of the Instantaneous to Steady-State Photocurrent. J.
Phys. Chem. 1985, 89, 3863–3869.
(25) Tijare, S. N.; Joshi, M. V.; Padole, P. S.; Mangrulkar, P. A.; Rayalu, S. S.; Labhsetwar, N. K.
Photocatalytic hydrogen generation through water splitting on nano-crystalline LaFeO3
perovskite. Int. J. Hydrogen Energy 2012, 37, 10451–10456.
(26) Díaz-García, A. K.; Lana-Villarreal, T.; Gómez, R. Sol–gel copper chromium delafossite thin
films as stable oxide photocathodes for water splitting. J. Mater. Chem. A 2015, 3, 19683–19687.
(27) Yang, Z.; Huang, Z.; Ye, L.; Xie, X. Influence of parameters U and J in the LSDA+U method on
electronic structure of the perovskite LaMO3 (M=Cr,Mn,Fe,Co,Ni). Phys. Rev. B 1999, 60,
15674–15682.
(28) Kleibeuker, J. E.; Zhong, Z.; Nishikawa, H.; Gabel, J.; Müller, A.; Pfaff, F.; Sing, M.; Held, K.;
Claessen, R.; Koster, G.; Rijnders, G. Electronic reconstruction at the isopolar LaTiO 3/LaFeO3
interface: An X-ray photoemission and density-functional theory study. Phys. Rev. Lett. 2014,
113, 237402.
242
LaFeO3 photocathodes: metal doping as way to enhance the photoelectrochemical behavior
(29) Idrees, M.; Nadeem, M.; Atif, M.; Siddique, M.; Mehmood, M.; Hassan, M. M. Origin of colossal
dielectric response in LaFeO3. Acta Mater. 2011, 59, 1338–1345.
(30) Cardon, F.; Gomes, W. P. On the determination of the flat-band potential of a semiconductor in
contact with a metal or an electrolyte from the Mott-Schottky plot. J. Phys. D. Appl. Phys. 1978,
11, L63–L67.
(31) Cots, A.; Cibrev, D.; Bonete, P.; Gómez, R. Hematite Nanorod Electrodes Modified with
Molybdenum: Photoelectrochemical Studies. ChemElectroChem 2017, 4, 585–593.
(32) Leng, W. H.; Zhang, Z.; Zhang, J. Q.; Cao, C. N. Investigation of the Kinetics of a TiO 2
Photoelectrocatalytic Reaction Involving Charge Transfer and Recombination through Surface
States by Electrochemical Impedance Spectroscopy. J. Phys. Chem. B 2005, 109, 15008–15023.
(33) Newhouse, P. F.; Park, C.-H.; Keszler, A.; Tate, J.; Nyholm, P. S. High electron mobility W-
doped In2O3 thin films by pulsed laser deposition. Appl. Phys. Lett. 2005, 87, 112108.
(34) Yamada, N.; Tatejima, T.; Ishizaki, H.; Nakada, T. Effects of postdeposition annealing on
electrical properties of Mo-doped indium oxide (IMO) thin films deposited by RF magnetron
cosputtering. Jpn. J. Appl. Phys. 2006, 45, L1179–L1182.
(35) Abe, Y.; Ishiyama, N. Titanium-doped indium oxide films prepared by d.c. magnetron sputtering
using ceramic target. J. Mater. Sci. 2006, 41, 7580–7584.
(36) Okuda, T.; Nakanishi, K.; Miyasaka, S.; Tokura, Y. Large thermoelectric response of metallic
perovskites: Sr1−xLaxTiO3 (0≤x≤0.1). Phys. Rev. B 2001, 63, 113104.
243
Chapter 6
Photoelectrochemistry of compact and
nanostructured YFeO3 photocathodes
6.1. Introduction ........................................................................................................... 247
6.2. Experimental section ............................................................................................. 248
6.2.1. Synthesis of doped and undoped nanoparticulate thin film
electrodes (NP) .............................................................................................................. 248
6.2.2. Synthesis of compact thin film electrode (C) synthesis ............................... 248
6.2.3. Instrumentation .............................................................................................. 249
6.3. Photoelectrochemical response of YFeO 3 photoelectrodes ................................ 249
6.3.1. Comparison between the two electrode fabrication routes ........................ 249
6.3.2. Effect of Mg-doping in NP-film electrodes .................................................. 262
6.4. Conclusions............................................................................................................. 266
6.5. Bibliography ........................................................................................................... 267
Photoelectrochemistry of compact and nanostructured YFeO3 photocathodes
6.1. Introduction
As stated in the above chapters, photoelectrochemical (PEC) water splitting can
potentially offer a viable and scalable approach for solar energy conversion into fuels.
Tandem devices featuring dimensionally stable photoanodes and photocathodes
represent the key challenge in this field. Therefore, it is crucial to extend current
research activities beyond conventional oxide photoanodes (such as TiO2, ZnO, Fe2O3,
BiVO4 and WO3)1–4 into materials capable of generating hydrogen under illumination.
Ferrite perovskites constitute a particularly interesting class of materials featuring band
gap values in the range of 2.3–2.4 eV. As reviewed in Chapter 1, Celorrio et al.
investigated the photoelectrochemical properties of phase-pure LaFeO3 nanoparticles
sintered on FTO electrodes,5 while compact thin films have been fabricated by pulsed
laser deposition.6 Photoelectrochemical properties critically depend on the electrode
configuration and, consequently on the synthetic route.
Doping with metals has been extensively implemented to enhance photocurrents in
photocathodes (see Chapter 1, section 1.7.3). In Chapter 5, doping with Mg has been
demonstrated to be a successful strategy to increase the material conductivity and
consequently, increase the photocurrents ascribed to oxygen reduction. Regarding
YFeO3, some reports are dealing with Ca and Mn doping to improve the magnetic,
dielectric and electrical properties of YFeO3.7–10
In this work, the behavior of YFeO3 thin-film electrodes under illumination is
investigated for the first time. We describe the synthesis and properties of YFeO3 as a
photocathode for hydrogen evolution reaction under alkaline conditions. Despite several
reports about YFeO3 as a photocatalyst for water remediation11–14 and photochemical H2
production,15 very little is known about key properties such as band edge energy
positions and charge transport properties. On the other hand, we have investigated two
different synthesis routes, leading to sintered nanostructured (NP-Films) and compact
films (C-Films) supported on FTO. The rationale for investigating these two different
morphologies is to assess the effect of grain boundaries and material disorder on the
dynamics of minority charge carrier transfer, hole collection and carrier recombination.
On the other hand, the introduction of Mg2+ as a dopant in the nanostructured films (NP)
in a stoichiometry YFe0.95Mg0.05O3 will be investigated. The materials are characterized
247
Chapter 6
by X-ray diffraction, X-ray photoelectron spectroscopy, diffuse reflectance and electron
microscopy.
6.2. Experimental section
6.2.1. Synthesis of doped and undoped nanoparticulate thin film electrodes (NP)
YFeO3 nanoparticles were synthesized following the ionic liquid protocol.5 1 mL of an
aqueous solution of 0.05 M of Y(NO3)3·6H2O and 0.05 M of Fe(NO3)3·9H2O was added
to a vial containing 1 mL of 1-ethyl-3-methylimidazolium acetate (97%, Sigma
Aldrich). The solution was dehydrated at 80 °C for 3 h and 100 mg of cellulose was
added before calcination at 1000 °C for 2 h. This method yields phase-pure YFeO3
nanoparticles. Approximately 1 μm-thick films were fabricated via the doctor blade
method over FTO (F:SnO2) conductive glass. First, 100 μg of the YFeO3 powder was
suspended in 100 μL of water and sonicated for 15 min in an ultrasonic bath.
Acetylacetone (Sigma Aldrich) and Triton 100X (Fisher Scientific) were added in order
to obtain a homogeneous and viscous paste. Finally, the paste was spread over the FTO
and sintered at 400 °C for 1 h in air.
An identical procedure was followed for preparing Mg-doped YFeO3 films but the
precursor concentrations were modified in order to obtain the theoretical stoichiometry
of YFe0.95Mg0.05O3. Mg(NO3)2·6H2O was employed as a magnesium source.
6.2.2. Synthesis of compact thin film electrode (C) synthesis
The synthesis is based on a sol-gel method using citric acid as a chelating agent.
Y(NO3)3·6H2O (0.3 M) and Fe(NO3)3·9H2O (0.3 M) were dissolved in water and the
solution was stirred for 1 h. Citric acid monohydrate was added to a concentration of 0.6
M and stirred for 20 h. Then, 30 μL of acetylacetone per mL and 30 μL of Triton 100X
per mL were added. This solution was spin coated over the FTO substrate at 3000 rpm
for 20 s and calcined at 400 °C for 1 h. This procedure was repeated twice (2-layers),
followed by heating at 640 °C for 2 h, leading to approximately 80 nm-thick films. The
248
Photoelectrochemistry of compact and nanostructured YFeO3 photocathodes
same heating procedure was applied in a crucible to generate powder samples for XRD
analysis.
6.2.3. Instrumentation
X-ray diffraction (XRD) was recorded using a Bruker AXS D8 Advance diffractometer
with a θ–θ configuration, using the Cu Kα radiation (λ = 0.154 nm). Transmission
electron microscopy (TEM) and high resolution TEM were carried out on a JEOL JEM-
1400-Plus and a JEOL JEM-2010 microscopes, respectively. Field emission scanning
electron microscopy (FE-SEM) images were obtained by a ZEISS Merlin VP Compact
microscope. Energy-dispersive X-ray (EDX) analysis was performed with a SEM
instrument JEOL SEM 5600 LV. A Shimadzu UV-2401PC spectrophotometer equipped
with an integrating sphere coated with BaSO4 was used to measure UV-visible diffuse
reflectance spectra. Core level photoemission spectra was collected in normal emission
at room temperature with a K-Alpha Thermo-Scientific X-ray Photoelectron
Spectrometer (XPS) using an Al Kα X-ray source. Electrochemical measurements were
performed in a three-electrode cell equipped with a fused silica window using a
computer-controlled Ivium CompactStat potentiostat. A Ag/AgCl/KCl(sat) electrode
was used as a reference, while a platinum wire was used as a counter electrode. The
electrolyte solution used in all the experiments was 0.1 M NaOH purged with high
purity Ar. Measurements under illumination were carried out using a LED with a
narrow emission centered at 404 nm LED (Thorlabs), driven by a waveform generator
(Stanford Research Systems). Photon flux was measured employing a calibrated silicon
photodiode (Newport Corporation).
