University of Southern California Dissertations and Theses

Proton kinetics in electrochemistry: new directions and mechanistic analysis

(USC Thesis Other)

Proton kinetics in electrochemistry: new directions and mechanistic analysis

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content PROTON KINETICS IN ELECTROCHEMISTRY: NEW DIRECTIONS AND
MECHANISTIC ANALYSIS
by
Shima Haghighat
A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulllment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(MATERIALS SCIENCE)
December 2017
Copyright 2017 Shima Haghighat
To Mohsen,
the love of my life
and Milan,
the joy of my life
ii
Acknowledgments
The nishing touch on my dissertation is writing this note of appreciation. I would
not be able to complete my Ph.D without the help and the guidance of the ones
who in one way or another contributed and extended their valuable assistance.
First and foremost, I would very much like to thank my thesis advisor Prof.
Jahan Dawlaty. It was truly an honor to work under his supervision and I felt very
lucky every single day to have his endless support. Thank you for all the advices,
ideas, moral support and patience in guiding me through this research. Jahan
is a wonderful mentor. I can not nd words to describe his joy and enthusiasm
for science. Through him, I learned how to critically evaluate a research idea by
always trying to explain it in a simple basic way. What I like about him the most
is his ability to inspire you by discussing new research ideas in such a way that
you would go to the lab and try to implement the idea . He always gave me the
freedom to explore my own paths toward experiencing new ideas, while at the same
time helping me to stay on the right track. I am very grateful to have him as my
doctoral advisor.
I would also like to thank the other members of both my proposal and disser-
tation committees for providing me valuable comments on my thesis: Professor
Steve Nutt, and Professor Jayakanth Ravichandran. Their feedback and support
is greatly appreciated.
Special thanks to my amazing group members who always helped me with any
problems in the lab and gave me their valuable feedbacks on my research updates
and presentations, Shayne Sorenson, Eric Driscoll, Joel Partrow, Aaron Rury and
Ryan Hunt.
To all my friends in LA, thank you all for all the fun and incredible moments we
had together. Thanks to Payman, Soheil, Marjan, Nima, Sepideh and Amirsoheil
iii
for your unconditional help and support which means a lot to me. Special thanks
goes to Tahereh for being a wonderful friend in every way, keeping in touch from
miles away and making me laugh when I did not even want to smile.
I would like to thank my parents for always being there for me. I never thank
you enough for supporting me in my academic pursuits and all other aspects of life.
Without the inspiration, drive and support that my dad has given me I might not
be the person that I am today. Dad, thank you for believing in me. To my Mom,
Fariba , thank you for always loving me and for helping me at every turn. You are
a wonderful, caring, and kind-hearted mom. Without your sacrices, nishing up
my PhD work would have not been possible. Thanks for staying away from home
and taking care of my baby while I was working on my thesis. Thank you my
brothers, Reza, and Mohammad for your love and support. The way you watch
out for me and protect me makes me condent that I can always count on you.
Last but certainly not least, thanks to my wonderful, encouraging and patient
husband Mohsen. Thank you for being my biggest supporter. You always help
me push my boundaries and motivate me when I am feeling down. Thanks for
letting me complain and putting up with my mood swings. I am grateful for your
invaluable advices and devotion towards my success. Loving you is one of the best
things that happen to me and I will cherish every moment I spent with you.
iv
Table of Contents
Dedication ii
List of Figures viii
Abstract xv
1 Introduction 1
1.1 Water Splitting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.1.1 Hydrogen Evolution Reaction (HER) . . . . . . . . . . . . . 4
1.1.2 Oxygen Evolution Reaction (OER) . . . . . . . . . . . . . . 7
1.2 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.3 Outline of the Thesis . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2 Oxygen Evolution Reaction on the surface of Manganese Oxide
and Iron Oxide Electrocatalysts 14
2.1 Experimental . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.1.1 Thin Film Preparation . . . . . . . . . . . . . . . . . . . . . 15
2.1.2 Electrochemical Set Up . . . . . . . . . . . . . . . . . . . . . 16
2.2 MnO
x
and Fe
2
O
3
Thin Films . . . . . . . . . . . . . . . . . . . . . 16
2.3 Electrocatalytic Activity of MnO
x
and Fe
2
O
3
Thin Films . . . . . . 21
2.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3 Photoelectrochemical Activity of Iron Oxide 24
3.1 Photoelectrochemical Water Splitting . . . . . . . . . . . . . . . . . 25
3.2 Experimental . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
3.2.1 Iron Oxide Thin Film Photoanode Preparation . . . . . . . 30
3.2.2 Incorporation of Manganese Oxide to the Iron Oxide Thin
Film Photoanode as a Co-Catalyst . . . . . . . . . . . . . . 30
3.2.3 Photoelectrochemical (PEC) Set Up . . . . . . . . . . . . . 32
3.3 Hydrothermally Grown Iron Oxide Photoanode . . . . . . . . . . . 33
3.4 Photoelectrochemical Water Splitting by Iron Oxide Photoanode . . 34
3.4.1 The Thickness Eect on the Photoactivity . . . . . . . . . . 36
v
3.5 Applying a Manganese Co-Catalyst on the Iron Oxide Photoanode . 39
3.5.1 Photoactivity of Fe
2
O
3
+ MnO
x
Co-Catalyst Photoanodes . 39
3.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
4 Kinetic Studies of Oxygen Evolution Reaction on the Surface of
Iron Oxide Electrode 45
4.1 Introducing the Kinetic Parameters: Electron Transfer Coecient
() and Reaction Order () . . . . . . . . . . . . . . . . . . . . . . 46
4.1.1 Electron Transfer Parameter . . . . . . . . . . . . . . . . . 46
4.1.2 Reaction Order . . . . . . . . . . . . . . . . . . . . . . . . 49
4.2 Experimental . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
Ohmic Drop Correction . . . . . . . . . . . . . . . . . . . . 50
4.3 Constructing the 2-D Current-Voltage Plots . . . . . . . . . . . . . 51
4.4 Constructing the (pH,E) and (pH,E) Plots . . . . . . . . . . . . . 53
4.5 Langmuir Surface Adsorption Model to Interpret the Observed . . 55
4.6 Surface Adsorbed Intermediates Involved in the OER . . . . . . . . 57
4.7 Interpreting the Reaction Order at dierent pH regions . . . . . . 58
4.7.1 Acidic Pathway . . . . . . . . . . . . . . . . . . . . . . . . . 58
4.7.2 Basic Pathway . . . . . . . . . . . . . . . . . . . . . . . . . 59
4.7.3 Intermediate Pathway . . . . . . . . . . . . . . . . . . . . . 59
4.8 Electron Transfer Parameter at dierent pH regions . . . . . . . . 60
4.9 The Mass Transport Eect on the Mechanistic Scheme . . . . . . . 63
4.10 Participation of Ions in the Reaction Mechanism . . . . . . . . . . . 65
4.11 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
5 Kinetic Studies of Hydrogen Evolution Reaction on the Surface
of Metallic/Semiconductor Electrodes 68
5.1 Experimental . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
5.2 A 4-State Potential Energy Surface for Proton Reduction: . . . . . 70
5.2.1 Prediction of the model for the pH-dependence of . . . . . 78
5.3 Experimental Result and Discussion . . . . . . . . . . . . . . . . . . 81
5.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
6 Controlling Proton Conductivity with Light: A Scheme based on
Photoacid Doping of Materials 89
6.1 Experimental . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
6.2 Using Photoacid as a Proton-bearing Molecule . . . . . . . . . . . . 92
6.3 Electrochemical Impedance Spectroscopy (EIS) . . . . . . . . . . . 93
Resistor (R) . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
Capacitor (C) . . . . . . . . . . . . . . . . . . . . . . . . . . 97
Constant Phase Element (CPE) . . . . . . . . . . . . . . . . 97
Warburg Element (W) . . . . . . . . . . . . . . . . . . . . . 98
vi
6.4 Results of Impedance Measurements on Photo-acid HPTS Solutions
in the Light and Dark Conditions . . . . . . . . . . . . . . . . . . . 100
6.5 Fitting the Impedance data to an equivalent Circuit Model . . . . . 101
6.6 The in
uence of adding NaOH on the conductivity of Photo-acid
HPTS Solutions in the Light and Dark Conditions . . . . . . . . . . 103
6.6.1 Relation of the real part of impedance to the magnitude of
the Warburg element . . . . . . . . . . . . . . . . . . . . . . 104
6.7 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
6.8 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
7 Future Suggested Steps 112
Mechanistic studies of one electron, one proton transfer sys-
tem at the electrode surface: . . . . . . . . . . . . . 112
In-situ studies of reaction intermediates in water oxidation
reaction: . . . . . . . . . . . . . . . . . . . . . . . . 113
Nanostructuring the co-catalyst on the surface of the pho-
toanode: . . . . . . . . . . . . . . . . . . . . . . . . 113
Protonic photoconductivity in bipolar membranes: . . . . . . 113
Bibliography 115
vii
List of Figures
1.1 Illustration of a water electrolysis system. The electron
ow is from
the positive terminal (anode) to the negative terminal (cathode).
Oxygen is generated at the anode while hydrogen is evolving in
the cathode. The diaphragm is for collecting oxygen and hydrogen
separately at the receivers[Zeng and Zhang, 2010]. . . . . . . . . . . 3
1.2 The square scheme for one proton-coupled one electron transfer reac-
tion (PCET) (M +e

+H
+
! MH), M is the surface active site
. ET = electron transfer, PT = proton transfer, and CPET = con-
certed protonelectron transfer. [Koper, 2013b] . . . . . . . . . . . . 6
1.3 The square scheme for hydrogen evolution reaction (two proton two
electron transfer reaction) [Koper, 2013b; Shinagawa and Takanabe,
2017]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.4 A volcano plot. The data points are experimentally measured exchange
current density, log (i
0
) plotted as a function of the calculated free
energy of H adsorption at 0 V. The line is a prediction by a kinetic
model in which all input parameters are taken from DFT calcu-
lations. The dashed line indicates that the metals which bind H
stronger than 0.2 eV/H usually form oxides at 0 V. The open circles
are single crystal data whereas the lled circles are polycrystalline.
[Sklason et al., 2010] . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.5 The square scheme for oxygen evolution reaction (four proton four
electron transfer reaction) that occurs at an active surface site (M)
and possible sequential protonelectron transfers along the path. The
assumption is that the charge of the reaction intermediates does not
exceed one electron. The positive and negative intermediates are
indicated in blue and pink, respectively [Giordano et al., 2016]. . . . 10
2.1 Schematic illustration of photochemical metal organic deposition.
The M stands for a metal (Fe or Mn) and L
n
is an organic ligand,
2-ethylhexanoate [Buono-Core et al., 1999]. . . . . . . . . . . . . . . 17
2.2 Thin lm of Manganese oxide and iron oxide on a glass slide. From
left to right is as-deposited, UV-treated and heat-treated. (a) Man-
ganese oxide thin lms, (b) iron oxide thin lms. . . . . . . . . . . . 18
viii
2.3 Infrared spectra of a thin lm of Fe(III) 2-ethylhexanoate just deposited
and subjected to UV irradiation. After 10 hr of UV radiation, char-
acteristic organic vibrations (CH) are eliminated. . . . . . . . . . . 18
2.4 Infrared spectra of a thin lm of Mn(II) 2-ethylhexanoate just deposited
and subjected to UV irradiation. After 10 hr of UV radiation, char-
acteristic organic vibrations (CH) are decreased signicantly. . . . . 19
2.5 X-ray diraction pattern of iron oxide thin lm on FTO substrate.
Two asterisk signs denote the Fe2O3 peaks. The rest is from FTO. 19
2.6 X-ray diraction pattern of iron oxide thin lm on FTO substrate.
Two asterisk signs denote the Fe2O3 peaks. The rest is from FTO. 20
2.7 Energy Dispersive X-ray (EDX) spectrum of iron oxide thin lm on
silicon substrate, showing atomic ratio of oxygen to iron of1.7. . . 21
2.8 Energy Dispersive X-ray (EDX) spectrum of manganese oxide thin
lm on silicon substrate, showing atomic ratio of oxygen to man-
ganese of2.8. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.9 Linear sweep voltammograms (after correction for ohmic drop) for
MnO
x
, Fe
2
O
3
and bare FTO at pH=13. The inset plot is Log(current
density) vs. potential. . . . . . . . . . . . . . . . . . . . . . . . . . 22
3.1 Band model of semiconductor-solution interface(a) Before contact -
the Fermi level of the semiconductor is given by E
F
, the electrochem-
ical potential of the solution containing a redox pair is E(X/X
-
)(b)
At contact- because the E(X/X
-
) is lower than E
F
the electrons
ow
from semiconductor to the solution until reaching charge equilib-
rium (c) After the contact - a positively charged region of width W
is established in the semiconductor and corresponding to that the
band bending occurs [Tan et al., 2007]. . . . . . . . . . . . . . . . . 27
3.2 Energy diagram for photoelectrochemical water splitting with an n-
type semiconductor photoanode. The involved processes are absorp-
tion of a photon of light, electron-hole pair generation, transport of
hole toward the interface of photoanode-electrolyte, oxygen evolu-
tion at the interface, and nally transport of electron through an
external wire and reduction of water at the metallic cathode surface
[Sivula et al., 2011]. . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
3.3 (a) The Thorlabs high power light source and the lens (b) The PEC
cell including hematite working electrode, Ag/AgCl reference elec-
trode and platinum counter electrode in a 0.1 M KOH solution. . . 32
3.4 Schematic illustration of hydrothermal synthesis of iron oxide thin
lms followed by heat treatment [Ling et al., 2014]. . . . . . . . . . 33
ix
3.5 X-ray diraction pattern of iron oxide thin lm hydrothermally
grown for 24 hr on FTO substrate. Then heat treated at 550
0
C
for 2 hr. Three asterisk signs denote the Fe
2
O
3
peaks. The rest is
from FTO. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
3.6 Current density vs. potential plots under dark (dashed line) and
under illumination (solid line) for iron oxide photoanode obtained
after 6 h of hydrolysis. . . . . . . . . . . . . . . . . . . . . . . . . . 35
3.7 Energy diagram for water oxidation on hematite surface. The rela-
tive band levels of hematite with respect to the water oxidation and
reduction reaction is shown [Sivula et al., 2011]. . . . . . . . . . . 36
3.8 UV-vis spectroscopy on hematite thin lms on FTO substrate at dif-
ferent thickness. The thicker the hematite the more light absorption
it has. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
3.9 In
uence of lm thickness on measured photocurrent at constant
voltage (0.95 V vs NHE) for hematite photoanodes. The highest
photocurrent is observed in the 6 h synthesized hematite with 70
nm thickness. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
3.10 a) and b) Current density vs. potential plots under dark (dashed
line) and under illumination (solid line) for Fe
2
O
3
photoanode (H-
6h) and Fe
2
O
3
modied with MnO
x
(H-6h-MnCl
2
spin-coated) and
(H-6h-Mn(NO
3
)
2
spin-coated), respectively. c) and d) chopped light
chronoamperometry measurements of Fe
2
O
3
photoanode and Fe
2
O
3
electrodes modied with MnO
x
. . . . . . . . . . . . . . . . . . . . . 41
3.11 a) and b) Current density vs. potential plots under dark (dashed
line) and under illumination (solid line) for Fe
2
O
3
photoanode (H-
6h) and Fe
2
O
3
modied with MnO
x
(H-6h-MnCl
2
SILAR-3cycles)
and (H-6h-Mn(NO
3
)
2
SILAR-3cycles), respectively. c) and d) chopped
light chronoamperometry measurements of Fe
2
O
3
photoanode and
Fe
2
O
3
electrodes modied with MnO
x
. . . . . . . . . . . . . . . . . 42
3.12 a) and b) Current density vs. potential plots under dark (dashed
line) and under illumination (solid line) for Fe
2
O
3
photoanode (H-
6h) and Fe
2
O
3
modied with MnO
x
(H-6h-MnCl
2
SILAR-10cycles)
and (H-6h-Mn(NO
3
)
2
SILAR-10cycles), respectively. c) and d) chopped
light chronoamperometry measurements of Fe
2
O
3
photoanode and
Fe
2
O
3
electrodes modied with MnO
x
. . . . . . . . . . . . . . . . . 43
4.1 Simplied diagram of electrochemical Gibbs energy along the reac-
tion coordinate for a single electron oxidation reaction R!O +e

. 49
4.2 An example of ohmic drop correction by extrapolation of the
dV
dI
vs.
1
I
the value of R
u
on the iron oxide electrode at pH 7 is70
. . . 52
x
4.3 Measured linear sweep voltammograms after compensation for lm
resistance at dierent pH values from pH=7 to pH=13. Although
dierences in the behavior of I-E curves as a function of pH exist,
it is hard to visualize the in
uence of pH using this representation.
This is better addressed in the two-dimensional representation in
gure 4.4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
4.4 The electrochemical current logI(pH;E) for water splitting on the
surface of iron oxide as a function of pH and potential. The Pour-
baix line for water stability (solid line) is overlaid for reference. The
double arrows indicate the overpotential for water oxidation, and
demonstrate its variation as a function of pH. The hashed region
corresponds to no data, since the compensated voltage axis is dif-
ferent for various values of pH (see experimental section). . . . . . 54
4.5 The proton reaction order plot (pH;E) for water splitting on the
surface of iron oxide. The reaction order shows the sensitivity of the
electrochemical current to changes in pH as dened in equation 4.6.
In the acidic region, the electrochemical current, and correspond-
ingly the water oxidation reaction rate, is practically insensitive to
pH, while for the basic mechanism 6= 0. The largest sensitivity to
pH is observed in the region where a switch between the two mech-
anisms is expected. This is explained in the text using a surface
protonation model. . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
4.6 The electron transfer coecient plot(pH;E), with mechanistically
distinct regions. The electron transfer coecient is a dimension-
less measure of the sensitivity of the reaction rate to applied poten-
tial as shown in equation 4.2. A noteworthy feature is that in the
basic region is distinctly larger than in the acidic region. The dotted
arrow in the intermediate region shows a reaction that starts with
the basic mechanism, but due to electrochemical generation of pro-
tons near in the interface, switches to acidic mechanism, displaying
a characteristic smaller . These features are further explained in
the text. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
4.7 Hypothesized paths for two-electron, two-proton generation of sur-
face MO, where M represents a surface active site (i.e. an iron
ion). At zero applied potential, the pH determines the extent of the
equilibrium MOH
2
MOH

, and hence the starting ingredients
for oxidation. These paths provide a reasonable explanation of the
observed and in the acidic and basic limits. . . . . . . . . . . . 60
xi
4.8 The kinetic diagrams under stirring condition. (a) The electrochem-
ical current logI(pH;E). (b) The electron transfer coecient dia-
gram (pH;E) and (c) proton reaction order diagram (pH;E).
The characteristic intermediate region is shifted to lower pH by
about 0.5 pH value. . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
4.9 The kinetic diagrams for an experiment where phosphoric acid,
instead of hydrochloric acid, was used for pH adjustment of the
electrolyte. (a) The electrochemical current logI(pH;E). (b) The
electron transfer coecient diagram (pH;E) and (c) proton reac-
tion order diagram (pH;E). . . . . . . . . . . . . . . . . . . . . . . 66
5.1 Four states participating in the Volmer reaction . . . . . . . . . . . 71
5.2 Example of an adiabatic potential energy surfaces of the four states
as a function of electron and proton transfer coordinates q
e
and q
p
.
The parameters employed are: = 0 V,
e
=
p
= 1 eV,

= 0:5
eV, G
prot
= 0:1 eV, and G

prot
=0:2 eV. . . . . . . . . . . . . 74
5.3 Free energy surfaces for CPET (a) and stepwise paths (b). The
transitions states are indicated by. As explained in the text, pH
controls the height of the surface in the corner 2. At high pH proto-
nation is unfavorable, corner 2 is raised and the PES shown in (a)
determines the reaction. At low pH, protonation is spontaneous,
corner 2 is lowered and the PES shown in (b) controls the reaction. 80
5.4 Electrochemical current logI(pH;E) for hydrogen evolution reac-
tion on the surface of dierent electrodes as a function of pH and
potential a) Pt b) Au c) Cu d) Ni e) Ti f) FeS
2
and g) glassy carbon
(GC). The color scale is kept the same in all gures for ease of com-
parison. The Pourbaix line for water stability (solid line) is overlaid
for reference. As expected, Pt exhibits almost no overpotential (the
potential dierence between the Pourbaix line and the onset of cur-
rent) at low pH values, while GC has the highest overpotential in
the same range. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
5.5 The measured electron transfer coecient(pH;E) for Pt, Au, Cu,
Ni, Ti, FeS
2
, and glassy carbon (GC) electrodes. The color scale is
kept the same in all gures for ease of comparison. . . . . . . . . . . 82
5.6 Experimentally measured values of as a function of pH for dierent
type of electrodes a) copper, gold, platinum, and iron pyrite. b)
glassy carbon, nickel, and titanium. The two sets of data are shown
separately for clarity. c) calculated values of as a function of pH
which is obtained from the model for selected values of parameter :

14
=0:4 V,
24
=0:1 V,
e
=
p
= 0:2 eV,

= 0:15 eV, and
pKa = 3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
xii
5.7 The in
uence of rotation on the dynamic potential pH diagrams
and electron transfer coecient for gold rotating disk electrode.
(a,b) The logI(pH;E) and (pH;E) with no rotation. (c,d) The
logI(pH;E) and (pH;E) at 2000 rpm. . . . . . . . . . . . . . . . . 85
5.8 The in
uence of rotation on the dynamic potential pH diagrams
and electron transfer coecient for glassy carbon rotating disk
electrode. (a,b) The logI(pH;E) and (pH;E) with no rotation.
(c,d) The logI(pH;E) and (pH;E) at 2000 rpm. . . . . . . . . . . 86
5.9 The in
uence of rotation on the dynamic potential pH diagrams and
electron transfer coecient for platinum rotating disk electrode.
(a,b) The logI(pH;E) and (pH;E) with no rotation. (c,d) The
logI(pH;E) and (pH;E) at 1600 rpm. . . . . . . . . . . . . . . . . 87
5.10 Levich plot obtained at E=0.5 V vs NHE for the HER at dierent
pH (from 3.01 to 1.57) and dierent rotation rates. . . . . . . . . . 87
6.1 Schematic 3D open view of the liquid cell used for the impedance
study. The material is sandwiched between two transparent con-
ductive electrodes. . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
6.2 Photoacids (HA) are organic molecules which are signicantly more
acidic in the excited electronic state than in the ground state (HA
*
) H
+
+
A
{
). The generated conjugate base (A
{
) can bound with the free
protos and recover the original HA at the dark condition. . . . . . . 93
6.3 Structures of 8-hydroxypyrene-1,3,6-trisulfonic, commonly referred
to as HPTS and its photochemical reaction. [Spry et al., 2007; Wen
et al., 2010] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
6.4 Optical absorption measured using 1mM HPTS solution shows two
distinct absorptions at 405 nm and 385 nm. after adding few drops
of 0.1 M NaOH the HPTS gets deprotonated and the absorption
peaks disappear. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
6.5 principles of impedance spectroscopy (a) by applying a sinusoidal
AC voltage to the system under the steady-state we can measure
the oscillating current response. Here, V
0
is the DC voltage, V is
the amplitude of the probing signal (typically V < 10mV ), and
! is the angular frequency, t is time, I
0
is the direct current, I is
the amplitude of the AC response, and is the phase shift between
the probing signal and the response. (b) In a linear system, the
current response to a sinusoidal potential will be a sinusoid at the
same frequency but shifted in phase. . . . . . . . . . . . . . . . . . 95
6.6 A typical Nyquist plot which results from plotting the imaginary
impedance component (Z
00
) against the real impedance component
(Z
0
) at each excitation frequency. The impedance has a magnitude
jZj and phase-shift at each frequency. . . . . . . . . . . . . . . . . 96
xiii
6.7 The most common equivalent circuit models with their representa-
tive Nyquist plots (a) Randles circuit (b) constant phase element
(CPE) in parallel with charge transfer resistant (c) mass transfer
dominant circuit model with Warburg diusive element (W). . . . 100
6.8 (a) Typical Nyquist plot (the negative of the imaginary part ver-
sus the real part of the impedance) for 1 mM HPTS in distilled
water solution over the frequency range of 0.1 Hz to 1 MHz in the
presence and absence of light. (b) The impedance spectra of 1 mM
HPTS in PEG from 0.1 Hz to 1 MHz. The inset shows the high
frequency region. The solid lines represent the impedance of the
model equivalent circuit shown in gure 6.10. . . . . . . . . . . . . 101
6.9 Nyquist plot of pure distilled water over the frequency range of 0.1
Hz to 1 MHz in the presence and absence of light. . . . . . . . . . . 102
6.10 The equivalent circuit model for (a) 1 mM HPTS in distilled water
solution (b) 1 mM HPTS in PEG, in the presence and absence of
light. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
6.11 Data demonstrating that the contrast between the dark and light
EIS spectra is signicantly diminished upon addition of NaOH in
solution of HPTS in both water (a) and PEG (b), thus supporting
that the change in impedance is related to photoreleased protons. . 105
6.12 The interplay between diusion and recapture in photoinduced free
proton generation. After light absorption, a proton is released from
the photoacid which diuses away from the parent ion. When the
molecule returns back to the ground state the drive to recapture the
proton is established again. However, if the proton has managed to
exit the zone of in
uence of the parent ion, approximated here as
the electrostatic Bjerrum length
B
, it can escape and contribute
to conductivity as a free carrier. Thus a preliminary gure of merit
for understanding this process is comparison of the diusion length
L
D
and the recapture radius
B
. . . . . . . . . . . . . . . . . . . . 110
xiv
Abstract
Many (photo-)electrochemical processes involve the transfer of both protons and
electrons (R + nH
+
+ ne
{
*
) P) at the electrochemical interfaces. Fundamental
understanding of the complex mechanistic details of these processes is crucial to
improve their overall eciency. The operating conditions, such as applied electrode
potential and the electrolyte pH, can impact the electron transfer rate and the
proton transfer rate in a non-trivial way. In particular, pH in
uences the kinetic
of proton transfer between the electrolyte and the surface adsorbed intermediates
which can consequently modify the electron transfer rate to/from the electrode
surface. In this research, the importance of proton transfer kinetics is addressed.
Toward this goal, we rst study the kinetics of electrochemical water splitting
half reactions, oxygen evolution reaction (OER) and hydrogen evolution reaction
(HER). We demonstrate the variations of two important kinetic parameters as a
function of pH and potential for the OER on metal oxide electrodes. The elec-
tron transfer parameter () and the proton reaction order () are represented
in 2-dimensional contour plots in the pH-potential space. We propose dierent
mechanism pathways for oxygen evolution based on the variation of and as a
function of pH and potential. To study the kinetic of the HER, we measure on
several electrode materials over a wide range of pH. We compare our experimental
results with a 2D potential energy surface model that has been investigated before.
xv
There are several similarities between the pH dependence of derived from the
model and the empirical that is obtained from the experiment.
To further explore the proton transfer kinetics, we take a dierent path and
study protonic photoconductivity. In electronic photoconductivity, a material
becomes more electrically conductive due to generation of mobile electrons with
light. Here we report, for the rst time a protonic analogue of photoconductivity.
When a material is doped with photoacid, light excitation generates extra mobile
protons that change the low-frequency conductivity of the material. We measure
and model such change both in polyethylene glycol (PEG) and in water sandwiched
between two transparent electrodes and doped with a well-known photoacid 8-
hydroxypyrene-1,3,6-trisulfonic acid (HPTS). We anticipate that this scheme can
be employed in protonic circuits where direct transduction of energy from light to
protonic gradients or protonic currents is necessary.
xvi
Chapter 1
Introduction
Transitioning from fossil fuels to clean and sustainable energy resources is an ulti-
mate goal for our current society. Electrocatalytic conversion of chemicals to pro-
duce energy carriers is a promising route toward this goal. Hydrogen is widely
considered to be the targeted energy carrier owing to its extremely high energy
density (Hydrogen: 120 MJ/kg and gasoline: 44 MJ/kg). Hydrogen can be pro-
duced from water, which is one of the most abundant sources in our planet, through
electrochemical water splitting.
Electrochemical water splitting reaction involves two half reactions, namely,
the hydrogen evolution reaction (HER, 2 H
+
+ 2 e
{
! H
2
) and oxygen evolution
reaction (OER, 2 H
2
O! O
2
+4 H
+
+4 e
{
). Both reactions involve multiple trans-
fer of electrons from or to the electrode and exchange of multiple protons between
the electrolyte and the reaction intermediates. To minimize the energy losses, we
need to use electrocatalys. Understanding the reaction mechanism of these elec-
trocatalytic multistep processes is a key step to revolutionize the development of
clean energy systems.
This chapter begins with a review of water splitting reaction principals. Then
we will explain separately the hydrogen evolution reaction (HER) and oxygen
evolution reaction (OER) and their reaction mechanisms .
1
1.1 Water Splitting
Water is one of the most abundant and cheapest resources for making hydrogen.
It is also a thermodynamically stable compound thus breaking it into its elements
is an uphill battle. It requires the minimum +237 kJ/mol input energy or 2.46 eV
for making one mole of H
2
and
1
2
mole of O
2
out of one mole of H
2
O. Two electrons
(total charges of 192928 C (=2F)) has to cross the electrocatalytic interface for
this conversion. Splitting of water using electricity is electrolysis and has been
known for around 200 years [Zeng and Zhang, 2010]. It is one of the easiest ways
to make ultra-pure hydrogen (i.e. 99:999%). However, only 4% of total world
hydrogen production is coming from water electrolysis. This is due to the lack of
cheap and ecient water oxidation catalysts to increase the rate of the reaction
[Galn-Mascars, 2015].
A basic electrolysis cell consists of an anode, a cathode, an alkaline electrolyte
solution and a power supply, as depicted in gure 1.1. The overall reaction consists
of two half reactions. The oxidation reaction (oxygen evolution reaction,OER) at
the anode and the reduction reaction (hydrogen evolution reaction, HER) at the
cathode [Kanan and Nocera, 2008].
O
2
+ 4 H
+
+ 4 e