6.3. Photoelectrochemical response of YFeO3 photoelectrodes
6.3.1. Comparison between the two electrode fabrication routes
Figure 6.1A contrasts the XRD patterns of YFeO3 obtained from ionic liquid and sol-gel
routes. Full profile Rietveld refinement on XRD pattern of the sample resulting from the
249
Chapter 6
ionic liquid route is shown in Figure 6.1A, confirming the formation of YFeO 3 in
orthorhombic phase (Pnma) with lattice parameters a = 5.5936 Å, b = 7.6023 Å and c =
5.2796 Å. The unit cell (inset Figure 6.1A) is composed of Fe3+ centered octahedra,
with oxygen atoms occupying non-symmetric axial and equatorial positions. The Fe
atom is found to be off-centered leading to two different bond lengths between Fe and
equatorial oxygen (Oeq) atoms: 196.35 and 203.29 pm, while Fe to axial O (Oax) bond
lengths have the same value equal to 198.89 pm. Similarly, the bond angles formed
between Oeq and Fe and the Oax-Fe-Oeq angles deviate from the ideal 90° by up to 2.72°
and 11.62°, respectively. Overall bulk composition is slightly metal deficient,
promoting p-type conductivity. No other peaks due to secondary phases comprising
tetrahedrally coordinated Fe, such as in the Y3Fe5O12 garnet frequently formed during
the synthesis of YFeO3, are observed within the measurement limit.16,17
Figure 6.1. Powder XRD patterns of YFeO3 particles obtained by the ionic liquid method at
1000 °C and sol-gel route at 640 °C.
The C-film shows no clear diffraction peaks, indicating a lack of crystallinity (Figure
6.1, bottom). The key difference in the synthesis method is the temperature used for
promoting the orthorhombic YFeO3 phase. Zhang et al.13 reported the crystallization of
YFeO3 powder from 700 °C using a sol-gel route with citric acid as chelating agent,
which is slightly above the temperature limit set by the stability of the substrate
(approximately 640 °C). The crystallization temperature reflects the enthalpy of
formation of the perovskite phase which is influenced by the nature of the A and B
250
Photoelectrochemistry of compact and nanostructured YFeO3 photocathodes
sites.18,19 For instance, phase pure LaFeO3 synthesized by the same ionic-liquid
approach can be achieved at 900 °C, while LaMnO3 can be obtained at 700 °C.20
A. 80 B.
4
Reflectance / %
2
60
(F(R) h) / eV
2
40
2
20
Eg= 2.45
0 0
400 500 600 700 800 2.0 2.2 2.4 2.6 2.8 3.0
/ nm h/ eV
Figure 6.2. (A) UV-visible diffuse reflectance spectra and (B) Tauc plot (direct transition)
corresponding to NP (red line) and C (blue line) films (10 layers by spin coating were deposited
for the C-film). A band gap of 2.45 eV was estimated for the NP film.
The diffuse reflectance spectra in Figure 6.2A show absorption edges at 600 and 650 nm
for the NP and C films, respectively. Figure 6.2B displays Tauc plots corresponding to
the NP and C-films obtained by operating the Kubelka-Munk function (F(R)) on the
reflectance spectra. The Tauc plot representation shows a direct band gap transition for
the NP-film at 2.45 eV, which is in close agreement with previous studies in the
literature.12,21 On the other hand, C-films do not show a clearly defined linear region,
which is consistent with the amorphous nature of the material. This behavior can be
explained in terms of potential fluctuations of the band edge energies, which leads to the
so-called Urbach tails.22,23 We shall come back to this point further below.
SEM images in Figure 6.3A and B contrast the smooth nature of the C-films against the
nanoscale corrugated NP-films. TEM images in Figure 6.3C show that the particles
obtained after calcination at 1000 °C exhibit a size distribution between 100 and 200
nm. The lattice fringes displayed in the high-resolution TEM image (Figure 6.3D)
correspond to a d-spacing of 0.27 nm associated with the {121} planes, which is
consistent with the most prominent feature in XRD (Figure 6.1A). The structural
251
Chapter 6
features of the TEM images further demonstrate the high crystallinity and phase purity
of the particles obtained by the ionic-liquid method.
Figure 6.3. FE-SEM images for (A) C-film (B) the NP-film. (C) TEM micrograph of the
nanoparticles obtained by the ionic-liquid method at 1000 °C. Inset: Particle size distribution
determined from TEM images (over 100 nanoparticles were counted). (C) HRTEM image
highlighting the lattice fringes associated with the {121} plane of the YFeO 3 orthorhombic
phase.
Both films are characterized by homogeneous Y and Fe distributions as shown in Figure
6.4. EDX mapping does not show regions in which a single element is segregated, i.e.
the Y/Fe ratio is constant throughout the sample. This experimental evidence is
particularly important in the case of the C-films, in which XRD and reflectance data do
not provide fully conclusive evidence on the formation of YFeO3.
252
Photoelectrochemistry of compact and nanostructured YFeO3 photocathodes
A.
100 mm
Y La1 Fe Ka1
100 mm 100 mm
B.
500 mm
Y La1 Fe Ka1
500 mm 500 mm
Figure 6.4. SEM micrographs (upper images) and the corresponding elemental mapping of Y
and Fe for NP (A) and C (B) thin films.
High resolution XPS spectra of both films in the regions of Y 3d, O 1s and Fe 2p are
shown in Figure 6.5. The Y 3d spectra contain contributions from Y 3d3/2 and Y 3d5/2
that can be deconvoluted in two further components. The main Y 3d5/2 component at
156.5 eV (labelled as 1) is related to the formation of Y2O324,25 as a consequence of the
Y surface segregation, whereas the component at 157.6 eV (peak 2) corresponds to Y3+
in the perovskite lattice.26 The atomic percentage of Y as Y2O3 is 66.5% and 80% in the
C and NP-film, as displayed in Table 6.1. This suggests a higher extent of A-site
segregation at the surface of the NP-film. Four different components were considered in
the deconvolution of the O 1s line (Table 6.1). The lower binding energy component at
529.1 eV is assigned to oxygen in the perovskite lattice. The second component at 530.6
eV can be attributed to hydroxyl groups, whereas the third component at 531.8 eV is
linked to carbonyl groups. The component with the highest binding energy is associated
with adsorbed molecular water. For the NP-YFeO3 sample, a strong contribution of
253
Chapter 6
SnO2 from the FTO substrate was observed. In this case, the component 2 of the O 1s
line also includes the contribution from SnO2.
Figure 6.5. XPS spectra of the Y 3d, Fe 2p, O 1s orbitals for the NP and C-films. The Y 3d and
O 1s spectra were deconvoluted, while interference from the Sn 3p 3/2 line in the Fe 2p region
prevented a fully quantitative analysis. Both films exhibited a surface Y:Fe ratio of 70:30.
254
Photoelectrochemistry of compact and nanostructured YFeO3 photocathodes
Table 6.1. Analysis of the single chemical components of the O 1s region. For each single
chemical component, the binding energy (eV) and atomic percentage are given. Contributions to
the O 1s spectra were assigned to: (1) oxygen in the lattice, (2) hydroxyl species OH and SnO2,
(3) carbonyl groups and (4) adsorbed molecular water.
Peak
Sample
1 2 3 4
529.2 eV 530.7 eV 531.8 eV 533.3 eV
C-film
55.7 % 25.3 % 16.9 % 2.1 %
529.1 eV 530.6 eV 531.8 eV 533.0 eV
NP-film
34.5 % 50.5 % 13.6 % 1.4 %
The Fe 2p region of the photoemission spectrum (Figure 6.5) provides further insights
into the surface composition of the films. The Fe 2p3/2 and Fe 2p1/2 peaks are located at
710.2 eV and 724.4 eV, respectively, which is consistent with the presence of Fe3+. The
satellite peak located at 718.4 eV further confirms this oxidation state. In the case of the
NP-YFeO3 sample, a new component at 715.9 eV is observed which can be linked to
the Sn 3p from the FTO. Based on the contributions of the Fe 2p and Y 3d regions, the
Y:Fe surface ratio can be estimated taking into account the corresponding sensitivity
factors. A Y:Fe ratio of 70:30 is found for the C and NP-films, confirming a Y-surface
enrichment in accordance with a previous report on perovskite oxides.27 However, it is
interesting that the analysis of the Y 3d region shows a larger content of the binary
(Y2O3) oxide at the surface in the case of the NP-films, which are synthesized at
significantly higher temperatures in comparison to C-films. This analysis suggests that,
although XRD shows a remarkable degree of phase purity, the nanoparticle surface
appears to contain a complex mixture of secondary oxide phases.
255
Chapter 6
2 NP-film
-2
0.1
j / mA cm
0.0
1
-0.1
-0.2
-0.4 -0.2 0.0
0
-1
4 C-film
0.0
-2
j / mA cm 2 -0.2
-0.4
-0.4 -0.2 0.0
0
-2
-0.4 -0.2 0.0 0.2 0.4 0.6
E / V (vs. Ag/AgCl)
Figure 6.6. Cyclic voltammetry at 5 mV s−1 under transient illumination (404 nm and photon
flux of 1.0·1016 cm−2 s−1) for NP and C-films. Illumination was performed from the back side
(SE).
Figure 6.6 displays the photoelectrochemical responses of the NP- and C-YFeO3
electrodes under square-wave light perturbation at 5 mV s−1. The voltammograms are
characterized by a potential independent capacitive response at negative potentials and a
large increase of the current toward more positive values. The NP-films show a
substantial increase of the current at potentials above 0.4 V vs Ag/AgCl (1.35 V vs
RHE), which is associated with hole accumulation at the interface. This behavior
suggests that the potential associated with the valence band edge is located in this
potential range. On the other hand, carrier accumulation can already be seen from 0.1 V
in the case of C-films. This is also a manifestation of band tails arising from the
amorphous nature of the material. Consequently, there is an effective density of states
spreading from the valence band edge into the gap.
256
Photoelectrochemistry of compact and nanostructured YFeO3 photocathodes
2 Imaginary
-2
0
j / nA cm
-2
Real
-4
NP-film
10
Imaginary
-2
j / nA cm 0
-10
Real
-20
C-film
-30
-0.6 -0.4 -0.2 0.0 0.2
E / V (vs. Ag/AgCl)
Figure 6.7. Real and imaginary components of the photocurrent responses as a function of the
potential measured under square-wave photon flux perturbation (2 Hz) and lock-in detection for
the NP and C-films.
In Figure 6.7, lock-in detection experiments for NP- and C-films display a photocurrent
onset potential located close to 0.1 V vs Ag/AgCl (1.05 V vs RHE), which is in
accordance with that observed in Figure 6.6. This onset potential is close to values
reported for other ternary iron oxides, namely LaFeO3 (see Chapter 5),5,6 CaFe2O4 (see
Chapter 4)28 or CuFeO2.29 Interestingly, the intensity and potential dependence of the
photocurrent are similar for both films despite the large difference in crystallinity.
257
Chapter 6
4
NP-film
2
-2
j / mA cm
Ar-purged
0
-2 O2-purged
4 0.4
0.0 Ar-purged
-2
j / mA cm
-2
-0.4
j / mA cm
2 -0.8 O2-purged
-0.4 -0.2 0.0
E / V (vs. Ag/AgCl)
-2 C-film
-0.4 -0.2 0.0 0.2 0.4 0.6
E / V (vs. Ag/AgCl)
Figure 6.8. Cyclic voltammograms at 5 mV s1 under chopped illumination at 404 nm and a
photon flux of 1.0 · 1016 cm2 s1 for NP and C-films in 0.1 M NaOH Ar-purged (black lines)
and O2-purged (red lines) electrolytes. (Inset) magnification of the response in the main
figure.