*
) 2H
2
O E
anodic
= 1:23V 0:059(pH)V vsNHE (1.1)
4 H
+
+ 4 e

*
) 2H
2
E
cathodic
= 0V 0:059(pH)V vsNHE (1.2)
2 H
2
O!O
2
+ 2 H
2
E
rxn
=1:23V (1.3)
The thermodynamic voltage that is required to split the water to hydrogen and
oxygen is 1.23 V which is called the reversible volatge. According to the Nernst
equation (equation1.4) the potential is dependent on the concentration of H
+
in
2
anode cathode
diaphragm
-
-
-
-
-
-
OH
-
+
- +
+
+
+ +
+
H
+
O
2
H
2
electrolyte electrolyte
hydrogen
receiver
oxygen
receiver
Electron
flow
+ -
DC power
Figure 1.1: Illustration of a water electrolysis system. The electron
ow is from the
positive terminal (anode) to the negative terminal (cathode). Oxygen is generated
at the anode while hydrogen is evolving in the cathode. The diaphragm is for
collecting oxygen and hydrogen separately at the receivers[Zeng and Zhang, 2010].
the solution. Hence for a one unit of pH the potential change would be 0.059 V
[Gong and Dai, 2015].
E

E
0

RT
nF
ln
[a
R
]
[a
O
]
(1.4)
In this equation (equation1.4), the E is the cell potential, E
0
is the cell potential at
the standard condition, R is the gas constant, T is the temperature in Kelvin, n is
the number of electron transferred, F is the Faraday constant ( 96500 C mol
-1
),
and the a
R
and a
O
are the activity of reduced and oxidized reagents in a typical
half reaction like O + ne
{
*
) R.
As we discussed earlier, the minimum theoretical potential to split the water is
1.23 V at room temperature. However, to make the reaction proceed at appreciable
3
rates we need additional potential above 1.23 V. The term 'overpotentail' () is
indeed the dierence between the applied cell potential and the thermodynamic
reversible potential:
= E
0
E (1.5)
This overpotential determines the energy losses to drive the water oxidation reac-
tion. The job of an electrocatalyst is to minimize the and thereby an ecient
electrocatalyst is the one that operates close to the Nernstian potential (E).
1.1.1 Hydrogen Evolution Reaction (HER)
Hydrogen evolution reaction (HER) is the simplest, most studied multi-step elec-
trochemical reaction. Its overall reaction involves discharging of hydrated protons
at the electrode surface that generates gaseous hydrogen. Depending on the reac-
tion condition (acidic vs alkaline) , we can write the general reaction as hydronium
ion reduction (equation 1.6) or water molecule reduction (equation 1.7). In neutral
solutions both reactions can occur.
2 H
3
O
+
+ 2 e

! H
2
+ 2 H
2
O (1.6)
2 H
2
O + 2 e

! H
2
+ 2 OH

(1.7)
Two dierent mechanism have been established for HER, namely, Volmer-Tafel
and Volmer-Heyrovsky. The rst step is always the Volmer reaction which is elec-
trochemical hydrogen adsorption (MH). In acidic media, MH forms via discharge
of protons, where M designates an active site on the electrode surface:
M + H
3
O
+
+ e

! MH + H
2
O (Volmer) (1.8)
4
Then for the second step there are two possibilities, Tafel (chemical desorption
reaction) or Heyvrosky step (electrochemical desorption):
2 MH! H
2
+ 2 M (Tafel) (1.9)
MH + H
3
O
+
+ e

! H
2
+ H
2
O + M (Heyrovsky) (1.10)
In alkaline solutions, the Volmer and Heyvrosky reactions are:
M + H
2
O + e

! MH + OH

(Volmer) (1.11)
MH + H
2
O + e

! H
2
+ OH

+ M (Heyrovsky) (1.12)
These reactions involve transfer of both proton and electron. Such reactions
are termed proton-coupled electron transfer (PCET) and they are very common
in electrocatalysis. The transfer of proton and electron can happen sequentially,
for which proton transfer (PT) either precedes or follows electron transfer (ET), or
the transfer can take place in a concerted way, namely concerted proton-electron
transfer (CPET) [Costentin and Savant, 2017]. These dierent pathways are com-
monly illustrated in a so-called square scheme for a single proton-electron transfer
reaction (gure 1.2)[Koper, 2013b].
Many PCET reactions favor concerted pathways to avoid the high energy inter-
mediates associated with the ET and PT. However, for the molecular and oxide
catalysts, the stepwise electron and proton transfer is the likely scenario to hap-
pen since they can hold more localized charge which leads to formation of charged
5
M+ e
-
+ H
+
MH
+
+ e
-
MH
M
-
+ H
+
CPET
ET
PT
ET
PT
Figure 1.2: The square scheme for one proton-coupled one electron transfer reaction
(PCET) (M +e

+H
+
! MH), M is the surface active site . ET = electron
transfer, PT = proton transfer, and CPET = concerted protonelectron transfer.
[Koper, 2013b]
reaction intermediates. The HER on metal electrode surfaces is usually consid-
ered to proceed via a CPET pathway because the extended band structure in a
metal makes a delocalized charge distribution over the entire metal [Jackson and
Surendranath, 2016]. Considering the ET, PT and CPET possible pathways for
a two proton two electron transfer reaction (i.e. HER), gure 1.3 summarizes all
the plausible pathways and intermediates for the HER reaction. pH is an impor-
tant electrolyte property which will aect the rate of these PCET type reactions.
In particular, the stepwise paths in which proton and electron do not transfer
simultaneously are pH dependent.
The Ability of a given catalyst to catalyze the HER is usually measured by
the exchange current density, which is proportional to the reaction rate at the
equilibrium potential. Sabatier's principle in catalysis suggests, that for hydrogen
evolution reaction the plot of exchange current density vs. hydrogen adsorption
energy should follow a volcano plot (gure 1.4). Such plots show an optimum
hydrogen binding energy, in which hydrogen is held neither too strongly nor too
weakly. Thus at the optimum point the HER catalyst is able to adsorb the reactant
6
M + 2e
-
+ 2H
+
MH
+
+ 2e
-
+ H
+

MH + e
-
+ H
+
M
-
+ e
-
+ 2H
+
CPET
ET
MH
2
2+
+ 2e
-
MH
2
+
+ e
-

M
2-
+ 2H
+
MH
-
+ H
+

MH
2

ET
ET ET
ET ET
PT
PT
PT PT
PT PT
CPET
Figure 1.3: The square scheme for hydrogen evolution reaction (two proton two
electron transfer reaction) [Koper, 2013b; Shinagawa and Takanabe, 2017].
and desorb the product eciently. The platinum group metals fall near the apex of
volcano and they are known to be the best heterogenous catalysts for HER. Metals
on the left adsorb hydrogen atoms too strongly, and on the right too weakly.
1.1.2 Oxygen Evolution Reaction (OER)
As described above, the water splitting has two half reactions, OER and HER
(equations 1.1 and 1.2). Among these two reactions, OER is the more complex
and demanding one. To make a double O-O bond from two water molecules, con-
secutive transfer of four electrons and four protons needs to be complete. Since
it is generally accepted that one electron transfer at the interface can occur at
a time, the OER is a multi-step reaction with a slow kinetic [Surendranath and
Nocera, 2011]. Despite years of research on the OER, the mechanistic pathways
for evolving oxygen (from a surface of a catalyst) still remains controversial and
not fully understood in atomic scale. Over the past ve decades, dierent possi-
ble mechanistic schemes have been proposed based on a classical electrochemical
studies (current-potential relationship, tafel slope and steady state polarization
7
Figure 1.4: A volcano plot. The data points are experimentally measured exchange
current density, log (i
0
) plotted as a function of the calculated free energy of H
adsorption at 0 V. The line is a prediction by a kinetic model in which all input
parameters are taken from DFT calculations. The dashed line indicates that the
metals which bind H stronger than 0.2 eV/H usually form oxides at 0 V. The open
circles are single crystal data whereas the lled circles are polycrystalline. [Sklason
et al., 2010]
measurements) [Doyle et al., 2013; Dau et al., 2010]. Here we mention two com-
mon proposed pathways for acidic and basic environment by Bockris [Bockris,
1956]. The acidic mechanistic hypothesis is outlined by equations 1.13 to 1.15. In
these equations M denotes a catalytically active surface site. The equation 1.13 is
considered as the rate limiting step (RDS).
M + H
2
O! MOH + H
+
+ e

(1.13)
MOH! MO + H
+
+ e

(1.14)
MO + MO! O
2
+ 2 M (1.15)
8
The reaction mechanism for a basic environment is brought by equations 1.16
to 1.19. The desorption of OH from the surface of the catalyst is considered to be
the RDS (equation 1.17)
M + OH

! MOH + e

(1.16)
MOH + OH

! MO

+ H
2
O (1.17)
MO

! MO + e

(1.18)
MO + MO! O
2
+ 2 M (1.19)
Therefore, all of the early mechanistic schemes starts with formation of surface
adsorbed intermediates (i.e. MOH and MO). Then the surface specious could
react with electrolyte specious or other intermediates to nally liberate oxygen.
The binding energy of these adsorbed intermediates to the surface and hence the
relative activation barrier at each step denes the rate limiting step of the OER
[Doyle et al., 2013; Dau et al., 2010].
As discussed earlier, the oxygen evolution reaction on the surface of a catalyst
consist of transfer of multiple protons and electrons. Hence the OER is an example
of the proton-coupled electron transfer (PCET) reactions and we can extend the
general square reaction scheme which was presented in gure 1.2 to describe the
possible mechanistic pathways (gure 1.5). The scheme shows how dierent reac-
tion paths, consisting concerted proton and electron transfer and sequential paths
can be imagined to generate oxygen molecule on the surface. Based on the relative
stability of reaction intermediates at a certain applied voltage and the pH of the
solution, these hypothesized paths are favored or disfavored [Giordano et al., 2016;
Shinagawa and Takanabe, 2017].
9
MOO

MOH
2
[MOH
2
]
+
MOH

[MOH]
-
-H
+
-e
-
-H
+
[MOH]
+
[MO]
-
+ H
2
O

MO + H
2
O
-e
-
-H
+
-H
+
-e
-
-e
-
-e
-
-H
+
-e
-
-H
+
-H
+
-e
-
-H
+
-e
-
[MO]
+
+ H
2
O
-e
-
-H
+
-e
-
[MOOH]
+

[MOOH]

[MOOH]
-

-e
-
-H
+
-H
+
-e
-
-H
+
-e
-
[MOO]
-

-H
+
-H
+
-H
+
-H
+
-e
-
-H
+
-e
-
-H
+
-e
-
-e
-
Increasing Deprotonation
Increasing Electron
Transfer into Electrode
M + O
2
-e
-
Figure 1.5: The square scheme for oxygen evolution reaction (four proton four
electron transfer reaction) that occurs at an active surface site (M) and possible
sequential protonelectron transfers along the path. The assumption is that the
charge of the reaction intermediates does not exceed one electron. The positive
and negative intermediates are indicated in blue and pink, respectively [Giordano
et al., 2016].
In order to generate oxygen at an appreciable rate at the OER, developing an
ecient catalyst is essential. Using an OER catalyst can lower the overpotential
required for water oxidation. This can be done by lowering the activation barrier
energy of the surface intermediate related to the rate determining step (RDS). A
large family of extended solid state catalyst have been evaluated for water oxidation
reaction including metals, metal oxides, spinels, and perovskites [Bassi et al., 2014;
Cook et al., 2010]. Metal oxides are a good candidate since they are stable in
electrolysis operational conditions which is highly oxidizing. The most ecient
and active OER catalysts are among the rare and precious based metal oxides
such as Ir, and Ru. However, the scarcity and high cost are the major limitation
for those catalyst to be used in large scale systems. As such much eort has
10
been focused on looking for water oxidation catalyst made from inexpensive, earth
abundant materials such as Co, Ni, Fe, and Mn based oxides.
1.2 Motivation
Electrochemical conversion of water toH
2
(clean fuel) andO
2
(by-product) is a fun-
damental step in solar to fuel energy technologies. Its half reactions, oxygen evolu-
tion reaction (OER) and hydrogen evolution reaction (HER) involve transferring of
multiple electrons and protons at electrochemical interfaces (electrode/electrolyte).
To improve the rates of these electron-proton redox reactions we need to gain more
mechanistic insight about the kinetics of electron transfer and proton transfer at
the interfaces. The applied electrode potential can modulate the electron transfer
kinetics; while, the pH of the electrolyte can impact the rate of proton transfer
reaction. However, this is a simplistic picture, and in reality the kinetics of elec-
tron and proton transfer are intimately coupled to each other, often in non-trivial
ways. Despite the central role of pH on the kinetics of electron-proton redox
reactions, many pH related studies have focused on the extreme pH conditions
(highly acidic/alkaline) to maximize the reaction eciency. The role of pH over
a wide range especially at close to neutral conditions are poorly understood. We
are inspired to study the kinetics of OER and HER reactions on the surface of
variety of metal oxides and metallic/semiconductor electrodes to unravel the pH
dependence on these fundamental electrocatalytic reactions.
11
1.3 Outline of the Thesis
This thesis is divided into four parts. First, we focus on (photo)electrochemical
water oxidation reaction on metal oxide surfaces (chapter 2 and 3). In chapter 2,
we prepare iron oxide and manganese oxide for electrochemical water oxidation.
They are made as thin lms using a facile photochemical technique. We measure
and compare the electrocatalytic activities of these lms in an electrochemical cell
in alkaline (pH=13) solution. Then in chapter 3, our intention moves toward the
light driven water oxidation reaction (photoelectrochemical) using iron oxide (-
Fe
2
O
3
) as the photoanode. For this purpose, we prepare a photoactive iron oxide
thin lm via a hydrothermal method. The in
uence of the lm thickness of -
Fe
2
O
3
on water oxidation photocurrent is shown. To improve the photoactivity
of the iron oxide, we use a coating layer of MnO
x
as a co-catalyst to increase the
kinetics of the water oxidation reaction.
In the second part (chapter 4), we study the kinetic of water oxidation reaction
(OER) on the surface of iron oxide thin lm electrode. We explicitly, identify the
role of pH on the rate of OER reaction. A new type of 2-dimensional contour plot
is used to represent the electrochemical current as a function of pH and applied
potential I(pH,E). We discuss the advantages of this kind of representation to
convey kinetic and thermodynamic information at the same time. We calculate
two fundamental kinetic parameters, the empirical electron transfer coecient ,
and the empirical reaction order with respect to protons . These parameters
are obtained from the experimental current-voltage measurements on the surface
of the iron oxide thin lm in a large range of pH (pH = 7 to 13) and potential.
Hence, by presenting the (pH,E) and (pH,E) in a 2-dimensional contour plot,
we gain a global view of the sensitivity of the reaction to pH and potential in the
pH-potential space.
12
In the third part (chapter 5), we investigate the kinetic of water reduction
reaction (HER) on several electrode materials over a wide range of pH (pH = 1.5
to 6.5). We use the 2D representation of empirical electron transfer coecient
to draw attention to the variation of both pH and potential. We also analyze a
parametrized potential energy surface based on a model for the reduction of one
proton and calculate the electron transfer coecient and its pH dependence.
Then we compare our experimental data to the predicted based on the model.
Finally, we study the proton conductivity of a photoacid-doped material in the
absence and presence of light in chapter 6. We show that when a material is doped
with photoacids, light excitation generates extra mobile protons that can change
the low-frequency conductivity of the material. We measure such changes in both
poly(ethylene glycol) (PEG) and in water sandwiched between two transparent
electrodes and doped with a photoacid. Chapter 7 includes the summary and
future work.
13
Chapter 2
Oxygen Evolution Reaction on
the surface of Manganese Oxide
and Iron Oxide Electrocatalysts
The oxygen evolution (2 H
2
O
*
) O
2
+ 4 H
+
+ 4 e
{
), one half of the overall water
splitting, provides the electron required to reduce H
+
and produce H
2
as a fuel.
This reaction is considered a major bottleneck with a signicant overpotential
to generate measurable current densities. Consequently, it demands application
of highly active catalysts that enables water oxidation at potentials close to the
thermodynamic limit. Metal oxides specially the precious based (i.e. IrO
2
and
RuO
2
) are known to be the most stable and active catalysts for the OER. The
limited supply and high cost has made the urge to investigate and develop highly
active earth-abundant metal oxides [Singh and Spiccia, 2013]. Inspired by natural
water splitting catalysts in photosystem II (PSII), which is a unique Manganese
cluster (Mn
4
CaO
5
), OER catalysts based on Mn oxide have attracted considerable
research interest [Najafpour et al., 2016]. Also iron oxide is another appealing
catalyst for water oxidation because of the abundance and low cost of iron.
This section begins with preparation of two earth-abundant metal oxide elec-
trocatalysts, iron oxide (Fe
2
O
3
) and manganese oxide (MnO
x
) thin lms and their
characterization. Then we use electrochemical measurements to evaluate their
14
catalytic activities in electrochemical water oxidation reaction at basic (alkaline)
condition.
2.1 Experimental
2.1.1 Thin Film Preparation
The iron oxide and manganese oxide thin lms are prepared by the photochem-
ical metal organic deposition (PMOD) technique [Smith et al., 2013; Buono-
Core et al., 1999] . In this method, a photosensitive precursor solution is spin-
coated onto a substrate and then exposed to UV radiation. The metal precur-
sors used are Iron(III)ethylhexanoate (50% in mineral spirits, Alfa Aesar) and
Mn(II)ethylhexanoate (40% in mineral spirits, Alfa Aesar). All precursors are used
without any further purication. A 0.058 g of Iron(III)ethylhexanoate solution is
mixed with 0.142 g hexane to produce a 15% w/w of metal complex. Similarly,
0.0852 g of Mn(II)ethylhexanoate solution is mixed with 0.142 g hexane (15% w/w
metal complex). Flourin-doped tin oxide (FTO) coated glass (MTI Corporation,
TEC70) substrates are cleaned ultrasonically in acetone followed by methanol and
DI water then dried under air streaming. A few drops of the solution are dis-
pensed on to the FTO by spin coating at a rate of 3000 rpm for 1 min. The lm
is irradiated under a UV lamp (254 nm) for 10 hr until removal of organic ligands
was complete, leaving an amorphous metal oxide lm on the substrate. FT-IR
spectroscopy is performed on the thin lm deposited on glass substrates to ensure
elimination of the 2-ethylhexanoate signature vibrational peaks. The thin lms
are annealed at T=600

C for 1 hr in air using a Vulcan A-550 oven. The X-ray
diraction (XRD) technique (A Rigaku diractometer(CuK

, = 1.54118 A

)
is used and energy dispersive X-ray spectroscopy (EDX) (Spectrum was collected
15
using a JEOL JSM 7001F LV scanning electron microscope operating at 15 kV
and equipped with an EDAX Apollo X detector) also performed on the samples
for further characterization. The lm thickness is measured by prolometry.
2.1.2 Electrochemical Set Up
The electrodes are prepared by contacting the FTO layer with a copper wire and
silver paint in a region that is deliberately not covered by the metal oxide lm, and
fully covered and secured using epoxy glue. Electrochemical data were obtained
using a Gamry Reference 3000 potentiostat and a three electrode cell. A Ag/AgCl
reference electrode and Pt mesh counter electrode are used in all the experiments.
The electrolyte solution was 0.1 M KOH (pH 13). The cell was purged with N
2
gas prior to each scan and then was closed to the ambient air. Linear sweep
voltammetry experiments were carried out at a 10 mV/s scan rate, between 0 and
2.4 V versus Ag/AgCl electrode. The indicated potentials are always converted
to the normal hydrogen electrode (NHE) using E
NHE
= E
Ag/AgCl
+ 197mV . By
measuring the electrodes surface area in cm
2
, we calculate the current density by
simply dividing the current values by the surface area.
2.2 MnO
x
and Fe
2
O
3
Thin Films
The photochemical metal organic deposition (PMOD) technique is a simple method
for preparing metallic oxide thin lms. A photo-sensitive precursor which has the
starting metal, ML
n
(M is a metal and L
n
is the organic ligand), is used to make
a lm on a substrate. Then by exposure to the UV light, organic ligands are
ejected from the metal center leaving an amorphous metal lm on the substrate.
Due to the presence of oxygen in the environment, a partial oxidation of the metal
16
ML
n
(thin film) M + nL (gas) M
x
O
y
(thin film)
hν

O
2
Spin coating
(3000rpm-60 s)
Fe(C
8
H
15
O
2
)
3
or
Mn(C
16
H
30
O
4
)
UV irradiation
(10 hr)
Heat treatment
(1 hr at 600°C)
Fe
2
O
3
or MnO
x
Figure 2.1: Schematic illustration of photochemical metal organic deposition. The
M stands for a metal (Fe or Mn) and L
n
is an organic ligand, 2-ethylhexanoate
[Buono-Core et al., 1999].
center occurs making a M
x
O
y
metal oxide thin lm. This process is illustrated
schematically in gure 2.1.
The as-deposited, UV treated and heat treated manganese oxide and iron oxide
thin lms on glass slide substrate are shown in a digital picture in gure 2.2. The
trend of changing color is obvious from this picture. The pale color of the Mn
and Fe organometallic compounds turned to a darker color after exposure to the
UV light. They turned to a dark lm after heat treatment at T=600
0
. The lm
thickness is measured by prolometry and was determined to be 40 nm and
46 nm for iron oxide and manganese oxide respectively. The infrared spectra of the
metal oxide lms before UV irradiation and after 10 hour UV treatment reveals
the liberation of ligand and formation of metal oxide. This is depicted in gure
2.3 and 2.4.
The X-ray diraction on metal oxide lms reveles that they are mostly amor-
phous. Although for iron oxide lm there are two characteristic peaks for Fe
2
O
3
located at 37
0
and 65
0
, for manganese oxide there is no crystallin peak related to
any kind of manganese oxide structure. Hence, we assigne MnO
x
structure to our
manganese oxide lms.
17
(a)
(b)
Deposited

UV treated Heat treated
Figure 2.2: Thin lm of Manganese oxide and iron oxide on a glass slide. From
left to right is as-deposited, UV-treated and heat-treated. (a) Manganese oxide
thin lms, (b) iron oxide thin lms.
2000 2500 3000 3500 4000
0.8
0.9
1
1.1
1.2
1.3
Wavenumber (cm
−1
)
Transmittance

Fe2O3 Fresh
Fe2O3 after 10 hr UV
Figure 2.3: Infrared spectra of a thin lm of Fe(III) 2-ethylhexanoate just deposited
and subjected to UV irradiation. After 10 hr of UV radiation, characteristic organic
vibrations (CH) are eliminated.
18
2500 3000 3500
0.95
1
1.05
1.1
1.15
1.2
1.25
Wavenumber (cm
−1
)
Transmittance

MnOx Fresh
MnOx after 10 hr UV
Figure 2.4: Infrared spectra of a thin lm of Mn(II) 2-ethylhexanoate just deposited
and subjected to UV irradiation. After 10 hr of UV radiation, characteristic organic
vibrations (CH) are decreased signicantly.
20 30 40 50 60 70
0
200
400
600
800
1000
2 Θ (degree)
Intensity

*
*
Fe
2
O
3
*
FTO #
#
#
#
#
#
#
#
Figure 2.5: X-ray diraction pattern of iron oxide thin lm on FTO substrate.
Two asterisk signs denote the Fe2O3 peaks. The rest is from FTO.
19
20 30 40 50 60 70
0
100
200
300
400
500
600
700
800
2 Θ (degree)
Intensity
FTO #
#
#
#
#
# #
#
Figure 2.6: X-ray diraction pattern of iron oxide thin lm on FTO substrate.
Two asterisk signs denote the Fe2O3 peaks. The rest is from FTO.
Energy Dispersive X-ray (EDX) spectrum of metal oxide thin lms on silicon
substrate, show us the ratio of Fe (and Mn) to oxygen (gures 2.7, 2.8). The
atomic ratio of oxygen to iron is1.7, which is close to the composition of Fe
2
O
3
.
The slight excess of oxygen is likely due to the underlying silicon oxide layer. In
the manganese oxide lm, the ratio of oxygen to manganese is2.8 that if we
consider the excess of oxygen rising from the silicon oxide layer, it might be close
to the MnO
2
structure.
20
Figure 2.7: Energy Dispersive X-ray (EDX) spectrum of iron oxide thin lm on
silicon substrate, showing atomic ratio of oxygen to iron of1.7.
Figure 2.8: Energy Dispersive X-ray (EDX) spectrum of manganese oxide thin lm
on silicon substrate, showing atomic ratio of oxygen to manganese of2.8.
2.3 Electrocatalytic Activity of MnO
x
and Fe
2
O
3
Thin Films
The performance of MnO
x
and Fe
2
O
3
lms for water oxidation are tested by linear
sweep voltammetry at pH 13 (0.1 M KOH) electrolyte. We also check any possible
21
electrocatalytic activity of FTO substrate by taking the I-V curve at the same
condition as the metal oxide electrodes. As shown in gure 2.9, the onset potential
for water oxidation is lower in MnO
x
electrode compare to Fe
2
O
3
by 0.1 V. The
inset plot shows the current density in logarithmic scale and thereby the 100 mV
dierence of onset potential is more visible. The bare FTO in comparison was found
to be practically inactive. This result shows the MnO
x
is a better electrocatalyst
in oxidizing water compare to the Fe
2
O
3
electrode. Although, at higher applied
potential the capability of Fe
2
O
3
electrode to generate higher currents is noticeable.
This could be related to higher conductivity of Fe
2
O
3
compare to MnO
x
which
makes the electron transfer to the FTO easier.
The reasons that why manganese oxide thin lm showed a better electrocat-
alytic activity compare to iron oxide is unclear from our understanding. It has
been suggested that manganese has multiple oxidation states so in catalysis cycle
it can hold multiple charges and form stable intermediates.
0 0.5 1 1.5 2
0
0.005
0.01
0.015
0.02
Potential (V vs. RHE)
Current Density (A/cm
2
)

MnOx
Fe2O3
FTO
0 0.5 1 1.5
−4.5
−4
−3.5
−3
−2.5
−2
−1.5
Potential (V vs. RHE)
Log [Current Density (A/cm
2
)]

0.1 V
Figure 2.9: Linear sweep voltammograms (after correction for ohmic drop) for
MnO
x
, Fe
2
O
3
and bare FTO at pH=13. The inset plot is Log(current density) vs.
potential.
22
2.4 Conclusion
In summary, we have made two metal oxide thin lms, manganese oxide and
iron oxide, as electrocatalysts in oxygen evolution reaction. The X-ray diraction
revealed that the thin lms were mostly amorphous. These thin lm electrodes
are very stable in alkaline conditions which they were tested for water oxidation
reaction. Manganese oxide has shown lower onset potential for oxygen evolution
reaction compare to the iron oxide.
23
Chapter 3
Photoelectrochemical Activity of
Iron Oxide
Iron(III)oxide in the form of hematite is an attractive material for photoelectro-
chemical water splitting. Natural abundance, nontoxicity, low cost, visible light
absorption and aqueous stability are unique combination to make hematite a
promising candidate toward solar hydrogen production. While in theory with
a band gap of2.1 eV hematite can reach the eciency of solar to hydrogen
conversion of 15 %, in practice only a quarter of the theoretical limit has been
achieved. This is due to charge carrier recombination losses, slow kinetic of water
oxidation at the surface and short diusion length of holes [Klahr et al., 2012;
Sivula et al., 2011]. Part of these challenges can be addressed by nanostructuring
and surface treatment [Bassi et al., 2014]. In this section of our study, we investi-
gated the photoactivity of hematite in a photoelectrochemical cell. The hematite
photoanode was prepared by a convenient hydrothermal method that previously
has been shown to be photoactive [Carvalho and Souza, 2014; Dalle Carbonare
et al., 2014a]. The hematite photoanode served as our test bed for further surface
treatment using MnO
x
deposited layer.
In this chapter, after a brief introduction about photoelectrochemical water
splitting systems, we will describe the behavior of hematite photoanode in dark
and light condition. Also, we show the in
uence of the lm thickness of hematite
24
on water oxidation photocurrent. We also deposited a co-catalyst of manganese
oxide on the iron oxide photoanode to investigate any photoactivity improvement.
3.1 Photoelectrochemical Water Splitting
The basics of conversion of sunlight energy into chemical energy deals with light
stimulation of a chemical reaction and generation of high-energy chemical prod-
ucts. General principle of this type of conversion could be found in photoelec-
trochemical cells in which shining light onto an electrode in solution will cause
an electrochemical reaction to happen. The energy gaps in semiconductors makes
them the suitable electrodes in photoelectrochemistry. The incoming light with
high energy photons pumps the electrons from the valence band to the conduction
band of the semiconductor. This is followed by an electrochemical interfacial reac-
tion that uses electrons from ions in the interface to ll the holes in the valence
band created by the departure of the excited electron [Bockris and Reddy, 2001;
Pleskov, 1990].
Solar water splitting using a semiconductor is one of the most attractive appli-
cation of photoelectrochemistry that may lead to a sustainable hydrogen-based
energy economy. A semiconductor as an electrode material has some distinguished
features such as low concentration of free charge carriers and the two dierent types
of free carriers, electrons and positive holes. By denition, the highest occupied
band is called the valence band of the semiconductor and the lowest vacant band
is called the conduction band. The dierence between the energy of the bottom
of the conduction band (the next unlled band) and the top of the valence band
(upper of the lled bands) is the width of the semiconductor band gap, E
g
. By
absorption of photons of light with hv >E
g
(v is the frequency of incoming light
25
andh is the Plank's constant) by a semiconductor, the electron transition from the
valance band to the conduction band occurs and a hole is left in the valance band.
The illuminated semiconductor has an excess concentration of carriers, electrons
and holes.
All phenomena associated with photoelectrochemical systems rely on the inter-
face that is formed between a semiconductor and an electrolyte. When a semicon-
ductor is immersed in an electrolyte solution, charge transfer will occur between the
two phases because of the dierence in electrochemical potential of the two phases.
The phase with more negative electrochemical potential will lose electrons to the
phase with more positive electrochemical potential. The charge keeps
owing until
the establishment of an equilibrium between the electrode and the electrolyte. At
equilibrium, one phase has excess negative charge and the other has excess positive
charge. If the electrolyte solution contains a redox system (X/X
-
) with the follow-
ing reaction (equation3.1), the electrochemical potential of the electrolyte can be
determined from the Nernst equation (equation3.2). In this equation, E(X/X
-
) is
the actual electrochemical potential of the electrolyte and E
0
(X/X
-
) is the standard
electrochemical potential of the redox couple (X/X
-
), n is the number of electron
transferred, and [X] and [X
-
] are the respective concentrations of the acceptor and
donor species [Tan et al., 2007].
X + ne

! X

(3.1)
E(X=X

)

E
0
(X=X

)
RT
nF
ln
[X

]
[X]
(3.2)
The electrochemical potential of the electrons of a semiconductor is known as
Fermi level (E
F
). The Fermi energy is dened from the the Fermi distribution
function and it corresponds to an energy level where the probability of nding an
26
(a) Before contact (b) At contact
Semiconductor
E
vb
E
cb
E
F
E(X/X
-
)
Solution
E
F
E(X/X
-
)
E
cb
E
vb
e
-
E
F
E(X/X
-
)
E
cb
E
vb
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
+++
+++
+++
+++
+++
+++
+++
+++
+++
W
(c) After contact
Figure 3.1: Band model of semiconductor-solution interface(a) Before contact - the
Fermi level of the semiconductor is given by E
F
, the electrochemical potential of the
solution containing a redox pair is E(X/X
-
)(b) At contact- because the E(X/X
-
) is
lower than E
F
the electrons
ow from semiconductor to the solution until reaching
charge equilibrium (c) After the contact - a positively charged region of width W
is established in the semiconductor and corresponding to that the band bending
occurs [Tan et al., 2007].
electron per available state is
1
2
. The Fermi energy for an intrinsic semiconductor
(extremely pure semiconductor) lies almost in the middle of the band gap. At
absolute zero, the probability of nding an electron in the valance band is 1 and 0
in the conduction band and would reach a
1
2
value in the middle of the gap. In a
semiconductor with impurity (doped) the Fermi level will be shifted towards the
edge of the majority carriers band. In an n-type semiconductor, dopant atoms are
donors (readily ionized to make an electron in the conduction band and a positive
charge on the dopant atom). Therefore, the Fermi level lies closer to the bottom of
the conduction band. Similarly, in a p-type semiconductor which dopants are the
acceptors (get ionized to produce holes in the valence band and negative charges on
the dopant site) the Fermi level is closer to the top of the valence band. The Fermi
level determines the tendency of the semiconductor to transfer charges to other
phases. In order to describe the distinct features of a semiconductor/electrolyte
interface at this point we will use a situation in which an n-type semiconductor
with a Fermi level (E
F
) is brought in to contact with an electrolyte containing a
27
redox pair (X/X
-
)(gure 3.1). To reach the interfacial charge equilibrium between
these two phases, the electron will
ow from a higher electrochemical potential
point to the lower one. In our specic example, the electron will
ow from the
n-type semiconductor to the electrolyte. The electron will be accepted by the
electrolyte which has a lower electrochemical potential. After neutralizing the
electrochemical potential dierence , the excess positive charge is formed in the
semiconductor and an excess negative charge in the electrolyte. The semiconductor
and the electrolyte are no longer electrically neutral phases. The net result of the
depletion of electrons at the surface of the semiconductor will be formation of an
electric eld between the bulk of the semiconductor and its surface. This electric
eld acts like a barrier against further electron transfer to the electrolyte. In this
positive space-charge region ( or a depletion region, since the majority carriers are
depleted from it) with width W an electrical energy is added to the electrons in
the band and thereby the energy of the bands has to change with distance, the
bands bend. The electric potential gradient at the interface will cause both the
conduction and valance bands to bend. The distance between the conduction band
and the valence band remains constant.
Now, if we consider a photo-induced water splitting on an n-type semiconduc-
tor surface (gure 3.2) the above discussion on band bending will come to play an
important role. When the semiconductor/electrolyte interface is under illumina-
tion with light (hv>E
g
), electron-hole pairs are generated in the semiconductor.
The presence of the electric eld at the space-charge region separates the photo-
generated charges. The electron moves toward the bulk of the semiconductor and
the hole moves toward the surface. Now if the solution contains a species, in the
case of water oxidation reaction OH
-
, which has an energy level above that of the
photogenerated hole at the surface, the electron transfer reaction can happen and
28
oxygen will evolve. The photogenerated electrons at the conduction band can be
transferred through an external wire connected to a counter electrode. In the metal
electrode, the electrons reduced water to evolve hydrogen. For the overall photo-
induced water splitting reaction, four electron-hole pairs are required to break two
molecules of water to one molecule of oxygen and two molecules of hydrogen [Sivula
et al., 2011].
E
X
h
+
e
-
Depletion layer
n-type
semiconductor
photoanode
E
F
H
2
O/O
2
H
+
/H
2
1.23 eV

e
-
Cathode
hν

4H
2
O

+
4e
-‐

2H
2

+
4OH
-‐

E
F
V
bias
e
-
4OH
-‐

+
4h
+

2H
2
O
+
O
2

E
c
E
v
H
2

O
2

Figure 3.2: Energy diagram for photoelectrochemical water splitting with an n-
type semiconductor photoanode. The involved processes are absorption of a photon
of light, electron-hole pair generation, transport of hole toward the interface of
photoanode-electrolyte, oxygen evolution at the interface, and nally transport of
electron through an external wire and reduction of water at the metallic cathode
surface [Sivula et al., 2011].
29
3.2 Experimental
3.2.1 Iron Oxide Thin Film Photoanode Preparation
The iron oxide photoanaodes are prepared via a hydrothermal method [Vayssieres
et al., 2001]. In this technique, we do the hydrolysis of an iron chloride precur-
sor solution at boiling temperature using a re
ux column. A mixture of 0.15 M
FeCl
3
.6H
2
O and 1 M NaNO
3
at pH=1.5 (adjusted by HCl) is poured in a 100
ml round
ask at T=100
0
C for variety of time ranging from 4 hour to 24 hour.
A clean FTO (F:SnO
2
) is used as a substrate (ultrasonically washed in acetone
followed by methanol and DI water). A portion of the substrate is masked with
a KAPTON tape to keep it clean for further electrical contacts. The substrate is
placed horizontally (the FTO coated side is exposed to the solution from top) in a
100 ml round bottom
ask prior to pouring the solution into the
ask. After the
hydrolysis reaction is accomplished, a light orang lm is formed on the substrate.
The coated lm is washed ultrasonically to remove all the unreacted chemicals
from its surface. A subsequent annealing at T=550
0
C for 2 hour is done for a
complete oxidation of iron which nally makes a dark orange lm. The thin lms
are characterized by X-ray diraction (XRD) and UV-vis spectroscopy.
3.2.2 Incorporation of Manganese Oxide to the Iron Oxide
Thin Film Photoanode as a Co-Catalyst
To study the catalytic eect of manganese oxide on iron oxide photoanode in
photoelectrochemical water splitting, the manganese oxide is deposited on the
iron oxide by dierent growth methods, spin coating and successive ionic layer
adsorption and reaction (SILAR). Also we use two dierent types of manganese
solution precursor in each method, Manganese chloride (MnCl
2
) and Manganese
30
Nitrate (Mn(NO
3
)
2
) to study any possible in
uence of the starting material on the
catalytic activity. These methods are explained in the following:
Spin Coating of a Mn Precursor Solution on an Iron Oxide Electrode:
A 50mM solutions of MnCl
2
and Mn(NO
3
)
2
are prepared in ethanol separately.
A 6 hour hydrothermally treated iron oxide electrode which has undergone the heat
treatment at 550
0
C is used as a substrate. Spin coating of the starting solution
on to the substrate is done at the speed of 3000 rpm for 1 minute repeatedly for 5
times. After air drying at room temperature, subsequent heat treatment at 550
0
C
for two hours is performed on the sample.
Successive Ionic Layer Adsorption and Reaction (SILAR) with High
Concentration of Mn Solution and Small Number of Cycles: Two types
of solutions are needed here, a precursor solution of Mn and an aqueous ionic
solution (NaOH). The manganese oxide thin lm deposition is completed in three
repeated cycles. For each cycle the prepared iron oxide electrode is immersed in a
0.1 M aqueous solution of MnCl
2
(or Mn(NO
3
)
2
) for 10 seconds , followed by blow
drying with compressed oxygen. Then the electrode is dipped in 1 M NaOH for
another 10 seconds, blow-dried with compressed oxygen.
Successive Ionic Layer Adsorption and Reaction (SILAR) with Low
Concentration of Mn Solution and Large Number of Cycles: The prepared
iron oxide electrode is dipped in a 50 mM aqueous solution of MnCl
2
(or Mn(NO
3
)
2
)
for 10 seconds, followed by immersing in a 0.1 M NaOH solution (10 s). After
each cycle the electrode is rinsed with DI water thoroughly to remove weakly
bound deposits from its surface. This cycle is repeated 10 times. The as-deposited
electrodes are heat treated at 300
0
C for 30 min.
31
3.2.3 Photoelectrochemical (PEC) Set Up
The prepared thin lms are fabricated into a photoanode by contacting the
uncoated area of FTO with a copper wire and silver paint followed by covering
up the whole contact area with epoxy glue. The photoanodes are tested for pho-
toelectrochemical activity using a Gamry Reference 3000 potentiostat and a three
electrode cell. A Ag/AgCl reference electrode (Ag/AgCl vs. NHE = +197 mV)
and Pt mesh counter electrode are used in all the experiments. A 0.1 M KOH
solution is used as an electrolyte in all the PEC tests. A Thorlabs high power
light source (HPLS-30-03) is used for illumination purposes. In gure 3.3 the PEC
set up with the illumination accessory are shown. Linear sweep voltammetery was
performed in darkness and under illumination. The applied potential was from
0 to 2.4 V at a scan rate of 10 mV/s . For the photocurrent measurements the
photoanodes are held in the electrolyte at a xed potential (0.9 vs. Ag/AgCl) for
120 seconds. We manually chopp the light by holding a dark aluminum foil in
front of the light source in 10 s intervals.
(a) (b)
Figure 3.3: (a) The Thorlabs high power light source and the lens (b) The PEC cell
including hematite working electrode, Ag/AgCl reference electrode and platinum
counter electrode in a 0.1 M KOH solution.
32
3.3 Hydrothermally Grown Iron Oxide Photoan-
ode
The hydrothermal growth of iron oxide has some advantageous over other meth-
ods of thin lm preparation. It is a simple and convenient technique to make a
reasonable photoactive material from abundant and nontoxic source of salt and
solvent (water). It does not need any complicated equipment and there is an easy
control of variables such as the solution's concentration and the temperature. Also,
it can be scaled up to be used for large surfaces [Dalle Carbonare et al., 2014b].
The preparation steps are shown schematically in gure 3.4. The X-ray diraction
(XRD) pattern of the iron oxide thin lm on the FTO substrate displays three
crystalline peaks identifying -Fe
2
O
3
structure (gure 3.5).
0.15 M FeCl
3
.6H
2
O
1 M NaNO
3
pH=1.5 (adjusted by HCl)
100°C, 6h
550°C, 2h
KAPTON
Tape
FTO
Air
β-FeOOH
Fe
2
O
3
Figure 3.4: Schematic illustration of hydrothermal synthesis of iron oxide thin lms
followed by heat treatment [Ling et al., 2014].
33
20 30 40 50 60 70
0
200
400
600
800
1000
1200
2 Θ (degree)
Intensity

Hydothermal Fe
2
O
3
Fe
2
O
3
*
*
*
*
Figure 3.5: X-ray diraction pattern of iron oxide thin lm hydrothermally grown
for 24 hr on FTO substrate. Then heat treated at 550
0
C for 2 hr. Three asterisk
signs denote the Fe
2
O
3
peaks. The rest is from FTO.
3.4 Photoelectrochemical Water Splitting by
Iron Oxide Photoanode
The hematite photoanode with the 6 hour synthesis time is tested for PEC water
splitting. The plots of current density vs. potential under the illumination and in
the dark are shown in gure 3.6. The iron oxide photoanode yielded a photocurrent
(0.7 mA cm
-2
at 0.95 V vs. NHE). Compared to the dark condition this indicates
photoactivity of iron oxide photoanode. Fe
2
O
3
is an n-type semiconductor with
a band gap of2.1. The position of conduction and valence band energy levels
in hematite with respect to the electrochemical potential of oxygen and hydrogen
evolution are illustrated in gure 3.7. While hematite has a lower valance band
34
with respect to the (H
2
O/O
2
), it has a lower conduction band compare to (H
2
/H
+
).
Therefore, an external electrical bias is required for reduction of water to hydrogen.
The photocatalytic oxygen evolution by Fe
2
O
3
involves absorption of photons of
light, electron-hole pair generation, traveling of the holes to the surface and nally
oxidizing water to make oxygen as represented by the following [Katz et al., 2012;
Peter and Upul Wijayantha, 2014];
hv +Fe
2
O
3
! 4 h
+
(Fe
2
O
3
) + 4 e

(3.3)
4 h
+
(Fe
2
O
3
) + H
2
O! 4 H
+
+ O
2
(3.4)
Light
Dark
Photocurrent
Figure 3.6: Current density vs. potential plots under dark (dashed line) and under
illumination (solid line) for iron oxide photoanode obtained after 6 h of hydrolysis.
35
E
X
h
+
e
-
Hematite photoanode
E
F
E
c
E
v
H
2
O/O
2
H
+
/H
2
1.23 eV

2.1 eV

4OH
-‐

+
4h
+

2H
2
O
+
O
2

Figure 3.7: Energy diagram for water oxidation on hematite surface. The relative
band levels of hematite with respect to the water oxidation and reduction reaction
is shown [Sivula et al., 2011].
3.4.1 The Thickness Eect on the Photoactivity
To explore the eect of thickness of iron oxide thin lms on the photocurrent
generation, the thickness of the lms were varied by changing the hydrolysis time
from 4 hr to 24 hr. The thickness values obtained by prolometry are 240, 120,
70 and 50 nm for 24, 11, 6 and 4 hour prepared hematite lms respectively. UV-
vis absorption data is presented in gure 3.8. From a bare FTO substrate to a
24 hr synthesized iron oxide thin lm the absorption is increasing by increasing
the thickness according to the Beer's law (equation 3.5). Where f is the fraction
of photon that will be absorbed, is the absorption coecient and l is the lm
thickness [Hamann, 2012].
f

1exp(l) (3.5)
The spectra's shape for all the hematite samples exhibited similar behavior with
a maximum absorbance around 400 nm and a decrease in absorption until the 600
36
nm wavelength. The photocurrent measurements at constant potential (0.95 V vs
NHE) for hematite photoanode with dierent thickness are shown in gure 3.9.
The 6 hour synthesized hematite shows the highest photocurrent while the 24 h
and 4 h samples exhibit lowest photo current. Thinner lms collect less light so
deliver less photocurrent. On the other hand, increasing the thickness of hematite,
more light gets absorbed into the bulk but at the same time the recombination
rate of electron and hole increases. The photo-generated holes have to travel a
longer distance to reach the surface of the electrode-electrolyte and participate in
water oxidation reaction.
400 500 600 700 800
0
0.5
1
1.5
2
2.5
Wavelength (nm)
Absorbance

24h (240 nm)
11h (120 nm)

6h (70 nm)

4h (52 nm)
FTO
Figure 3.8: UV-vis spectroscopy on hematite thin lms on FTO substrate at dif-
ferent thickness. The thicker the hematite the more light absorption it has.
37
0 20 40 60 80 100 120
0
0.2
0.4
0.6
0.8
1
1.2
x 10
−3
Time (s)
Current Density / A cm
−2

6h
11h
FTO
4h
24h
52 70 120 240
0
0.2
0.4
0.6
0.8
1
1.2
1.4
Thickness (nm)
photocurrent (mA) at 0.95 V vs. NHE
6h
4h
11h
24h
Figure 3.9: In
uence of lm thickness on measured photocurrent at constant volt-
age (0.95 V vs NHE) for hematite photoanodes. The highest photocurrent is
observed in the 6 h synthesized hematite with 70 nm thickness.
38
3.5 Applying a Manganese Co-Catalyst on the
Iron Oxide Photoanode
The need for large applied positive bias on hematite electrode makes the water
oxidation inecient. In order to increase the rate of water oxidation, we need
to apply a co-catalyst on hematite surface. This co-catalyst has to be a good
oxygen evolution catalyst by itself like cobalt-phosphate (Co-Pi), Iridium oxide,
manganese oxides and cobalt oxides [Barroso et al., 2012; Wiechen et al., 2014; Xi
et al., 2012]. Since the nature choose a manganese compound to oxidize the water
in PSII, manganese oxides stand as promising candidates to be the co-catalyst in
hematite photoelectrochemical cell.
As we have already discussed in the chapter 2, the catalytic activity of man-
ganese oxide in water oxidation is better compare to the iron oxide. The MnO
x
water oxidation onset potential is lower than Fe
2
O
3
. Hence, in this section of pho-
toelectrochemistry we intend to make a hybrid structure with Fe
2
O
3
photoanodes
and MnO
x
and investigate any possible photocurrent enhancement and overpoten-
tial reduction on the hematite photoanode. Because the quality of MnO
x
lm is
dependent on preparation method, we apply several dierent methods of deposit-
ing MnO
x
on Fe
2
O
3
and investigate any possible in
uences on the photoactivity
of the hybrid structure.
3.5.1 Photoactivity of Fe
2
O
3
+ MnO
x
Co-Catalyst Pho-
toanodes
After depositing MnO
x
on Fe
2
O
3
thin lm, our intention is to observe whether
the incorporation of MnO
x
(known as a better catalyst for water oxidation) would
enhance the activity of bare Fe
2
O
3
in the dark and in the light separately. The
39
electrochemical and photoelectrochemical measurements are performed on all the
hybrid lms. The bare Fe
2
O
3
electrode performance is overlaid on the plots as the
reference point. The ndings for all the three dierent ways of preparation (spin-
coating, SILAR high concentration and few cycles, and SILAR low concentration
and more cycles) are discussed in the following.
Photocurrent Measurements for the Spin-Coated MnO
x
on Fe
2
O
3
Electrode:
The onset potential of water oxidation in the dark experiment is lower in Fe
2
O
3
electrodes that are treated with MnO
x
via spin-coating technique. This is inde-
pendent of the type of precursor that we used (MnCl
2
) or Mn(NO
3
)
2
as shown in
gure 3.10 a) and b). However this is not true for the light experiment. Under
light illumination the treated electrodes with MnO
x
show almost no photoactivity
(the dashed and solid line are overlaying) whereas the bare Fe
2
O
3
electrode has a
photocurrent even at very low applied potential (< 0.4 V) with also a lower onset
potential. In part c) and d) of gure 3.10 the chopped photocurrent measure-
ments of the hybrid electrode versus bare hematite is illustrated .Again, there is
no signicant photocurrent for Fe
2
O
3
treated MnO
x
electrode.
Photocurrent Measurements for the SILAR Deposited(high concen-
tration of Mn solution and few cycles) MnO
x
on Fe
2
O
3
Electrode:
The onset potential of water oxidation in dark experiment is lower in Fe
2
O
3
treated MnO
x
via SILAR-few cycles technique. This is independent of the type of
precursor that we used (MnCl
2
) or Mn(NO
3
)
2
as shown in gure 3.11 a) and b).
The small bump in the curves could be related to some unnished redox processes
on the surface of the electrode since these electrodes have not been heat treated
after deposition of manganese solution. Under light illumination, however, the
treated electrodes with MnO
x
shows a small photoactivity (0.1 mA cm
-2
at 0.95 V
40
0.2 0.4 0.6 0.8 1
0
0.2
0.4
0.6
0.8
1
1.2
1.4
x 10
−3
Potential / V vs. NHE
Current Density / A cm
−2

H6hr −Dark
H6hr −Light
H6h −MnCl2 SpinCoated − Dark
H6h −MnCl2 SpinCoated − Light
light onset
dark
Fe
2
O
3
Fe
2
O
3
+ MnO
x
0.2 0.4 0.6 0.8 1
0
0.2
0.4
0.6
0.8
1
1.2
1.4
x 10
−3
Potential / V vs. NHE
Current Density / A cm
−2