Cyclic voltammograms under transient illumination was also acquired in the presence of
O2, as shown in Figure 6.8. Oxygen is an electron scavenger, and one would expect that
the photocurrent increases when it is present. Surprisingly, no substantial changes in the
photocurrent were observed with respect to the Ar-purged solutions in both NP and C-
films, which could be linked to a poor charge transport across the film. The NP-film
showed larger dark currents at negative potentials. To get insights into the origin of this
dark current, a voltammetric analysis was performed with a bare FTO electrode, as
shown in Figure 6.9. In the presence of O2, a dark current appears below 0.25 V for
both the bare FTO and the NP-film electrode. Then, this current can be associated with
pin-holes in the NP-film, allowing oxygen reduction to take place at the FTO electrode.
258
Photoelectrochemistry of compact and nanostructured YFeO3 photocathodes
FTO
-2
Ar-purged
j / mA cm
0
-4 O2-purged
-8
-0.4 -0.2 0.0 0.2 0.4 0.6
E (vs. Ag/AgCl) / V
Figure 6.9. Cyclic voltammograms in the dark of the NP-film in 0.1 M NaOH either Ar-purged
(black) or O2-purged (red) electrolyte. The voltammogram of the FTO substrate under O 2
purged electrolyte (blue) is also shown for the sake of comparison.
Figure 6.10 displays the photocurrent transients at different potentials for the NP- and
C-YFeO3 electrodes under back illumination. The overall photocurrent increases toward
more negative potentials. NP-films exhibit a photocurrent decay after the initial
response toward a photostationary current value, followed by a sharp decay to zero upon
switching off the light. A similar behavior is observed for the C-films, although a rather
interesting feature is observed at 0.1 V. This transient exhibits a positive instantaneous
photocurrent, a negative photostationary value and a negative photocurrent overshoot
upon switching off illumination. Again, this response coincides with the onset of
electron depopulation of tail states from the valence band observed in the voltammetric
measurements (see Figure 6.6). In contrast to the responses at more negative potentials,
the transient at 0.1 V in the case of the C-film shows evidence of surface
recombination.30,31 The decay observed on the second timescale can be described in
terms of a redistribution of the potential drop across the film/electrolyte interface,
leading to a time dependent recombination. This type of phenomena is observed in
materials with low carrier mobility.32
259
Chapter 6
0.00
0.1 V
-2
j / mA cm
-0.02
-0.5 V
-0.04
NP-film
-0.06
-2 0.00
0.1 V
j / mA cm
-0.02
-0.04 -0.5 V
-0.06 C-film
10 20 30 40 50
t/s
Figure 6.10. Photocurrent transient responses under illumination at 404 nm and photon flux of
1.0·1016 cm−2 s−1 at different potentials for the NP and C-films. Electrolyte solution is Ar-
saturated 0.1 M NaOH. Illumination was performed from the back side (SE).
Figure 6.11 shows the photostationary current after 20 s (j SS) as a function of the photon
flux under illumination through either the back contact (back) or the semiconductor-
electrolyte boundary (front). C-films show very similar photocurrent responses under
back and front illumination, thus only the former is shown. The photocurrent increases
with photon flux in a rather non-linear relationship. This is a clear indication of
substantial carrier recombination, which also manifests itself by rather low IPCE values
(below 0.01%). Despite substantial differences in film thickness, surface roughness and
crystallinity, the photocurrent magnitude and photon-flux dependence is comparable for
NP and C-films. This behavior provides a clear indication that only carriers generated
close to the back contact (FTO) are effectively collected at the back contact. This is
supported by the fact that photocurrent responses are higher under back illumination in
the case of NP-films, while not such a contrast is observed in the case of C-films due to
the significantly lower film thickness (80 nm).
260
Photoelectrochemistry of compact and nanostructured YFeO3 photocathodes
0.020
C-film (back)
NP-film (back)
0.015
-2
-jph / mA cm
0.010 NP-film (front)
0.005
0.000
0 2 4 6 8 10 12
-15 -2 -1
10 I0 / cm s
Figure 6.11. Photostationary current (jSS) obtained for NP and C-films as a function of the
photon flux. Measurements carried out under back and front illumination are shown in the case
of NP-films. Only back illumination is shown for the C-film, as little difference was observed
with respect to front illumination. The experiments were performed under potentiostatic
conditions at −0.5 V vs Ag/AgCl.
The low collection efficiency of carriers can be associated with two key factors, carrier
recombination and hole-capture by water. As mentioned previously, the potential
associated with the valence band edge is located at approximately 1.35 V vs RHE,
which raises the possibility that the poor hole collection efficiency is due to hole capture
by water to generate adsorbed hydroxyl radicals. This carrier loss mechanism appears
consistent with the photocatalytic performance reported for YFeO3 and the very low
rates of generation of oxidized species.11–15 On the other hand, recombination losses can
be linked to A-site surface segregation as shown by XPS measurements, leading to an
insulating yttrium oxide layer irrespective of the calcination temperature. In the case of
NP-films, this insulating layer will not only affect electron transfer to evolve hydrogen,
but also interparticle carrier transport in the film. In the case of the C-film, Y-
segregation may have a weaker effect on charge transport in the film. However, the
presence of tail states due to the amorphous nature of the oxide will have a strong effect
on carrier mobility. The fact that the photocurrent onset potential is close to the valence
band edge reveals the suitability of YFeO3 as a photocathode for hydrogen generation. It
should be mentioned that IPCE values for hydrogen evolution under alkaline conditions
261
Chapter 6
are typically lower than in acid solutions,33 consequently, improvement in
photoresponses will require optimized surface passivation and deposition of co-catalysts
as recently reported for a variety of semiconductors.34–37
6.3.2. Effect of Mg-doping in NP-film electrodes
In this section, the introduction of divalent Mg as a dopant in the NP-films is
investigated. Doped and undoped nanostructured films were fabricated by the ionic
liquid synthesis route described above. XRD of the Mg-doped YFeO3 is shown in
Figure 6.12A. The diffraction peaks are attributed to the orthorhombic phase of the
YFeO3 (Pnma), as in the case of the undoped powder (in Figure 6.1A). Only a small
peak corresponding to Y2O3 phase was detected. Importantly, peaks ascribed to
impurities of Mg oxide or hydroxide are absent, suggesting that Mg is incorporated into
the YFeO3 lattice as a dopant. No apparent changes in the TEM images for the Mg-
doped sample (shown in Figure 6.12B) with respect to the undoped case (Figure 6.3C)
are observed. The particles exhibit similar shape and size in both samples.
A. B.
Intensity / a.u.
*
500 nm
20 30 40 50 60 70 80
2 / degrees
Figure 6.12. (A) Powder XRD patterns (the * peak is ascribed to Y2O3 impurities) and (B) TEM
micrograph of YFe0.95Mg0.05O3 particles obtained by the ionic liquid method at 1000 °C.
262
Photoelectrochemistry of compact and nanostructured YFeO3 photocathodes
250 mm
Y La1 Fe Ka1 Mg Ka1
100 mm 100 mm 100 mm
Figure 6.13. SEM micrographs (upper image) and the corresponding elemental mapping of Y,
Fe and Mg for an YFe0.95Mg0.05O3 NP-film.
EDX mapping analysis in Figure 6.13 shows the Y, Fe and Mg distribution. The Mg-
doped film exhibits a homogeneous distribution of Y and Fe, as in the case of the
undoped NP film. Importantly, Mg also appears evenly distributed across the analyzed
region, not being aggregated in specific regions. This supports the idea that Mg is
incorporated into the YFeO3 lattice.
A. B.
1.0 10 Imaginary
0.5
-2
0
j / mA cm
-2
j / nA cm
0.0
-10
-0.5 Real
-20
-1.0
-0.4 -0.2 0.0 0.2 0.4 0.6 -0.6 -0.4 -0.2 0.0 0.2 0.4
E (vs Ag/AgCl) / V E (vs Ag/AgCl) / V
Figure 6.14. (A) Cyclic voltammogram at 5 mV s−1 under transient SE -illumination (404 nm
and photon flux of 1.0 · 10 16 cm−2 s−1) and (B) real and imaginary components of the
photocurrent responses of an YFe0.95Mg0.05O3 NP-film as a function of the potential measured
under square-wave photon flux perturbation (2 Hz) and lock-in detection.
263
Chapter 6
Cyclic voltammetry under transient illumination in Figure 6.14A reveals cathodic
photocurrents, indicating that the Mg-doped LaFeO3 also presents a p-type character.
The onset of photocurrent is located at about 0.1 V vs Ag/AgCl, and below this value,
the photocurrent magnitude grows as the potential decreases. Figure 6.14B displays
lock-in measurements for the YFe0.95Mg0.05O3 film, revealing an onset potential near 0.1
V vs Ag/AgCl in agreement with that deduced from cyclic voltammetry (Figure 6.14A).
The potential onset of photocurrent does not appreciably change with the introduction of
Mg.
0.05
0.1 V
0.00
-2
j / mA cm
-0.05
-0.10 -0.5 V
-0.15
10 20 30 40 50
t/s
Figure 6.15. Photocurrent transient responses under SE-illumination at 404 nm and photon flux
of 1.0 · 1016 cm−2 s−1 at different potentials for an YFe0.95Mg0.05O3 NP-film. Electrolyte solution
is Ar-saturated 0.1 M NaOH.
The photocurrent transients at different potentials are presented in Figure 6.15. They
show a continuous increase of both the initial photocurrent spike and the
photostationary currents as the applied potential decreases. Even for the highest
potential values (close to the onset potential), the transient presents a cathodic peak just
when light is turned on, and an anodic spike upon light interruption. This can be linked
to a substantial electron trapping in surface states upon illumination, and the subsequent
recombination of the trapping carriers when light is turned off. The photocurrent spikes
also increase with potential. The larger the potential bias, the higher the driving force
for transporting the carriers to the surface.
264
Photoelectrochemistry of compact and nanostructured YFeO3 photocathodes
2
Ar
1 O2
-2
j / mA cm
0
-1
-2
-0.4 -0.2 0.0 0.2 0.4 0.6
E (vs Ag/AgCl) / V
Figure 6.16. Cyclic voltammograms at 5 mV s−1 for an YFe0.95Mg0.05O3 NP-film under chopped
illumination at 404 nm and a photon flux of 1.0 · 1016 cm−2 s−1 in 0.1 M NaOH Ar-purged (black
line) and O2-purged (red line) electrolyte.
Figure 6.16 contrasts the photoelectrochemical response in the presence and the absence
of O2 for an YFe0.95Mg0.05O3 electrode. As opposed to the undoped electrodes, the
photocurrent increases in the presence of O2. This suggests an improved charge
transport induced by Mg doping, in accordance with the higher photocurrents observed
in the absence of oxygen with respect to the NP-YFeO3 films. It is important to mention
that Tauc plots (not shown) from diffuse reflectance values display similar direct band
gap for the doped and undoped films.
It has been reported that Ca2+ doping of YFeO3 dramatically increases the conductivity
up to x = 0.05 in Y1-xCaxFeO3-.9 This effect was ascribed to the formation of holes
localized at Fe3+ ions due to charge compensation upon substituting Y3+ species by Ca2+
species. In Chapter 5, the improvement of the conductivity in Mg- and Zn-doped
LaFeO3 were assigned to the formation of Fe4+ by the introduction of the divalent ions.