H6hr −Dark
H6hr −Light
H6h −Mn(NO3)2 SpinCoated − Dark
H6h −Mn(NO3)2 SpinCoated − Light
light onset dark
Fe
2
O
3
Fe
2
O
3
+ MnO
x
20 40 60 80 100
0
0.2
0.4
0.6
0.8
1
x 10
−3
Time (s)
Current Density / A cm
−2

H6h −MnCl2 Spincoated
H6h
Fe
2
O
3
Fe
2
O
3
+ MnO
x
20 40 60 80 100
0
0.2
0.4
0.6
0.8
1
x 10
−3
Time (s)
Current Density / A cm
−2

H6h −Mn(NO3)2 Spincoated
H6h
Fe
2
O
3
+ MnO
x
Fe
2
O
3
a)
c)
b)
d)
Figure 3.10: a) and b) Current density vs. potential plots under dark (dashed
line) and under illumination (solid line) for Fe
2
O
3
photoanode (H-6h) and Fe
2
O
3
modied with MnO
x
(H-6h-MnCl
2
spin-coated) and (H-6h-Mn(NO
3
)
2
spin-coated),
respectively. c) and d) chopped light chronoamperometry measurements of Fe
2
O
3
photoanode and Fe
2
O
3
electrodes modied with MnO
x
.
vs. NHE) compare to the bare Fe
2
O
3
electrode (0.7 mA cm
-2
at 0.95 V vs. NHE).
This means a signicant suppression of the photocurrent that is even more clear
in the chopped photocurrent measurements in part c) and d) of gure 3.11.
Photocurrent Measurements for the SILAR Deposited(low concen-
tration of Mn solution and many cycles) MnO
x
on Fe
2
O
3
Electrode:
41
0.2 0.4 0.6 0.8 1
0
0.5
1
1.5
x 10
−3
Potential / V vs. NHE
Current Density / A cm
−2

H6hr −Dark
H6hr −Light
H6h −MnCl2 dipping3cycles −Dark
H6h −MnCl2 dipping3cycles −Light
light onset
Fe
2
O
3
Fe
2
O
3
+ MnO
x
0.2 0.4 0.6 0.8 1
0
0.5
1
1.5
x 10
−3
Potential / V vs. NHE
Current Density / A cm
−2

H6hr −Dark
H6hr −Light
H6h −Mn(NO3)2 dipping3cycles −Dark
H6h −Mn(NO3)2 dipping3cycles −Light
light onset
Fe
2
O
3
Fe
2
O
3
+ MnO
x
20 40 60 80 100
0
0.2
0.4
0.6
0.8
1
x 10
−3
Time (s)
Current Density / A cm
−2

H6h −MnCl2 SILAR3cycles
H6h
Fe
2
O
3
Fe
2
O
3
+ MnO
x
20 40 60 80 100
0
0.2
0.4
0.6
0.8
1
x 10
−3
Time (s)
Current Density / A cm
−2

H6h −Mn(NO3)2 SILAR3cycles
H6h
Fe
2
O
3
Fe
2
O
3
+ MnO
x
a)
c)
b)
d)
Figure 3.11: a) and b) Current density vs. potential plots under dark (dashed
line) and under illumination (solid line) for Fe
2
O
3
photoanode (H-6h) and Fe
2
O
3
modied with MnO
x
(H-6h-MnCl
2
SILAR-3cycles) and (H-6h-Mn(NO
3
)
2
SILAR-
3cycles), respectively. c) and d) chopped light chronoamperometry measurements
of Fe
2
O
3
photoanode and Fe
2
O
3
electrodes modied with MnO
x
.
Similar to our previous ndings, the modied hematite electrodes with MnO
x
using the SILAR-many cycles deposition method show lower onset potential in the
dark (gure 3.12 a) and b)). Under illumination, it has an anodic onset potential
shift (light onset shifts to the right). Unlike the previous modied electrodes
(spin-coated and SILAR few cycles), here the treated electrodes with MnO
x
via
SILAR-many cycles technique show a small photoactivity (0.15 mA at 0.95 V
42
vs. NHE) compare to the bare Fe
2
O
3
electrode (0.7 mA at 0.95 V vs. NHE).
This photocurrent eect is more obvious in the chopped light chronoamperometery
measurements in gure 3.12 c) and d).
0.2 0.4 0.6 0.8 1
0
0.2
0.4
0.6
0.8
1
1.2
1.4
x 10
−3
Potential / V vs. NHE
Current Density / A cm
−2

H6hr −Dark
H6hr −Light
H6h −MnCl2 dipping10cycles −Dark
H6h −MnCl2 dipping10cycles −Light
light onset dark
Fe
2
O
3
Fe
2
O
3
+ MnO
x
0.2 0.4 0.6 0.8 1
0
0.2
0.4
0.6
0.8
1
1.2
1.4
x 10
−3
Potential / V vs. NHE
Current Density / A cm
−2

H6hr −Dark
H6hr −Light
H6h −Mn(NO3)2 dipping10cycles −Dark
H6h −Mn(NO3)2 dipping10cycles −Light
light onset
dark
Fe
2
O
3
Fe
2
O
3
+ MnO
x
40 60 80 100 120
0
0.2
0.4
0.6
0.8
1
x 10
−3
Time (s)
Current Density / A cm
−2

H6h −Mn(NO3)2 SILAR10cycles
H6h
Fe
2
O
3
+ MnO
x
Fe
2
O
3
20 40 60 80 100
0
0.2
0.4
0.6
0.8
1
x 10
−3
Time (s)
Current Density / A cm
−2

H6h −MnCl2 SILAR10cycles
H6h
Fe
2
O
3
+ MnO
x
Fe
2
O
3
a)
c)
b)
d)
Figure 3.12: a) and b) Current density vs. potential plots under dark (dashed
line) and under illumination (solid line) for Fe
2
O
3
photoanode (H-6h) and Fe
2
O
3
modied with MnO
x
(H-6h-MnCl
2
SILAR-10cycles) and (H-6h-Mn(NO
3
)
2
SILAR-
10cycles), respectively. c) and d) chopped light chronoamperometry measurements
of Fe
2
O
3
photoanode and Fe
2
O
3
electrodes modied with MnO
x
.
43
3.6 Conclusion
In conclusion, we investigate the photoelectrochemical water oxidation on the sur-
face of hydrothermally grown Fe
2
O
3
photoanode. The photoanode with the opti-
mal thickness of 70 nm shows 0.7 mA cm
-2
photocurrent at 0.95 V vs. NHE.
To improve the water oxidation eciency at the surface of Fe
2
O
3
photoanode, we
coat it with a layer of MnO
x
co-catalyst. While in the dark experiment, the MnO
x
layer increase the kinetic of water oxidation and shows lower onset potential than
bare Fe
2
O
3
, it exhibits anodic onset potential shift and signicant suppression of
photocurrent in the light.
44
Chapter 4
Kinetic Studies of Oxygen
Evolution Reaction on the
Surface of Iron Oxide Electrode
Understanding the mechanisms of multielectron and multiproton electrochemical
reactions, particularly in the context of solar-to-fuel water splitting, is an outstand-
ing challenge. Historically, Pourbaix diagram are used to show the in
uence of
potential and pH on the thermodynamic stability of electrode-electrolyte systems.
These diagrams do not carry kinetic or mechanistic information, often restricting
their use to cases where the thermodynamic limit can be assumed.
In this section, we introduce and construct from experimental data two new
types of diagrams that demonstrate the kinetic variations of electrochemical reac-
tions as a function of potential and pH. These diagrams show the variation of the
electron transfer parameter and the proton reaction order in a wide range of
potential and pH (pH = 7 to 13). We present (pH,E) and (pH,E) for water
electrolysis on an iron oxide electrode.
45
4.1 Introducing the Kinetic Parameters: Elec-
tron Transfer Coecient () and Reaction
Order ()
The core of any electrode kinetic studies is the correspondence between the rate of
the processes occurring at the electrode interface and the electrochemical current
that is measured experimentally. Current has units of charge per unit time (A =
Cs
1
) so for a given electrode reaction one immediately knows the rate of the
charge consumption. The electrochemical current is a function of applied voltage
as the driving force for the reaction. Hence, nding the relationship between
the potential and current provides insight into the mechanistic pathways of the
electrode processes.
In the context of electrocatalysis (can be broadly dened as the catalysis of
redox reactions) which involves transfer of electrons and protons to and from the
surface of electrode, elucidation of kinetic parameters such as electron transfer
parameter () and the reaction order () of mechanistically signicant reactants
are the key points. In the following section, we will have a general overview of
these parameters.
4.1.1 Electron Transfer Parameter
The electron transfer parameter, , is originally introduced in electrochemistry
by Butler and Volmer[Gileadi, 1993]. It is dened as the fraction of the applied
potential aecting the reduction rate in an electrode reaction, with the remaining
fraction (1) accounting for the corresponding oxidation rate. For a single
46
cathodic step (O + ne! R), there is an exponential relationship between the
cathodic current I
c
and the applied potential E:
I
c
/e
(cnFE=RT )
(4.1)
In this equation (4.1),
c
is the cathodic electron transfer parameter, F is
the Faraday constant (96500 C/mol), n is the total number of transferred elec-
trons, R and T are the universal gas constant and thermodynamic temperature,
respectively. In this expression the reactant concentration at the electrode sur-
face is assumed to be constant (in the absence of depletion and diusion layer).
The dimensionless cathodic electron transfer parameter
c
is immediately obtained
from the slope of the plot of logI vs E :

c
=2:3

RT
nF
!
@ logI
@E
!
(4.2)
The derivative
@ logI
@E
is considered as the cathodic Tafel slope and it is a mea-
surable quantity. This indicates that the electron transfer parameter is simply the
reciprocal Tafel slope in dimensionless form [Gileadi, 1993]. For example, the value
of = 0:5 corresponds to the Tefel slope of 119 mV/decade at T = 298K for a
single electron transfer process. Similarly, for an anodic reaction (R! O + ne)
we have:
I
a
/e
(anFE=RT )
(4.3)

a
= 2:3

RT
nF
!
@ logI
@E
!
(4.4)
47
We should note that both equations 4.2 and 4.4 hold strictly for a single ele-
mentary step involving the simultaneous release or uptake of n electrons by the
electrode.
For a single-electron process, the electron transfer parameter is known as the
symmetry factor . It is a fundamental parameter in electrode kinetics and its
value is related to the shape of free-energy barrier to the position of the activated
complex along the reaction coordinate. To be specic, we will use a simplied
reaction coordinate diagram for a single electron oxidation reaction (R!O +e

)
and justify the denition of from there. The vertical distance between the
minima of two curves (R and O +e

) is the electrochemical Gibbs energy of the
electrode reaction, G = G
0
+F (EE
0
), where G
0
is the electrochemical
Gibbs energy at the standard potential E
0
and F (EE
0
) is the shift from this
value when an overpotential is applied. The activation barrier height G

for
this oxidation reaction is the vertical distance between the minimum of R curve
and the intersection of the two curves.
The driving force for electron transfer reaction is referred to the overpotential.
As the changes to more positive values, both the electrochemical Gibbs energy,
G, and the activation energy, G

, are altered. As can be graphically seen
(gure 4.1), the O +e

curve lowers with respect to the R curve resulting to the
new dashed curve, while the new activation energy get reduced by some fraction of
this overpotential (). The is the well known symmetry factor [Gileadi, 1993;
Guidelli et al., 2014-01-18; Surendranath and Nocera, 2011].
48
ΔG
ΔG
*
R

O + e
-
Rxn Coord

Gibbs energy

R O + e
-
-βη

η

Figure 4.1: Simplied diagram of electrochemical Gibbs energy along the reaction
coordinate for a single electron oxidation reaction R!O +e

.
4.1.2 Reaction Order
The reaction order in chemical kinetics is dened as the sensitivity of the reac-
tion rate () on the concentration of one species in the solution (C
i
), keeping the
concentrations of all others constant (at a constant pressure and temperature):
=

@ log
@ logC
i
!
T;P;C
j 6=i
(4.5)
In electrochemistry, the reaction order is determined in a similar way but in
addition to keeping theT ,P andC
j6=i
constant the potential (E) needs to also be
held constant. We recall that the current is a very sensitive measure of the rate
of any electrochemical reaction therefore the following expression is the electro-
chemical reaction order with respect to the concentration of species C
i
[Gileadi,
1993]:
49
=

@ logI
@ logC
i
!
E;T;P;C
j 6=i
(4.6)
The partial derivatives are very important in understanding the mechanism of
electrocatalytic reactions. We can only access to the slowest step in the electrocat-
alytic cycle in steady-state measurements. Hence, these derivatives in equations
4.2, 4.3 and 4.6 are associated with the potential and concentration dependence of
the rate limiting steps.
4.2 Experimental
This kinetic study performs on the iron oxide electrode that prepared by the
PMOD technique as discussed in section 2.1.1. The same electrochemical set up
as explained in section 2.1.2 is used. The electrolyte solution is 0.1 M KOH and
its pH is adjusted by adding a dilute HCl solution. To avoid in
uences from the
specic adsorption of multivalent anions, no agent for pH buering is added. Since
reaction order is a derivative, its continuous representation is limited by the num-
ber of discrete points along the pH axis. In our experiment, current-voltage curves
were obtained for 14 values of pH between pH=7 to pH=13. The uncompensated
resistance correction has been performed on all the measured current-voltage data
with the following method that will be explained here.
Ohmic Drop Correction
The resistance of the electrolyte solution and the electrode lm is evaluated using
a common method as outlined by Krstajic and Trasatti [Krstajic and Trasatti,
1998]. It starts from the Tafel law and assumes that the solution ohmic resistance
50
is constant independent of current (equation 4.7). Taking the derivitive of V with
respect to I gives the relationship in equation 4.8.
V

a + b lnI + R
u
I (4.7)
dV
dI
=
b
I
+ R
u
(4.8)
Now, the uncompensated resistance appears as the intercept of a plot of
dV
dI
versus
1
I
. Where dV and dI are the change in potential and current for two con-
secutive data points respectively and I is the mean current for the same two data
points. When current I
ows through the cell, this uncompensated resistance R
u
contributes to the measured voltage by an amount I R
u
. Hence at large currents,
the measured voltage is in
uenced by the uncompensated resistance giving rise to
linear I-V curves. An example of this ohmic drop correction and determination of
the R
u
value for iron oxide electrode in pH=7 is shown in gure 4.2. The value of
R
u
is determined to be 70
. For various values of pH and also dierent elec-
trodes of iron oxide, the R
u
value nds to be also close to 70
. Therefore, the
uncompensated resistance mostly arises from the lm rather than the electrolyte
solution. To account for this resistance, for each pair of I-V data, the value IR
u
is
subtracted from the voltage V, giving rise to the corrected voltage. The obtained
value for manganese oxide electrodes is 100
.
4.3 Constructing the 2-D Current-Voltage Plots
The resistance-compensated I-E curves for pH values ranging from 7 to 13 are
shown in gure 4.3. Although dierences in the behavior of I-E curves between
the acidic and basic sides are visible, it is hard to quantify the pH dependence of
51
0 200 400 600
40
60
80
100
120
1/I (A
−1
)
dV/dI (VA
−1
)

pH=7 (unbuffered) iron oxide electrode
Figure 4.2: An example of ohmic drop correction by extrapolation of the
dV
dI
vs.
1
I
the value of R
u
on the iron oxide electrode at pH 7 is70
.
the I-E curves from this gure. In particular, the dierences in the slop of I-E
curves, and the relatively
at region between pH10-12 demand explanation. For
that reason, we prepare the two-dimensional representation of the I-E curves as
a function of potential and pH. The plot of logI(pH;E) for water oxidation on
a hematite thin lm electrode is shown in Figure 4.4. The thermodynamic line
for water stability (the Pourbaix line) with the Nernstian slope of 59 mV/pH is
overlaid on the plot for reference. An immediate insight can be already gleaned
from this diagram. While on the basic side the onset potential follows a Nernstian
slope, on the acidic side it does not show any dependence on pH. This feature, along
with others, will be explained in section 4.5 in the context of surface protonation
equilibrium and dierences in electron transfer parameter in the basic and acidic
sides. We emphasize that in this work the words acidic and basic do not mean
below and above pH=7, but rather refer to the two limits of the pH range that we
have studied (pH = 7 and pH =13).
52
0.8 1 1.2 1.4 1.6 1.8
0
0.002
0.004
0.006
0.008
0.01
0.012
0.014
0.016
Potential / V vs. NHE
Current Density / A cm
−2

pH 13
12.8
12.6
12.3
12
11.8
11.5
10.9
8.4
7.7
7.2
7
10
9.2
Figure 4.3: Measured linear sweep voltammograms after compensation for lm
resistance at dierent pH values from pH=7 to pH=13. Although dierences in
the behavior of I-E curves as a function of pH exist, it is hard to visualize the
in
uence of pH using this representation. This is better addressed in the two-
dimensional representation in gure 4.4.
4.4 Constructing the(pH,E) and(pH,E) Plots
We construct the kinetic diagrams by taking the gradient of thelogI along the pH
and potential axes, and calculate (pH;E) and (pH;E) according to equations
4.6 and 4.2. In calculating the derivatives, we set the onset current to 0.1 mA and
ignore the regions of low overpotential where the current is too low for a derivative
to be meaningful. The resulting plots are shown in gures 4.5 and 4.6.
The proton reaction order plot in gure 4.5 shows that at low pH values, the
electrochemical current has no sensitivity to pH ( = 0). However, at high pH
values a clearly non-zero value of is observed. This indicates that the reaction
mechanism in the basic region is dierent from the acidic region. The highest
value of reaction order is observed in the intermediate region ( 1:2) and
corresponds to the region where a switch between the two mechanisms is expected.
To understand these observations, rst we brie
y review a model that accounts for
53
pH
Potential / V vs. NHE

8 9 10 11 12 13
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
−3.5
−3
−2.5
−2
Log(i/A cm
−2
)
Overpotential
Overpotential
Water Stability Line
Figure 4.4: The electrochemical current logI(pH;E) for water splitting on the
surface of iron oxide as a function of pH and potential. The Pourbaix line for
water stability (solid line) is overlaid for reference. The double arrows indicate the
overpotential for water oxidation, and demonstrate its variation as a function of
pH. The hashed region corresponds to no data, since the compensated voltage axis
is dierent for various values of pH (see experimental section).
both surface and bulk contributions of protons to the the reaction rate. Then we
explain the observed variation of based on this model. It is known that even
non-adsorbed molecules within a few monolayers of an interface do not behave like
bulk molecules and often assume anisotropic structure. This work is not concerned
with distinguishing that type dierence, and the word bulk here is used to refer to
all non-adsorbed species.
54
pH
Potential / V vs. NHE

8 9 10 11 12
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
−1
−0.8
−0.6
−0.4
−0.2
0
ρ
Basic Mechanism
≠0
Transition Between
Mechanisms
Acidic Mechanism
=0
Figure 4.5: The proton reaction order plot (pH;E) for water splitting on the
surface of iron oxide. The reaction order shows the sensitivity of the electrochem-
ical current to changes in pH as dened in equation 4.6. In the acidic region, the
electrochemical current, and correspondingly the water oxidation reaction rate, is
practically insensitive to pH, while for the basic mechanism6= 0. The largest sen-
sitivity to pH is observed in the region where a switch between the two mechanisms
is expected. This is explained in the text using a surface protonation model.
4.5 Langmuir Surface Adsorption Model to
Interpret the Observed
The rate of a reaction that consumes an adsorbed species is expected to depend on
the coverage function of the adsorbed species, rather than its bulk concentration.
In this case, as in conventional kinetics, a rate law of the form I /
n
can be
written, where n depends on the stoichiometry of the adsorbed species [Gileadi,
1993; Bard and Faulkner, 2001; Tilak and Conway, 1992]. For example, if a reaction
consumes adsorbed OH
{
with a stoichiometry of 1, its rate will depend on surface
coverage of OH
{
to the rst power. The coverage, in turn, is a function of the bulk
concentration of that species via the commonly used adsorption functions, such
55
pH
Potential / V vs. NHE

8 9 10 11 12 13
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
α
Acidic Mechanism
Transition
Between
Mechanisms
Basic Mechanism
Figure 4.6: The electron transfer coecient plot (pH;E), with mechanistically
distinct regions. The electron transfer coecient is a dimensionless measure of
the sensitivity of the reaction rate to applied potential as shown in equation 4.2.
A noteworthy feature is that in the basic region is distinctly larger than in the
acidic region. The dotted arrow in the intermediate region shows a reaction that
starts with the basic mechanism, but due to electrochemical generation of protons
near in the interface, switches to acidic mechanism, displaying a characteristic
smaller . These features are further explained in the text.
as the Langmuir or Temkin isotherms [Tilak and Conway, 1992]. An adsorption
function(pH) that shows the variation of surface coverage of a species with respect
to bulk pH can be used to re-write equation 4.6 as:

ads
=

@ logI
@ (log)
!
E

@ (log)
@pH
!
E
(4.9)
The above is best understood in the limiting cases of adsorption. When the surface
is completely saturated with an adsorbed species (! 1, as in the
at region of the
Langmuir isotherm), any small change in the bulk pH will not change the surface
coverage signicantly. The second factor in the above equation, and consequently
, vanishes. For example, in the pH range where the surface is fully covered with
56
OH
{
, a dierential change in the bulk pH will not in
uence the surface coverage of
OH
{
any further. In the limit of low coverage ! 0, lateral interactions between
adsorbates vanishes and adsorption models converge to the low coverage region
of the Langmuir isotherm. In this limit the coverage increases linearly with
increasing bulk concentration and the second factor in equation 4.9 becomes unity.
For intermediate coverage, a sublinear dependence of coverage on concentration is
expected. In that case, the value of the second factor is between 0 and 1 and can, in
principle, be determined from the adsorption isotherm model. When the reaction
involves both an adsorbed species x and the same species in the bulk (such as the
reaction of bulk OH
{
with adsorbed OH
{
), the rate law will depend on both the
surface coverage and the bulk concentration [x] as I/
n
[x]
m
. In this case the
the reaction order is:
=
ads
+
bulk
(4.10)
where
ads
is given by equation 4.9, and
bulk
=m.
4.6 Surface Adsorbed Intermediates Involved in
the OER
For a metal oxide in equilibrium with an aqueous electrolyte the following surface
protonation equilibrium is justied in the absence of potential:
MOH
2
MOH

+H
+
(4.11)
whereM stands for a surface active site, namely an available iron ion[Daghetti
et al., 1983]. At low (high) pH values, the equilibrium is shifted far to the left (right)
and the surface is covered completely with adsorbed water (hydroxyls). Thus the
57
initial ingredients of the oxidation process at low and high pH are adsorbed water
and adsorbed hydroxyls respectively. Figure 4.7 presents a scheme that follows the
fate of the adsorbed water or hydroxyl species as they lose electrons and protons
and become surface MO, which is considered an important intermediate of water
oxidation. We point out that various proposed schemes for water oxidation mech-
anisms in the acidic and basic limits have existed for several decades [Damjanovic
et al., 1966; Lodi et al., 1978; Mehandru and Anderson, 1989; Bockris, 1956]. The
scheme in gure 4.7 is the one that reasonably explains both the and data
observed in this work, both in the acidic and basic limits.
4.7 Interpreting the Reaction Order at dier-
ent pH regions
4.7.1 Acidic Pathway
In the hypothesized acidic pathway, as the potential is increased and an electron is
transferred into the electrode, the adsorbed water (MOH
2
) is concurrently depro-
tonated. The generation of the highly acidic and unstableMOH
+
2
is avoided. Such
concerted electron-proton transfer (CPET) step is hypothesized for heterogeneous
reactions previously [Yamaguchi et al., 2014; Koper, 2013c; Bediako et al., 2013a;
Schrauben et al., 2012; Venkataraman et al., 2008]. Prior to electron transfer,
a surface bound proton is likely already hydrogen bonded to a bulk hydroxyl or
water. Transfer of electrons into the electrode renders the surface bound proton
acidic, such that the proton is immediately lost to the bulk deprotonating species.
A fully concerted pathway means that as the electron transfer approaches the tran-
sition state, concurrently the hydrogen bonding of the proton to the bulk species
58
strengthens and the proton transfer also approaches transition state. A second
electron transfer is also associated with proton transfer, giving rise to the surface
MO species.
At low pH range, this path is not sensitive to the initial adsorbed MOH
2
coverage, since the surface is fully covered and any small change in pH will not
change the surface coverage appreciably. The second term in equation 4.9 is zero in
this case, and the reaction rate loses its sensitivity to the bulk pH changes
ads
= 0.
Since bulk water is the deprotonating agent, no bulk reaction order will arise and
an overall proton reaction order of = 0 is expected. This is exactly observed in
the acidic region of the plot of (pH;E) in gure 4.5.
4.7.2 Basic Pathway
In the basic regime, the surface is fully hydroxylated at zero applied potential.
Only the rst electron transfer is associated with deprotonation, and the second
electron transfer yields MO. In this regime, the deprotonating agent is a bulk
hydroxide. Hence, even when the surface is fully saturated with hydroxyls and

ads
= 0, a non-zero reaction order with respect to bulk hydroxides or bulk pH is
expected. This matches our observations, where is distinctly non-zero just above
the onset potential in gure 4.5 in the basic region. We also point out that the
slope of the onset potential in the basic range is very close to the Nernstian value
of 59 mV/pH, suggesting that bulk hydroxides play a role in lowering the kinetic
barrier for the basic path.
4.7.3 Intermediate Pathway
In the intermediate pH values, the surface protonation equilibrium (equation 4.11)
is not favored heavily to either of the two extremes. The surface is covered with
59
Increasing Deprotonation
Increasing Electron
Transfer into Electrode
MOH
2
MOH

-

MOH

MO

-

MO

Figure 4.7: Hypothesized paths for two-electron, two-proton generation of surface
MO, where M represents a surface active site (i.e. an iron ion). At zero applied
potential, the pH determines the extent of the equilibrium MOH
2
MOH

,
and hence the starting ingredients for oxidation. These paths provide a reasonable
explanation of the observed and in the acidic and basic limits.
both hydroxyls and adsorbed water. Unlike the pH extremes, the second factor
in equation 4.9 is not zero, and thus a large
ads
is expected. In fact, a dieren-
tial change in the pH signicantly changes the population of the surface adsorbed
species, and tips the balance between acidic and basic mechanisms easily. There-
fore, a large sensitivity to pH is expected in this range, which is indeed observed
in our data as 1:2 in gure 4.5.
4.8 Electron Transfer Parameter at dierent
pH regions
Now we turn to the empirical electron transfer plots. As we discussed in section
4.1 for a single-electron process, the electron transfer parameter is also known as
the symmetry factor . It is interpreted as the proportionality factor that relates
60
the thermodynamic stabilization of the transferred electron in the electrode to
the electron transfer barrier height (i.e. the energy of the transition state). For
example, = 0:5 means that if the electrode potential is raised by 0:1V , the
electron transfer barrier will reduce by 0:05V . It has been given a relatively simple
geometric interpretation, which is related to the crossing angle of the free energy
surfaces for electron transfer [Bard and Faulkner, 2001]. While this is reasonable
for single electron processes, the potential dependence of multielectron processes
is more complicated. A multistep kinetic model [Bard and Faulkner, 2001; Bockris
and Reddy, 1973; Matthews and Bockris, 1971] for multielectron transfer processes
assumes a distinct rate-determining step (RDS) and pre-equilibrium prior to RDS.
Following standard approaches the electron transfer parameter is derived as =

+r. Here,
is the number of electrons transferred prior to the RDS, is the
number of times that the RDS is repeated per catalytic cycle, r is the number of
electrons transferred at the RDS, and is the familiar symmetry factor. Although
the above relation for is known for several decades, it has not yielded unique
and unambiguous identication of the mechanism of complicated reactions such as
water oxidation. The above relation still holds qualitative value and we will use it
in that capacity. If we assume that the RDS is the oxygen-oxygen bond formation,
as is suggested in several works [Bediako et al., 2013b; Norskov et al., 2002], then
both the acidic and basic pathways will have the same number of steps prior to
the RDS and will have the same
and the same . Then the observed dierences
in may be attributed to dierences in the symmetry factor for the acidic and
basic regions.
Several previous works [Venkataraman et al., 2008; Ludlow et al., 2010; Cos-
tentin et al., 2006] suggest that the electron transfer parameter for a concerted
electron-proton transfer reaction is dierent from that of single electron transfer.
61
In particular smaller values of were measured and justied for the CPET pro-
cess[Ludlow et al., 2010]. Although it is much desirable, the exact simulation of
the parameter for the concerted and stepwise mechanisms is beyond the scope
of this work. It is reasonable to postulate that when a single step involves both
electron and proton transfer, the transition state will correspondingly have partial
electron and proton transfer character. Variation of the electrode potential has
a larger in
uence on the electron transfer character of the transition state. Thus
the transition state of one electron transfer is expected to be more sensitive to
electrode potential variation, compared to the the transition state of a step that
involves both electron and proton transfer. Therefore, a smaller is expected for
the steps that involve both electron and proton transfer, compared to the steps
that are only electron transfer. In the acidic regime, two electron-proton transfer
steps are required, while in the basic regime only one of the steps is combined pro-
ton and electron transfer. Correspondingly, a smaller is expected in the acidic
side compared to the basic side. Our observations are in conformity with this
explanation as shown in gure 4.6. A smaller value of 0:1 0:2 is observed in
the acidic region, while in the basic region is as high as 0:7.
A valuable feature of the continuous representation of and in the two-
dimensional plot is easy identication of the transition in mechanism between the
two regimes. This is most prominently seen in the intermediate pH values near
pH=10.5. In this region as the potential is scanned to higher values (see dotted
line in 4.6), the reaction starts out with the basic mechanism. Since the reaction
produces protons near the interface, the interfacial pH gradually changes to acidic,
and the intermediate behavior is observed. Finally, after a certain potential, the
reaction rate is large enough that the produced protons near the surface enforce a
fully acidic mechanism.
62
4.9 The Mass Transport Eect on the Mechanis-
tic Scheme
In the regions of large current, mass transfer of ions limits the current and com-
plicates kinetic studies. The current no longer obeys an exponential relation with
respect to potential as in equation (I/a

H
+
e
(FE=RT )
). We have compensated the
measured voltage for resistance, and have identied the electrode thin lm as the
main contributor to the resistance. However, the in
uence of mass transfer is still
identiable in the reported data, especially in the intermediate pH range. The
protons that are released as reaction products, acidify the interface to pH values
far lower than that of the bulk. This enforces an eective acidic mechanism with
the characteristic smaller , as observed for large potentials in gure 4.6. As the
bulk pH is made more basic, surface acidication due to the reaction is more eas-
ily compensated by bulk hydroxides. Thus at higher pH a larger electrochemical
current will be required to overwhelmingly acidify the interface. This is evident
in the boundary between the acidic and the transition region in gure 4.6, which
shifts to larger potentials as the pH is raised. We have performed experiments
under stirring conditions (gure 4.8). In that case, we observe a shift of the inter-
mediate region to lower pH by about 0.5 pH value, with all of the salient features
of other regions remaining the same. This observation is consistent with the fact
that any nite current acidies the region near the interface with respect to the
bulk, resulting into an eective shift of the pH axis of the diagram. Thus, the
interpretation of the horizontal axis of such diagrams should always be performed
with the experimental conditions in mind.
63
pH
Potential / V vs. NHE

8 9 10 11 12 13
0.5
1
1.5
2
2.5
−4
−3.5
−3
−2.5
−2
Log(i/A cm
−2
)
thermodynamic water stability line
overpotential
(a)
pH
Potential / V vs. NHE

8 9 10 11 12 13
0.5
1
1.5
2
2.5
−0.2
−0.1
0
0.1
0.2
0.3
0.4
0.5
α
0.5 pH
(b)
pH
Potential / V vs. NHE

8 9 10 11 12
0.5
1
1.5
2
2.5
−1
−0.5
0
0.5
1
ρ
0.5 pH
(c)
Figure 4.8: The kinetic diagrams under stirring condition. (a) The electrochemical
current logI(pH;E). (b) The electron transfer coecient diagram (pH;E) and
(c) proton reaction order diagram(pH;E). The characteristic intermediate region
is shifted to lower pH by about 0.5 pH value.
64
4.10 Participation of Ions in the Reaction Mech-
anism
It is well-known that the choice of buer can in
uence the mechanism either via
adsorbed buer anions (specic adsorption), or bulk buer anions acting as proton
acceptors or donors [Minguzzi et al., 2012; Surendranath et al., 2010; Kushner-
Lenho et al., 2013]. In this study, we have chosen a relatively simple electrolyte
composed of KOH and HCl to adjust the pH from high values to lower values, and
our discussions only apply to this choice of electrolyte. We have also performed
experiments with multivalent anions (phosphate) electrolytes. In that case, some
of the essential features of the diagrams remain the same. In particular, we have
observed the 59 mV/pH slope for the onset potential in the basic side and that
the coecient in the basic side is larger that of the acidic side. However, the
ranges and shapes of the mechanistic regions dier from the case of the KOH and
HCl system (gure 4.9). Their interpretation requires accounting for adsorption
of the buer anions on the surface, and the buer protonation equilibrium. It is a
matter of future studies.
4.11 Conclusion
In conclusion, we have evaluated the electron transfer parameter and the proton
reaction order as continuous variables in the pH and potential space for water
oxidation on hematite. This treatment of and is useful and insightful since it
gives a global view of the sensitivity of the reaction to potential and pH. Simul-
taneous consideration of and sheds light on the basic and acidic pathways of
formation of surfaceMO species, which is considered a precursor for oxygen bond
65
pH
Potential / V vs. RHE

8 9 10 11 12 13
0.5
1
1.5
2
2.5
−3.8
−3.6
−3.4
−3.2
−3
−2.8
−2.6
−2.4
−2.2
−2
Log(j/A cm
−2
)
(a)
pH
Potential / V vs. NHE

8 10 12
0.5
1
1.5
2
2.5
0
0.1
0.2
0.3
0.4
0.5
0.6
α
(b)
pH
Potential / V vs. NHE

8 9 10 11 12
0.5
1
1.5
2
2.5
−0.8
−0.6
−0.4
−0.2
0
ρ
(c)
Figure 4.9: The kinetic diagrams for an experiment where phosphoric acid, instead
of hydrochloric acid, was used for pH adjustment of the electrolyte. (a) The
electrochemical currentlogI(pH;E). (b) The electron transfer coecient diagram
(pH;E) and (c) proton reaction order diagram (pH;E).
66
formation. The evaluated and for the acidic and basic regions are reasonably
justied for the two pathways shown in gure 4.7. We hope that the hypothesized
pathways inspire future experimental studies such as surface specic sum-frequency
generation, in-situ EXAFS, and transient FTIR measurements, for identication
of intermediates.
We anticipate that our work will inspire further investigation in two directions.
First, we hope that continuous representation of kinetic parameters, especially in
the potential-pH space, will be used more extensively in investigations of hetero-
geneous catalysis reactions as complementary tools. Putative mechanisms should
be checked to consistently explain observed values of and over a large range
of potential and pH. Second, we hope that theoretical models of surface equilibria,
such as adsorption isotherms, can benet from the observations in these diagrams.
67
Chapter 5
Kinetic Studies of Hydrogen
Evolution Reaction on the Surface
of Metallic/Semiconductor
Electrodes
The empirical electron transfer coecient is a valuable electrochemical observ-
able which bridges the thermodynamics and kinetics of redox reactions. For reac-
tions that involve protons, the value of is expected to be pH dependent. However,
even for the simplest redox processes (i.e. HER) the nature of this dependency
remains unclear.
Towards clarifying this problem, in this chapter we follow two goals. First
we calculate the electron transfer coecient and its pH-dependence based on a
two-dimensional potential energy surface model for proton reduction. Second, we
report our experimentally measured for hydrogen evolution on several electrode
materials over a wide pH range.
5.1 Experimental
Electrochemical data were obtained using a Gamry Reference 3000 potentiostat
and a three electrode cell. A Ag/AgCl reference electrode (Ag/AgCl vs. NHE =
68
+197 mV) and Graphite rod counter electrode were used in all the experiments.
The working electrodes were seven dierent materials as follows: Platinum gauze
(52 mesh woven with surface area of 8.3 cm
2
-Alfa Aesar), Ti and Ni (surface area
: 1cm
2
- MTI Corporation), Gold (as a thin lm on silicon, surface area : 1.9 cm
2
- LGA Thin Films, Inc.) Iron pyrite- FeS2 (surface area : 1.5 cm
2
-Geollector),
Copper(surface area : 1 cm
2
- Metalliferous), and Glassy carbon (surface area : 1
cm
2
, Glas 11 Grade-SPI supplies). The electrolyte solution was 0.1 M KCl and
its pH was adjusted by adding a dilute HCl solution. To avoid in
uences from
the specic adsorption of multivalent anions, no agent for pH buering was added.
The cell was purged with N
2
gas prior to each scan and then was closed to the
ambient air. Current-voltage curves were obtained for at least 12 to 15 values of
pH between pH=6.5 to pH=1.5. The indicated potentials are always referenced
to the normal hydrogen electrode. Linear sweep voltammetry experiments were
carried out at 10 mV/s scan rate, between 0 and -2 V versus Ag/AgCl electrode.
To construct the (pH;E) diagram we took the gradient of the logI along the
potential axes for each pH, and calculated (pH;E). In determining the deriva-
tives, we set the onset current to 0.1 mA and ignored the regions of low overpoten-
tial where the current is too low for a derivative to be meaningful. Also to evaluate
the average value of for each electrode material at each pH value, we calculated
the average value of s over a range of 50 mV right after the onset current (0.1
mA). The error in the value of depends on several parameters, including the
value of current at which is measured. Our average estimate for the error in
is in the order of 0.05. For the rotating disk electrode measurements, we used
Platinum, Gold and Glassy Carbon rotating disk electrodes (Gamry-E6 Series) all
with surface area of 0.196 cm
2
.
69
5.2 A 4-State Potential Energy Surface for Pro-
ton Reduction:
Schmickler and Koper have worked on a two-dimensional potential energy sur-
face for the electrochemical electron-proton transfer and, in particular, the proton
reduction reaction.[Grimminger and Schmickler, 2007; Grimminger et al., 2005;
Koper, 2013b,a] Other models for the electrochemical electron-proton transfer also
exist and have their own value [Costentin et al., 2005; Costentin, 2008; Venkatara-
man et al., 2008; Hammes-Schier and Stuchebrukhov, 2010] . We will use the
Schmickler-Koper approach since it introduces experimentally controllable param-
eters into the model in a convenient way. We will calculate electron transfer
coecients based on this model. For notational consistency, and in particular
to emphasize the in
uence of pH on reaction paths, we will brie
y introduce the
model. Then we will proceed with identifying the stepwise and concerted limits of
proton transfer according to this model and calculating their respective electron
transfer coecients.
We consider the in
uence of pH and potential on the thermodynamics of the
Volmer reaction.
M + H
+
+ e

! MH (5.1)
Here, M is a surface site that catalyzes the reaction of a proton (abstracted from
the electrolyte) with an electron to create chemically adsorbed hydrogen MH. The
thermodynamics of this reaction is clearly a function of the electrode potential (i.e.
chemical potential of electron) and the electrolyte pH (i.e. the chemical potential
of protons in the electrolyte). It is conventional to consider four states for the
reaction 5.1 as shown in gure 5.1. States 1 and 4 are the reactants and products,
while states 2 and 3 are proton transfer and electron transfer intermediates. We
70
assume the potential to be E which is relative to an arbitrary reference. Soon we
will choose the equilibrium potential as the reference.
e
-
H
+ M

H
+ M
-
e
-
H
+
M

H

M

state 1
state 2
state 3
state 4
Figure 5.1: Four states participating in the Volmer reaction
In assigning thermodynamic free energy to each one of the four states, we
follow the conventional understanding of the electron and proton free energies in
the electrode and in the electrolyte respectively. The free energy of electron in the
electrode is proportional to the potential E as G(e

) =eN
A
E =FE, where
e is the charge of the electron and F is the Faraday constant. The free energy
of solvated protons G(H
+
) is related to pH via G(H
+
) = G
0
(H
+
) 2:3RTpH.
At pH = 0, we set the G
0
(H
+
) to be zero as a reference point. Then we have
G(H
+
) =2:3RTpH. Thus, the free energies of the four states shown in gure
5.1 are:
G
1
= FE +G(M) 2:3RTpH (5.2)
G
2
= FE +G(MH
+
) (5.3)
G
3
= G(M

) 2:3RTpH (5.4)
G
4
= G(MH) (5.5)
71
For convenience, we nd the potential that brings the reaction to equilibrium
G(1) =G(4) at a given pH.
E
eq
=
1
F
(G(M)G(MH) 2:3RTpH) (5.6)
Using this value as a reference, we dene the overpotential as
=EE
eq
(5.7)
Next, we consider the protonation free energies for M and M

from the fol-
lowing two reactions
M + H
+
! MH
+
(5.8)
M

+ H
+
! MH (5.9)
and dene:
G
prot
=G(MH
+
)G(M) + 2:3RTpH (5.10)
G

prot
=G(MH)G(M

) + 2:3RTpH (5.11)
Equations 5.7, 5.10 and 5.11 can be inserted into the free energies in equations
5.2-5.5. Given that G(MH) is common to all energies, it is most convenient to
consider relative energies and subtract it from all, leading to:
G
1
= F (5.12)
G
2
= F + G
prot
(5.13)
G
3
= G

prot
(5.14)
G
4
= 0 (5.15)
72
The above is not dierent from what we started from (equations 5.2-5.5). However,
it is cast in a convenient way using only three intuitive quantities: the overpotential
, the protonation free energies of the neutral G
prot
, and negatively charged
G

prot
surface site.
Next, we consider transitions between the 4 states described above. To describe
transitions, a potential energy surface that connects these states must be con-
structed. Schmickler [Grimminger et al., 2005; Grimminger and Schmickler, 2007]
has calculated adiabatic potential energy surfaces that connect the four states
described above. The distance of a proton from the surface represents the proton
transfer coordinate and the electronic charge on the surface site represents the
electron transfer coordinate. Then the energies of the four states are smoothly
connected with each other through variations along these coordinates, producing a
two-dimensional potential energy surface. The resulting surface is analogous to the
Marcus electron transfer parabolas in two dimensions, which have been discussed
in the literature [Venkataraman et al., 2008]. We shall refer to the coordinate
aligned with proton and electron transfer as q
p
and q
e
respectively. Built on this,
Koper [Koper, 2013b] has considered the four states as diabatic surfaces, each with
a minimum energy shown in equations 5.12-5.15 and quadratic increase in energy
as a function of q
p
and q
e
.
G
1
(q
e
;q
p
) =
e
q
2
e
+
p
q
2
p
+ 2

q
e
q
p
F (5.16)
G
2
(q
e
;q
p
) =
e
q
2
e
+
p
(q
p
1)
2
+ 2

q
e
(q
p
1)F + G
prot
(5.17)
G
3
(q
e
;q
p
) =
e
(q
e
1)
2
+
p
q
2
p
+ 2

(q
e
1)q
p
G

prot
(5.18)
G
4
(q
e
;q
p
) =
e
(q
e
1)
2
+
p
(q
p
1)
2
+ 2

(q
e
1)(q
p
1) (5.19)
73
The coecient
e
is the the familiar reorganization energy along the electron
transfer coordinate, and
p
is the natural extension of that for proton transfer.
The parameter

is referred to as the solvent overlap parameter and accounts
for those types of motion in the environment that in
uence both electron and
proton transfer. The minima corresponding to equations 5.12-5.15 are located at
fq
e
=q
p
= 0g,fq
e
= 0; q
p
= 1g,fq
e
= 1; q
p
= 0g andfq
e
=q
p
= 1g respectively.
To construct an approximate adiabatic PES, the coupling between the diabatic
surfaces near crossing points is considered strong enough to allow transitions from
one to the other, but weak enough such that energy splitting is minimal. Thus the
approximate adiabatic free energy surface is constructed as the minimum of the
four diabatic surfaces above.
G(q
e
;q
p
) =minfG
1
;G
2
;G
3
;G
4
g (5.20)

(q
p
)
(q
e
)
G
AH
+
+ e
-

A+ e
-
+ H
+
AH
A
-
+ H
+
G
1
G
2
G
3
G
4
Figure 5.2: Example of an adiabatic potential energy surfaces of the four states as
a function of electron and proton transfer coordinates q
e
and q
p
. The parameters
employed are: = 0 V,
e
=
p
= 1 eV,

= 0:5 eV, G
prot
= 0:1 eV, and
G

prot
=0:2 eV.
74
An example of such a surface is shown in gure 5.2 with the parameters men-
tioned in the gure caption. To allow smooth representation in the gure, a small
coupling between the surfaces is introduced. From the gure, it is evident that the
form of the potential energy surface is a function of the three parameters deter-
mining the minima of the diabatic states F, G
prot
, and G

prot
as shown in
equations 5.12-5.15, and the three parameters
e
,
p
, and

determine how the
energies of the states change upon moving along the electron and proton transfer
coordinates.
Now we consider the case in which G

prot
is much more negative than G
prot
.
This is a likely scenario, since protonation ofM

is reasonably much more favorable
than protonation ofM. In this case theG
3
(q
e
;q
p
) surface will be shifted very high
and a passage from 1 to 3 and then to 4 will not be permitted. Thus we consider
transitions from 1 to 4 either directly across the diagonal, or from 1 to the proton
transfer intermediate 2 and then to 4. We will refer to these as the concerted and
stepwise paths respectively.
Next, we consider the in
uence of pH on the surface. For simplicity, we consider
= 0 and notice from equation 5.13 that the minimum energy of G
2
is G
prot
,
which according to equation 5.10 depends on pH. Thus lowering the pH makes
protonation of M more favorable and lowers the corner of the potential energy
surface corresponding to the minimum of G
2
. Raising the pH, on the other hand,
makes proton transfer less favorable and raises this corner. Thus the height of
surface at thefq
e
= 0; q
p
= 1g corner is controlled by pH.
A consequence of this observation is that for large values of pH, for which
proton transfer via 2 is not favorable, the only available path for proton transfer is
the concerted path through the diagonal. At the other extreme, for low pH values,
the corner corresponding toG
2
is signicantly lowered, and the electrode is readily
75
protonated. The transition state, in that case, lies along the electron transfer
coordinate betweenG
2
andG
4
. Following Koper, the transition state for each case
can be found by rst nding the crossing of the respective diabatic surfaces and
then nding the point of minimum energy along that line. To begin with, we nd
the transition state between states 2 and 4.
G
2
(q
e
;q
p
) =G
4
(q
e
;q
p
) (5.21)

e
q
2
e
+
p
(q
p
1)
2
+2

q
e
(q
p
1)F+G
prot
=
e
(q
e
1)
2
+
p
(q
p
1)
2
+2

(q
e
1)(q
p
1)
(5.22)
q
p
=
G
prot
+F + 2

+
e
2
e
q
e
2

(5.23)
We know thatq
p
for transition between state "2" to "4" is 1 so we will nd the
q
e
at the crossing point:
q
e
=
G
prot
+F +
e
2
e
(5.24)
Therefore, at the crossing line of the G
2
and G
4
the minimum of energy is
located atfq
m
e
=
Gprot+F+e
2e
;q
m
p
= 1g and the corresponding transition state
energy relative to the minimum of G
2
is:
G
z
24
=G
2
(q
m
e
;q
m
p
)G
2
(0; 1) =G
prot
+F +
(G
prot
F +
e
)
2
4
e
(5.25)
Similarly, for the crossing line of G
1
and G
4
we have:
76
G
1
(q
e
;q
p
) =G
4
(q
e
;q
p
) (5.26)

e
q
2
e
+
p
q
2
p
+ 2

q
e
q
p
F =
e
(q
e
1)
2
+
p
(q
p
1)
2
+ 2

(q
e
1)(q
p
1) (5.27)
q
p
=
F + 2

+
e
+
p
2

q
e
2
e
q
e
2(

+
p
)
(5.28)
Now we can insert the crossing point (q
e
;q
p
) at either ofG
1
(q
e
;q
p
) orG
4
(q
e
;q
p
)
and nd the surface energy at the transition.
G
1
(q
e
;
F+2

+
e
+
p
2

q
e
2
e
q
e
2(

+
p
)
) =::: (5.29)
F +
e
q
e
2
+

q
e
(F+2

+
e
+
p
2

q
e
2
e
q
e
)

+
p
+

p
(F+2

+
e
+
p
2

q
e
2
e
q
e
)
2
4(

+
p
)
2
By taking the derivative of equation 5.29 with respect to q
e
we will nd the
minima (derivative=0):
q
m
e
=
F + 2

+
e
+
p
2(2

+
e
+
p
)
(5.30)
Inserting this expression back into equation 5.28 we realize that the q
m
p
is also
equal to q
m
e
. Hence, the minimum energy is located atfq
m
e
=q
m
p
=
F+2

+e+p
2(2

+e+p)
g
and the corresponding transition state energy relative to the minimum of G
1
is:
G
z
14
=G
1
(q
m
e
;q
m
p
)G
1
(0; 0) =
(F + 2

+
e
+
p
)
2
4(2

+
e
+
p
)
(5.31)
77
5.2.1 Prediction of the model for the pH-dependence of
We will use the formal denition of the electron transfer coecient as the sensitivity
of the transition state to the applied overpotential[Guidelli et al., 2014-01-18]:
=
@G
z
@(F)
(5.32)
Using the above, we arrive at the following two expressions for the electron
transfer coecient for the two transition states:

14
=
F
2(2

+
e
+
p
)
+
1
2
(5.33)

24
=
G
prot
+F
2
e
+
1
2
(5.34)
The striking observation in the above two results is that the electron transfer
coecient for the concerted pathway
14
is independent of G
prot
. To make this
more explicit, we will re-write the above at the limit of small overpotential 0
and use the denition of G
prot
from equation 5.10.