In the investigated Mg-doped YFeO3, we assume that Mg is substituting Fe
(analogously to the case of the LaFeO3 perovskite), although Y3+ substitution by Mg2+
cannot be discarded. Unfortunately, Mott-Schottky plots were not conclusive for
determining the charge carrier density, as the measured data were inconsistent, probably
because of (a) the absence of a well-defined accumulation region in the nanostructure or
more probably and (b) a poor contact among particles or of particles with the substrate.
265
Chapter 6
6.4. Conclusions
YFeO3 thin film electrodes were prepared by two different methodologies: (i) sintering
of phase-pure nanoparticles (NP) and (ii) sol-gel leading to compact (C) films. The
former was obtained after calcination of an ionic-liquid based precursor at 1000 °C,
while the temperature of the C-film thermal treatment was limited to 640 °C. The NP-
films were characterized by an orthorhombic perovskite structure with no secondary
phases, while C-films were essentially amorphous. The high crystallinity of the NP
films manifested itself by a sharp band gap transition at 2.45 eV and an onset for charge
accumulation at 0.4 V vs Ag/AgCl (1.35 V vs RHE). On the other hand, the amorphous
C-film exhibited a distribution of states tailing into the band gap region as probed by
spectroscopic and electrochemical analysis. XPS indicates that +3 is the main Fe
oxidation state present in both films. Despite the significant difference in crystallinity,
XPS clearly shows that the surface of both films is Y-rich (70:30 Y:Fe atomic ratio)
with a significant presence of the binary Y oxide at the NP-films surface.
Photoelectrochemical studies in Ar-saturated 0.1 M NaOH solution were characterized
by potential dependent responses with an onset potential close to 0.1 V vs Ag/AgCl
(1.05 V vs RHE) in the absence of any catalyst. The photocurrent responses show
complex dynamic features revealing hindered charge transport. Evidence of surface
recombination was observed in the C-films at potentials in which valence band tail
states are partially depopulated. The photocurrent dependence on photon flux at 404 nm
was highly non-linear, with rather low IPCE values.
To the best of our knowledge, this is the first time that Mg is used as a dopant in YFeO3
electrodes. Mg-doped YFeO3 presents the same orthorhombic phase after the
introduction of Mg with Mg/Fe ratio of 0.05 (without detecting impurities related with
Mg rich phases), which suggests that Mg has been incorporated into the YFeO3 lattice.
In 0.1 M NaOH, the photocurrent increases more than twice in Ar-purged solutions with
respect to the undoped material, while the onset of photocurrent remains almost
unaltered upon doping. Moreover, the presence of O2 produces an increase of the
photocurrent, opposite to the case of the undoped electrodes. These results point to an
increase of the conductivity by the introduction of Mg, inducing an improvement of
charge transport across the electrode, which produces an enhancement of the PEC
performance.
266
Photoelectrochemistry of compact and nanostructured YFeO3 photocathodes
6.5. Bibliography
(1) Cristino, V.; Caramori, S.; Argazzi, R.; Meda, L.; Marra, G. L.; Bignozzi, C. A. Efficient
photoelectrochemical water splitting by anodically grown WO 3 electrodes. Langmuir 2011, 27,
7276–7284.
(2) Jankulovska, M.; Barceló, I.; Lana-Villarreal, T.; Gómez, R. Improving the Photoelectrochemical
Response of TiO2 Nanotubes upon Decoration with Quantum-Sized Anatase Nanowires. J. Phys.
Chem. C 2013, 117, 4024–4031.
(3) Tallarida, M.; Das, C.; Cibrev, D.; Kukli, K.; Tamm, A.; Ritala, M.; Lana-Villarreal, T.; Gómez,
R.; Leskelä, M.; Schmeisser, D. Modification of Hematite Electronic Properties with Trimethyl
Aluminum to Enhance the Efficiency of Photoelectrodes. J. Phys. Chem. Lett. 2014, 5, 3582–
3587.
(4) Quiñonero, J.; Lana–Villarreal, T.; Gómez, R. Improving the photoactivity of bismuth vanadate
thin film photoanodes through doping and surface modification strategies. Appl. Catal. B
Environ. 2016, 194, 141–149.
(5) Celorrio, V.; Bradley, K.; Weber, O. J.; Hall, S. R.; Fermín, D. J. Photoelectrochemical Properties
of LaFeO3 Nanoparticles. ChemElectroChem 2014, 1, 1667–1671.
(6) Yu, Q.; Meng, X.; Wang, T.; Li, P.; Liu, L.; Chang, K.; Liu, G.; Ye, J. Highly Durable p-
LaFeO3/n-Fe2O3 Photocell for Effective Water Splitting under Visible Light. Chem. Commun.
2015, 51, 3630–3633.
(7) Yoo, H.-I.; Kim, C.-S. Electrical properties and defect structure of Y1−xCaxFeO3 orthoferrites.
Solid State Ionics 1992, 53–56, 583–591.
(8) Cao, X.; Kim, C.-S.; Yoo, H.-I. Effect of Substitution of Manganese for Iron on the Structure and
Electrical Properties of Yttrium Ferrite. J. Am. Ceram. Soc. 2001, 84, 1265–1272.
(9) Jacob, K. T.; Rajitha, G.; Dasgupta, N. Electrical conductivity of Ca-doped YFeO3. Indian J. Eng.
Mater. Sci. 2012, 19, 47–53.
(10) Deka, B.; Ravi, S.; Perumal, A.; Pamu, D. Effect of Mn doping on magnetic and dielectric
properties of YFeO3. Ceram. Int. 2017, 43, 1323–1334.
(11) Lü, X.; Xie, J.; Shu, H.; Liu, J.; Yin, C.; Lin, J. Microwave-assisted synthesis of nanocrystalline
YFeO3 and study of its photoactivity. Mater. Sci. Eng. B 2007, 138, 289–292.
(12) Tang, P.; Chen, H.; Cao, F.; Pan, G. Magnetically recoverable and visible-light-driven
nanocrystalline YFeO3 photocatalysts. Catal. Sci. Technol. 2011, 1, 1145–1148.
(13) Zhang, Y.; Yang, J.; Xu, J.; Gao, Q.; Hong, Z. Controllable synthesis of hexagonal and
orthorhombic YFeO3 and their visible-light photocatalytic activities. Mater. Lett. 2012, 81, 1–4.
267
Chapter 6
(14) Chen, Y.; Yang, J.; Wang, X.; Feng, F.; Zhang, Y.; Tang, Y. Synthesis YFeO 3 by salt-assisted
solution combustion method and its photocatalytic activity. J. Ceram. Soc. Japan 2014, 122, 146–
150.
(15) Khraisheh, M.; Khazndar, A.; Al-Ghouti, M. A. Visible light-driven metal-oxide photocatalytic
CO2 conversion. Int. J. Energy Res. 2015, 39, 1142–1152.
(16) Geller, S.; Gilleo, M. A. The crystal structure and ferrimagnetism of yttrium-iron garnet,
Y3Fe2(FeO4)3. J. Phys. Chem. Solids 1957, 3, 30–36.
(17) Cherepanov, V.; Kolokolov, I.; L’vov, V. The saga of YIG: Spectra, thermodynamics, interaction
and relaxation of magnons in a complex magnet. Phys. Rep. 1993, 229, 81–144.
(18) Navrotsky, A. Thermochemistry of complex perovskites. In AIP Conference Proceedings; 2000;
Vol. 535, pp 288–296.
(19) Cheng, J.; Navrotsky, A.; Zhou, X.-D.; Anderson, H. U. Enthalpies of Formation of LaMO 3
Perovskites (M = Cr, Fe, Co, and Ni). J. Mater. Res. 2005, 20, 191–200.
(20) Celorrio, V.; Dann, E.; Calvillo, L.; Morgan, D. J.; Hall, S. R.; Fermin, D. J. Oxygen Reduction at
Carbon-Supported Lanthanides: The Role of the B-Site. ChemElectroChem 2016, 3, 283–291.
(21) Wang, W.; Li, S.; Wen, Y.; Gong, M.; Zhang, L.; Yao, Y.; Chen, Y. Synthesis and
Characterization of TiO2/YFeO3 and Its Photocatalytic Oxidation of Gaseous Benzene. Acta
Physico-Chimica Sin. 2008, 24, 1761–1766.
(22) Skumanich, A.; Frova, A.; Amer, N. M. Urbach tail and gap states in hydrogenated a-SiC and a-
SiGe alloys. Solid State Commun. 1985, 54, 597–601.
(23) Zanatta, A. R.; Chambouleyron, I. Absorption edge, band tails, and disorder of amorphous
semiconductors. Phys. Rev. B 1996, 53, 3833–3836.
(24) de Rouffignac, P.; Park, J.-S.; Gordon, R. G. Atomic Layer Deposition of Y2O3 Thin Films from
Yttrium Tris(N,N‘-diisopropylacetamidinate) and Water. Chem. Mater. 2005, 17, 4808–4814.
(25) Wu, L.; Yu, J. C.; Zhang, L.; Wang, X.; Li, S. Selective self-propagating combustion synthesis of
hexagonal and orthorhombic nanocrystalline yttrium iron oxide. J. Solid State Chem. 2004, 177,
3666–3674.
(26) Li, J.; Singh, U. G.; Schladt, T. D.; Stalick, J. K.; Scott, S. L.; Seshadri, R. Hexagonal
YFe1−xPdxO3−δ: Nonperovskite host compounds for Pd2+ and their catalytic activity for CO
oxidation. Chem. Mater. 2008, 20, 6567–6576.
(27) Druce, J.; Téllez, H.; Burriel, M.; Sharp, M. D.; Fawcett, L. J.; Cook, S. N.; McPhail, D. S.;
Ishihara, T.; Brongersma, H. H.; Kilner, J. A. Surface termination and subsurface restructuring of
perovskite-based solid oxide electrode materials. Energy Environ. Sci. 2014, 7, 3593–3599.
(28) Cao, J.; Kako, T.; Li, P.; Ouyang, S.; Ye, J. Fabrication of p-type CaFe2O4 nanofilms for
photoelectrochemical hydrogen generation. Electrochem. Commun. 2011, 13, 275–278.
268
Photoelectrochemistry of compact and nanostructured YFeO3 photocathodes
(29) Prévot, M. S.; Guijarro, N.; Sivula, K. Enhancing the Performance of a Robust Sol-Gel-Processed
p-Type Delafossite CuFeO2 Photocathode for Solar Water Reduction. ChemSusChem 2015, 8,
1359–1367.
(30) Peter, L. M. Dynamic aspects of semiconductor photoelectrochemistry. Chem. Rev. 1990, 90,
753–769.
(31) Li, J.; Peat, R.; Peter, L. M. Surface recombination at semiconductor electrodes. Part II.
Photoinduced “near-surface” recombination centres in p-GaP. J. Electroanal. Chem. 1984, 165,
41–59.
(32) Gomis-Berenguer, A.; Celorrio, V.; Iniesta, J.; Fermin, D. J.; Ania, C. O. Nanoporous
carbon/WO3 anodes for an enhanced water photooxidation. Carbon 2016, 108, 471–479.
(33) Backris, J. O.; Uosaki, K. Rate of the photoelectrochemical generation of hydrogen at p-type
semiconductors. J. Electrochem. Soc. 1977, 124, 1348–1355.