14
=
1
2
(5.35)

24
=
G(MH
+
) +G(M) 2:3RTpH
2
e
+
1
2
(5.36)
Thus the model presents a pH-independent for the concerted path and a pH-
dependent for the stepwise path. More specically, the negative sign of the pH
term in the equation 5.36 means that a decrease in pH should increase with a
linear slope of
2:3RT
2e
. These predictions are, in principle, experimentally veriable.
As explained before, the model also predicts that at high pH the concerted path
78
1!4 is preferred since protonation of the surface is unfavorable, while at low pH the
stepwise path 1!2!4 becomes favorable. Thus an experimental test of this model
will require measuring for a range of pH values and studying its pH dependence
at high and low pH limits. The experimental work that is presented in the next
section was motivated by this hypothesis. While we will show results that partly
agree with this hypothesis, we will also discuss the experimental impediments and
nuances that interfere with the measurement of and can deter from unequivocal
verication of this model.
The model also predicts the transition pH, lower than which the stepwise and
higher than which the concerted mechanism is at play. The pH
trans
may be esti-
mated by setting the minimum of the G
2
surface (equation 5.13) equal to the
energy of the 1! 4 transition state (equation 5.31). At that point, accessing
the minimum of G
2
from the minimum of G
1
will require the same free energy as
accessing the 1! 4 transition state. The pH value that achieves this is found to
be:
pH
trans
=
1
2:3RT

(F + 2

+
e
+
p
)
2
4(2

+
e
+
p
)
+G(M)G(MH
+
)

(5.37)
which based on the G
prot
= 0 (equation 5.10) for the point of protonation equi-
librium, can be written in terms of surface pKa as follows:
pH
trans
=
1
2:3RT

(F + 2

+
e
+
p
)
2
4(2

+
e
+
p
)

+pKa (5.38)
Note that while the pH
trans
depends on the pKa of the surface, they are not
the same, as implied by the rst term in the above equation.
79
(q
e
)
(q
p
)
A+ e
-
+ H
+
AH
CPET
(q
e
)
ET
AH
AH
+
+ e
-

(q
p
)
a) high pH b) low pH
*
*
1
2
4
1
2
4
Figure 5.3: Free energy surfaces for CPET (a) and stepwise paths (b). The tran-
sitions states are indicated by. As explained in the text, pH controls the height
of the surface in the corner 2. At high pH protonation is unfavorable, corner 2 is
raised and the PES shown in (a) determines the reaction. At low pH, protonation
is spontaneous, corner 2 is lowered and the PES shown in (b) controls the reaction.
Finally, gure 5.3 graphically represents the PES for select values of parameters.
The variation of as a function of pH can be understood from this gure in
a relatively intuitive way. First, it is noted that at the high pH limit (gure
5.3.a), the transition state is in the middle of the PES and the G
2
corner is raised
signicantly higher than that. Under this condition, a small change in the height
of G
2
(i.e. pH) will not in
uence the saddle point height. The overpotential will
change the relative heights of the G
1
and G
4
states and hence will in
uence the
height of the transition state that is at their crossing point. Consequently, the
transition state energy will be sensitive to the overpotential (6= 0), but it will be
insensitive to pH. In the low-pH limit (gure 5.3.b), the height ofG
2
is signicantly
lower compared to the saddle point and proton transfer is spontaneous. In that
situation, the transition state is along the electron transfer coordinate and will be
sensitive to the relative heights of G
2
and G
4
. The height of G
2
is controlled by
both potential and pH (equation 5.13). Thus the transition state will be sensitive
to overpotential (6= 0) and will also be pH-dependent.
80
a) Pt b) Au c) Cu d) Ni
e) Ti g) GC

f) FeS
2
pH
Potential/ V vs NHE

2 3 4 5 6
−1.5
−1
−0.5
0
−4
−3.5
−3
−2.5
−2
−1.5
Log(i/A cm
−2
)
pH
Potential/ V vs NHE

2 3 4 5 6
−1.5
−1
−0.5
0
−4
−3.5
−3
−2.5
−2
−1.5
Log(i/A cm
−2
)
pH
Potential/ V vs NHE

2 3 4 5 6
−1.5
−1
−0.5
0
−4
−3.5
−3
−2.5
−2
−1.5
−1
Log(i/A cm
−2
)
pH
Potential/ V vs NHE

2 3 4 5 6
−1.5
−1
−0.5
0
−4
−3.5
−3
−2.5
−2
−1.5
Log(i/A cm
−2
)
pH
Potential/ V vs NHE

2 3 4 5 6
−1.5
−1
−0.5
0
−4
−3.5
−3
−2.5
−2
−1.5
Log(i/A cm
−2
)
pH
Potential/ V vs NHE

2 3 4 5 6
−1.5
−1
−0.5
0
−4
−3.5
−3
−2.5
−2
−1.5
−1
Log(i/A cm
−2
)
pH
Potential/ V vs NHE

2 3 4 5 6
−1.5
−1
−0.5
0
−3
−3.5
−2
−2.5
−1
−1.5
Log(i/A cm
−2
)
Figure 5.4: Electrochemical currentlogI(pH;E) for hydrogen evolution reaction on
the surface of dierent electrodes as a function ofpH and potential a) Pt b) Au c)
Cu d) Ni e) Ti f) FeS
2
and g) glassy carbon (GC). The color scale is kept the same
in all gures for ease of comparison. The Pourbaix line for water stability (solid
line) is overlaid for reference. As expected, Pt exhibits almost no overpotential
(the potential dierence between the Pourbaix line and the onset of current) at
low pH values, while GC has the highest overpotential in the same range.
5.3 Experimental Result and Discussion
To emphasize both the pH and potential dependence of the electrochemical rate,
we present our results in 2-dimensional plots of current as a function of pH
and potential I(pH;E) which have been named dynamic potential-pH diagrams
(DPPDs)[Haghighat and Dawlaty, 2015; Minguzzi et al., 2012]. Figure 5.4 (a to
g) show such measurements. The electron transfer coecient is presented in
the form of contour plots (pH;E) for dierent types of electrodes in gure 5.5.
We chose a variety of electrodes to nd out if the pH dependence of varied for
dierent materials. While dierent electrodes do show variations in onset potential
and total current, we rst concentrate on the common features observed in all of
the electrodes.
81
pH
Potential/ V vs NHE

2 3 4 5 6
−1.5
−1
−0.5
0
0
0.1
0.3
0.5
0.7
0.9
α
a) Pt
pH
Potential/ V vs NHE

2 3 4 5 6
−1.5
−1
−0.5
0
0
0.1
0.3
0.5
0.7
α
b) Au
pH
Potential/ V vs NHE

2 3 4 5 6
−1.5
−1
−0.5
0
0
0.1
0.3
0.5
0.7
α
c) Cu
pH
Potential/ V vs NHE

2 3 4 5 6
−1.5
−1
−0.5
0
0
0.05
0.1
0.15
0.2
0.25
α
d) Ni
pH
Potential/ V vs NHE

2 3 4 5 6
−1.5
−1
−0.5
0
0.1
0.2
0.3
α
e) Ti
pH
Potential/ V vs NHE

2 3 4 5 6
−1.5
−1
−0.5
0
0.1
0.2
0.3
0.4
0.5
α
g) GC

pH
Potential/ V vs NHE

2 3 4 5 6
−1.5
−1
−0.5
0
0.1
0.2
0.3
0.4
0.5
0.6
α
f) FeS
2
Figure 5.5: The measured electron transfer coecient (pH;E) for Pt, Au, Cu,
Ni, Ti, FeS
2
, and glassy carbon (GC) electrodes. The color scale is kept the same
in all gures for ease of comparison.
The features that are common to all electrodes are the following. Both from the
I(pH;E) graphs and the(pH;E) plots (gures 5.4 and 5.5) two distinct regions in
pH with dierent behavior are identiable. At higher pH values the overpotential
is clearly larger while the empirical is smaller and somewhat pH-independent.
At lower pH values a smaller overpotential is observed with linear dependence on
pH which is close to the Nernstian slope for some of the electrodes. The value of
is distinctly larger in the acidic region compared to the basic region. It also shows
a linear increase with decreasing pH value for many of the electrodes in the acidic
range. The high and low pH regions seem to have an abrupt transition pH in the
range of 2.5 and 4.
At large overpotentials, corresponding to large currents, the empirically deter-
mined becomes small and nearly vanishes, pointing to mass transfer limita-
tions on the rate. It should be noted that whenever the current becomes mass
transfer limited, the empirical electron transfer can no longer be interpreted as
82
2 3 4 5 6
0
0.2
0.4
0.6
0.8
1
pH
α (unitless)

Cu
Au
Pt
FeS
2
2 3 4 5 6
0
0.1
0.2
0.3
0.4
0.5
0.6
pH
α (unitless)
2 3 4 5 6
−0.1
0
0.1
0.2
0.3
0.4
0.5
pH
α (unitless)

GC
Ni
Ti
a) b) c)
experiment experiment model
α
24
α
14
Figure 5.6: Experimentally measured values of as a function of pH for dierent
type of electrodes a) copper, gold, platinum, and iron pyrite. b) glassy carbon,
nickel, and titanium. The two sets of data are shown separately for clarity. c)
calculated values of as a function of pH which is obtained from the model for
selected values of parameter :
14
=0:4 V,
24
=0:1 V,
e
=
p
= 0:2 eV,

= 0:15 eV, and pKa = 3.
=(2:3RT=F )(@ logI=@E). For that reason, it is most reasonable to only con-
centrate on the value of for low values of current, corresponding to the current
onset edge in the I(pH;E) plots. A convenient way to see the variation of alpha
as a function of pH is to isolate the the average value of near the onset for each
pH value as discussed in the experimental section.
In gure 5.6.a-b, we show the variation of as a function of pH for several
electrodes. The overall observation in gure 5.6.a-b is that rises with decreasing
pH at low pH range and is nearly constant at high pH range. Two observations in
the data are interestingly similar to the predictions of the model discussed above
and shown in gure 5.6.c. The rst is the existence of two regions, the high and
low pH regions, with distinct kinetic behavior. The second is the nearly constant
value of for the high pH region and the linearly increasing with decreasing
pH in the low pH region. Figure 5.6 shows the qualitative similarity between the
experimental data and the predicted trend based on the model. Reporting this
similarity is an important message of our work.
83
However, there are several features in the data that do not seem to plausibly
follow from the model. First is the similarity in the transition range of pH for
various types of electrodes. The model predicts the transition pH to depend on
the surface pKa as shown in equation 5.38. Since the surface pKa of dierent
materials are expected to be dierent, the transition pH should vary for dierent
electrodes. The data, on the other hand, shows a narrow range ofpH
trans
= 2:54
for all of the electrodes studied. This aspect of the ndings contradicts the model.
Neglect of mass transfer considerations may be the source of this discrepancy.
Mass transfer limitations in
uence the measurement of electrochemical rates.
In particular, at high overpotentials the rate becomes large enough that the con-
centration of reactants and products near the surface largely deviate from those
of the bulk. Thus at large current values the value of pH near the surface will be
markedly more basic compared to that of the bulk. The known technique to under-
stand and isolate the in
uence of mass transfer limits on the rate is the rotating
disk electrode (RDE) method[Schmidt et al., 1998; Gasteiger et al., 1995]. At high
rotation speeds, a larger
ux of material to the electrode can be maintained and
mass transfer limitation can be partially mitigated.
To nd out how rotation in
uences the behavior described so far, we performed
the experiments for three electrode materials, gold, platinum, and glassy carbon,
for which we had suitable rotating disk electrodes available. The results of the RDE
experiments are shown in gures 5.7 , 5.8, and 5.9. An important observation is
that the apparent transition pH values seen in the experiments performed without
rotating electrode (gures 5.4 and 5.5 ) move to higher pH values upon rotation.
This dependence on rotation speed indicates that the apparent transition pH,
which was pointed out earlier as one of the inconsistent points between the model
and the experiment, is a function of mass transfer. A series of models, such as the
84
pH
Potential/ V vs NHE

2 3 4 5
−1.5
−1
−0.5
0
−4
−3.5
−3
−2.5
−2
Log(i/A cm
−2
)
pH
Potential/ V vs NHE

2 3 4 5
−1.5
−1
−0.5
0
0
0.1
0.3
0.5
0.7
α
pH
Potential/ V vs NHE

2 3 4 5
−1.5
−1
−0.5
0
−4
−3.5
−3
−2.5
−2
−1.5
Log(i/A cm
−2
)
a)
b)
c)
pH
Potential/ V vs NHE

2 3 4 5
−1.5
−1
−0.5
0
0
0.05
0.1
0.15
0.2
0.25
0.3
α
d)
Figure 5.7: The in
uence of rotation on the dynamic potential pH diagrams
and electron transfer coecient for gold rotating disk electrode. (a,b) The
logI(pH;E) and (pH;E) with no rotation. (c,d) The logI(pH;E) and (pH;E)
at 2000 rpm.
Levich equation, systematically relate the rate to rotation speed. We have found
that in the transition region the current correlates with the Levich equation to
some extent.Figu re 5.10 shows the relationship between inverse of the current and
the inverse square root of rotation!

1
2
. Such plots are known as Levich-Koutecky
plots. Normally a linear relationship between theI
1
versus!

1
2
is expected under
diusive mass transport. Our experimental observation, however, shows somewhat
linear behavior at pH 3.01, partially non-monotonic dependence at intermediate
pH values, and almost no dependence at low pH values.
Thus we suggest that while mass transfer may be a factor that in
uences the
value of the apparent transition pH, an underlying chemical mechanism may still be
85
pH
Potential/ V vs NHE

2 3 4 5
−1.5
−1
−0.5
0
−4
−3.5
−3
−2.5
−2
Log(i/A cm
−2
)
pH
Potential/ V vs NHE

2 3 4 5
−1.5
−1
−0.5
0
0
0.1
0.2
0.3
0.4
0.5
0.6
α
pH
Potential/ V vs NHE

2 3 4 5
−1.5
−1
−0.5
0
−4
−3.5
−3
−2.5
−2
−1.5
Log(i/A cm
−2
)
pH
Potential/ V vs NHE

2 3 4 5
−1.5
−1
−0.5
0
0
0.05
0.1
0.15
0.2
α
a) b)
c) d)
Figure 5.8: The in
uence of rotation on the dynamic potential pH diagrams and
electron transfer coecient for glassy carbon rotating disk electrode. (a,b) The
logI(pH;E) and (pH;E) with no rotation. (c,d) The logI(pH;E) and (pH;E)
at 2000 rpm.
at play. In other words, we hypothesize that even if mass transfer limits are entirely
removed, the kinetic parameters at basic pH values would be distinctly dierent
from those at acidic pH values and the reaction would follow the potential energy
surface described before. To isolate the eects of mass transfer, further experiments
along with models that potentially go beyond the diusive assumptions of the
Levich equation are possibly necessary.
86
pH
Potential/ V vs NHE

2 3 4 5 6
−1.2
−1
−0.8
−0.6
−0.4
−0.2
0
−4
−3.5
−3
−2.5
−2
−1.5
Log(i/A cm
−2
)
pH
Potential/ V vs NHE

2 3 4 5 6
−1.2
−1
−0.8
−0.6
−0.4
−0.2
0
0
0.1
0.3
0.5
0.7
α
pH
Potential/ V vs NHE

2 3 4 5 6
−1.2
−1
−0.8
−0.6
−0.4
−0.2
0
−4
−3.5
−3
−2.5
−2
−1.5
−1
Log(i/A cm
−2
)
pH
Potential/ V vs NHE

2 3 4 5 6
−1.2
−1
−0.8
−0.6
−0.4
−0.2
0
0
0.1
0.2
0.3
0.4
0.5
α
a)
c) d)
b)
Figure 5.9: The in
uence of rotation on the dynamic potential pH diagrams and
electron transfer coecient for platinum rotating disk electrode. (a,b) The
logI(pH;E) and (pH;E) with no rotation. (c,d) The logI(pH;E) and (pH;E)
at 1600 rpm.
0.02 0.03 0.04 0.05 0.06 0.07 0.08
0
200
400
600
800
1000
ω
−0.5
(rpm
−0.5
)
I
−1
(A
−1
cm
2
) at 0.5 V vs NHE

2.76
2.48
2.23
1.57
1.97
pH 3.01
Figure 5.10: Levich plot obtained at E=0.5 V vs NHE for the HER at dierent
pH (from 3.01 to 1.57) and dierent rotation rates.
87
5.4 Conclusions
We have analyzed a parameterized potential energy surface (PES) for the reduc-
tion of one proton (Volmer reaction) which is based on the work of Schmickler
and Koper. The in
uence of pH on the geometry of the PES and possible reac-
tion pathways have been discussed thoroughly. We calculated the electron transfer
coecient based on this model and evaluated its pH-dependency. We showed
experimental results on the variation of empirical for hydrogen evolution reaction
on dierent types of electrodes over a wide pH range. There are several similari-
ties between the pH-dependence of derived from the model and the empirical
based on the experiments. The data suggests that the governing mechanisms may
be stepwise proton-electron transfer in the low-pH, and concerted proton-electron
transfer in the high pH range. We caution that the similarity between the obser-
vations and the model should not be taken as proof of the model. Some features of
our data are not fully explainable by the prediction of the model such as the tran-
sition pH between the acidic and basic region. We hope that our work can inspire
further theoretical and experimental studies on understanding proton thermody-
namics and kinetics near an electrode. In particular, theoretical work on explicit
pH-dependence of the transfer coecient is desirable. Also, new experimental
approaches and models that will isolate the mass transfer eects, and concentrate
on local spectroscopic measurements on the surface are expected to be useful in
further elucidation of this problem.
88
Chapter 6
Controlling Proton Conductivity
with Light: A Scheme based on
Photoacid Doping of Materials
Transducing light into changes in material properties is at the heart of a large
range of functional materials. Controlling charge carrier concentration in a mate-
rial is an essential functional feature of several classes of materials. In conventional
semiconductors with optical bandgaps, light excitation can generate carriers (elec-
trons and holes) which can change the electronic conductivity. Such electronic
photoconductivity is the functioning principle of several devices, including pho-
todetectors. Similarly, optical charge carrier generation in a region with a built-in
eld, as in the junction of two dierent materials, is the principle of photovoltaic
devices. Control of carrier concentrations with an external voltage is the working
principle of a transistor.
All of the above electronic eects can, in principle, be implemented in materi-
als where the mobile charge carriers are ions, such as protons. Proton conducting
materials is a mature and vibrant eld of study due to their utility in several tech-
nologies, in particular fuel cells. Ionic analogs of diodes have been demonstrated
several decades ago[Lovrecek et al., 1959]. However, only in recent years protonic
materials that are analogs of transistors and diodes have attracted a new surge in
attention [Yamamoto and Doi; Gabrielsson et al., 2012; Cayre et al., 2007; Deng
89
et al., 2013; Tributsch, 2000; Chun and Chung, 2015; Norby, 2001; Ordinario et al.,
2014; Leger et al., 2010]. A new class of protonic materials that are built in anal-
ogy to their electronic counterparts awaits development and implementation. It is
anticipated that such devices will open new possibilities such as interfacing with
biological ionic currents[Irimia-Vladu et al., 2011; Tarabella et al., 2013; Miyake
et al., 2015], articial proton pumps[Tributsch, 2000], interfacing light with fuel
cells and batteries, and light harvesting.
Here we demonstrate, to our knowledge for the rst time, an analog of this eect
for protons in an organic polymer solution and in water. We show that when a
material is doped with photoacids, light excitation generates extra mobile protons
that change the low-frequency conductivity of the material. We measure such
change both in polyethylene glycol (PEG) and in water sandwiched between two
transparent electrodes and doped with a well-known photoacid 8-hydroxypyrene-
1,3,6-trisulfonic acid (HPTS).
6.1 Experimental
We prepared 1 mM solutions of photo-acid (8-hydroxypyrene-1,3,6-trisulfonic acid
trisodium salt, HPTS) in distilled water and polyethylene glycol (PEG) separately.
All the chemicals were purchased from Sigma-Aldrich. To do the impedance mea-
surements, we used a super sealed liquid cell from PIKE Technologies which we
further modied for our experiments (see gure 6.1). Two conductive transparent
electrodes made of
uorine-doped tin oxide (FTO) coated glass (MTI Corpora-
tion, TEC70) were sandwiched together, having a 0.5 mm thick Te
on spacer in
between. Two holes were drilled in the top electrode as inlet ports for liquid. The
electrodes were cleaned ultrasonically in ethanol followed by acetone then dried
90
supporting back
plates
spacer
conductive
glass
Liquid
injection
inlet
contact
wire
Figure 6.1: Schematic 3D open view of the liquid cell used for the impedance study.
The material is sandwiched between two transparent conductive electrodes.
under air streaming. We made electrical contacts to these electrodes using copper
wire and silver paint which were glued to the outside tail with epoxy.
The sandwiched electrodes were held rmly by tightening the screws of two
metallic supporting back plates. The metallic plates had a central circular shape
window which permits optical access. We then injected the photo-acid solution
from the inlet port using a syringe to make a uniform liquid lm between two con-
ductive electrodes. We performed electrochemical impedance spectroscopy (EIS)
on photo-acid solution of liquid lms using a Gamry Reference 3000 potentiostat
with Gamry EIS300 software. EIS data was acquired at zero bias from 0.1 Hz to
10
6
Hz (10 points per decade, 20 mV AC voltage amplitude). We rst measured
the impedance at dark condition while covering the whole cell with black aluminum
foil. Then we used a blue light diode (Thorlabs 405 nm mounted LED with colli-
mation accessory) to conduct the EIS measurements under continuous illumination
91
with 44 mW of power over the 1.5 cm diameter window, after accounting for the
re
ection and transmission of the front window.
6.2 Using Photoacid as a Proton-bearing
Molecule
To optically control the concentration of free protons in the material, we introduce
photoacids into the material matrix. Photoacids are a well-known class of proton-
bearing small molecules that become acidic by several orders of magnitude in
the excited state and release their protons (gure 6.2)[Ireland and Wyatt, 1976;
Shizuka, 1985; Arnaut and Formosinho, 1993; Tolbert et al., 2002; Rini et al.,
2003; Leiderman et al., 2005; Cox et al., 2009; Spry et al., 2007]. They have been
implemented in several applications, including in chemical kinetics where creating
a rapid optical pH jump to initiate a proton-requiring process is desired [Irie, 1983;
Donten et al., 2015; Dempsey et al., 2010]. In our case, the photoacid molecules
embedded in the polymer release a mobile protons upon light absorption. The
photo-released proton is analogous to a free carrier in a semiconductor and the less
mobile conjugate base is analogous to a hole.
8-hydroxypyrene-1,3,6-trisulfonic (HPTS) is one of the earliest recognized pho-
toacids. It has three sulfonate groups, which make it soluble in water. The hydroxyl
group (OH) in the 8-position can release protons upon irradiation. At the ground
state, it has a pK
a
of7.7 and drops to pK
a
of1 after light excitation. This
photochemical reaction is shown in gure 6.3. The absorption spectra of the pro-
tonated HPTS (gure 6.4) indicates two apparent absorptions at 405 nm and 385
nm while after deprotonation with 0.1 M NaOH solution, these features disappear.
92
OH O
-
H
+
HA A
-
dark
light
Figure 6.2: Photoacids (HA) are organic molecules which are signicantly more
acidic in the excited electronic state than in the ground state (HA
*
) H
+
+ A
{
).
The generated conjugate base (A
{
) can bound with the free protos and recover
the original HA at the dark condition.
irradiation
darkness
OH
S
O
-
Na
+
O
O
S
O
-
O
O
Na
+
O
-
Na
+
O
O
S
S
S
O
-
O
O
Na
+
O
-
Na
+
O
O
O
-
Na
+
O
O
S O
-
pK
a
=7.7 pK
a
*
≅ 1
Figure 6.3: Structures of 8-hydroxypyrene-1,3,6-trisulfonic, commonly referred to
as HPTS and its photochemical reaction. [Spry et al., 2007; Wen et al., 2010]
6.3 Electrochemical Impedance Spectroscopy
(EIS)
Electrochemical impedance spectroscopy (EIS) is a well established method to
study the electrochemical systems. As shown in gure 6.5.a the fundamental
approach of the impedance spectroscopy is to apply a small amplitude sinusoidal
AC voltage to the system under the steady-state and measure the oscillating cur-
rent response. The experiment is carried at a wide range of frequencies. In a linear
(or pseudo-linear) system, the current response to a sinusoidal potential will be a
93
350 375 400 425 450 475 500
0
0.2
0.4
0.6
0.8
1
1.2
1.4
Wavelength (nm)
Absorbance

HPTS in distilled water
HPTS deprotonated with NaOH
Figure 6.4: Optical absorption measured using 1mM HPTS solution shows two
distinct absorptions at 405 nm and 385 nm. after adding few drops of 0.1 M
NaOH the HPTS gets deprotonated and the absorption peaks disappear.
sinusoid at the same frequency but shifted in phase (gure 6.5.b). Similar to resis-
tance, the impedance is a measure of the ability of a system to impede the
ow
of electrical current (electron or ions) and hence is the ratio of voltage-to-current
(equation6.1).
Z
!
=
V
!
I
!
=
V
0
sin(!t)
I
0
sin(!t +)
=Z
0
sin(!t)
sin(!t +)
(6.1)
In this equation (6.1), The V
0
and the I
0
are the amplitude of voltage and
current, respectively, ! = 2f is the radial frequency in rad/s (f is a frequency
in Hz), Z
0
is impedance magnitude and is the phase shift . We can use the Z
0
and the to represent the impedance as a complex number using Euler's formula
(exp(ix) = cosx +isinx), i is a complex unit i =
p
1. The voltage and current
are described as:
V
!
=V
0
exp(i!t) (6.2)
94
V
I
V
0
V
0
+ ΔV Sinωt
I
0
+ ΔI Sin(ωt + Φ)
I
0
Time
Magnitude

Applied Voltage
Measured Current
Phase shift
Voltage
Current
a) b)
Figure 6.5: principles of impedance spectroscopy (a) by applying a sinusoidal AC
voltage to the system under the steady-state we can measure the oscillating current
response. Here, V
0
is the DC voltage, V is the amplitude of the probing signal
(typically V < 10mV ), and! is the angular frequency, t is time, I
0
is the direct
current, I is the amplitude of the AC response, and is the phase shift between
the probing signal and the response. (b) In a linear system, the current response
to a sinusoidal potential will be a sinusoid at the same frequency but shifted in
phase.
I
!
=I
0
exp(i!t i) (6.3)
The impedance is then represented as a complex number:
Z
!
=Z
0
exp(i) =Z
0
(cos +isin) =Z
Re
+iZ
Im
(6.4)
The expression for Z
!
in equation 6.4 is composed of a real and an imagi-
nary part. Note that in the electrochemical literature real and imaginary parts of
impedance often are marked as Z
0
and Z
00
, respectively. If the real part is plot-
ted on the X-axis and the imaginary part is plotted on the Y-axis of a chart, we
get a Nyquist Plot (gure 6.6). On this plot the impedance can be represented
as a vector of lengthjZj and the angle between this vector and the x-axis is
(phase-shift). In this case, the frequency is implicit.
95
Real Impedance, Z’
Imaginary Impedance, Z”
ω
2
∞
|Z|
φ
ω
1
ω
4
ω
3
ω
small
Figure 6.6: A typical Nyquist plot which results from plotting the imaginary
impedance component (Z
00
) against the real impedance component (Z
0
) at each
excitation frequency. The impedance has a magnitudejZj and phase-shift at
each frequency.
The ultimate goal of the analysis of impedance data is to elucidate a physical
model of the system and estimate its parameters. To achieve that we need an
appropriate equivalent circuit model to t the impedance data. The circuit model
usually consists of passive elements that do not generate current or potential such
as resistors (R), capacitors (C), and inductors (L). Each of this element has to
have a physical meaning in the electrochemical system. For example, most models
contain a resistor that models the cells solution resistance. Thus in the following
we will look at the impedance expression for some common circuit elements.
Resistor (R)
Resistance is the opposition to the
ow of DC current. When a sinusoidal voltage
is applied to a resistor, it induces an alternating current that is in phase with
96
applied voltage. Therefore, the impedance of an ohmic circuit element is equal to
its resistance R:
Z
R
=
V
!
I
!
=
V
0
sin(!t)
I
0
sin(!t)
=R (6.5)
Capacitor (C)
The impedance of a capacitor can be calculated based on the current and voltage
relation:
I =C
dV
dt
(6.6)
This yields to the impedance for a capacitor:
Z
C
=
V
!
I
!
=
V
0
exp(i!t)
V
0
C (i!)exp(i!t)
=
i
C!
(6.7)
Constant Phase Element (CPE)
The constant phase element (CPE) is introduced as a non-intuitive circuit element
to model the real-world electrochemical phenomenon. In an ideal capacitor, we
assume that the electrode/electrolyte interface is homogeneous which is normally
not the case. It contains a large number of surface defects such as kinks, jags, and
ledges, local charge inhomogeneities, etc. To account for this eect, modeling the
equivalent circuit with CPE is often essential in EIS data interpretation.
The impedance of a CPE is given by:
Z
CPE
=
1
Y
0
(i!)

(6.8)
97
Where the parameter is an emirical constant ranges from 0 to 1. For = 1
the CPE element is an ideal capacitor and for = 0 is a pure resistor. Between 0
and 1 the CPE can be thought of as a fractional generalization of a conventional
capacitor.
Warburg Element (W)
Whenever an electrochemical reaction is under partial or complete diusional
transport of an electroactive species, the impedance is represented by Warburg
element (W). Mathematically, the Warburg impedance is the solution to the one-
dimensional diusion equation with oscillating concentration at the boundary. The
impedance is represented as Z
W
:
Z
W
=
A
W
p
i!
=
A
W
p
2!
(1i) (6.9)
WhereA
W
is the Warburg coecient and it depends on the diusion constant
and the carrier concentration which will be explained later in section 6.5. As
follows from equation 6.9:
tan =
Z
Im
Z
Re
=1 (6.10)
Therefore, the phase shift is equal to 45

, and the Nyquist plot has a diusion
tail which is inclined at 45

to the real axis.
The most common equivalent circuits with their representative Nyquist plots
are provided in gure 6.7. A simple circuit consisting of a capacitor and resistor
in parallel will result to a Nyquist plot (complex-plane representation) of a perfect
semicircle. The diameter of the semicircle is equal to the resistance and it is
independent of the capacitance. Adding a solution resistance (R
s
) to the RC
98
parallel circuit, we obtain a Nyquist plot shown in gure 6.7.a, which is a semicircle
but shifted to the right on the real axis about R
s
. This circuit model is called
Randles and is the simplest circuit that can model the interfacial electrochemical
reactions. It includes a solution resistance (R
s
), a double layer capacitor (C
dl
) and
a charge transfer resistance (R
ct
).
As mentioned above, the parallel connection of a resistor and a capacitor will
represent a perfect semicircle in the Nyquist plot. If we replace the capacitance
(C
dl
) in the equivalent circuit of gure 6.7.a by a constant phase element (CPE),
the Nyquist plot becomes
attened semicircle (gure 6.7.b). This type of semicir-
cles are often encountered in real measurements instead of the perfect ones. The
distribution of characteristic frequency could be a reasonable explanation for this
arc depression.
In electrochemical systems that are governed by mass transport of the elec-
troactive species, we need to add the Warburg diusive element next to the charge
transfer resistance part in the circuit model (gure 6.7.c). In its representative
Nyquist plot, there are two distinct regions: at high frequency, a semicircle shows
that the charge transfer resistance determines the impedance, while at low frequen-
cies diusion (or mass transfer) dominates and the Warburg impedance is observed
as a straight line with a slope of 45

in the complex plane.
99
Real Impedance, Z’
Imaginary Impedance, Z”
R
ct R
s
R
ct
R
s
C
dl
Real Impedance, Z’
Imaginary Impedance, Z”
R
ct
R
s
CPE
Real Impedance, Z’
Imaginary Impedance, Z”
R
ct
R
s
R
ct
R
s
C
dl
W
mass transfer
control

kinetic
control

45°

a)
b)
c)
Figure 6.7: The most common equivalent circuit models with their representative
Nyquist plots (a) Randles circuit (b) constant phase element (CPE) in parallel with
charge transfer resistant (c) mass transfer dominant circuit model with Warburg
diusive element (W).
6.4 Results of Impedance Measurements on
Photo-acid HPTS Solutions in the Light and
Dark Conditions
Impedance measurements on photo-acid HPTS solutions in distilled water and
polyethylene glycol in the absence and presence of light are reported in the form of
Nyquist plots in gure 6.8. In both dark and light cases, the low-frequency part of
the impedance spectra appear as a slanted line in the Nyquist plot and is due to the
diusive impedance of ions in the material. This feature is well-known under the
name of Warburg impedance [Grahame, 1952; Barsoukov and Macdonald, 2005]
and will be referred to shortly. The high frequency range for the HPTS dissolved
in PEG sample shows a semicircle, which is absent in the dissolved HPTS in
distilled water. From these measurements one can see a signicant change in both
the real and imaginary parts of the low-frequency impedance in the presence of
light. Upon shining light, mobile protons are released from photo-acid molecules
and act as free charge carriers. As will be shown shortly, this change of carrier
concentration explains the dierence in the observed light and dark impedance
100
0 1 2 3 4 5
x 10
4
0
5
10
15
x 10
4
Z
Real
( Ω)
− Z
Imag
( Ω)

Dark −Experiment
Dark −Model
Light −Experiment
Light −Model
0 1 2 3
x 10
4
0
0.5
1
1.5
2
x 10
5
Z
Real
( Ω)
− Z
Imag
( Ω)

Dark −Experiment
Dark −Model
Light −Experiment
Light −Model
2000 4000
0
1000
2000
3000

a)
b)
increasing ω
increasing ω
Figure 6.8: (a) Typical Nyquist plot (the negative of the imaginary part versus
the real part of the impedance) for 1 mM HPTS in distilled water solution over
the frequency range of 0.1 Hz to 1 MHz in the presence and absence of light. (b)
The impedance spectra of 1 mM HPTS in PEG from 0.1 Hz to 1 MHz. The inset
shows the high frequency region. The solid lines represent the impedance of the
model equivalent circuit shown in gure 6.10.
spectra. It was also veried that the electrodes on their own are not responsive to
light by conducting an experiment on a blank sample (only distilled water in the
cell) in the presence and absence of light in gure 6.9.
6.5 Fitting the Impedance data to an equivalent
Circuit Model
We t the experimental impedance data to an equivalent circuit model schemati-
cally shown in gure 6.10. The main element of the models is the Warburg element
W which as described in section 6.3 is very well-known to represent the diusive
impedance of ions in a medium and is often manifested in ionic conduction in
solids. The characteristic feature of the Warburg element is an impedance given
by equation 6.9.
101
0 0.5 1 1.5 2 2.5
x 10
4
0
0.5
1
1.5
2
x 10
5
Z
Real
( Ω)
− Z
Imag
( Ω)

Distilled Water − Dark
Distilled Water − Light
Figure 6.9: Nyquist plot of pure distilled water over the frequency range of 0.1 Hz
to 1 MHz in the presence and absence of light.
CPE
W
CPE
CPE
W
Z
CPE

dark
=Z
CPE light
= 10
5
/(jω)
0.95
Ω
Z
W

light
= 5×10
4
/(jω)
0.5
Ω
Z
W

dark
= 10
6
/(jω)
0.5
Ω R
dark
= 660 Ω
R
light
= 60000 Ω
R
dark
=R
light
= 320 Ω
Z
CPE

dark
=Z
CPE light
= 8×10
9
/(jω)
0.99
Ω
Z
W

dark
= 2×10
6
/(jω)
0.5
Ω
Z
W

light
= 83×10
4
/(jω)
0.5
Ω
R
dark
=R
light
= 4400 Ω
Z
CPE

dark
=Z
CPE light
=
13×10
4
/(jω)
0.95
Ω
a)
b)
Figure 6.10: The equivalent circuit model for (a) 1 mM HPTS in distilled water
solution (b) 1 mM HPTS in PEG, in the presence and absence of light.
An ideal Warburg element acting alone produces a 45

line in the Nyquist plot.
In conjunction with other elements, the overall impedance is modied from such
an idealized line. As will be discussed shortly, the magnitude of impedance A
W
is
a function of carrier concentration and changes with light in the photoacid doped
materials. To model the behavior of the entire circuit, it is necessary to introduce
constant phase elements (CPEs) and resistors.
102
We used the Gamry EIS300 impedance modeling software to t the EIS data.
All the parameters noted in the gure 6.10 were allowed to vary independently for
each experiment. For modeling the PEG data an extra parallel CPE was necessary.
This is mainly due to the dierences in the high-frequency region of the impedance
spectra, which show minimal dierence between the dark and light cases and thus
are not of central importance to our discussion. The expected error in the value
of each component is lower than 5% and is much smaller than the changes that
are induced by light. For PEG, the model reveals that both the light and dark
data share the same circuit element values and the only parameter that changes
is the Warburg impedance Z
W
. For the water data, both the Warburg impedance
Z
W
and an in-series resistance change in the presence of light. Also note that the
value of for the CPE elements is very close to 1, indicating that they represent
nearly ideal capacitors. The main nding of impedance measurements is a large
reduction of the Z
W
by more than 1 order of magnitude for both cases of water
and PEG upon continuous illumination.
6.6 The in
uence of adding NaOH on the con-
ductivity of Photo-acid HPTS Solutions in
the Light and Dark Conditions
To demonstrate that the contrast in conductivity between light and dark indeed
comes from protons, NaOH was added to the solution. Addition of the base shifts
the acid-base equilibrium for the photoacid in the ground state towards deproto-
nation. Thus in the presence of light, the number of released protons must sig-
nicantly reduce and consequently the light-dark contrast in conductivity should
103
be minimized. In conformity to this picture, we observe a signicant reduction in
the contrast between the dark and light impedance spectra after adding few drops
of 0.1 M of NaOH to the solutions (gure 6.11). The absorption spectrum of the
conjugate base is red-shifted with respect to the photoacid. However, signicant
absorption remains in the range of our blue light. The isosbestic point of HPTS
lies very close to the center of the LED source and the LED bandwidth spans
both the protonated and deprotonated forms (see Figure 2 in SI). Although the
deprotonated form still absorbs some of the light, the change in conductivity is
largely eliminated. This further supports the point that the conductivity change
observed in the absence of NaOH is indeed due to the light-induced release of pro-
tons. Comparing the deprotonated plots of water and PEG, we also note that the
real part of impedance has a large shift after adding NaOH in PEG compared to
the water case ( 50
for water and 4000
for PEG). The central nding
here is the change in contrast between the light and dark spectra in the presence
of the base and not the overall shift. The latter is expected due to the change in
the background ion concentration (extra Na
+
) upon addition of the base.
6.6.1 Relation of the real part of impedance to the magni-
tude of the Warburg element
In our experimental data the real part of the impedance of the circuit increases
upon shining light. We have modeled and interpreted the response of the entire
circuit to light and have arrived at the conclusion that the light illumination results
into reduction of the the Warburg impedance as supported and justied in the main
text. This may seem counterintuitive that reduction of the Warburg impedance
would result into increase in the real part of the impedance of the circuit. However,
it should be born in mind thatZ
w
is complex, and when put in parallel in a circuit
104
0 1 2 3 4
x 10
4
0
0.5
1
1.5
2
x 10
5
Z
Real
( Ω)
− Z
Imag
( Ω)

HPTS in PEG+ NaOH −Dark
HPTS in PEG+ NaOH −Light
HPTS in PEG −Dark
HPTS in PEG −Light
0 1 2 3 4 5
x 10
4
0
5
10
15
x 10
4
Z
Real
( Ω)
− Z
Imag
( Ω)

HPTS in Water+NaOH −Dark
HPTS in Water+NaOH −Light
HPTS in Water −Dark
HPTS in Water −Light
a) b)
increasing ω
increasing ω
Figure 6.11: Data demonstrating that the contrast between the dark and light EIS
spectra is signicantly diminished upon addition of NaOH in solution of HPTS
in both water (a) and PEG (b), thus supporting that the change in impedance is
related to photoreleased protons.
with a capacitor, it is possible that reduction of its magnitude would result into
increase in the real part of the circuit impedance. In the following we prove this
point.
Consider a Warburg element:
Z
w
=
(1i)A
w
p
2!
(6.11)
in parallel with a capacitor:
Z
c
=
1
i!C
(6.12)
The overall impedance becomes:
Z =
1
1=Z
w
+ 1=Z
c
=
1
(1+i)
p
!
p
2Aw
+ic!
(6.13)
105
The real part of the above is:
Re(Z) =
1
p
2!(c
2
!A
2
w
+1)
Aw
+ 2c!
(6.14)
Next we take the derivative of the above with respect to A
w
:
@Re(Z)
@A
w
=
p
2 (1c
2
!A
2
w
)
p
!
p
2c
2
!A
2
w
+ 2c
p
!A
w
+
p
2

2
(6.15)
The above can be negative when:
A
2
w
<
1
c
2
!
(6.16)
When the Warburg impedance magnitude and its parallel capacitance satisfy
the above relation a decrease inA
w
results into a increase inRe(Z). Thus justifying
the point made at the beginning of this section. Including the resistance in series
with Z
w
results into a more complicated equation for the conditions in which the
reductionA
w
can increaseRe(Z). However, it does not provide any further insight
and is not included here.
6.7 Discussion
Now we explain the origin of our main nding, which is the change in the diusive
(Warburg) impedance upon shining light. The magnitude A
W
of the Warburg
impedance Z
W
= A
W
(i!)
1=2
is related to carrier concentration via [Barsoukov
and Macdonald, 2005; Bard and Faulkner, 2001]:
A
W
=
k
B
T
C
0
Ae
2
p
D
(6.17)
106
where C
0
is the carrier concentration, k
B
T is the thermal energy, e is the
fundamental charge, andD is the diusion coecient for the carriers, andA is the
surface area of the electrode. It is clear that the the magnitude of the Warburg
impedance is inversely related to the carrier concentration, with larger carrier
concentrations resulting into smaller magnitude of impedance. This concentration
is exactly the parameter that is modied in our experiments by light. In the
absence of light a certain number of ions are present in the solution due to the
inherent ionic equilibrium in the medium. In the presence of light, protons are
released and increase C
0
. Thus it is quite natural to expect A
W
to go down upon
illumination, as observed in our experiments and shown in both gures 6.8 a and
c.
It is well-known that photoacids release protons in a solution[Mohammed et al.,
2005; Cox et al., 2009]. Even an example of release of protons by photoacids in
ice is reported [Timmer et al., 2010]. The nding in our experiment is that the
released protons are free enough that they contribute to low-frequency (quasi-DC)
conductivity. For this to be true, the released protons must escape the zone of
in
uence of the parent ion. After light absorption, the photoacidity of the molecule
lasts as long as the excited state lifetime, which is reported to be 5 ns for HPTS
[Spry et al., 2007]. When the photoacid returns back to the ground state, the
thermodynamic drive for the conjugate base and the proton to recombine is re-
established. However, during the excited state lifetime the released proton can take
advantage of the hydrogen bond network of the surrounding and can escape away
from the parent ion. Thus the diusion length of proton during the excited state
lifetime is an important parameter. For an ideal diusive transport, the diusion
length L
D
within time is related to the diusion coecient D by L =
p
D.
Water is known for its exceptionally large proton diusion coecient of 9 10

5
107
cm
2
=s or 9nm
2
=ns [Agmon, 1995]. Thus the diusion length for protons after
photoexcitation is estimated to be L
D
= 7nm. This is in conformity with several
ultrafast spectroscopy studies which have conrmed that photogenerated protons
can diuse over several molecular lengths and recombine with a target proton
acceptor [Spry and Fayer, 2007; Cox et al., 2009; Siwick et al., 2008; Douhal et al.,
1996]. The proton diusion length in PEG is expected to be smaller than water,
since it does not have as many protonatable groups.
The second important parameter that should be contrasted against the diusion
length is the distance that the proton must travel before it can be considered free
of the in
uence of the parent ion. If the proton diuses far enough to avoid
rapid recapture, then it can contribute to bulk low frequency conductivity as a
free carrier. One may estimate the radius of in
uence of the parent ion over the
escaped proton using electrostatic arguments. An electrostatic zone of in
uence
can be either dened by the Debye screening length, or more conveniently by
its closely related concept the Bjerrum length
B
[Robinson and Stokes, 1959].
The Bjerrum length is dened as the distance at which thermal
uctuations can
overwhelm the electrostatic attraction between two oppositely charged ions and is
given by:

B
=
e
2
4
0

r
k
B
T
(6.18)
When the charges are highly screened the Bjerrum length is smaller. This
length for pure water (
r
= 80) is
B
= 0:7nm. Comparing this to the diusion
length of photogenerated protons for water discussed above L
D
= 7nm, it is clear
that L
D
>
B
. Thus based on this rough electrostatic estimate, escape of protons
from the parent ions is permissible in water. For PEG, the dielectric constant
is smaller (
r
= 10) and leads to a larger value of
B
6nm. Given that the
108
diusion length of protons in PEG is already smaller than water, the protons are
less capable of diusing away from the zone of in
uence of the parent ion and
generate free carriers. This is in conformity with our observations, where under
identical concentration and illumination conditions, the HPTS doped PEG sample
shows much smaller contrast of conductivity between light and dark (gure 6.8 a
and c) compared to HPTS in water. The cartoon in gure 6.12 demonstrates the
interplay between diusion and recapture in the photogeneration of free protons.
We caution that the above line of reasoning is preliminary and should only
be used as a qualitative argument to understand this eect. Simple electrostatic
attractions alone does not take into account the intricate hydrogen bonding net-
works around the parent ion. Earlier work [Agmon et al., 1988; Pines et al.,
1988] that estimate the recapture radius of proton by tting a diusion model to
ultrafast time-resolved spectroscopic measurements inform us about a recapture
radius. The escape zone should not be taken literally as a sphere, but it should
be understood as an eective radius. The initial diusion away from the parent
ion is facilitated by chemical changes (i.e. increase in acidity of the photoacid due
to electronic charge redistribution upon optical excitation). During the excited
state lifetime the proton feels less bound to the parent molecule due to the local
changes in electronic charge density on the OH group of the photoacid and hence
manages to escape despite the overall negative charge of the parent molecule.
Understanding such local charge density re-distributions may be key in describ-
ing this phenomenon. Thus detailed molecular dynamics studies, and possibly
quantum mechanics/molecular mechanics (QM/MM) approaches, of the proton
diusion away from the parent ion is required to better understand the interplay
between diusion and recapture in such systems. We further emphasize that any
mention of the word free proton in this paper is a shorthand for solvated proton
109
H
+