(34) Paracchino, A.; Laporte, V.; Sivula, K.; Grätzel, M.; Thimsen, E. Highly active oxide
photocathode for photoelectrochemical water reduction. Nat. Mater. 2011, 10, 456–461.
(35) Hu, S.; Lewis, N. S.; Ager, J. W.; Yang, J.; McKone, J. R.; Strandwitz, N. C. Thin-Film Materials
for the Protection of Semiconducting Photoelectrodes in Solar-Fuel Generators. J. Phys. Chem. C
2015, 119, 24201–24228.
(36) Lin, C.-Y.; Lai, Y.-H.; Mersch, D.; Reisner, E. Cu2O/NiOx nanocomposite as an inexpensive
photocathode in photoelectrochemical water splitting. Chem. Sci. 2012, 3, 3482–3487.
(37) Chen, S.; Thind, S. S.; Chen, A. Nanostructured materials for water splitting - State of the art and
future needs: A mini-review. Electrochem. Commun. 2016, 63, 10–17.
269
Chapter 7
Enhancing water photooxidation on
nanostructured NiTiO3 and anatase
electrodes by electrochemical doping
7.1. Introduction ..................................................................................................... 273
7.2. Experimental section ....................................................................................... 275
7.2.1. Preparation of NiTiO3 Nanorods ............................................................ 275
7.2.2. Electrode Preparation .............................................................................. 276
7.2.3. Electrode Characterization ..................................................................... 276
7.3. Effect of the ultrasound on the synthesis of NiTiO 3 nanorods .................... 277
7.4. Electrochemical doping of NiTiO3 and comparison with the anatase case 279
7.4.1. Physical characterization ......................................................................... 279
7.4.2. Photoelectrochemical characterization of pristine NiTiO 3................... 282
7.4.3. Effect of the electrochemical doping in NiTiO3 electrodes ................... 284
7.4.4. Comparison of the electrochemical doping in NiTiO3 with anatase
electrodes ................................................................................................................. 286
7.5. Conclusions....................................................................................................... 298
7.6. Bibliography ..................................................................................................... 300
Enhancing water photooxidation on nanostructured NiTiO3 and anatase electrodes by electrochemical doping
7.1. Introduction
Significant effort is also being devoted to the pursuit of efficient photoanodes suitable
for the photoelectrochemical water splitting harvesting solar energy. Although some
authors are focused on the search of new materials, particularly ternary oxides, 1,2 other
researchers have dealt with the upgrade of well-known materials such as titanium
dioxide (TiO2). In this regard, today a number of general strategies are considered
suitable to improve the photocatalytic efficiency.3–7 Namely, we can highlight
nanostructuring, sensitization, generation of heterojunctions and doping with
heteroatoms. Most of these strategies have been optimized for TiO 2 and subsequently,
they have been applied to other metal oxide semiconductors.8
Over the last decade, doping TiO2 with heteroatoms, both metallic and non-metallic (e. g.
N, F, C), has been widely studied as it allows to expand the light absorption spectrum
into the visible region.3,5 However, such a modification also triggers a significant
enhancement in the number of structural defects, which undesirably promotes charge
carrier recombination. In this respect, self-doping with Ti3+ species is considered
particularly beneficial as it leads to visible light absorption and to an increased
conductivity while the density of expected defects is minimized.9 Consequently, an
enhancement in the photoelectrochemical activity has been reported by increasing the
content of Ti3+ species or in other words, the number of oxygen vacancies. Such a
doping procedure is traditionally carried out by chemically reducing TiO 2 at high
temperatures in a hydrogen atmosphere.10,11 Concomitantly, in the last years, several
authors have proposed a mild electrochemical reductive pretreatment that can also
trigger the generation of Ti3+ species (and/or oxygen vacancies12) in the structure,
promoting the TiO2 performance.13 In this case, the accumulation of negative charge is
compensated by the insertion/adsorption of cations coming from the electrolyte.
Although the induced reduction is partially reversible, it improves the photovoltaic
efficiency14 and leads to a photocurrent increase for water oxidation. 13 This behavior
has been ascribed to a more rapid electron transport that promotes electron collection,
hindering charge recombination.12,13 This electrochemical reduction has been
successfully applied to different TiO2 structures including single crystals,15
polycrystals,16 compact films,17 nanoparticles, nanowires and nanotubes18–21 in acidic
and neutral electrolytes. Furthermore, the effect of the cation nature (Li + or H+) present
273
Chapter 7
in the electrolyte has also been studied.18,20 To the best of our knowledge, the influence
of the electrolyte pH remains unexplored.
The electrochemical reductive doping has also been applied to other n-type
semiconductors such as WO322,23 and doped and undoped hematite electrodes,24 but in
the latter case, morphological changes upon 20 minutes of reductive treatment were
observed, which could be at least partially responsible for the observed changes. 24
Some authors have proposed that the electrochemical doping could be a general route
to enhance the performance of metal oxides for photoelectrochemical water oxidation.25
Indeed, in Chapter 3 an oxidative electrochemical pretreatment has also been revealed
as a successful strategy to promote the p-type behavior of electrodes based on CuFe 2O4
and a similar effect was found by Guijarro et al. 26 for zinc copper indium sulfide
(ZCIS).
Fueled by a growing interest in new materials that could be advantageous for the
photoelectrochemical production of hydrogen, ternary oxides have attracted particular
attention also for the development of the photoanodic electrode. In this context, NiTiO3
with a calculated band gap of 2.16 eV has been proposed as a promising candidate. 27
With an ilmenite structure, it has conduction band and valence band edges that
appropriately straddle the H+/H2 and O2/H2O redox potentials, respectively.28 This
28–31
allows to use this material as a photoanode for water splitting although some
authors have pointed out its limited (photo)(electro)chemical stability.28,29
Among the different methods for preparing semiconducting materials, those assisted by
high intensity ultrasound irradiation have been described as promising routes to prepare
nanomaterials without using bulk high temperatures and/or high pressures or long
reaction times. In fact, NiTiO3 has been prepared using ultrasonic assisted synthesis for
the (photo)catalytic degradation of pollutants. 32,33 Most authors consider this process
the main differential step in the ultrasonic assisted synthesis of nanomaterials.34–36 In
this regard, the generation of amorphous products is easily understood due to the very
high cooling rates (in the nanosecond time scale), which hinder atom organization, i.e.
crystallization. However, the creation of well-defined nanostructures in the presence of
ultrasound is not so well rationalized, although the volatile character of the solvent
seems to play a key role. The possible physical effects caused by ultrasound should also
274
Enhancing water photooxidation on nanostructured NiTiO3 and anatase electrodes by electrochemical doping
be mentioned. The most important ones are related to high-speed microjets and intense
shock waves induced by cavitation.
In this chapter, we investigate electrochemical doping in NiTiO3 nanorod
photoelectrodes under different conditions. Previously, the morphology, UV/vis
absorption, and (photo)electrochemical performance of nanorods, obtained by the
conventional hydrothermal and ultrasound assisted method, are compared.
Electrochemical reduction reversibly promotes the photoelectrochemical behavior of
ternary NiTiO3 electrodes, improving the photoanodic currents. Herein, the
electrochemical doping potential and time of the treatment are studied in NiTiO3
electrodes prepared by the conventional hydrothermal method. The effect of the
treatment is also analyzed by applying an analogous treatment to anatase nanoporous
electrodes.37
7.2. Experimental section
7.2.1. Preparation of NiTiO3 Nanorods
One-dimensional NiTiO3 nanorods were prepared using a sol-gel route. Namely, 2.48 g
of nickel acetate were dissolved in 60 mL of ethylene glycol. Subsequently, 3.4 mL of
titanium isopropoxide were added drop-wise under constant stirring at room
temperature. The resulting green color suspension was stirred in air for 2 h. After this
step, the suspension color turned into light blue, which was indicative that the reaction
was completed. The light blue precipitate was recovered by vacuum filtration, washed
with ethanol several times and dried under vacuum at 80 °C for 1 h. Finally, the
precipitate was calcined at 600 °C for 2 h in air to obtain the desired NiTiO 3 nanorods.
The obtained powder is named “Hy-NiTiO3”.
The sonochemical synthesis of NiTiO3 nanorods was achieved using an adaptation of
the above described procedure.32 The titanium isopropoxide was added dropwise to the
nickel(II) acetate/ethylene glycol solution under sonication (Ti-horn, 20 kHz), and the
irradiation was continued until the suspension turned light blue. The following steps
were identical, obtaining the NiTiO3 nanorods named “Sono-NiTiO3”.
275
Chapter 7
7.2.2. Electrode Preparation
Electrodes based on NiTiO3 nanorods were fabricated on F:SnO2-coated glass (FTO,
Pilkington TEC15) by the well-known doctor blade method. An aqueous slurry was
prepared by grinding 25 mg of the as-prepared NiTiO3 nanorod powder, 7 L
acetylacetone, 7 L Triton 100X with 125 L of ultrapure water. Some droplets of this
paste were spread over ~1 cm2 of the substrate, dried at ambient temperature, and
annealed at 600 °C for 1 h in air to sinter the nanorods among them and with the
substrate. This procedure is applied to fabricate the electrodes either from the “Hy-
NiTiO3” or “Sono- NiTiO3” powder.
In a similar way, TiO2 nanoporous electrodes based on commercial anatase
nanoparticles (Alfa Aesar, 99.9 %) were also prepared but fired at 450 °C for 1 hour.38
The thickness of all the resulting films was measured with an Alpha Step D-100
profilometer.
7.2.3. Electrode Characterization
A morphological characterization of the as-prepared NiTiO3 powders was carried out
by TEM with a JEM-2010 (JEOL) microscope. The electrode morphology was studied
by SEM with a JSM-840 (JEOL). The crystal structure was analyzed by XRD with a
Seifert JSO- Debyeflex 2002 diffractometer using the Cu K line ( = 1.5406 Å). A
Shimadzu UV-2401PC spectrophotometer equipped with an integrating sphere coated
with BaSO4 was used to measure the UV/vis diffuse reflectance spectra of the NiTiO3
films.
(Photo)electrochemical measurements were performed at room temperature in a three-
electrode cell equipped with a fused silica window. All the potentials were measured
against and are referred to an Ag/AgCl/KCl (3 M) reference electrode. A platinum wire
was used as a counter electrode and an N2-saturated 1 M NaOH solution as a working
electrolyte. Measurements were carried out with a computer-controlled Autolab
PGSTAT30 potentiostat. A 1000 W Hg (Xe) lamp (Newport) equipped with a water
filter and a cutoff filter (cutoff wavelength of 350 nm) Newport model FSR-KG3 were
276
Enhancing water photooxidation on nanostructured NiTiO3 and anatase electrodes by electrochemical doping
used as a UV-visible light source to illuminate the NiTiO3 from the substrate-electrode
side (SE illumination). The experimental conditions for the TiO2 electrodes are detailed
in the thesis of D. Monllor-Satoca.37
7.3. Effect of the ultrasound on the synthesis of NiTiO3 nanorods
Independently if the synthesis of NiTiO3 was carried out in the presence or absence of
ultrasound, the XRD (see section 7.4.1) reveals well-crystallized NiTiO3 phase. A
careful inspection by TEM (Figure 7.1A and B) revealed that in both cases the sample
is composed of nanorods, being slightly smaller those prepared by ultrasound-assisted
synthesis (Sono-NiTiO3). Moreover, no significant differences are observed in the
diffuse reflectance spectra of both Hy-NiTiO3 and Sono-NiTiO3 electrodes (Figure
7.1C).
A. B.
C. 80 Hy-NiTiO3
Reflectance / %
Sono-NiTiO3
60
40
20
0
300 400 500 600 700 800
/ nm
Figure 7.1. SEM micrographs of NiTiO3/FTO electrodes based on (A) Hy-NiTiO3 and (B)
Sono-NiTiO3 nanorods and (C) UV-visible diffuse reflectance spectra for Hy-NiTiO3 (black
line) and Sono-NiTiO3 electrodes (red line).
277
Chapter 7
Hy- and Sono-NiTiO3 nanorod electrodes were prepared on FTO and characterized by
cyclic voltammetry in the dark. As shown in Figure 7.2, the electrochemical response is
characterized in both cases by the presence of capacitive currents at potentials more
negative than 1 V, as expected for an n-type semiconductor. This capacitive region
can be physically associated to the filling with electrons of the conduction band and/or
a trap distribution below it. As this accumulation process is accompanied by the
insertion/adsorption of cations from the electrolyte, the charge in this region can be
considered as a direct measurement of the interfacial area of the electrode.38 The charge
accumulation region is slightly larger for the Sono-NiTiO3 sample than for the Hy-
NiTiO3. This result agrees well with the fact that the electrodes have approximately the
same thickness (Sono-NiTiO3, 2.7 μm and Hy-NiTiO3, 3.2 μm) but the Sono-NiTiO3 is
composed by smaller nanorods.
0.00
h on
-0.05 4
-2
h off
j / mA cm
-2
j / A cm
-0.10 2
-0.15 0
0 25 50 75 100
Tiempo / s
-0.20
-1.5 -1.0 -0.5 0.0 0.5
E (vs Ag/AgCl) / V
Figure 7.2. Cyclic voltammogram in the dark in N2-purged 1 M NaOH (20 mV/s) for Hy-
NiTiO3 (black line) and Sono-NiTiO3 (red line). Inset: photocurrent transient at 0.4 V vs
Ag/AgCl.
We have performed some preliminary experiments to verify the possibility of using
these materials to perform water photooxidation under white light illumination. At
sufficiently positive potentials, a small photocurrent was measured for both samples
(inset of Figure 7.2). In any case, the electrodes based on the Sono-NiTiO3
nanoparticles were more active. This could be related to the higher interfacial area.
Afterward, it was observed that an electrochemical treatment can improve the
photoelectrochemical response of both Sono- and Hy-NiTiO3 (conditions of the
278
Enhancing water photooxidation on nanostructured NiTiO3 and anatase electrodes by electrochemical doping
pretreatment are detailed in the next section). Preliminary experiments showed that the
development of the photocurrent after the electrochemical pretreatment under identical
conditions were larger for the Hy-NiTiO3 electrodes. Thus, the following focuses on the
electrodes fabricated with Hy-NiTiO3, and the “Hy-“ prefix will be omitted.
7.4. Electrochemical doping of NiTiO3 and comparison with the
anatase case
7.4.1. Physical characterization
NiTiO3 nanorods were prepared by the conventional hydrothermal method described in
the experimental section. The morphology of the electrodes was characterized by SEM
and a representative image of the top view is shown in Figure 7.3A. Randomly stacked
nanorods can be easily distinguished with diameters and lengths in the 1-1.5 m and 2-
5 m ranges, respectively. These one-dimensional structures are characterized by a
rough surface. A more detailed morphological analysis by TEM (Figure 7.3B) showed
that they are constructed by aggregated nanoparticles smaller than 100 nanometers in
diameter.
279
Chapter 7
A. B.
2 m 40 nm
C. (104)
D.
80
Reflectance / %
(110)
(116)
60 b
0.4
(012)
40 0.3
(024)
(113)
(300)
2
(124)
0.2 Eg
(hA)
20 a 0.1
(1010)
(018)
(220)
0.0
(006)
(208)
(202)
1.5 2.0 2.5 3.0 3.5
h (eV)
0
10 20 30 40 50 60 70 80 300 400 500 600 700
2 /º / nm
Figure 7.3. (A) SEM image of an FTO/NiTiO3 electrode; (B) TEM image of NiTiO3
nanoparticles obtained after detaching the NiTiO 3 nanorods from the FTO substrate; (C) X-ray
diffraction pattern for the as-prepared NiTiO3 powder with the identification of the different
crystal orientations (* peak ascribed to rutile impurities); and (D) UV/vis diffuse reflectance
spectrum for an NiTiO3 film. Arrows (a) and (b) indicate electronic transitions between Ni 3d
and Ti 3d. Inset: Determination of the direct optical band gap (the absorbance (A) was
estimated from the reflectance measurements using the Kubelka-Munk function).
On the other hand, the XRD analysis (Figure 7.3C) revealed the crystalline nature of
the sample. The different crystallographic plane diffractions could be indexed as the
rhombohedral ilmenite NiTiO3 structure in agreement with the JCPDS file 33-0960,
except for a diffraction peak located at 27.5°, which indicates some rutile impurities.
The ilmenite structure consists of distorted NiO6 and TiO6 octahedra distributed in
alternating layers along the c-axis. The relatively strong and narrow diffraction peaks
indicate high crystallinity. The size of the crystallites oriented along the {104} direction
was estimated by using the Scherrer equation (7.1):
h l (7.1)
cos
280
Enhancing water photooxidation on nanostructured NiTiO3 and anatase electrodes by electrochemical doping
where is the wavelength of the X-rays, is the Bragg angle, B is the XRD peak full
width at half maximum and K is a numerical factor that considers the crystal shape,
being 0.9 for spherical nanoparticles. According to this expression, the NiTiO 3
nanoparticles have a diameter of 22 nm in agreement with the TEM analysis. Figure
7.3D shows the UV/vis diffuse reflectance spectrum, which reflects the yellow-green
color of NiTiO3. Aside from an increasingly growing absorption (i.e. reflectance
diminution) at short wavelengths, two apparent absorption features at 450 and 510 nm
can be distinguished, in agreement with previous results. 39,40 These particular features
can be related to the electronic structure of NiTiO3 in which Ni-3d states are localized
above the valence band constituted by O-2p orbitals while the bottom of the conduction
band is based on Ti-3d orbitals.28,41,42 The bands at 450 nm (2.76 eV) and 510 nm (2.43
eV), indicated by arrows a and b in Figure 7.3D have been related to the Ni-3d orbitals
that split up into two sub-bands, allowing charge transfer between these sub-bands and
the Ti-3d band located at the bottom of the conduction band. 40 The broad absorption at
shorter wavelengths is attributed to the charge transfer between O-2p states at the top of
the valence band and Ti-3d states at the bottom of the conduction band. The
accompanying Tauc plot for this transition (inset in Figure 7.3D) reveals a direct band
gap42 of 3.05 eV. This value agrees well with the photon energy required to develop a
photocurrent28 or a photovoltage43 with NiTiO3 electrodes.
Anatase A. B.
+ Rutile
+
+ +
20 nm
10 20 30 40 50 60 70 80
2 (º)
Figure 7.4. (A) X-ray diffraction pattern for the commercial TiO 2 powder (Alfa Aesar)
calcined at 450 ºC for 1 h in air. (B) Representative TEM image of the TiO 2 nanoparticles
obtained after detaching them from the FTO substrate.
281
Chapter 7
Figure 7.4A shows the XRD of the Alfa Aesar nanopowder, revealing that it is mainly
composed of the TiO2 anatase phase. Low intensity peaks associated with rutile phase
are also detected. Figure 7.4B displays a TEM image of the Alfa Aesar nanopowder.
The sample is composed by spherical crystallites with particle diameters up to 100 nm.
Previous analysis of the commercial powder reveal that most nanoparticles exhibit a
diameter in the range 10-30 nm.13
7.4.2. Photoelectrochemical characterization of pristine NiTiO3
The electrochemical behavior of NiTiO3 electrodes has been studied in the dark and
under transient illumination in N2-purged 1 M NaOH. As shown in Figure 7.5, the
cyclic voltammogram in the dark is characterized by the presence of capacitive currents
at potentials more negative than 0.8 V, as expected for a nanoporous n-type
semiconductor. Only a minor faradaic contribution can be observed at potentials more
negative than 1.4 V, possibly related to water reduction. By analogy with the behavior
observed for nanoporous TiO2 electrodes,38 we can ascribe the capacitive currents to
the filling of band gap states and/or states at the bottom of the conduction band. The
cyclic voltammogram also exhibits a shoulder at about 0.95 V. Considering the
location of the band edges proposed in the literature for NiTiO3,44,45 it can tentatively
be ascribed to the existence of trap states energetically located in the band gap and
physically identified as grain boundaries. On the other hand, the currents appearing at
1.25 V can be tentatively associated with the filling of band gap states just below the
conduction band edge or directly of conduction band states.
282
Enhancing water photooxidation on nanostructured NiTiO3 and anatase electrodes by electrochemical doping
0.05
x10
A.
0.00
-2
j / mA·cm
-0.05 B.
-0.10
-0.15
-0.20
-1.5 -1.0 -0.5 0.0 0.5
E (vs. Ag/AgCl) / V
Figure 7.5. (A) Cyclic voltammogram in the dark for an FTO/NiTiO 3 electrode (4.5 m
thickness). Scan rate: 20 mV s1. (B) Linear sweep voltammogram (negative-going scan) under
transient illumination. The arrows indicate when the light is turned on. Their color indicates if
the photocurrent generated is anodic (black) or cathodic (blue). Scan rate: 5 mV s1.
The electron accumulation needs to be accompanied by the adsorption/insertion of
cations coming from the electrolyte to preserve electroneutrality. In view of the
aqueous nature of the electrolyte, sodium cation adsorption and/or proton
insertion/adsorption can occur. In this regard, although NiTiO 3 has been proposed as a
material for Li-ion battery anodes,46 the insertion of Na+ is unlikely in aqueous media
considering the crystalline nature of the nanorods and the large size of solvated Na +
ions.
Compact NiTiO3 electrodes have been previously studied as photoanodes. 28–31 At
sufficiently positive potentials, anodic photocurrents ascribed to water photooxidation
were measured. As shown in Figure 7.5, the behavior that we observe for the
nanoporous electrodes is clearly different. Under illumination, at potentials more
positive than 0.6 V, anodic photocurrents are developed, while at potentials more
negative than this value, cathodic photocurrents are recorded. Considering the
electrolyte composition, and the fact that the material is stable in 1 M NaOH in this
potential range, the only possible reactions associated with the photocurrents are water
photooxidation (oxygen evolution) and water photoreduction (hydrogen evolution).
Linear scan in the inverse direction (not shown) was also performed in order to
283
Chapter 7
corroborate that the cathodic currents are not related with O2 reduction previously
generated by the anodic photocurrents. In other words, photocathodic currents are
present independently of the scan direction. Then, it can be concluded that NiTiO3
nanorod electrodes behave as photoanodes or photocathodes depending on the applied
potential. This behavior agrees well with that expected for a nanoporous semiconductor
electrode with a limited charge mobility47 (this behavior was also observed in untreated
CuFe2O4 electrodes in Chapter 3). Under these conditions, the photogenerated charge
carrier separation and transport are dominated by the kinetics at the semiconductor-
electrolyte interface, which depends on the exposed facets48 and by the charge diffusion
inside the nanostructure, which is tuned by the substrate potential.
7.4.3. Effect of the electrochemical doping in NiTiO3 electrodes
In this work, the possibility of improving the photoelectrochemical behavior of NiTiO3
through an electrochemical reductive pretreatment has been studied similarly to that
applied on TiO2 electrodes based on commercial anatase nanoparticles in a previous
work.13
The application of an electrochemical pretreatment to the NiTiO3 electrodes consisting
in the polarization of the electrode at negative potentials produces an increase in the
photoanodic currents. In Figure 7.6, linear sweep voltammograms under transient
illumination are presented for an untreated electrode and after applying 1.5 V for 10
min. The photocathodic currents appearing at negative potential are present both before
and after the electrochemical pretreatment while the onset of the photocurrent shifts
around 0.25 V toward negative potentials.
284
Enhancing water photooxidation on nanostructured NiTiO3 and anatase electrodes by electrochemical doping
0.06
Before
After
0.04
-2
j / mA cm
0.02
0.00
-0.02
-0.04
-1.2 -0.8 -0.4 0.0 0.4
E (vs. Ag/AgCl) / V
Figure 7.6. Linear sweep voltammogram under transient illumination for a NiTiO3 electrode
before (black line) and after (red line) an electrochemical pretreatment at 1.5 V for 10 min.
Scan rate: 2 mV s1.
Open circuit measurements for NiTiO3 electrodes prior and after electrochemical
doping are shown in Figure 7.7. After the electrochemical pretreatment, the
photopotential under illumination is more negative, which indicates a diminished
recombination rate. In addition, the faster photopotential decay upon light interruption
is an indication of faster electron transport.
0.00
E vs. Ag/AgCl / V
-0.25
-0.50
Before
-0.75 After
0 200 400 600 800
Time / s
Figure 7.7. Photopotential transients in N2-purged 1 M NaOH obtained before (black line) and
after (red line) the electrochemical pretreatment (1.5 V for 4200 s) for NiTiO3 nanorod
electrodes.
285
Chapter 7
Some organic compounds are usually employed as hole scavengers to evaluate the
photoelectrochemical performance. The presence of methanol in the electrolyte is
studied for the electrochemically doped NiTiO3 electrodes. In Figure 7.8 the
photocurrent after the electrochemical doping was measured at 0.4 V vs Ag/AgCl after
the addition of different amounts of methanol. Surprisingly, the presence of methanol
produces only an increase of 15% of the photocurrent for concentrations above 2 M,
which can be related with a hindered charge transport through the film.
48
46
44
-2
jph / A cm
42
40
38
36
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
[CH3OH] / M
Figure 7.8. Photocurrents after 120 s of illumination for an electrochemical pretreated electrode
at 1.5 V for different concentrations of methanol.
7.4.4. Comparison of the electrochemical doping in NiTiO 3 with anatase
electrodes
In this section, the photoelectrochemical behavior of NiTiO 3 through the
electrochemical reductive pretreatment has been optimized and compared with the
behavior of the nanoporous TiO2 electrodes based on commercial anatase nanoparticles
with polyhedral shapes and with a diameter of 10-30 nm (Figure 7.4). These electrodes
have already been characterized in 0.1 M HClO4 in a previous work.13 They showed
small photocurrents toward water oxidation and the electrochemical reductive
pretreatment only induced a minor enhancement (around 20 % increase). 13 In this
study, we have employed N2-purged 1 M NaOH as electrolyte due to the chemical
stability restrictions of NiTiO3. The effect of the pH is analyzed based on the results
286
Enhancing water photooxidation on nanostructured NiTiO3 and anatase electrodes by electrochemical doping
published in the thesis of D. Monllor-Satoca37 using anatase TiO2 electrodes in N2-
purged 0.1 M NaOH. The experimental results coming from this work are referred in
the captions of the corresponding figures.
It is worth mentioning that although some authors have pointed out the possibility that
sodium ions from aqueous electrolytes could be inserted in amorphous TiO2
electrodes,49,50 this is unlikely if we consider other reports where a potential below 1 V
vs Na+/Na is established as a requirement for Na insertion in anatase nanoparticles.51
This implies that water reduction is thermodynamically more favorable than sodium
insertion. Hence, we have not considered the possible insertion of sodium ions as a way
to compensate electron accumulation. Obviously, sodium adsorption can occur.
The effect of the electrochemical doping strongly depends on a number of factors,
mainly the applied potential and the time, apart from the electrolyte and electrode
nature. In order to select the best pretreatment, we have optimized the experimental
conditions. First, the potential was amended by applying a step for 4200 s and
subsequently, the duration was optimized for the previously selected potential. Figure
7.9 shows the photocurrent recorded at 0.4 V after performing a potentiostatic step at
different potentials for 4200 s for NiTiO3 and TiO2 nanoporous electrodes. As can be
observed, in both cases the photocurrent basically remains unaltered until the
pretreatment attains a certain potential value. Once this potential has been reached, the
photocurrent steeply increases and then, it saturates. For longer periods of time, the
photoelectrochemical behavior seems to deteriorate. In any case, the photocurrent trend
is virtually identical for nanoporous anatase and NiTiO 3 electrodes. Only the optimized
potential is slightly different, being 1.6 V for TiO2 and 1.5 V for NiTiO3. It should
be noted that, in both cases, the photocurrent is exclusively promoted when the
pretreatment is performed at a potential negative enough to trigger charge
accumulation. This can be evinced by the correspondence between the curves of the
dependence of the photocurrent on the pretreatment potential (Figure 7.9A and C) and
the dark linear scan voltammograms (Figure 7.9B and D). This result clearly
demonstrates that the photocurrent enhancement after the reductive pretreatment is due
to charge accumulation and therefore, to the concomitant insertion of protons coming
from the electrolyte. The adsorption of protons/sodium ions is expected to be fully
287
Chapter 7
potential reversible (see below) and thus, the mere presence of these processes would
not explain the photoactivity enhancement.
TiO2 NiTiO3
0.08
A. 0.03 C.
0.06
-2
-2
jph / mA·cm
jph / mA·cm
0.02
0.04
0.01
0.02 0.00
0.0 0
-2
-2
jdark / mA·cm
jdark / mA·cm
-0.5
-1
-1.0
-2
B. D.
-1.5
-1.8 -1.5 -1.2 -0.9 -0.6 -1.8 -1.5 -1.2 -0.9 -0.6
E (vs. Ag/AgCl) / V E (vs. Ag/AgCl) / V
Figure 7.9. (A, C) Stationary photocurrent at 0.4 V measured after electrochemical doping at
each potential during 4200 s for (A) anatase37 (3 m thickness) and (C) NiTiO3 (3.5 m
thickness) electrodes. (B, D) Linear sweep voltammograms (negative-going scan) in the dark
for (B) anatase37 and (D) NiTiO3 electrodes. Scan rate: 20 mV s1.
Once the optimized potentials were selected, the step duration was refined. Figure 7.10
shows the dependence of the stationary photocurrent at 0.4 V measured after the
electrochemical doping, as a function of the pretreatment time. For both materials, the
photocurrent for water photooxidation is clearly promoted with the electrochemical
pretreatment in the dark, attaining a maximum enhancement after some time. Note that
the time scale employed for both materials is very different. NiTiO 3 requires longer
times to attain the maximum photocurrent. It should be mentioned that the charge
involved in this process is about 3.5 times larger. However, it cannot be considered a
direct measurement of the electron accumulation, because at 1.5 V and 1.6 V (the
optimized potentials), water reduction also occurs.
288
Enhancing water photooxidation on nanostructured NiTiO3 and anatase electrodes by electrochemical doping
0.03
A. TiO2 B. NiTiO3
0.06
0.02
-2
-2
j / mA·cm
j / mA·cm
0.04
0.01
0.02
0.00
0.00
0 200 400 600 800 1000 1200 0 2000 4000 6000 8000
Electrochemical doping time / s Electrochemical doping time / s
Figure 7.10. Stationary photocurrent measured at 0.4 V as a function of the electrochemical
doping time (A) at 1.6 V for anatase (3 m)37 and (B) at 1.5 V for NiTiO3 (3 m) electrodes.
In any case, we can consider that the process leading to the photocurrent enhancement
is highly demanding for NiTiO3, i.e., the required potential is similar and the optimized
time is about one order of magnitude longer than for TiO2.
As mentioned above, electrochemical doping can be ascribed to the accumulation of
electrons with the corresponding insertion of protons. Being the crystal structure of
NiTiO3 more open than that of anatase, it should facilitate the diffusion of cations.
However, charge accumulation occurs more easily in TiO 2. Obviously, apart from the
crystal structure, the nanoparticle size and shape can also play a role: the smaller the
building blocks constituting the electrode, the higher the specific area and the shorter
the length that the cations may diffuse to reach the particle bulk in order to compensate
the accumulated electrons and, subsequently, the easier charge accumulation, i.e.
electrochemical doping. NiTiO3 electrodes are constituted by rods composed of
relatively large nanoparticles. The space between these nanoparticles is too small for
the electrolyte to permeate, so we can consider the rods as the main building blocks of
these electrodes, being the active surface area mainly limited to the rough surface of the
nanorods, in agreement with the small capacitive currents observed in Figure 7.5. To
accumulate electrons in the whole structure, the protons should diffuse through such
voluminous structures in NiTiO3, while in the case of anatase, the diffusion pathway is
mainly limited to each individual nanoparticle or to small aggregates.
289
Chapter 7
On the other hand, the photocurrent after electrochemical doping is multiplied by a
factor of about 5 for anatase and 25 for NiTiO3 electrodes. This is probably linked to
the initial larger recombination existing at NiTiO 3 electrodes. In this regard, open
circuit measurements (Figure 7.7) for NiTiO3 electrodes prior and after the optimized
electrochemical doping have revealed, a diminished recombination upon the
pretreatment (as shown by the photopotential attained under illumination).
Importantly, in the case of anatase, the enhancement of the photocurrent after an
optimum electrochemical pretreatment is much larger than in acidic medium, where the
photocurrent is only a 20 % higher.13 This result clearly points to the key role of pH in
the electrochemical doping process. Interestingly, using the same duration for the
electrochemical pretreatment, the enhancement is larger in alkaline medium, where the
concentration of protons is smaller. The reason for this behavior is still far from being
completely understood, but the fact that the anatase surface is negatively charged in
alkaline media could favor the adsorption/insertion of protons. In this regard, there are
a few publications on the untreated material that have shown that it exists an
exponential relation between the charge corresponding to the grain boundary peak and
the electrolyte pH52,53 and these traps have been linked to recombination processes.
Based on these works, higher recombination rates in alkaline media are expected, and
subsequently any modification of the electrode leading to an improvement should be
more noticeable than in acidic media.
The comparison of the photoactivity of different electrode materials is an arduous task,
as it is difficult to have electrodes with the same real surface area or absorption spectra.
In our case, the TiO2 and NiTiO3 electrodes have the same thickness. Figure 7.11
shows the linear sweep voltammogramas under 100 mW cm2 standardized AM 1.5G
illumination. Although we are aware that we cannot discuss in detail the different
factors governing the photoactivity of both materials, from the photocurrents obtained
with AM 1.5G illumination it can be concluded that anatase is more efficient for water
photooxidation.
290
Enhancing water photooxidation on nanostructured NiTiO3 and anatase electrodes by electrochemical doping
40 TiO2
-2
j / A·cm
20
0
(A)
2 NiTiO3
-2
También podría gustarte
- Generalidades de La Psicología Del AprendizajeAún no hay calificacionesGeneralidades de La Psicología Del Aprendizaje13 páginas
- =?Utf8?b?Mza3mjnfy2fub190b3jyzxnfam9zzv9tyxldrweucgrm?=Aún no hay calificaciones=?Utf8?b?Mza3mjnfy2fub190b3jyzxnfam9zzv9tyxldrweucgrm?=291 páginas
- Síntesis de Nanoestructuras Semiconductoras Núcleo/Coraza/Coraza Con Potencial Aplicación en Sistemas para Conversión de EnergíaAún no hay calificacionesSíntesis de Nanoestructuras Semiconductoras Núcleo/Coraza/Coraza Con Potencial Aplicación en Sistemas para Conversión de Energía161 páginas
- Generación de Hidrógeno Mediante Fotoelectrocatálisis Con Fotoánodos de Circonato de Bario Dopado Con Tierras RarasAún no hay calificacionesGeneración de Hidrógeno Mediante Fotoelectrocatálisis Con Fotoánodos de Circonato de Bario Dopado Con Tierras Raras121 páginas
- Materiales Termoeléctricos de Alto ZT en Base de Si y Sus Aleaciones Con GeAún no hay calificacionesMateriales Termoeléctricos de Alto ZT en Base de Si y Sus Aleaciones Con Ge109 páginas
- Nanotubos de Carbono Estructura Porosa y SusAún no hay calificacionesNanotubos de Carbono Estructura Porosa y Sus256 páginas
- 20230123093906_Tesis V.Zapata ICVAún no hay calificaciones20230123093906_Tesis V.Zapata ICV253 páginas
- Síntesis Y Caracterización de Puntos Cuánticos de Cdse Con Aplicaciones en Celdas Fotovoltaicas Con ConfiguraciónAún no hay calificacionesSíntesis Y Caracterización de Puntos Cuánticos de Cdse Con Aplicaciones en Celdas Fotovoltaicas Con Configuración75 páginas
- Tesis Angel Heriberto Romero CorderoAún no hay calificacionesTesis Angel Heriberto Romero Cordero165 páginas
- Estudio de Catalizadores de Platino Soportados Mediante Adsorción Electrostática Selectiva en ÓxidosAún no hay calificacionesEstudio de Catalizadores de Platino Soportados Mediante Adsorción Electrostática Selectiva en Óxidos155 páginas
- Sinterización de Materiales Refractarios PDFAún no hay calificacionesSinterización de Materiales Refractarios PDF151 páginas
- TesisEmilianoMartnezPerin18 12 15Aún no hay calificacionesTesisEmilianoMartnezPerin18 12 15430 páginas
- Evaluación del proceso de secado del CarbónAún no hay calificacionesEvaluación del proceso de secado del Carbón92 páginas
- Escuela Técnica Superior de Ingenieros Industriales y deAún no hay calificacionesEscuela Técnica Superior de Ingenieros Industriales y de184 páginas
- Tesis Freddy-Ariel Sanchez Ruiz 2015Aún no hay calificacionesTesis Freddy-Ariel Sanchez Ruiz 2015209 páginas
- Gutiérrez Pineda - Eduart A. - Tesis - pdf-PDFA PDFAún no hay calificacionesGutiérrez Pineda - Eduart A. - Tesis - pdf-PDFA PDF251 páginas
- Tesis Final Gallegos MV 03-2018.PDF-pdfaAún no hay calificacionesTesis Final Gallegos MV 03-2018.PDF-pdfa295 páginas
- Tomografias para Plumas ContaminantesAún no hay calificacionesTomografias para Plumas Contaminantes55 páginas
- Efectividad Del Proceso Avanzado de OxidaciónAún no hay calificacionesEfectividad Del Proceso Avanzado de Oxidación203 páginas
- Integración de Un Reactor Electroquímico de Placas Paralelas para La Producción de Hipoclorito de Sodio.Aún no hay calificacionesIntegración de Un Reactor Electroquímico de Placas Paralelas para La Producción de Hipoclorito de Sodio.93 páginas
- Solidificación de Hierro Nodular PDFAún no hay calificacionesSolidificación de Hierro Nodular PDF87 páginas
- Centro de Investigación Y Desarrollo Tecnológico en ElectroquímicaAún no hay calificacionesCentro de Investigación Y Desarrollo Tecnológico en Electroquímica104 páginas
- PFC Jesus LaParra Albaladejo CastellanoAún no hay calificacionesPFC Jesus LaParra Albaladejo Castellano55 páginas
- Estudio Numérico-Experimental de La Generación de Gas de Síntesis en Un Reactor Híbrido de Medios Porosos Con Carbón Sub-Bituminoso y Vapor Con Aporte SolarAún no hay calificacionesEstudio Numérico-Experimental de La Generación de Gas de Síntesis en Un Reactor Híbrido de Medios Porosos Con Carbón Sub-Bituminoso y Vapor Con Aporte Solar132 páginas
- Termodinámica clásica. Protocolos experimentales sobre primera y segunda leyes, y sobre potenciales selectosDe EverandTermodinámica clásica. Protocolos experimentales sobre primera y segunda leyes, y sobre potenciales selectosAún no hay calificaciones
- Concepto de Educación - Pedagogía - DidácticaAún no hay calificacionesConcepto de Educación - Pedagogía - Didáctica14 páginas
- Anexo 1-Etapa 0 Milena Historia de La PsicologiaAún no hay calificacionesAnexo 1-Etapa 0 Milena Historia de La Psicologia5 páginas
- Ejercicio de Giving Directions en PDF OnlineAún no hay calificacionesEjercicio de Giving Directions en PDF Online4 páginas
- Guía de Trabajo Energia Cinetica, Potencial y Conservación de La EnergíaAún no hay calificacionesGuía de Trabajo Energia Cinetica, Potencial y Conservación de La Energía3 páginas
- Documento Estrategia para Una Educación de CalidadAún no hay calificacionesDocumento Estrategia para Una Educación de Calidad29 páginas
- BASES - CONCURSO DE DESFILE CÍVICO ESCOLAR POR FIESTAS PATRIAS 2024 I.E InmaculadaAún no hay calificacionesBASES - CONCURSO DE DESFILE CÍVICO ESCOLAR POR FIESTAS PATRIAS 2024 I.E Inmaculada5 páginas
- Solicitud de Autorizacion de EntrevistaAún no hay calificacionesSolicitud de Autorizacion de Entrevista1 página
- Bolívar, Antonio. Las Historias de Vida Del Profesorado PDFAún no hay calificacionesBolívar, Antonio. Las Historias de Vida Del Profesorado PDF24 páginas
- Desarrollo de Habilidades Comunicativas Cuadernillo de Apoyo 2012 Segundo GradoAún no hay calificacionesDesarrollo de Habilidades Comunicativas Cuadernillo de Apoyo 2012 Segundo Grado61 páginas
- =?Utf8?b?Mza3mjnfy2fub190b3jyzxnfam9zzv9tyxldrweucgrm?==?Utf8?b?Mza3mjnfy2fub190b3jyzxnfam9zzv9tyxldrweucgrm?=
- Síntesis de Nanoestructuras Semiconductoras Núcleo/Coraza/Coraza Con Potencial Aplicación en Sistemas para Conversión de EnergíaSíntesis de Nanoestructuras Semiconductoras Núcleo/Coraza/Coraza Con Potencial Aplicación en Sistemas para Conversión de Energía
- Generación de Hidrógeno Mediante Fotoelectrocatálisis Con Fotoánodos de Circonato de Bario Dopado Con Tierras RarasGeneración de Hidrógeno Mediante Fotoelectrocatálisis Con Fotoánodos de Circonato de Bario Dopado Con Tierras Raras
- Materiales Termoeléctricos de Alto ZT en Base de Si y Sus Aleaciones Con GeMateriales Termoeléctricos de Alto ZT en Base de Si y Sus Aleaciones Con Ge
- Síntesis Y Caracterización de Puntos Cuánticos de Cdse Con Aplicaciones en Celdas Fotovoltaicas Con ConfiguraciónSíntesis Y Caracterización de Puntos Cuánticos de Cdse Con Aplicaciones en Celdas Fotovoltaicas Con Configuración
- Estudio de Catalizadores de Platino Soportados Mediante Adsorción Electrostática Selectiva en ÓxidosEstudio de Catalizadores de Platino Soportados Mediante Adsorción Electrostática Selectiva en Óxidos
- Escuela Técnica Superior de Ingenieros Industriales y deEscuela Técnica Superior de Ingenieros Industriales y de
- Gutiérrez Pineda - Eduart A. - Tesis - pdf-PDFA PDFGutiérrez Pineda - Eduart A. - Tesis - pdf-PDFA PDF
- Integración de Un Reactor Electroquímico de Placas Paralelas para La Producción de Hipoclorito de Sodio.Integración de Un Reactor Electroquímico de Placas Paralelas para La Producción de Hipoclorito de Sodio.
- Centro de Investigación Y Desarrollo Tecnológico en ElectroquímicaCentro de Investigación Y Desarrollo Tecnológico en Electroquímica
- Estudio Numérico-Experimental de La Generación de Gas de Síntesis en Un Reactor Híbrido de Medios Porosos Con Carbón Sub-Bituminoso y Vapor Con Aporte SolarEstudio Numérico-Experimental de La Generación de Gas de Síntesis en Un Reactor Híbrido de Medios Porosos Con Carbón Sub-Bituminoso y Vapor Con Aporte Solar
- Química física del ambiente y de los procesos medioambientalesDe EverandQuímica física del ambiente y de los procesos medioambientales
- Termodinámica clásica. Protocolos experimentales sobre primera y segunda leyes, y sobre potenciales selectosDe EverandTermodinámica clásica. Protocolos experimentales sobre primera y segunda leyes, y sobre potenciales selectos
- Guía de Trabajo Energia Cinetica, Potencial y Conservación de La EnergíaGuía de Trabajo Energia Cinetica, Potencial y Conservación de La Energía
- Documento Estrategia para Una Educación de CalidadDocumento Estrategia para Una Educación de Calidad
- BASES - CONCURSO DE DESFILE CÍVICO ESCOLAR POR FIESTAS PATRIAS 2024 I.E InmaculadaBASES - CONCURSO DE DESFILE CÍVICO ESCOLAR POR FIESTAS PATRIAS 2024 I.E Inmaculada
- Bolívar, Antonio. Las Historias de Vida Del Profesorado PDFBolívar, Antonio. Las Historias de Vida Del Profesorado PDF
- Desarrollo de Habilidades Comunicativas Cuadernillo de Apoyo 2012 Segundo GradoDesarrollo de Habilidades Comunicativas Cuadernillo de Apoyo 2012 Segundo Grado