H
λ
B
λ
B
400 nm
after light
excitation
L
D
L
D
> λ
B
photo
acid
photo
acid
H
+

Figure 6.12: The interplay between diusion and recapture in photoinduced free
proton generation. After light absorption, a proton is released from the photoacid
which diuses away from the parent ion. When the molecule returns back to the
ground state the drive to recapture the proton is established again. However, if
the proton has managed to exit the zone of in
uence of the parent ion, approxi-
mated here as the electrostatic Bjerrum length
B
, it can escape and contribute to
conductivity as a free carrier. Thus a preliminary gure of merit for understanding
this process is comparison of the diusion lengthL
D
and the recapture radius
B
.
in its respective environment. Here free proton means free to move independent
of the parent ion, and of course does not imply lack of the interaction or solvation
of the proton by the solvent environment. This is in analogy to semiconductors
where bound electron-hole pairs are known as excitons while free carriers can still
interact with their environment and form polarons.
One should also consider the diusion of the parent ion after photoexciation.
After light absorption two free ions, the conjugate base and a proton are created.
However, the conjugate base or the parent ion is much larger than the proton and
consequently its mobility is expected to be much poorer. Thus it is plausible to
110
ascribe the increased conductivity to protons. It may be possible to design systems
where the mobilities of the photogenerated ions are closer to each other.
Finally, we note that an explanation for the increase in the real part of the
impedance (i.e. resistive element in gure 6.8.b) after photoexcitation of HPTS in
water does not follow from the diusive (Warburg) impedance analysis presented
above. A separate model may be required to reveal the microscopic origin of that
change. A possible reason could be accumulation of carriers near the electrode to
the concentration levels that is beyond the linear assumptions of diusive trans-
port.
6.8 Conclusions
In conclusion, we have demonstrated how light excitation can be transduced into
changes in protonic conductivity of a material. In particular, we have demonstrated
light induced protonic conductivity changes for solutions of a photoacid in water
and PEG and have explained the physical origin of this eect. We hope that this
work can trigger molecular dynamics and possibley QM/MM studies of this process
by theorists. In particular, it is hoped that the eective recapture radius is inferred
from a simulation rather than from an estimate as shown above. This will help
us understand the chemistry in the immediate vicinity of the proton release site
which is likely more complex than can be captured by the
B
andL
D
parameters.
We hope that this work can inspire a new way of controlling protonic and ionic
currents using light. Finally, this work may serve a starting point for interfacing
light with bulk protonic currents without the need to rely on electronic conduction.
111
Chapter 7
Future Suggested Steps
Besides the specic outlooks at the end of each chapter of this thesis, here a few
general recommendations are made for future research:
Mechanistic studies of one electron, one proton transfer system at the
electrode surface:
The mechanistic analysis of oxygen evolution and hydrogen evolution reactions are
complicated since they involve transfer of multiple electrons and multiple protons
across electrochemical interfaces. Having one electron one proton transfer redox
reactions (A + H
+
+ e
{
*
) AH) would make a simpler platform to unravel the
kinetics. In particular, self-assembled monolayers (SAMs) attached to an electrode
surface are interesting systems for this purpose. They act like a spacer between the
electrode and the redox center in the bulk electrolyte. Also, the tunneling across
the monolayer slows the electron transfer rates to experimentally accessible ranges.
Attaching an appropriate redox complex to the SAM on an electrode is an ideal
design. The complex needs to have one electron, one proton redox chemistry. The
rate of electron transfer can be measured over a wide range of potential and pH
for both the oxidation and the reduction steps. To obtain a deeper understanding
of the role of proton, examining dierent complex groups with dierent pKa could
also be useful. The kinetic isotope eect (KIE) study by replacing the proton with
the deuteron also would be informative in mechanistic analysis and identifying the
rate determining step.
112
In-situ studies of reaction intermediates in water oxidation reaction:
Identifying surface intermediates and molecular active sites in water oxidation
reaction will guide us to design a metal oxide catalyst with the best structure and
composition. The structure specic spectroscopy techniques such as sum-frequency
generation, FT-infrared, surface-enhanced Raman, and X-ray absorption can be
employed to monitor the catalytic steps under reaction conditions.
Nanostructuring the co-catalyst on the surface of the photoanode:
Photoelectrochemical water splitting can be promoted by loading co-catalyst on
the surface of the semiconductor. The co-catalyst will reduce the activation energy
and makes the kinetic of water oxidation faster. Morphology of the co-catalyst is
important for ecient photo-assisted water splitting. Increasing the surface area
by nanostructuring, reduces the diusion distance for the minority carriers to reach
the interface. This results into better overall performance of the hybrid structure.
The example would be deposition of nanoclusters co-catalyst on the surface of the
photoanode to promote the charge transfer across the interface and avoid electron-
hole recombination.
Protonic photoconductivity in bipolar membranes:
Bipolar membranes (BPM) are polymer membranes composed of a cation exchange
layer (blocks the anion transport) and an anion exchange layer (blocks the cation
transport). The interfacial layer is known as the space-charge region similar to pn
junction semiconductors. Due to their unique feature in maintaining a constant
pH gradient and separating the electrolytes, BPM has recently caught attention
for water splitting applications. The water oxidation and reduction reactions can
113
occur simultaneously in two dierent pH conditions across a BPM, which will
provide more stable conditions for lower-cost earth abundant catalyst.
Using photoacid in the interfacial layer as a source of protons will make an
interesting case study since they will change the pH of the interfacial layer. The
released protons upon light excitation can migrate through the cation exchange
membrane and favor one side of the water splitting reaction. By comparing the
current response of the dark and light experiments, the protonic conductivity eect
can be conrmed.
114
Bibliography
Agmon, Noam. The grotthuss mechanism. Chemical Physics Letters, 244(56):456
{ 462, 1995. ISSN 0009-2614.
Agmon, Noam; Pines, Ehud, and Huppert, Dan. Geminate recombination in
proton-transfer reactions. ii. comparison of diusional and kinetic schemes. The
Journal of Chemical Physics, 88(9):5631{5638, 1988.
Arnaut, Lus G. and Formosinho, Sebastio J. Excited-state proton transfer reactions
i. fundamentals and intermolecular reactions. Journal of Photochemistry and
Photobiology A: Chemistry, 75(1):1 { 20, 1993. ISSN 1010-6030.
Bard, Allen J. and Faulkner, Larry R. Electrochemical Methods: Fundamentals
and Applications. Wiley, 2nd edition, 2001.
Barroso, Monica; Mesa, Camilo A.; Pendlebury, Stephanie R.; Cowan, Alexan-
der J.; Hisatomi, Takashi; Sivula, Kevin; Grtzel, Michael; Klug, David R.,
and Durrant, James R. Dynamics of photogenerated holes in surface Modied
-Fe
2
O
3
Photoanodes for Solar Water Splitting. Proceedings of the National
Academy of Sciences, 109(39):15640{15645, 2012.
Barsoukov, E. and Macdonald, J.R. Impedance Spectroscopy: Theory, Experiment,
and Applications. Wiley, 2005. ISBN 9780471716228.
Bassi, Prince Saurabh; Gurudayal, ; Wong, Lydia Helena, and Barber, James. Iron
Based Photoanodes for Solar Fuel Production. Phys. Chem. Chem. Phys., 16:
11834{11842, 2014.
Bediako, D Kwabena; Costentin, Cyrille; Jones, Evan C; Nocera, Daniel G, and
Sav eant, Jean-Michel. Proton{Electron Transport and Transfer in Electrocat-
alytic Films. Application to a Cobalt-Based O
2
-Evolution Catalyst. J. Am.
Chem. Soc., 135(28):10492{10502, 2013a.
Bediako, D. Kwabena; Surendranath, Yogesh, and Nocera, Daniel G. Mechanistic
Studies of the Oxygen Evolution Reaction Mediated by a Nickel Borate Thin
Film Electrocatalyst. J. Am. Chem. Soc., 135(9):3662{3674, 2013b.
115
Bockris, J. O. M. Kinetics of Activation Controlled Consecutive Electrochemical
Reactions: Anodic Evolution of Oxygen. J. Chem. Phys., 24(4):817{827, 1956.
Bockris, J.O.M. and Reddy, A.K.N. Modern Electrochemistry: An Introduction to
an Interdisciplinary Area. Number v. 2 in A Plenum/Rosetta edition. Springer,
1973. ISBN 9780306250026.
Bockris, J.O.M. and Reddy, A.K.N. Modern Electrochemistry 2B: Electrodics in
Chemistry, Engineering, Biology and Environmental Science. Number v. 2 in
Modern electrochemistry. Springer, 2001. ISBN 9780306463259.
Buono-Core, G.E; Tejos, M; Lara, J; Aros, F, and Hill, R.H. Solid-State Photo-
chemistry of a Cu(II)-Diketonate Complex: the Photochemical Formation of
High Quality Films of Copper(II) Oxide. Mater. Res. Bull., 34(1415):2333 {
2340, 1999. ISSN 0025-5408.
Carvalho, Waldemir Moura de Jr. and Souza, Flavio Leandro. Recent advances
on solar water splitting using hematite nanorod lm produced by purpose-built
material methods. Journal of Materials Research, 29:16{28, 2014. ISSN 2044-
5326.
Cayre, Olivier J; Chang, Suk Tai, and Velev, Orlin D. Polyelectrolyte diode:
Nonlinear current response of a junction between aqueous ionic gels. Journal of
the American Chemical Society, 129(35):10801{10806, 2007.
Chun, Honggu and Chung, Taek Dong. Iontronics. Annual Review of Analytical
Chemistry, 8(1):441{462, 2015.
Cook, Timothy R.; Dogutan, Dilek K.; Reece, Steven Y.; Surendranath, Yogesh;
Teets, Thomas S., and Nocera, Daniel G. Solar Energy Supply and Storage for
the Legacy and Nonlegacy Worlds. Chemical Reviews, 110(11):6474{6502, 2010.
PMID: 21062098.
Costentin, Cyrille. Electrochemical approach to the mechanistic study of proton-
coupled electron transfer. Chem. Rev., 108(7):2145{2179, 2008.
Costentin, Cyrille and Savant, Jean-Michel. Theoretical and mechanistic aspects
of proton-coupled electron transfer in electrochemistry. Current Opinion in Elec-
trochemistry, 1(1):104 { 109, 2017. ISSN 2451-9103.
Costentin, Cyrille; Evans, Dennis H.; Robert, Marc; Sav eant, Jean-Michel, and
Singh, Pradyumna S. Electrochemical approach to concerted proton and electron
transfers. reduction of the water-superoxide ion complex. J. Am. Chem. Soc.,
127(36):12490{12491, 2005.
116
Costentin, Cyrille; Robert, Marc, and Sav eant, Jean-Michel. Electrochemical
Concerted Proton and Electron Transfers. Potential-Dependent Rate Constant,
Reorganization Factors, Proton Tunneling and Isotope Eects. J. Electroanal.
Chem., 588(2):197{206, 2006.
Cox, M. Jocelyn; Siwick, Bradley J., and Bakker, Huib J. In
uence of ions on
aqueous acid-base reactions. ChemPhysChem, 10(1):236{244, 2009. ISSN 1439-
7641.
Daghetti, A.; Lodi, G., and Trasatti, S. Interfacial Properties of Oxides Used as
Anodes in the Electrochemical Technology . Mater. Chem. Phys., 8(1):1 { 90,
1983. ISSN 0254-0584.
Dalle Carbonare, Nicola; Cristino, Vito; Berardi, Serena; Carli, Stefano; Argazzi,
Roberto; Caramori, Stefano; Meda, Laura; Tacca, Alessandra, and Bignozzi,
Carlo Alberto. Hematite Photoanodes Modied with an FeIII Water Oxidation
Catalyst. ChemPhysChem, 15(6):1164{1174, 2014a. ISSN 1439-7641.
Dalle Carbonare, Nicola; Cristino, Vito; Berardi, Serena; Carli, Stefano; Argazzi,
Roberto; Caramori, Stefano; Meda, Laura; Tacca, Alessandra, and Bignozzi,
Carlo Alberto. Hematite Photoanodes Modied with an FeIII Water Oxidation
Catalyst. ChemPhysChem, 15(6):1164{1174, 2014b. ISSN 1439-7641.
Damjanovic, A.; Dey, A., and Bockris, J.OM. Kinetics of Oxygen Evolution and
Dissolution on Platinum Electrodes. Electrochim. Acta, 11(7):791 { 814, 1966.
ISSN 0013-4686.
Dau, Holger; Limberg, Christian; Reier, Tobias; Risch, Marcel; Roggan, Stefan,
and Strasser, Peter. The Mechanism of Water Oxidation: From Electrolysis via
Homogeneous to Biological Catalysis. ChemCatChem, 2(7):724{761, 2010. ISSN
1867-3899.
Dempsey, Jillian L.; Winkler, Jay R., and Gray, Harry B. Mechanism of h2 evolu-
tion from a photogenerated hydridocobaloxime. Journal of the American Chem-
ical Society, 132(47):16774{16776, 2010.
Deng, Yingxin; Josberger, Erik; Jin, Jungho; Roudsari, Anita Fadavi; Helms,
Brett A.; Zhong, Chao; Anantram, M. P., and Rolandi, Marco. H
+
-type and oh
-
-
type biological protonic semiconductors and complementary devices. Scientic
Reports, 3:1{7, 2013.
Donten, Mateusz L.; Hassan, Shabir; Popp, Alexander; Halter, Jonathan; Hauser,
Karin, and Hamm, Peter. ph-jump induced leucine zipper folding beyond the
diusion limit. The Journal of Physical Chemistry B, 119(4):1425{1432, 2015.
117
Douhal, Abderrazzak; Lahmani, Franoise, and Zewail, Ahmed H. Proton-transfer
reaction dynamics. Chemical Physics, 207(23):477 { 498, 1996. ISSN 0301-0104.
Doyle, Richard L.; Godwin, Ian J.; Brandon, Michael P., and Lyons, Michael E. G.
Redox and Electrochemical Water Splitting Catalytic Properties of Hydrated
Metal Oxide Modied Electrodes. Phys. Chem. Chem. Phys., 15:13737{13783,
2013.
Gabrielsson, Erik O.; Tybrandt, Klas, and Berggren, Magnus. Ion diode logics for
ph control. Lab on a Chip, 12:2507{2513, 2012.
Galn-Mascars, Jos Ramn. Water Oxidation at Electrodes Modied with Earth-
Abundant Transition-Metal Catalysts. ChemElectroChem, 2(1):37{50, 2015.
ISSN 2196-0216.
Gasteiger, Hubert A.; Markovic, Nenad M., and Jr., Philip N. Ross. H
2
and co
electrooxidation on well-characterized pt, ru, and pt-ru. 1. rotating disk electrode
studies of the pure gases including temperature eects. J. Phys. Chem., 99(20):
8290{8301, 1995.
Gileadi, E. Electrode Kinetics for Chemists, Chemical Engineers, and Materials
Scientists. VCH, 1993. ISBN 9781560816263.
Giordano, Livia; Han, Binghong; Risch, Marcel; Hong, Wesley T.; Rao, Reshma R.;
Stoerzinger, Kelsey A., and Shao-Horn, Yang. ph dependence offOERg activity
of oxides: Current and future perspectives. Catalysis Today, 262:2 { 10, 2016.
ISSN 0920-5861. Electrocatalysis.
Gong, Ming and Dai, Hongjie. A Mini Review of NiFe-Based Materials as Highly
Active Oxygen Evolution Reaction Electrocatalysts. Nano Research, 8(1):23{39,
2015. ISSN 1998-0124.
Grahame, David C. Fiftieth anniversary: Mathematical theory of the faradaic
admittance: Pseudocapacity and polarization resistance. Journal of The Elec-
trochemical Society, 99(12):370C{385C, 1952.
Grimminger, J. and Schmickler, W. Dynamics of combined electron- and proton
transfer at metal electrodes. Chem. Phys., 334(13):8 { 17, 2007.
Grimminger, J.; Bartenschlager, S., and Schmickler, W. A model for combined
electron and proton transfer in electrochemical systems. Chem. Phys. Lett., 416
(46):316 { 320, 2005.
Guidelli, Rolando; Compton, Richard G.; Feliu, Juan M.; Gileadi, Eliezer; Lip-
kowski, Jacek; Schmickler, Wolfgang, and Trasatti, Sergio. Dening the transfer
118
coecient in electrochemistry: An assessment (iupac technical report). 2014-01-
18.
Haghighat, Shima and Dawlaty, Jahan M. Continuous representation of the proton
and electron kinetic parameters in the ph-potential space for water oxidation on
hematite. J. Phys. Chem. C, 119(12):6619{6625, 2015.
Hamann, Thomas W. Splitting Water with Rust: Hematite Photoelectrochemistry.
Dalton Trans., 41:7830{7834, 2012.
Hammes-Schier, Sharon and Stuchebrukhov, Alexei A. Theory of coupled electron
and proton transfer reactions. Chem. Rev., 110(12):6939{6960, 2010.
Ireland, J.F. and Wyatt, P.A.H. Acid-base properties of electronically excited
states of organic molecules. volume 12 of Advances in Physical Organic Chem-
istry, pages 131 { 221. Academic Press, 1976.
Irie, Masahiro. Light-induced reversible ph change. Journal of the American
Chemical Society, 105(7):2078{2079, 1983.
Irimia-Vladu, Mihai; Sariciftci, Niyazi Serdar, and Bauer, Siegfried. Exotic mate-
rials for bio-organic electronics. Journal of Materials Chemistry, 21:1350{1361,
2011.
Jackson, Megan N. and Surendranath, Yogesh. Donor-dependent kinetics of inter-
facial proton-coupled electron transfer. Journal of the American Chemical Soci-
ety, 138(9):3228{3234, 2016. PMID: 26862666.
Kanan, Matthew W. and Nocera, Daniel G. In Situ Formation of an Oxygen-
Evolving Catalyst in Neutral Water Containing Phosphate and Co2+. Science,
321(5892):1072{1075, 2008.
Katz, Michael J.; Riha, Shannon C.; Jeong, Nak Cheon; Martinson, Alex B.F.;
Farha, Omar K., and Hupp, Joseph T. Toward Solar Fuels: Water Splitting
with Sunlight and "Rust" . Coordination Chemistry Reviews, 256(21?22):2521
{ 2529, 2012. ISSN 0010-8545. Solar Fuels- by invitation only.
Klahr, Benjamin; Gimenez, Sixto; Fabregat-Santiago, Francisco; Bisquert, Juan,
and Hamann, Thomas W. Electrochemical and photoelectrochemical investi-
gation of water oxidation with hematite electrodes. Energy Environ. Sci., 5:
7626{7636, 2012.
Koper, Marc T. M. Theory of multiple proton-electron transfer reactions and its
implications for electrocatalysis. Chem. Sci., 4:2710{2723, 2013a.
119
Koper, Marc T. M. Theory of the transition from sequential to concerted elec-
trochemical proton-electron transfer. Phys. Chem. Chem. Phys., 15:1399{1407,
2013b.
Koper, Marc TM. Theory of Multiple Proton{Electron Transfer Reactions and its
Implications for Electrocatalysis. Chem. Sci., 4(7):2710{2723, 2013c.
Krstajic, N. and Trasatti, S. Cathodic Behaviour of RuO
2
-doped Ni/Co
3
O
4
Elec-
trodes in Alkaline Solutions: Hydrogen Evolution. J. Appl. Electrochem., 28
(12):1291{1297, 1998.
Kushner-Lenho, Maxwell N.; Blakemore, James D.; Schley, Nathan D.; Crabtree,
Robert H., and Brudvig, Gary W. Eects of Aqueous Buers on Electrocatalytic
Water Oxidation with an Iridium Oxide Material Electrodeposited in Thin Lay-
ers from an Organometallic Precursor. Dalton Trans., 42:3617{3622, 2013.
Leger, J.; Berggren, M., and Carter, S. Iontronics: Ionic Carriers in Organic
Electronic Materials and Devices. CRC Press, 2010. ISBN 9781439806890.
Leiderman, Pavel; Genosar, Liat; , and Huppert, Dan. Excited-state proton trans-
fer: Indication of three steps in the dissociation and recombination process. The
Journal of Physical Chemistry A, 109(27):5965{5977, 2005.
Ling, Yichuan; Wang, Gongming; Wang, Hanyu; Yang, Yi, and Li, Yat. Low-
Temperature Activation of Hematite Nanowires for Photoelectrochemical Water
Oxidation. ChemSusChem, 7(3):848{853, 2014. ISSN 1864-564X.
Lodi, G.; Sivieri, E.; De Battisti, A., and Trasatti, S. Ruthenium Dioxide-Based
Film Electrodes. J. Appl. Electrochem., 8(2):135{143, 1978. ISSN 0021-891X.
Lovrecek, B.; Despic, A., and Bockris, J. O. M. Electrolytic junctions with recti-
fying properties. The Journal of Physical Chemistry, 63(5):750{751, 1959.
Ludlow, Michelle K; Soudackov, Alexander V, and Hammes-Schier, Sharon. Elec-
trochemical Proton-Coupled Electron Transfer of an Osmium Aquo Complex:
Theoretical Analysis of Asymmetric Tafel Plots and Transfer Coecients. J.
Am. Chem. Soc., 132(4):1234{1235, 2010.
Matthews, D.B. and Bockris, J.O.M.. The Mechanism Of Charge Transfer From
Metal Electrodes To Ions In Solution. Modern Aspects of Electrochemistry
Series. Springer, 1971. ISBN 9780306464621.
Mehandru, S. P. and Anderson, Alfred B. Oxygen Evolution on a SrFeO
3
Anode:
Mechanistic Considerations from Molecular Orbital Theory. J. Electrochem.
Soc., 136(1):158{166, 1989.
120
Minguzzi, Alessandro; Fan, Fu-Ren F.; Vertova, Alberto; Rondinini, Sandra, and
Bard, Allen J. Dynamic Potential-pH Diagrams Application to Electrocatalysts
for Water Oxidation. Chem. Sci., 3:217{229, 2012.
Miyake, Takeo; Josberger, Erik E.; Keene, Scott; Deng, Yingxin, and Rolandi,
Marco. An enzyme logic bioprotonic transducer. APL Materials, 3(1):014906,
2015.
Mohammed, Omar F.; Pines, Dina; Dreyer, Jens; Pines, Ehud, and Nibbering, Erik
T. J. Sequential proton transfer through water bridges in acid-base reactions.
Science, 310(5745):83{86, 2005.
Najafpour, Mohammad Mahdi; Renger, Gernot; Hoyska, Magorzata; Moghaddam,
Atefeh Nemati; Aro, Eva-Mari; Carpentier, Robert; Nishihara, Hiroshi; Eaton-
Rye, Julian J.; Shen, Jian-Ren, and Allakhverdiev, Suleyman I. Manganese com-
pounds as water-oxidizing catalysts: From the natural water-oxidizing complex
to nanosized manganese oxide structures. Chemical Reviews, 116(5):2886{2936,
2016.
Norby, Truls. The promise of protonics. Nature, 410(6831):877{878, 04 2001.
Norskov, J.K.; Bligaard, T.; Logadottir, A.; Bahn, S.; Hansen, L.B.; Bollinger, M.;
Bengaard, H.; Hammer, B.; Sljivancanin, Z.; Mavrikakis, M.; Xu, Y.; Dahl, S.,
and Jacobsen, C.J.H. Universality in Heterogeneous Catalysis . J. Catal., 209
(2):275 { 278, 2002. ISSN 0021-9517.
Ordinario, David D.; Phan, Long; Walkup IV, Ward G.; Jocson, Jonah-Micah;
Karshalev, Emil; H usken, Nina, and Gorodetsky, Alon A. Bulk protonic con-
ductivity in a cephalopod structural protein. Nature Chemistry, 6(7):596{602,
07 2014.
Peter, Laurence M. and Upul Wijayantha, K. G. Photoelectrochemical Water
Splitting at Semiconductor Electrodes: Fundamental Problems and New Per-
spectives. ChemPhysChem, 15(10):1983{1995, 2014. ISSN 1439-7641.
Pines, Ehud; Huppert, Dan, and Agmon, Noam. Geminate recombination
in excited-state proton-transfer reactions: Numerical solution of the debye-
smoluchowski equation with backreaction and comparison with experimental
results. The Journal of Chemical Physics, 88(9):5620{5630, 1988.
Pleskov, I.U.V. Solar Energy Conversion: A Photoelectrochemical Approach.
Springer-Verlag, 1990. ISBN 9780387514741.
Rini, Matteo; Magnes, Ben-Zion; Pines, Ehud, and Nibbering, Erik T. J. Real-time
observation of bimodal proton transfer in acid-base pairs in water. Science, 301
(5631):349{352, 2003.
121
Robinson, R. A. and Stokes, R. H. Dover Publications, 1959. ISBN 978-0-486-
42225-1.
Schmidt, T. J.; Gasteiger, H. A.; St ab, G. D.; Urban, P. M.; Kolb, D. M., and
Behm, R. J. Characterization of high-surface-area electrocatalysts using a rotat-
ing disk electrode conguration. J. Electrochem. Soc., 145(7):2354{2358, 1998.
Schrauben, Joel N; Hayoun, Rebecca; Valdez, Carolyn N; Braten, Miles; Fridley,
Lila, and Mayer, James M. Titanium and Zinc Oxide Nanoparticles are Proton-
Coupled Electron Transfer Agents. Science, 336(6086):1298{1301, 2012.
Shinagawa, Tatsuya and Takanabe, Kazuhiro. Towards versatile and sustainable
hydrogen production through electrocatalytic water splitting: Electrolyte engi-
neering. ChemSusChem, pages n/a{n/a, 2017. ISSN 1864-564X.
Shizuka, Haruo. Excited-state proton-transfer reactions and proton-induced
quenching of aromatic compounds. Accounts of Chemical Research, 18(5):141{
147, 1985.
Singh, Archana and Spiccia, Leone. Water Oxidation Catalysts Based on Abundant
1st Row Transition Metals . Coordination Chemistry Reviews, 257(17?18):2607
{ 2622, 2013. ISSN 0010-8545. 40th International Conference on Coordination
Chemistry, Valencia, Spain, Sept. 9-13th 2012.
Sivula, Kevin; LeFormal, Florian, and Grtzel, Michael. Solar Water Splitting:
Progress Using Hematite (-Fe
2
O
3
) Photoelectrodes. ChemSusChem, 4(4):432{
449, 2011. ISSN 1864-564X.
Siwick, B. J.; Cox, M. J., and Bakker, H. J. Long-range proton transfer in aqueous
acid-base reactions. The Journal of Physical Chemistry B, 112(2):378{389, 2008.
Sklason, Egill; Tripkovic, Vladimir; Bjrketun, Mrten E.; Gudmundsdttir, Sigr-
dur; Karlberg, Gustav; Rossmeisl, Jan; Bligaard, Thomas; Jnsson, Hannes, and
Nrskov, Jens K. Modeling the electrochemical hydrogen oxidation and evolution
reactions on the basis of density functional theory calculations. The Journal of
Physical Chemistry C, 114(42):18182{18197, 2010.
Smith, Rodney D. L.; Pr evot, Mathieu S.; Fagan, Randal D.; Zhang, Zhipan;
Sedach, Pavel A.; Siu, Man Kit Jack; Trudel, Simon, and Berlinguette, Curtis P.
Photochemical Route for Accessing Amorphous Metal Oxide Materials for Water
Oxidation Catalysis. Science, 340(6128):60{63, 2013.
Spry, D. B. and Fayer, M. D. Observation of slow charge redistribution preceding
excited-state proton transfer. The Journal of Chemical Physics, 127(20):204501,
2007.
122
Spry, D. B.; Goun, A.; , and Fayer, M. D. Deprotonation dynamics and stokes
shift of pyranine (hpts). The Journal of Physical Chemistry A, 111(2):230{237,
2007.
Surendranath, Yogesh and Nocera, Daniel G. Oxygen Evolution Reaction Chem-
istry of Oxide-Based Electrodes. pages 505{560, 2011.
Surendranath, Yogesh; Kanan, Matthew W., and Nocera, Daniel G. Mechanistic
Studies of the Oxygen Evolution Reaction by a Cobalt-Phosphate Catalyst at
Neutral pH. J. Am. Chem. Soc., 132(46):16501{16509, 2010.
Tan, Ming X.; Laibinis, Paul E; Nguyen, Sonbinh T.; Kesselman, Janet M.; Stan-
ton, Colby E., and Lewis, Nathan S. Principles and Applications of Semiconduc-
tor Photoelectrochemistry, pages 21{144. John Wiley & Sons, Inc., 2007. ISBN
9780470166420.
Tarabella, Giuseppe; Mahvash Mohammadi, Farzaneh; Coppede, Nicola; Barbero,
Francesco; Iannotta, Salvatore; Santato, Clara, and Cicoira, Fabio. New oppor-
tunities for organic electronics and bioelectronics: Ions in action. Chemical
Science, 4:1395{1409, 2013.
Tilak, B.V. and Conway, B.E. Analytical Relations Between Reaction Order and
Tafel Slope Derivatives for Electrocatalytic Reactions Involving Chemisorbed
Intermediates. Electrochim. Acta, 37(1):51 { 63, 1992. ISSN 0013-4686.
Timmer, R. L. A.; Cox, M. J., and Bakker, H. J. Direct observation of proton trans-
fer in ice ih using femtosecond spectroscopy. The Journal of Physical Chemistry
A, 114(5):2091{2101, 2010.
Tolbert, Laren M.; , and Solntsev, Kyril M. Excited-state proton transfer: From
constrained systems to \super" photoacids to superfast proton transfer. Accounts
of Chemical Research, 35(1):19{27, 2002.
Tributsch, Helmut. Light driven proton pumps. Ionics, 6(3-4):161{171, 2000. ISSN
0947-7047.
Vayssieres, Lionel; Beermann, Niclas; Lindquist, Sten-Eric, and Hagfeldt, Anders.
Controlled Aqueous Chemical Growth of Oriented Three-Dimensional Crys-
talline Nanorod Arrays: Application to Iron(III) Oxides. Chemistry of Materials,
13(2):233{235, 2001.
Venkataraman, Charulatha; Soudackov, Alexander V, and Hammes-Schier,
Sharon. Theoretical Formulation of Nonadiabatic Electrochemical Proton-
Coupled Electron Transfer at Metal- Solution Interfaces. J. Phys. Chem. C,
112(32):12386{12397, 2008.
123
Wen, Liping; Hou, Xu; Tian, Ye; Zhai, Jin, and Jiang, Lei. Bio-inspired photo-
electric conversion based on smart-gating nanochannels. Advanced Functional
Materials, 20(16):2636{2642, 2010. ISSN 1616-3028.
Wiechen, M.; Najafpour, M. M.; Allakhverdiev, S. I., and Spiccia, L. Water
Oxidation Catalysis by Manganese Oxides: Learning from Evolution. Energy
Environ. Sci., 7:2203{2212, 2014.
Xi, Lifei; Tran, Phong D.; Chiam, Sing Yang; Bassi, Prince Saurabh; Mak,
Wai Fatt; Mulmudi, Hemant Kumar; Batabyal, Sudip K.; Barber, James;
Loo, Joachim Say Chye, and Wong, Lydia Helena. Co3o4-decorated hematite
nanorods as an eective photoanode for solar water oxidation. The Journal of
Physical Chemistry C, 116(26):13884{13889, 2012.
Yamaguchi, Akira; Inuzuka, Riko; Takashima, Toshihiro; Hayashi, Toru;
Hashimoto, Kazuhito, and Nakamura, Ryuhei. Regulating Proton-Coupled Elec-
tron Transfer for Ecient Water Splitting by Manganese Oxides at Neutral pH.
Nat. Commun., 5:1{6, 2014.
Yamamoto, Tetsuya and Doi, Masao. Electrochemical mechanism of ion current
rectication of polyelectrolyte gel diodes. Nature Communications, pages 1{8,
06 .
Zeng, Kai and Zhang, Dongke. Recent Progress in Alkaline water Electrolysis for
Hydrogen Production and Applications . Progress in Energy and Combustion
Science, 36(3):307 { 326, 2010. ISSN 0360-1285.
124

Abstract (if available)

Abstract Many (photo-)electrochemical processes involve the transfer of both protons and electrons (R + nH⁺ + ne⁻ ⇌ P) at the electrochemical interfaces. Fundamental understanding of the complex mechanistic details of these processes is crucial to improve their overall efficiency. The operating conditions, such as applied electrode potential and the electrolyte pH, can impact the electron transfer rate and the proton transfer rate in a non-trivial way. In particular, pH influences the kinetic of proton transfer between the electrolyte and the surface adsorbed intermediates which can consequently modify the electron transfer rate to/from the electrode surface. In this research, the importance of proton transfer kinetics is addressed. ❧ Toward this goal, we first study the kinetics of electrochemical water splitting half reactions, oxygen evolution reaction (OER) and hydrogen evolution reaction (HER). We demonstrate the variations of two important kinetic parameters as a function of pH and potential for the OER on metal oxide electrodes. The electron transfer parameter (α) and the proton reaction order (ρ) are represented in 2-dimensional contour plots in the pH-potential space. We propose different mechanism pathways for oxygen evolution based on the variation of α and ρ as a function of pH and potential. To study the kinetic of the HER, we measure α on several electrode materials over a wide range of pH. We compare our experimental results with a 2D potential energy surface model that has been investigated before. There are several similarities between the pH dependence of α derived from the model and the empirical α that is obtained from the experiment. ❧ To further explore the proton transfer kinetics, we take a different path and study protonic photoconductivity. In electronic photoconductivity, a material becomes more electrically conductive due to generation of mobile electrons with light. Here we report, for the first time a protonic analogue of photoconductivity. When a material is doped with photoacid, light excitation generates extra mobile protons that change the low-frequency conductivity of the material. We measure and model such change both in polyethylene glycol (PEG) and in water sandwiched between two transparent electrodes and doped with a well-known photoacid 8- hydroxypyrene-1,3,6-trisulfonic acid (HPTS). We anticipate that this scheme can be employed in protonic circuits where direct transduction of energy from light to protonic gradients or protonic currents is necessary.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Excited state proton transfer in quinoline photobases

PDF

Interfacial polarization and photobasicity: spectroscopic studies of the electrode environment and photochemical effects

PDF

Ultrafast dynamics of excited state intra- and intermolecular proton transfer in nitrogen containing photobases

PDF

Thermoelectric and transport studies of low dimensional materials & hot electron-driven photocatalysis on plasmon-resonant grating nanostructures

PDF

Electrochemical pathways for sustainable energy storage and energy conversion

PDF

Semiconductor devices for vacuum electronics, electrochemical reactions, and ultra-low power in-sensor computing

PDF

Understanding the mechanism of oxygen reduction and oxygen evolution on transition metal oxide electrocatalysts and applications in iron-air rechargeable battery

PDF

Design and characterization of flow batteries for large-scale energy storage

PDF

Developing efficient methods for the manufacture and analysis of composites structures

PDF

Understanding the factors affecting the performance of iron and nickel electrodes for alkaline nickel-iron batteries

PDF

Sustainable manufacturing of out-of-autoclave (OoA) prepreg: processing challenges

PDF

Hot carriers in bare metals and photocatalytically active defect sites in dielectric/metal structures

PDF

Monitoring local electric fields, charge densities at electrode surfaces using in-situ graphene/surface enhanced Raman spectroscopy (GERS/SERS) based Stark-shifts

PDF

Nano-fabricated devices in electrochemistry and cancer therapy

PDF

Hydrogen peroxide vapor for small satellite propulsion

PDF

Particle-in-cell simulations of kinetic-scale turbulence in the solar wind

PDF

Iridium and ruthenium complexes for catalytic hydrogen transfer reactions

PDF

Enhanced photocatalysis on titanium oxide passivated III-V semiconductors

PDF

Water partitioning between the bulk and an electrode surface

PDF

The use of carbon molecule sieve and Pd membranes for conventional and reactive applications

Asset Metadata

Creator Haghighat, Shima (author)

Core Title Proton kinetics in electrochemistry: new directions and mechanistic analysis

School Viterbi School of Engineering

Degree Doctor of Philosophy

Degree Program Materials Science

Publication Date 11/07/2017

Defense Date 04/28/2017

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag electron transfer kinetics,hydrogen evolution reaction,OAI-PMH Harvest,oxygen evolution reaction,pH,proton transfer kinetics,protonic photoconductivity

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Dawlaty, Jahan (committee chair), Nutt, Steve (committee member), Ravichandran, Jayakanth (committee member)

Creator Email shaghigh@usc.edu,shima.haghighat@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c40-451169

Unique identifier UC11264008

Identifier etd-HaghighatS-5881.pdf (filename),usctheses-c40-451169 (legacy record id)

Legacy Identifier etd-HaghighatS-5881.pdf

Dmrecord 451169

Document Type Dissertation

Rights Haghighat, Shima

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA