Interplay between Electronic Structure and Catalytic Activity
in Transition Metal Oxide Model System
ARCHIVES
MASSACHUSETTS INST MITE
O TECHNOLOGY
By
JUW
Jin Suntivich
0 S 2 012
2RARIES
B.A. Integrated Science, B.S. Materials Science and Engineering
Northwestern University, 2006
SUBMITTED TO THE DEPARTMENT OF MATERIALS SCIENCE AND
ENGINEERING IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF SCIENCE
AT THE MASSACHUSETTS INSTITUTE OF TECHNOLOGY
JUNE 2012
C Massachusetts Institute of Technology 2012. All rights reserved.
The author hereby grants to MIT permission to reproduce and to distribute publicly paper
and electronic copies of this thesis document in whole or in part in any medium now
known or hereafter created.
Signature of Author:
Departmentof Materials Science and Engineering
May 25 th2012
Certified by:
Yang Shao-Horn
Gale E. Kendall rofessor of Mechanical Engineering and Materials Science
Accepted by:
Gerbrand Ceder
R.P. Simmons Professor of Materials Science
Chairman, Department Committee on Graduate Theses
2
Interplay between Electronic Structure and Catalytic Activity
in Transition Metal Oxide Model System
By
Jin Suntivich
Submitted to the Department of Materials Science and Engineering on May
2 5 th,
2012 in
partial fulfillment of the requirement for the degree of Doctoral of Philosophy
ABSTRACT
The efficiency of many energy storage and conversion technologies, such as hydrogen
fuel cells, rechargeable metal-air batteries, and hydrogen production from water splitting,
is limited by the slow kinetics of the oxygen electrochemical reactions. Transition-metal
oxides can exhibit high catalytic activity for oxygen electrochemical reactions, which can
be used to improve efficiency and cost of these devices. Identifying a catalyst "design
principle" that links material properties to the catalytic activity can accelerate the
development of highly active, abundant transition metal oxide catalysts fore more
efficient, cost-effective energy storage and conversion system.
In this thesis, we
demonstrate that the oxygen electrocatalytic activity for perovskite transition metal oxide
catalysts primarily correlates to the a* orbital ("eg") occupation. We further find that the
extent of B-site transition metal-oxygen covalency can serve as a secondary activity
descriptor. We hypothesize that this correlation reflects the critical influences of the a*
orbital and transition metal-oxygen covalency on the ability of the surface to displace and
stabilize oxygen-species on surface transition metals. We further propose that this ability
to stabilize oxygen-species reflect as the rate-limiting steps of the oxygen electrochemical
3
reactions on the perovskite oxide surfaces, and thus highlight the importance of electronic
structure in controlling the oxide catalytic activity.
Thesis Supervisor: Yang Shao-Horn
Title: Gale E. Kendall Professor of Mechanical Engineering and Materials Science
4
To my parents, Vorathai and Chungcharoen,my brother,Jate, and my teachers
5
Contents
List of Figures
8-13
List of Tables
14
List of Schemes
15
Chapter 1 - Oxygen Electrocatalysis
1.1
Motivation
1.2
Introduction to oxygen electrocatalysis
17-22
1.3
Scope of the thesis
22-23
Chapter 2 - Molecular Orbital Approach to Perovskite Oxides
2.1
17
24-40
Simplified molecular orbital theory for perovskite oxide
24-25
2.2 Transition metal - oxygen covalency in perovskite oxide
26-27
2.3
Experimental procedure
2.4
Covalency trend in model series of perovskite oxides
28-33
2.5
Chapter 2 conclusion
35-36
Chapter 3 - Thin-film Rotating Disk Electrode
27
37-53
3.1
Thin-film rotating disk electrode methodology
37-38
3.2
Experimental procedure
39-43
3.3
Results from thin-film rotating disk electrode
43-52
3.4
Chapter 3 conclusion
52-53
Chapter 4 - Activity Descriptor for ORR
6
17-23
54-76
4.1
Introduction
54-55
4.2
Experimental procedure
55-63
4.3
Method for ORR activity comparison
63-67
4.4
ORR activity descriptor identification
67-69
4.5
Proposed ORR mechanism
72-74
4.6
Influence of B-O covalency on ORR activity
4.7
Chapter 4 conclusion
74
75-76
Chapter 5 - Activity Descriptor for OER
77-92
77
5.1
Introduction
5.2
Experimental procedure
77-83
5.3
OER activity descriptor identification
83-87
5.4
Proposed OER mechanism
87-88
5.5
Application of the activity descriptor
90
5.6
Chapter 5 conclusion
92
Chapter 6 - Influence of the Cation Species
93-104
6.1
Introduction
93-95
6.2
Experimental procedure
95-97
6.3
Influence of Li' cation on the ORR
97-98
6.4
Influence of Li* cation on the OER
99
6.5
Proposed cation mechanism
101-102
6.6
Chapter 5 conclusion
103-104
Chapter 7 - Summary and Outlook
7.1
Thesis summary
7.2
Outlook
Reference
105-108
105
106-107
109-125
Acknowledgement
7
List of Figures
Figure 1.1.
Illustration of the oxygen electrode in PEMFC ("conversion") and PEM
electrolyzer ("storage").
The oxygen electrochemical reaction occurs at the ionomer-
catalyst interface. Image was adapted from the initial drawing of Dr. Eva Mutoro.
Figure 1.2. Illustration of the oxygen electrode in AMFC ("conversion") and anionexchanged membrane electrolyzer ("storage").
Image was adapted from the initial
drawing of Dr. Eva Mutoro.
Figure 1.3. Illustration of the Sabatier Principle - "Volcano plot", where varying the
bond strength between the catalyst surface and the reactant/product to moderate bond
strength (neither too strong nor too weak) can lead to maximum kinetics activity.
Figure 1.4. Illustration of the perovskite structure, where the blue atom represents the
transition metal position (B-site), the red atom represents the oxygen, and the yellow
atom represents rare earth, alkaline earth, or mix therefore (A-site).
Figure 2.1. 0 K-edge X-ray Absorption Spectroscopy of LaBO 3 (B = Cr, Mn, Fe, Co,
Ni). The shaded region centered at -530 eV represents the feature assigned to B 3d - 0
2p hybridization. Spectra were collected in electron yield mode at room temperature.
Figure 2.2. Evolution of LaBO 3 covalency paremeter (Abs./(holeeg + 1/4 holet2g) for B
Cr, Mn, Fe, Co, Ni with respect to d-electron count. The absorbance was taken to the
shade region shown in Figure 2.1 (corresponding to the unoccupied state with the 0 2p
symmetry,) corrected with linear background. The dashed gray line serves only as guide
to the trend.
Figure 2.3. 0 K-edge X-ray Absorption Spectroscopy LaBB'0
Ni. 5 Mno. 5).
(Top) Raw spectra.
reference. (Bottom) LaBB'0
3
3
(BB'
=
The dashed grey lines show a LaMn0
3
5
and
0 K-edge
covalency parameter v.s. d-electron count, with LaBO 3
data as references. The dashed gray line serves only as guide to the trend.
8
Cuo. 5 Mn.
Figure 2.4.
0 K-edge X-ray Absorption Spectroscopy LaxCai..BO 3 and La1+,NiO 3+..
(A) Lao. 5 Cao. 5 CrO 3 and Lao.5Cao.5 CoO 3 .6. LaCrO 3 and LaCoO 3 are in shown dashed grey
lines as references. (B) La 0 .5Cao. 5MnO 3 and LaMnO 3+6 . LaMnO 3 are shown in dashed
grey lines. (C) Lao. 5 Cao. 5 FeO 3 and La 0 .7 5 Ca 0 .2 5 FeO 3. LaFeO 3 are shown in dashed grey
lines. (D) La 4Ni 3010 and La 2NiO 4. LaNiO 3 are shown in dashed grey lines.
Figure 2.5.
xCaxBO 3
Evolution of covalency with respect to different B oxidation using Lai.
and Lai+xNiO 3+x as model compounds. The covalency enhancement factor is
calculated from a ratio of a covalency parameter of either La1.xCaxBO3 or La1+xNi0 3+x
with respect to LaBO 3 material.
Figure 3.1. X-ray diffraction spectra of the oxides studied in this chapter.
Figure 3.2.
Cyclic voltammograms of LaNiO 3 at 10 mV s- 1 in 0.1 M KOH at room
temperature in N2-saturated electrolyte at 0 rpm (blue) or in 0 2 -saturated electrolyte at
1600 rpm. Sweep directions are shown by the arrow.
Figure 3.3.
ORR current densities (capacity-corrected negative-going scans) of GC-
supported and Nafion*-bonded (Na* ion exchanged) thin-film LaNiO 3, LaCuo.5 Mno. 5 0 3 ,
Lao. 7 5 Cao.2 5 FeO 3 electrodes at 1600 rpm in 0 2-saturated 0.1 M KOH at 10 mV s-'. The
background ORR activity deriving from either the GC substrate or a thin-film Nafion*bonded acetylene black (AB) thin-film electrode is shown for reference.
Figure 3.4.
Polarization curves of GC supported thin-film LaCuo.5Mno. 50 3 and Pt/C
electrodes using a Na+ ion-exchanged Nafion* binder: 1600 rpm in 0 2-saturated 0.1 M
KOH at 10 mV s~1 .
Figure 3.5.
Koutecky-Levich analyses for GC-supported thin-film LaCuo.5 Mno.5 0 3 ,
LaNiO 3, and Pt/C electrodes (02 saturated 0.1 M KOH at 10 mV s-1).
Figure 3.6. ORR specific activities, is, of of LaNiO 3, LaCuo.5 Mno. 50 3, La0. 7 5 Cao. 2 5FeO 3
(using specific surface areas listed in table 1) and of Pt/C versus potential, obtained after
iR and 02 mass-transport correction (from capacity-corrected negative-going scans at
10 mV s-1 in 0.1 M KOH for oxides, from positive-going scans at 10 mV s~1 in 0.1M
9
KOH for Pt/C). Error bars represent standard deviations from at least three independent
repeat measurements.
Figure 3.7. ORR mass activities, im, of of LaNiO 3, LaCuo. 5Mno. 50 3, Lao. 7 5 Cao. 2 5 FeO 3
(using specific surface areas listed in table 1) and of Pt/C versus potential, obtained after
iR and 02 mass-transport correction ((from capacity-corrected negative-going scans at
10 mV s4 in 0.1 M KOH for oxides, from positive-going scans at 10 mV s- in 0.1M
KOH for Pt/C). Error bars represent standard deviations from at least three independent
repeat measurements.
Figure 3.8.
Projected kinetically limited cathode performance under an idealized
scenario of full utilization of the catalyst with minimum mass transport loss for electrodes
with different ORR catalysts and catalyst loadings: LaNiO 3 at 36 mgoxide cm- 2 electrode or
commercial Pt/C at 0.4 mgpt cm 2 electrode. Projections are based on the mass activity data
in figure 7 (100 kPaabs 02 at room temperature).
Figure 4.1. X-ray Absorption Spectra LaMnO 3 vs. La0 .5Cao.5MnO3
Figure 4.2. X-ray Absorption Spectra LaFeO 3 vs. Lao. 5Cao.5FeO3
Figure 4.3. X-ray Absorption Spectra LaCoO 3 vs. Lao.5 Cao. 5 CoO3 6
Figure 4.4.
Determination of B-site oxidation state
of LaNio. 5Mno. 50 3 and
LaCuo. 5Mno. 50 3. (Top) Mn oxidation state for LaNio. 5Mno. 50 3 (red circle) and
LaCuo. 5 Mno.5 0
3
(blue triangle) is estimated to be ~3.7+. (Bottom) Ni oxidation state in
LaNio.5 Mno.50 3 (red circle) is estimated to be ~2.3+.
Figure 4.5.
Oxygen reduction activity of LaCuo.5Mno. 50 3 electrode in 0 2-saturated
0.1 M KOH at 10 mV/s scan rate at rotation rates of 100, 400, 900, and 1600 rpm
respectively.
The Koutecky-Levich analysis (inset) of the limiting currents (0.4 V)
indicates a 4e~ transfer reaction.
Figure 4.6. Specific activities of LaCuo.5 Mno.50 3, LaMnO 3, LaCoO 3 , and LaNiO 3. The
potential at 25 pA/cm 2ox is used as a benchmark for comparison (shown as the
intersection between the activity and a horizontal gray dashed line).
10
Figure 4.7. Tafel plot of the ORR specific activity of oxides studied in this chapter using
thin film RDE.
Figure 4.8. Tafel plot of the ORR specific activity of LaMnO 3+6 vs. Pt/C from the thin
film RDE experiments. These two catalysts have very similar ORR activities (~50 mV
shift, corresponding to less than an order of activity difference).
Error bars represent
standard deviations of at least three independent measurements.
Figure 4.9.
The potentials at 25 pA/cm 2 ex of the perovskite oxides have M-shaped
relationship with the d-electron number. The data symbols vary with the type of B ions
(square red for Cr, orange circle for Mn, beige upside up triangle for Fe, green upside
down triangle for Co, blue diamond for Ni, and purple left triangle for mixed compounds),
where x = 0 and 0.5 for Cr, and 0, 0.25, and 0.5 for Fe. Error bars represents standard
deviations of at least 3 measurements.
The potentials at 25 gA/cm 2
Figure 4.10.
ox
as a function of eg-orbital in perovskite-
based oxides. The data symbols vary with the type of B ions (square red for Cr, orange
circle for Mn, beige upside up triangle for Fe, green upside down triangle for Co, blue
diamond for Ni, and purple left triangle for mixed compounds), where x = 0 and 0.5 for
Cr, and 0, 0.25, and 0.5 for Fe. Error bar represents standard deviations of at least 3
measurements.
Figure 4.11.
Proposed ORR mechanism on perovskite oxide catalysts.
The ORR
proceeds via 4 steps: 1 surface hydroxide displacement, 2 surface peroxide formation, 3
surface oxide formation, and 4 surface hydroxide regeneration.
Figure 4.12. The role of B-O covalency on the ORR activity of perovskite oxides. The
potentials at 25 pA/cm 2 ex as a function of the B-O covalency at eg filling = 1. Error bars
represents standard deviations of at least 3 measurements.
Figure 5.1.
X-ray Diffraction Pattern of Ba 0 .5 Sro.5 Coo.8Feo.2O3-6.
The experiment was
conducted at room temperature using Cr source. The pattern revealed a space group of P
m -3 m (simple cubic) with a lattice parameter of 3.99 A.
11
Figure 5.2.
Ohmic and capacitive corrections of the as-measured OER activity of
example perovskite oxide catalyst. The as-measured OER activity of the Lao.5 Cao.5 CoO 3 6
thin-film composite catalyst ("raw", solid red line) is capacity-corrected by taking an
average of forward and backward (positive and negative-going)
scans.
The
capacity-corrected OER current (short-dash blue line) is then ohmically corrected with
the measured ionic resistance (~45 Q) to yield the final electrode OER activity (long-dash
green line).
Figure 5.3.
Co K-edge XANES of CoO, LaCoO 3, and BSCF.
(A) The Co K-edge
spectra. (B) The edge position (Eo, defined to be the energy at the highest first derivative
of the absorbance) of BSCF fits right in between CoO and LaCoO 3 . By using the known
metal oxidation vs. E0 relationship, we estimate the Co oxidation state to be ~2.8+.
Figure 5.4. Example OER currents of LaixCaxCoO 3 and LaCoO 3 thin-films on GCE in
O2 -saturated 0.1 M KOH at 10 mV s- scan rate at 1600 r.p.m, which were
capacitance-corrected by taking an average of the positive and negative scans. The
contributions from AB and binder (Nafion*) and GCE are shown for comparison.
Figure 5.5.
The relation between the OER catalytic activity, defined by the
overpotentials at 50 piA cm
ox
of OER current, and the occupancy of the eg-symmetry
electron of surface B transition metal. Data symbols vary with type of B ions (Cr, red;,
Mn, orange; Fe, beige; Co, green; Ni, blue; mixed compounds, purple), where x = 0, 0.25,
and 0.5 for Fe. Error bars represent standard deviations of at least three independent
measurements. The dash volcano lines are shown for guidance only.
Figure 5.6. The role of B-O covalency on the OER activity of perovskite oxides. The
potentials at 50 gA/cm 2 x as a function of the B-O covalency, which is estimated by the
normalized soft X-ray absorbance at eg filling = 1.
Error bars represent standard
deviations of at least 3 measurements. The B-O covalency data was from chapter 2.
Figure 5.7. Evidence of 02 generated from BSCF using rotating ring disk electrode
measurements. The 02 gas generated from BSCF on GCE disk (OER current given as
idisk)
12
is reduced at the Pt ring at a constant potential of 0.4 V vs. RHE.
Figure 5.8. The OER specific activity of BSCF (square) and IrO2 (circle) as extracted
from the polarization.
Error bars represent standard deviations of at least three
independent measurements.
Figure 5.9. Galvanostatic (constant current) measurement of mass activities of BSCF,
ball-milled BSCF (BM-BSCF), and IrO 2 .
Figure 6.1. The ORR activity of thin-film Pt/C RDE in the positive potential-going scan
direction at 10 mV/s in various alkaline electrolytes at 900 r.p.m.
Figure 6.2. The influence of Nafion* on the ORR specific activity of thin-film Pt/C
RDE in KOH and LiOH.
Figure 6.3. The ORR activity of thin-film LaMnO3 +6 (supported on acetylene black and
neutralized Nafion*) RDE in the negative potential-going scan direction at 10 mV/s in
various alkaline electrolytes at 1600 r.p.m.
Figure 7.1.
Schematic
drawing
of an integrated
catalyst-p-n junction
for
photoelectrochemical water splitting.
Figure 7.2. Schematic drawings of a proposed catalytically active oxide semiconductorliquid junction for photoelectrochemical water splitting. The catalytically active eg state
can either be integrated as a Fermi-level pinner at the surface (Top), or as a valence band,
where hole can be generated upon excitation (Bottom).
13
List of Tables
SEM characterization of the oxides studied in this chapter.
Table 3.1.
averaged diameter,
dnumber,
The number
with standard deviation (in parentheses), the volume-area
averaged diameter, d/a, and the specific surface area of the oxide (see equation 1), As.
were obtained from particle size distribution measurements.
Table 4.1. The derived lattice parameters of the oxide model compounds in this work
Table 4.2. Characterizations of the oxides studied in this chapter. The number averaged
diameter,
dnumber,
the volume-area averaged diameter, dv/a, and the specific surface area,
As. were obtained from particle size distribution measurements.
Methodology for
calculating each variable is given elsewhere (Ref. 14).
Table 4.3. Summary of literature information on the spin state of the perovskite oxides
Table 5.1. The number averaged diameter,
dnumber,
the volume-area averaged diameter,
d/a, and the specific surface area, As, were obtained from particle size distribution
measurements as described in the previous chapter. Note that we also measured the BET
area for BSCF, which were found to be within a factor of 2 of the specific area
determined from SEM.
Table 5.2.
catalysts.
Benchmark OER activities of various state-of-the-art perovskite oxide
All the OER activities were ohmic-corrected and normalized to the actual
surface area at 0.4 V overpotential in basic electrolyte (>13 pH).
14
List of Schemes
Scheme 4.1.
Molecular orbital of B0
6
interaction.
The energy levels of the
d-electrons in free B ion, octahedral B0 6 , and surface B0 6 configurations. The initial d
manifold degeneracy is split into the antibonding eg and t2g levels. Mn 3* (d*) in high spin
configuration is shown as an example of a cation with an antibonding eg-orbital
degeneracy with a single eg electron. At a surface cation, the degeneracy is removed by a
2
tetragonal site symmetry (C4 v) in which the occupied eg electron occupying a z orbital
directed toward a surface OH~ is lowered in energy relatively to the eg electron occupying
the x 2 _y2 orbital.
Scheme 5.1. Proposed OER mechanism on perovskite transition-metal oxide catalysts.
Both (top) the OER and (bottom) the ORR proceeds via 4 electron transfer steps. In the
OER case, the RDS are the 0-0 bond formation (2 in OER) and the proton extraction of
the oxy-hydroxide group (3 in OER).
For the ORR, the RDS are either the surface
hydroxide displacement (4 in OER) or the surface hydroxide regeneration (1 in OER).
Scheme 6.1. Graphic illustration of Strmcnik and co-workers' model for cation's noncovalent interaction, at which the cation species interacts with OHads group to interfere
the open active Pt site by "blocking" mechanism'o
m
Scheme 6.2: Proposed cation effect on the ORR/OER on transition metal oxide surfaces.
(Top) cation or (Bottom) its hydration interacts with the hydroxo/superoxo/peroxo
species and its energies during the turnover condition. The kinetics is affected by the
intermediate energy modifications. The B label represents active transition metal atom,
which we previously proposed as the active site for oxygen electrocatalysis on transition
metal oxide surfaces.
15
This page is intentionally left blank
16
Chapter 1 - Oxygen Electrocatalysis
1.1 Motivation
Driven by growing concerns about global warming and the depletion of petroleum
resources, developing renewable energy production and storage technologies represents
one of the major scientific challenges of the
1
2 1s' century '
2
. A critical element in pursuit
of this quest is the discovery of efficient and cost-effective catalysts for energy storage
and conversion processes. Prominent examples of reactions requiring catalysts are the
hydrogen evolution and the oxygen evolution reaction (OER) in solar-driven watersplitting concepts"' and in the temporary storage of renewable energy by electrolytic
hydrogen, 2, 5,
6
the hydrogen oxidation and the oxygen reduction reaction (ORR) in
hydrogen fuel cells7' 8, and the ORR and OER in metal-air batteries9'".
Over the past
decade, substantial efforts have been devoted to discovering catalytic surfaces with high
activities to facilitate these reactions. However, even with the state-of-the-art catalysts,
the electrocatalytic performance is insufficient - the slow reaction kinetics and the high
price of required precious metal catalyst impart significant inefficiency and cost to the
system. The goal of this thesis aims to assist as parts of the effort to help design novel,
highly active, low-cost materials that can effectively help reduce the cost and the
inefficiency of specifically the oxygen electrochemical reactions (ORR and OER) by
building a structure-property relation capable of identifying suitable activity descriptor
for a model series of catalytic materials - in this case, a class of transition metal oxides
known as the "perovskites".
1.2 Introduction to Oxygen Electrocatalysis
Oxygen electrochemical reactions are a class of electrochemical reactions ("redox") in
which a chemical transformation of oxygen is coupled to electron transfer.
The
ubiquitous nature of oxygen makes mastery of its electrochemistry an essential building
block for a variety of energy and environmental technologies. For example, in fuel cells,
where chemical fuels are converted to electricity, oxygen electrochemistry is an integral
17
component of the cathode, where oxygen is reduced to form either water (in protonexchanged membrane fuel cells, PEMFCs, 02 + 4H* + 4e- -> 2H 20) 7 "1 or hydroxide (in
alkaline fuel cells, AFCs, anion-exchanged membrane fuel cells, AMFCs, and metal-air
batteries, 02 + 2H 20 + 4e- -> 40H~)'.
Similarly, in storage devices, where electrical
energy is stored in form of chemical bonds, the OER is an enabling chemistry for the
anode in proton-exchanged membrane (PEM, 2H 20
->
02 +
4H* + 4e-), alkaline, anion-
exchanged membrane electrolyzer and metal-air batteries (40H- -> 02 + 2H 20 + 4e-).
Figure 1.1 and 1.2 demonstrate cartoon drawings of the oxygen electrode (where the
ORR and the OER occur) in both proton-exchanged and anion-exchanged membrane
devices.
Research efforts have been devoted to mastering and accelerating the kinetics of the
oxygen electrochemical reactions as, from the technology standpoint, the prohibitively
sluggish kinetics of the oxygen electrochemical reactions is one of the largest sources of
inefficiencies and cost in the electrochemical energy storage and conversion devices 7'14.
At present, platinum nanoparticles supported on carbon blacks ("Pt/C") are used as
electrocatalyst to catalyze the ORR for both PEMFCs 7 ' and AMFCs14 . The utilization
of platinum catalyst for the ORR is however far from ideal: more than 0.3 V of
overpotential is still required before any appreciable kinetics current can be obtained7 '",
and the majority fraction of the price of PEMFC is attributed to the platinum cost 15 . The
kinetics limitation of the oxygen electrochemical reaction also occurs in the case of the
OER. A PEM electrolyzer generally uses either iridium oxide or oxide electrocatalyst,
either option requires non-negligible overpotential (generally more than 0.3 V) and also
adds significant material cost to the system 6 '17
18
Figure 1.1.
Illustration of the oxygen electrode in PEMFC ("conversion") and PEM
electrolyzer ("storage").
The oxygen electrochemical reaction occurs at the ionomer-
catalyst interface. Image was adapted from the initial drawing of Dr. Eva Mutoro.
To address the need for a superior oxygen electrocatalyst, significant efforts have been
focused toward discovering new or more active form of ORR/OER catalytic surfaces.
However, over many decades, the development of new catalysts for ORR and OER has
been largely empirical. Although Sabatier's principle provides a qualitative argument for
tuning catalytic activity by varying the bond strength between the catalyst surface and the
reactant/product (neither too strong nor too weak, leading to maximum activity at
moderate bond strength, Figure 1.3), it has no predictive power to find catalysts with
enhanced activity. The ability to predict new catalyst has been only recently enabled by
fundamental ORR surface science research'" aided by ab-initio modeling 9 , a combination
of which have identified a unique descriptor for the surface electronic properties of
platinum-based catalysts, viz., the d-band center relative to the Fermi level. Specifically,
the d-band center has been shown to govern the metal-oxygen bond strength by
controlling the degree of antibonding population of the surface-oxygen interaction and
19
thus exhibit a volcano relationship with the ORR activity 2 0- 2 2 , reflecting the fact that the
maximum catalytic activity is attained at a moderate metal-oxygen bond strength (neither
too strong nor too weak).
Controlling this descriptor has led to the discovery of
promising new Pt-based catalysts,
21
02
Electrolyte
Figure 1.2. Illustration of the oxygen electrode in AMFC ("conversion") and anionexchanged membrane electrolyzer ("storage").
drawing of Dr. Eva Mutoro.
20
Image was adapted from the initial
Kinetics
Too strong
A
Too weak
Intermediate binding strength
Figure 1.3. Illustration of the Sabatier Principle - "Volcano plot", where varying the
bond strength between the catalyst surface and the reactant/product to moderate bond
strength (neither too strong nor too weak) can lead to maximum kinetics activity.
While advanced platinum-based catalysts might enable fuel cell commercialization,
abundant, non-precious-metal-containing catalytic surfaces such as transition metal
26
oxides" 24, graphitic carbons", and enzyme-mimicking supramolecular complex , 27
might offer an alternative, more economical option to energy storage and conversion
devices. However, a suitable descriptor for the oxygen electrocatalytic activity of these
non-precious-materials has not been identified, thereby hampering the development of
more active oxygen electrocatalysts. The objective of thesis seeks to address this open
question by developing design principles for non-precious-metal-containing catalytic
materials. Motivated by the success of the d-band center theory, this work will focus on
transferring the concept of the activity descriptor in the case of platinum metals to the
non-precious-metal-containing catalytic materials, specifically, transition metal oxides.
To accomplish the identification of the activity descriptor, transition metal oxides with
perovskite structure (Figure 1.4), which has a formula of A,-xA'xByB' 1yO3, where A or A'
is a rare-earth or alkaline-earth metal, and B or B' is a transition-metal, are chosen as
21
model series of catalytic compounds because of its flexibility in both physical-chemical
and catalytic properties. The ORR and the OER catalytic activities of the model oxoperovskite compounds are then measured in alkaline environment to take advantage of
the oxide stability in high pH environment, which allows wide array of perovskite
chemistries to be investigated systematically. In combination with the electrocatalytic
measurement, spectroscopic tools such as X-ray absorption spectroscopy will be
conducted to study the electronic structure of these oxide materials, from which, the
design principles will be developed from.
Oxygen
Transition metal (B-site)
Lanthanide (A-site)
Figure 1.4. Illustration of the perovskite structure, where the blue atom represents the
transition metal position (B-site), the red atom represents the oxygen, and the yellow
atom represents rare earth, alkaline earth, or mix therefore (A-site).
1.3 Scope of the thesis
This thesis will cover the methodology of the electrocatalytic measurement, the electronic
structure model used for establishing the activity descriptor, and the resulting design
principles of transition metal oxide catalyst materials.
Chapter 2 is devoted to the molecular orbital model for electronic structure of perovskite
transition metal oxide that will be used in this work as a guideline toward rationalizing
22
the activity descriptor. This chapter will also include a method to estimate the mixing
between the cationic and the anionic states ("covalency"), which is one of the key
features (vide infra) for understanding and developing the electronic structure descriptor
for the transition metal oxide catalysts.
Chapter 3 provides the basis for the experimental approach for quantifying the
electrocatalytic performance of the oxide materials. The technique, known as the thinfilm rotating disk electrode technique, takes advantage of the convective transport, which
can eliminate the contribution from the diffusion and allow electro-kinetics extraction,
will be applied to assess a variety of transition metal oxide catalysts. Comparison and
projection to the actual electrochemical device performance will be discussed.
Chapter 4 focuses on the electrocatalytic trend for the ORR on the perovskite transition
metal oxide model catalysts. Trend as derived using the rotating disk electrode technique
described in Chapter 3, will be evaluated using the electronic structure model described
in Chapter 2. The hypothesis on the activity descriptor and the ORR mechanism on
perovskite oxide surfaces will be presented.
Chapter 5 is devoted to the application of the activity descriptor found for the ORR to the
descriptor for the OER. We also present the hypothesis of the OER mechanism given the
information on the ORR side, and demonstrate the use of a design principle to help search
for a new compound for the OER, from which comparison to state-of-the-art oxide
materials will be given.
Chapter 6 presents the results to support the proposed mechanisms for ORR and OER by
examining the role of the cation in the alkaline electrolyte. Mechanism based on recent
finding of non-covalent interaction between surface oxygen and the cation species, and
its effect on the electro-kinetics will be present.
Chapter 7 provides an outlook for future work in the area of oxygen electrocatalysis with
particular attention paid to integration with the photoelectrochemistry setup, which may
allow development of future design principles for integrated catalyst-semiconductor
structure for solar-driven water-splitting.
23
Chapter 2 - Molecular Orbital Approach
to Perovskite Oxides
2.1 Simplified Molecular Orbital Model for Perovskite Oxide
The electronic structure complexity of perovskite transition metal oxides underpin many
phenomena such as superconductivity23, 29 , ferroelectricity30,31, magnetoresistance ,33, and
metal-insulator transition34 . In the scope of this thesis, we focus on the simplest treatment
of the transition metal oxide electronic using the linear combination of atomic orbital
approach from the molecular orbital theory. For oxo-perovskites, this wavefunction is
constructed from the interaction between the 3d (and 4s/4p) states of the B transition
metal and the 2p states of the oxygen in octahedral geometry, where crystal field removes
the degeneracy of the B 3d manifold. The general result of the B0 6 molecular orbital
model can be crudely summarized as the formation of the bonding (predominantly of 0
character) and the antibonding (predominantly of B character) level (see Figure 2.1).
Notably, the B 3d state is now split into eg and
t2g
states (the drawing of the eg and
t2g
states in B0 6 octahedra is shown in Figure 2.2.)
We note that this model is overly simplistic. We have not taken into account effects such
as bandwidth, spin-orbit coupling, electron-electron interaction, and etc. Nonetheless, the
model serves as a starting point for oxygen electrocatalysis model, which focuses mainly
on the interaction between B transition metal and adsorbed oxygen-like species. On the
zero-order approximation, the surface B0 5 with an adsorbed oxygen-like species
resembles B0 6 locally although, strictly speaking, the symmetry changes from Oh to C4 v.
In this work, we have assumed that the splitting is small due to the similarity in the
electronegativity between the 02- and the oxygen adsorbate species. In this assumption,
the eg and
t2g
states are assigned the antibonding, which is generally used as a determinant
of the B-O strength ("bond order"). This hypothesis has been discussed 35 ,36 and is similar
to the concept of the d-band center in precious metal surfaces19 37' 38 , where the position of
the d-band center is postulated to correlate to the degree of the antibonding occupancy of
24
the adsorbed state. This will be further discussed in Chapter 3. For the remaining of this
chapter, we will focus on the nature of the eg and t2g states, which generally is not the
same as the crystal field state due to hybridization with the 0 2p states ("covalency").
4p
3d
a (s - Pz)
ag (jt)
Figure 2.1. Molecular orbital diagram for B0 6 geometry. The drawing is adapted from a
well-known textbook by Ballhausen and Gray39 . The antibonding level is denoted with
an asterisk.
eg*
t2g.33
Figure 2.2. Drawing of the eg and t2g states. The oxygen axis shows the direction of the
B-O bond. Notably, the eg state points directly toward 0, whereas the t2g points away.
25
2.2 Transition metal - oxygen covalency in perovskite oxide
Covalency or hybridization has received substantial attention as an underpinning
parameter for various transition metal oxides behaviors including localized-to-itinerant
transition40 , intercalation potential 41'
42,
optical 43, magnetic 44' 45, and other unique physical
properties at oxide interfaces 46'47. This effect has been discussed in a range of transition
metal oxides including perovskite 4 0'
44'
45' 48,
spinel 49, pyrochlore5 0 , and layered
structures 42, and has been proposed to be key to many functionalities including
intercalation potential41 ,
42
and oxygen electrocatalytic activities5 1'
52.
Several
experimental methods have been applied to estimate the degree of covalency, ranging
from X-ray Photoemission Spectroscopy (XPS)'54 , Electron Energy Loss Spectroscopy
(EELS) 55' 56, X-ray Absorption Spectroscopy (XAS)57 , and neutron diffraction5 8.
However, to date, there has not been a study to systematically compare perovskite oxides
with different transition metals and cation substitution. The lack of such study is partially
due to the fact that there is currently no simple method for systematically comparing
covalency of different transition metal compositions. An approach that has been reported
in literature relies on an extraction of the p-d transfer integral of the cluster-type chargetransfer calculation of the photoemission spectra59 . In this chapter, we report a simpler
method as part of efforts to quantify the transition metal oxide covalency. We focus on a
structurally similar system of perovskites (formula: AA'BB'0 3 , where A and A' are rareearth and/or alkaline earth metals, B and B' are 3d transition metals, and 0 is oxygen) as
model compounds using the 0 K-edge XAS, which takes advantage of the first-order
transition from 0 Is to 0 2p state to selectively measure the unoccupied state with
dominant 0 2p symmetry ("0 2p hole") and consequently probe covalency5 7 . This
selective excitation is as a result of the dominating dipole transition matrix and can
therefore be indicative of the electron sharing between 0 2p and B 3d states. De Groot et
al., has previously examined 0 K-edge XAS in a series of model oxides in the context of
the localized eg and t2g states 57. Their work has indirectly revealed evidence for B 3d - 0
2p state mixing but has not lead to a conclusive finding about how the mixing evolves
with 3d electron. Abbate and co-workers also have systematically studied 0 K-edge
XAS on Lai-xSrxFe0 3 and LaiSrxMnO 3, where they found that the hole doping induced
26
by Sr substitution lies in a heavily mixed 0 2p and B 3d states 60 . Their study however
does not lead to a conclusive mean of quantifying covalency between different transition
metals. In this contribution, we demonstrate a method to extract the degree of 0 2p - B
3d covalency from 0 K-edge XAS in perovskite oxides using Golden Rule and
molecular-orbital assumptions.
We apply our method to study how the degree of
covalency evolves in in Lai.xCaxBO 3 (where B = 3d transition metals) perovskite, and the
Lai+xNiO 3+xRuddlesden-Popper oxides. From which, we explain the trend on the basis
of an electronegativity model.
2.3 Experimental Procedure
The perovskites samples used in this study were synthesized with a co-precipitation
method.
Briefly, rare- and/or alkaline-earth nitrate, and transition-metal nitrate (both
99.98% Alfa Aesar) in a 1:1 mole ratio were mixed in Milli-Q water (18 MQ-cm) at 0.2
M metal concentration. The solution containing rare and alkaline earth, and transion
metals was titrated to 1.2 M tetramethylammonium hydroxide (100% Alfa Aesar). The
precipitate was filtered, collected, and dried. The powder samples were subjected to heat
treatment at 1000*C under Ar atmosphere for the La1 -CaxMnO
3
(x = 0, 0.5) and
LaO5 CaO5 CrO 3 samples, at 1000*C under dried air atmosphere for LaCrO3 , La-xCaxFeO 3
(x = 0, 0.25, 0.5), La1 -CaCo0 3 (x
=
0, 0.5), La3 Ni 2 0 7 , La4 Ni 3 0
10
samples, and at 800*C
under 02 atmosphere for the LaNiO 3 sample. LaMnO 3, was prepared from an 800*C
heat treatment of LaMnO 3 in air. La 2NiO 4 was synthesized in 2 steps: 1000*C under Air
atmosphere and then heated at 800'C under Ar atmosphere. All gases had ultra-highgrade purity (Airgas). All samples were characterized by X-ray diffraction to determine
phase purity with a Rigaku High-Power Rotating Anode X-ray Powder Diffractometer at
a scan rate of 0.8 degree /min. All X-ray diffraction data and refined lattice parameters
had single-phase purity, except LaNiO 3, where <2% of NiO was present. 0 K-edge Xray Absorption spectra were collected at Saga-LS (BL-12), Japan. All measurements
were collected in electron-yield mode at 10-7 Pa pressure (corresponding to 10 nm
penetration depth61 ). All spectra were normalized to the oxygen absorption background
at ~550 eV.
27
2.4 Covalency Trend in Model Series of Perovskite Oxides
The results of the 0 K-edge spectra of LaBO 3 (where B = Cr, Mn, Fe, Co, Ni) are shown
in Figure 2.1. Our results have similar features to the spectra that have been reported in
literature, where the first peak (- 530 eV) corresponds to the excitation from the 0 Is
state to the hybridized 0 2p - B 3d state as a result of the covalency between the two
levels, and the second peak (- 535 eV) and third peak (- 543 eV) correspond to the
excitations of 0 1s to 0 2p - La 5d, and 0 2p - B 4sp respectively 60. The second and
third features remain largely unchanged throughout 3d transition metal substitution, in
contrast to the first feature, which undergoes significant transformation in both shape and
energy in going from Cr to Ni.
This observation is expected as the second feature
corresponds to the La 5d state, which is chemically identical to all the LaBO 3 samples,
and the third feature corresponds to the 4sp state, which generally has a wider bandwidth
and therefore is largely similar across different B transition metals. The first peak of the
O K-edge spectra, which contains information on the covalency, or the hybridization of
the 0 2p and B 3d system, however, changes drastically across different 3d transition
metal substitution. As Figure 2.1 shows, this feature varies significantly in both energy
and line shape. We hypothesize that this is a result of a dissimilar covalency and number
of hole across different LaBO 3 compounds.
To isolate the covalency contribution to the hybridized 0 2p and B 3d excitation in the 0
K-edge spectra from the contribution from the varying number of hole with different 3d
transition metal and consequently enable a quantitative comparison of the 0 2p - B 3d
excitation lines, we assume the simplest form of the molecular orbital construct, where
covalency arises from the mixing between 0 2p with transition metal ("B") 3d states.
Driven by the proximity of the electronegativities, the 0 2p and B 3d states overlap and
induce formation of hole in the nominally 0 2p level. This has a direct consequence of a
charge transfer ("redox") from oxygen to the transition metal. In the case of perovskite
oxides, B 3d states undergo field splitting to the eg and
t2g
symmetry. By assuming that
these are the two essential energy levels in the perovskite oxides, we can write a
hybridized 0 2p - B 3d state construct as the following:
28
'P=leg = Ceg
Tunoccupied
Here, C (=
(neo(1
-
tn2
t
2T
+ 1/n
e 2
I~egl
g
t7L
1 /2
T,|e
+ nto
eg tflt2g~l
tg2g
|T,| 2 ) + nto(1 - |T12))1/ 2 ) is a normalization coefficient, L
denotes the ligand hole, neO and nto are the nominal hole numbers in the eg and
t2g
symmetry, and T, and T,, represent the matrix element in orbital mixing for the eg and
t2g
symmetry respectively. From this simple molecular model, a systematic comparison
of IT, 12 and IT,|12 can describe the degree of covalency between different perovskite
oxides, from which the trend will be estimated from the 0 K-edge spectra. To enable
such comparison, we examine the 0 K-edge spectral feature, which is a culmination of
various many-body effects that generally requires multi-scattering-type calculations. In
this work, we assume that the spectra can be estimated based on the simple dipole
selection rule, which restricts the 0 Is transition to the 0 2p hole state (state with ligand
hole) only. Thus:
oc neo|Tc|
2
+ nto|7,1 2
In this construct, I corresponds to the intensity of the B 3d - 0 2p peak, which we use
linear background between local minima to correct for the background absorption.
Then, using T, (~ T,/2), this equation simplifies to 62 :
a120neo
|Tr'|2
__
__
__
+ nto/4
The covalency parameters (|Ts12) of the model LaBO 3 compounds as extracted from the
O K-edge data (Figure 2.1) are plotted against the number of d-electron in Figure 2.2.
Our analysis shows that, as the d-electron number increases, the covalency increases in a
monotonic fashion. This finding is in agreement with the previous trend reported derived
from p-d transfer integral extraction in the case of binary oxides and sulfides by Bocquet
and co-workers but not ternary oxides 59 . As the extraction of the p-d transfer integral is a
highly convoluted process, we caution that a sensitivity analysis would be essential to
understand this difference. Nonetheless, to explain our experimental trend, we use the
following argument: the 3d electron trend follows that of the electronegativity for LaBO 3
system, and thus the degree of hybridization increases with it 63 '64
In the context of
29
molecular orbital theory, the transition metal moves to the right of the periodic table as
the 3d electron increases, and, as a result, the electronegativity of 3d transition metal
becomes closer to that of oxygen, which results in an overlap in their energy levels that
consequently lead to stronger hybridization.
The agreement between our results and
intuitive concept such as electronegativity gives us confidence in further applying this
model to examine a more complicated scenario such as in the case of the cation
substitution.
5
4
3
*O)t
2
1
520
530
540
550
560
Energy (eV)
Figure 2.1. 0 K-edge X-ray Absorption Spectroscopy of LaBO 3 (B = Cr, Mn, Fe, Co,
Ni). The shaded region centered at -530 eV represents the feature assigned to B 3d - 0
2p hybridization. Spectra were collected in electron yield mode at room temperature.
30
W
I
W
I
W
I
V
1
9
1 '09
0. 64
LaNIO
7
A
1! 0.4
SOaMnO
5 0.2
LaCoO 3
LaFeO 3
3
LaCrO 3
2
3
4
5
-
-
-
-
-
0.0
6
7
8
d-electron
Figure 2.2. Evolution of LaBO 3 covalency paremeter (Abs./(holeeg + 1/4 holeag) for B
=
Cr, Mn, Fe, Co, Ni with respect to d-electron count. The absorbance was taken to the
shade region shown in Figure 2.1 (corresponding to the unoccupied state with the 0 2p
symmetry,) corrected with linear background. The dashed gray line serves only as guide
to the trend.
In this section, we apply this covalency model to understand the role of B-site
substitution on covalency in LaNio. 5Mno.5 0 3 and LaCuo.5Mno.5 0 3 compounds. The 0 Kedge spectra are shown in Figure 2.3 (Top), where LaMnO 3 is shown in a light gray
dashed line as reference. By assuming that mixing of Ni-Mn and Cu-Mn results in an
averaged d-electron transition metal, the covalency trend is in good agreement with that
obtained from the LaBO 3 case as shown in Figure 2.3 (Bottom). This finding is believed
to be as a result of the fact that the electronegativity, and consequently the covalency of
LaNio.5 Mno.5 0 3 and LaCuo.5Mno. 50 3 compounds can be approximated by an average of
the two transition metal cations.
We caution however that these findings may be
coincidental and further experiments with regarding to cation ordering (consequently
valency, vide infra) would be essential to understand the impact of B-site substitution on
covalency.
31
1
2
LaNi0. Mn 0.O
0
560
550
540
530
520
Energy (eV)
0.6
-
0.4 -:4
LaCu
0
+
Mn
05 0.5
LaNi 5Mn
0 ,O303
0 0.2
.0
<0.0
2
3
4
5
-
-
-
-
-
-
6
7
8
d-electron
Figure 2.3. 0 K-edge X-ray Absorption Spectroscopy LaBB'0
Ni.5Mno.5).
(Top) Raw spectra.
reference. (Bottom) LaBB'0
3
3
(BB' = Cuo. 5 Mno. 5 and
The dashed grey lines show a LaMnO 3 0 K-edge
covalency parameter v.s. d-electron count, with LaBO 3
data as references. The dashed gray line serves only as guide to the trend.
We then examine the role of Ca-substitution for La in the series of La1.xCaxBO3 (where B
=
Cr, Mn, Fe, Co). The raw 0 K-edge spectra are shown in Figure 2.4A-C, where the un-
substituted LaB0
3
spectra of the same B transition metal are also displayed in dashed
grey lines as reference.
32
With Ca-substitution, all three features of the spectra change
significantly, in agreement with the previous literature report with the similar case of
divalent Sr-substitution 60 ' 65' 66 . This is as a result of the fact that Ca-substitution modifies
the electronic structure of excitations to both the A-site and the B-site. As our interest is
in the B 3d - 0 2p hybridization, we again focus on the first feature which corresponds to
the excitation of the hybridized state. Figure 2.5 demonstrates how the Ca-substitution
influences the covalency parameter. Here, by normalizing the covalency parameter of the
compounds with aliovalent substitution to the parent LaBO 3 compounds, we found that
the oxidation of B transition metal from the initial, un-substituted 3+ state to oxidized
(>3+) state, generally results in an increase in covalency. This finding is consistent with
the fact that oxidation of B metal leads to a reduced electron-electron interaction, which
consequently lowers the 3d electron energy level, bringing the 3d level closer to the 0 2p
energy state
67
7.
We note that interestingly the covalency parameters of LaO. 5CaO. 5 Co0 3.-
and LaCoO 3 are quite similar, which is likely a result of the oxygen-vacancy-mediated
charge neutralization, leaving Co valence unchanged and thereby does not affect the Co0 covalency.
To further explore the opposite of this oxidation effect, where the B metal is reduced, we
use Lai+xNiO3+x as model compounds. The 0 K-edge spectra for La 4Ni 30 1 o and La 2NiO 4
are shown in Figure 2.4D, where LaNiO 3 0 K-edge spectra is also shown in grey line as
baseline for comparison. We then conduct the covalency analysis for the RuddlesdenPopper phases, which serve as model compounds for reduced B transition metal in Figure
2.5. In this system, the reduced Ni-containing Ruddlesden-Popper phases are found to
have lowered covalency parameter.
In the context of electron-electron interaction
mentioned in the case of La1.xCaxB03, this finding is in an excellent agreement. We
therefore propose that in the limit of the oxo-perovskites, the factor that dominates the
covalency of the compound is the choice of the B transition metal and its oxidation state.
33
3
3
2 -
2-
La0 5Ca 0 Coa- -
0
530
520
3
C
-.
,
550
540
Energy (eV)
.
,.3.
560 520
530
550
540
Energy (eV)
.
.
560
..
La Ca FeO
La Ni31
-
2 -2
0
520
LaMnO3 +8
0
La 0.5Ca0.2,FeO 3530
540
550
Energy (eV)
0
560 520
530
540
Energy (eV)
La2NiO4
560
550
Figure 2.4. 0 K-edge X-ray Absorption Spectroscopy LaxCai.xBO 3 and Lai+.Ni0 3+x. 0
Lao.5 Cao.5 CrO3 and Lao. 5 Cao. 5 Co0 3 .3. LaCrO 3 and LaCoO 3 are in shown dashed grey
lines as references. (B) Lao.5Cao. 5MnO 3 and LaMnO 3+8. LaMnO 3 are shown in dashed
grey lines. (C) Lao. 5Cao. 5FeO 3 and Lao.7 5 Cao.25 FeO 3 . LaFeO 3 are shown in dashed grey
lines. (D) La 4Ni 30 1 o and La 2NiO 4 . LaNiO 3 are shown in dashed grey lines.
34
2.0
,
.
0"
...
,
.
La0 Ca MnO 3 0
La Ca CrO
1.5
0.5
? %#OO0.5
0 1.0
0.5
.
La Ca 0.5 CoO 3-68
0
La Ca FeO 3
*
I1-x
LaMnO 3 +6
4
La Nis0
> 0.5
T0 >
o 0.0
O
4~ La2NiO
44
2
'
2.0
2.5
3.0
3.5
B oxidation state
Figure 2.5. Evolution of covalency with respect to different B oxidation using Lai.
xCaxBO 3
and Lai+xNiO 3+xas model compounds. The covalency enhancement factor is
calculated from a ratio of a covalency parameter of either Lai..xCaxBO 3 or Lai+xNiO 3+x
with respect to LaBO 3 material.
2.4 Chapter 2 Conclusion
This chapter describes a simplified picture of the transition metal oxide electronic
structure using linear combination of atomic orbitals.
This approach focuses on the
interaction between transition metal 3d and oxygen 2p states. We place emphasis on the
antibonding levels of this interaction, which has two symmetry states: triply degenerated
t2g
and doubly degenerated eg. To better understand the nature of the antibonding states,
we deconstruct them to the original transition metal 3d and oxygen 2p states and extract
the extent of hybridization between the twos from the 0 K-edge XAS Spectra. This
procedure relies on the molecular orbital and Golden Rule assumptions, which, albeit is
arguably too simplistic, results in a covalency trend that is sensible from both chemical
and physical intuition viewpoints. We found that the covalency increases with d-electron
number for LaBO 3 (B = Cr, Mn, Fe, Co, Ni) compounds, which we attribute to the
increasing electronegativity as d-electron number increases. The d-electron vs. covalency
35
trend persists even in the case of B-site substitution, which we demonstrate for
LaCuo. 5Mno.5 0 3 and LaNio. 5Mno.5 0
3
cases. We further report that covalency is highly
sensitive to the oxidation state of B transition metal, where from combined studies on Casubstitution and Ruddelsden-Popper phases, we found that increasing oxidation state can
further increase covalency.
We assign this observation to the sensitivity of the
electronegativity on the electron-electron interaction.
36
Chapter 3 - Thin-film Rotating Disk
Electrode
3.1 Thin-film Rotating Disk Electrode Methodology
One of the largest cost and efficiency limitations of electrochemical energy conversion
devices which have an air electrode, such as PEMFCs, AFCs, AMFCs, electrolyzers, and
metal-air batteries, lies in the slow kinetics of the ORR at the cathode (air electrode),
which must be facilitated by an electrocatalyst' 68,
69.
For example, viable large-scale
commercialization of PEMFCs requires either to substitute the current carbon-supported
platinum ORR catalyst (Pt/C) with non-precious metal catalysts or to improve the
intrinsic activity of platinum8 . Therefore, a large number of novel catalysts have been
developed by industrial and academic laboratories, and are typically screened for their
ORR activity using the thin-film rotating disk electrode (RDE) method 70 -72 , where a
catalyst suspension is drop-cast as a thin-film onto a glassy carbon rotating disk
electrode, using small amounts of Nafion* as binder70 . The measured ORR currents are
then corrected for 02 mass transport resistances using the Levich equation to obtain the
ORR activity of the catalyst. Compared to ORR activity measurements in fuel cells,
which require time-consuming optimization of membrane electrode assemblies (MEAs)
for each catalyst, the thin-film RDE method is a fast screening tool to obtain ORR
activities which are consistent with those obtained in MEAs 73 , and it has led to the
discovery of highly active novel ORR catalysts'is
2 1 ,74
While noble metal catalysts with significantly
enhanced activity
compared to
conventionally used Pt/C may enable commercialization of PEMFCs, AMFCs and
metal/air batteries offer an alternative route toward low cost, noble metal-free fuel cells.
Here, oxide-based ORR catalysts are possible substitutes for platinum. In the case of
oxide-based ORR catalysts for either alkaline fuel cells or metal/air batteries, the ORR
activity is most frequently tested in PTFE-bonded gas diffusion electrodes, where
elimination of 02 mass transport resistances is difficult 69'75-81.
Therefore, despite the fact
that many different oxide materials have been examined in these studies, it is not possible
37
to extract absolute ORR activities referenced to either oxide surface area (i.e., specific
activity) or mass (i.e., mass activity). A more systematic evaluation of the ORR activities
of a wide range of perovskites was conducted by Bockris and Otagawa 35 ,
82,
using
immersed porous pellets, whose very high internal surface area however introduces
significant 02 mass transport resistances within the pellets so that meaningful ORR
kinetics can only be obtained at very low current densities (on the order of
s0.1 tA cm
oxide).
The same restrictions apply to the RDE method using a recessed
porous pellet as electrode (recessed RDE)8 3'84, since the thickness of these porous disk
electrodes (on the order of 100 [m) is much larger than the RDE diffusion boundary
layer thickness (on the order of 10 tm8 5). One method to avoid undefined 02 mass
transport resistances when measuring the ORR activity of oxide catalysts is the adoption
of the thin-film RDE method, where a very thin film (on the order of 1 tm) of oxide
particles is attached to a glassy carbon disk electrode using an ionomeric binder,
analogous to the method used in acid electrolytes described above70 . This method was
adopted recently in studying the ORR activity of LalxSrxMn0
3
perovskites86 , 87 , but the
use of highly acidic Nafion* ionomer poses the risk that the oxide catalyst might be
partially dissolved in the acid environment during film casting.
This can be avoided
either by using an alkaline ionomeric binder (i.e, anion-exchange ionomers8 8 ), or by the
neturalization of the protons in Nafion*9 solution with alkali hydroxides, producing alkali
metal exchanged Nafion* binder.
In the following, we will introduce the latter approach, viz., the thin-film RDE
characterization of the ORR activity of several perovskite catalysts using sodium ion
exchanged Nafion* solutions as binder. We demonstrate that absolute specific and mass
activities can be obtained, which enables a more rigorous and relatively rapid activity
comparison of well-characterized model oxide ORR catalysts.
Furthermore, we will
compare the ORR activity of these model oxides with that of conventional Pt/C and
provide an estimate of the kinetically limited ORR activity in an optimized alkaline fuel
cell cathode.
38
3.2 Experimental Procedure
Synthesis and characterization. The perovskite transition metal oxides studied in this
chapter (Lao. 75Ca 0.25FeO 3 , LaCuo.5 Mno.5 O3 , and LaNiO 3) represent the full ORR activity
range obtained for various oxides measured in our laboratory, which were synthesized
using a co-precipitation method described elsewhere8 9 . Briefly, rare earth and alkali earth
nitrates and transition metal nitrates (all 99.98% purity, Alfa Aesar) at the respective
stoichiometric ratio were mixed in Milli-Q water (18 Mn cm) at metal concentrations on
the order of 0.1 M. This solution was then titrated with 1.2 M tetramethylammonium
hydroxide (Alfa Aesar) until a solid precipitate was formed. The precipitate was filtered,
collected, and dried at 200'C for 12 hours, and then subjected to heat-treatment at
1000 0 C in dry air (ultra-high purity grade, Airgas) for 12 hours for Lao. 7 5 Cao. 2 5 FeO 3 and
LaCuo.5Mno. 5 0 3. To ensure the oxygen stoichiometry of LaNiO 3 , this sample was
synthesized at 800'C in pure 02 (ultra-high purity grade, Airgas) for 8 hours and
thermogravimetric analysis (TGA, Perkin Elmer) was used to confirm that its oxygen
content was approximately 98% of the nominal value.
X-ray Diffraction (Rigaku)
revealed these materials to be single-phase, as shown in Figure 3.1. Particle sizes and
surface areas were quantified via Philips XL30 FEG Environmental SEM (FEI-Philips),
determining mass specific surface areas, As, by approximating spherical geometry:
As~ (14xr2)/(Z(4/3)prr 3 ) =6/p_[(Eyd 2 )/(7d3 )] = 6/(d/a-p)
where p is the oxide bulk density, d is the diameter of individual particles determined by
SEM, and dv/a is the volume/area averaged diameter 90 . While the value of As for nonsupported oxide catalysts can also be determined from nitrogen BET area (Brunauer,
Emmett, and Teller) measurements, we did not have sufficient material for most of our
samples to conduct meaningful BET analysis. Sufficient material for BET analysis was
only available for LaNiO 3, and the measured BET area of -5.6 m 2 g1 (Micromeritics
ASAP2020)
is reasonably consistent with the SEM-based value of ~3.5 m 2 g
Therefore, we believe that a comparison of the specific ORR activity of the various
oxides using the SEM-based As-values listed in table 3.1 is reasonably accurate.
39
space group
dnumber (pm)
dy/a (pm)
As (m2 -1
LaNiO3
R -3 c
0.20 (±0.06)
0.24
3.5
LaCuo. 5 Mno. 5 0 3
P n ma
0.58 (±0.28)
0.80
1.1
Lao. 7 5Cao. 2 5FeO 3
R -3 c
0.36 (±0.22)
0.59
1.8
Table 3.1.
SEM characterization of the oxides studied in this chapter.
The number
averaged diameter, dnumber, with standard deviation (in parentheses), the volume-area
averaged diameter, dy/a, and the specific surface area of the oxide (see equation 1), As.
were obtained from particle size distribution measurements.
Ion-exchanged Nafion* solution preparation. Because perovskite samples in this work
have on the order of hundreds of nanometer in size, oxide coatings on a glassy carbon
disk for thin-film RDE measurements with these relatively large and heavy catalyst
particles require an immobilizing binder that at the same time facilitates the transport of
dissolved 02 to the catalyst surface. To avoid possible corrosion of the oxide due to the
strong acidity of commercially available Nafion* solution (DE520, Ion Power, USA), we
ion-exchanged the solution with sodium hydroxide which, similar to an earlier report on
lithiated Nafion 91 , was made by adding a defined aliquot of 0.1 M NaOH to the Nafion*
solution: for an equivalent weight of 1050 gpolymer/molH+ of the 5%wt. Nafion* solution,
corresponding to a ~0.05 M proton concentration, replacement of H+ with Na* requires a
-2/1 volume ratio of Nafion* solution and 0.1 M NaOH.
The thus obtained Na*-
exchanged Nafion* solution had a pH of -11, indicating a slight molar excess of NaOH
(ca. 1 mM). This solution was then used to prepare oxide coatings for thin-film RDE
measurements.
40
LaNiO 3
C3
La
_
20
30
40
50
Ca0.25FeO3
60
70
80
20 (degree)
Figure 3.1. X-ray diffraction spectra of the oxides studied in this chapter.
Oxide electrode preparation. While all of the here examined samples had reasonably
high electronic conductivities (-0. 1 S cm-1
92,
~0.5 S cm-1 93, and -1500 S cm-1 94 for
Lao.7 5Cao. 25FeO 3, LaCuo.5Mno. 50 3, and LaNiO 3 , respectively), the perovskite oxides were
mixed with acetylene black (AB) carbon at a 5/1 mass ratio of oxide/carbon in order to
eliminate any concerns regarding electronic conductivity limitations within the thin
electrodes. The AB (Chevron) was treated in nitric acid overnight at 80*C, then filtered
and finally dried at 100*C overnight).
Catalyst inks were prepared by horn sonication
(Bransonic 3510, Branson) of perovskite oxides, AB, and an appropriate amount of the
Na*-exchanged Nafion* solution with tetrahydrofuran (THF, 99.9+% Sigma-Aldrich),
yielding
inks with
final
concentrations
of 5 mgoxide mlink1 ,
1 mgAB mlink' , and
1 mgNafion mlink 1. Next, 10 g1 of catalyst ink were drop-cast onto a glassy carbon (GC)
electrode (0.196 cm2 area, Pine, USA) polished to a mirror finish with 0.05 Im alumina
slurry (Buehler). The catalyst layer on the GC substrate was dried overnight in a sealed
glass jar which had been pre-saturated with THF vapor, enabling slow ink drying rates
which were found to be required for homogenous thin-film coatings on the GC disk
electrode. The electrode had a final composition of 250 pgoxide cm-2 disk, 50 ggAB cm
disk,
and 50 9gNafion cm 2disk at an estimated film thickness of -2 ptm. Based on the density of
water-immersed recast Nafion* of ~1.4 g cm~3 95, the Nafion film thickness based on the
41
geometric area of the disk would correspond to -0.4 rim, sufficiently thin so that the 02
transport resistance within the ionomer phase is negligible 6 .
Assuming uniform
Nafion* coverage on the oxide and carbon particles in the electrodes, the actual Nafion*
film thickness is more than one order of magnitude lower, due to the high electrode
roughness
factors
of the oxide
250 pgoxide cm-2 disk and ~1 m2
g-oxide)
and the
carbon
of
~2.5 cm 2oxide cm-2disk
and ~25 cm 2oxide cm-2 disk (at 50 pgAB cm
disk
(at
and
50 m2 -1AB), respectively 31.
Platinum electrode preparation. Glassy carbon disk electrodes coated with a thin film of
Pt/C catalyst (46 %wt. Pt on high-surface area carbon, TKK, Japan) were prepared as
described elsewhere 70 .
Briefly, 15 ptl of an aqueous Pt/C ink of a composition of
0.15 mgcatalyst ml-1 and 0.04 mgNafion ml~1 was drop-cast onto a GC disk elecktrode and
allowed to dry overnight, yielding a Pt loading of 5 pgpt cm-2disk. and 3 pgNafion cm-2disk.
The Pt/C surface area was estimated from the electrochemical H2 underpotential
deposition as described elsewhere9 6' 97 and yielded a Pt specific surface area of
48 m 2 gp -I
Electrochemicalcharacterization. All electrochemical measurements were conducted in
a three-electrode glass cell (Pine Instrument) and using a rotator (Pine) to which the thinfilm RDE working electrodes were attached; the potential was controlled using a
VoltaLab PST050 potentiostat. The 0.1 M KOH electrolyte was prepared from Milli-Q
water (18 MA cm) and KOH pellets (99.99% purity, Sigma-Aldrich). All measurements
were conducted at 10 mV s-1 in either N 2 or 02 (ultra-high purity grade, Airgas) at room
temperature. A saturated calomel electrode (SCE) reference electrode (Pine Instrument)
was used and was
calibrated in the same electrolyte by measuring hydrogen
oxidation/evolution currents on a Pt-RDE and defining the potential of zero current as the
reversible hydrogen electrode potential (RHE); 0 V vs. RHE in this case corresponds to
0.998±0.005 V vs. RHE. All the potentials in this study were referenced to the RHE
potential scale and correspond to the applied potentials, Eapplied, unless they are stated to
be iR-corrected potentials, EiR-corrected, calculated by the following equation:
EiR-corrected = Eapplied -
iR
where i is the current and R is the uncompensated ohmic electrolyte resistance (~45 Q)
42
measured via high-frequency AC impedance in 0 2 -saturated 0.1 M KOH. ORR activities
reported in the work were obtained from the negative-going scans in pure 02 at 1600 rpm
and were corrected for capacitive currents in pure N2 . Error bars represent standard
deviations from at least three independent repeat measurements, whereby the change in
activity over subsequent potential cycles is less than 3%. To achieve this steady-state
response, the potential range was kept within 0.7 to 1 V vs. RHE in order to prevent
degradation of the oxide catalysts by oxidation/reduction 76, with the exception of
LaCuo.5Mno.5 0 3, which displayed remarkable stability down to 0.4 V vs. RHE. The mass
transport correction was performed using the well-known RDE correction (Levich
equation) 97imeasured
ik
+
D1
where imeasured is the measured 02 reduction current density, ik is the mass-transport
corrected kinetic ORR current density, and iD represents the limiting current density
which was obtained experimentally for LaNiO 3 and LaCuo.5 Mno. 50 3 electrodes (vide
infra).
To avoid impurities from the corrosion of the glass cell 98 ' 99, all experiments were
collected within 2 hours of the initial exposure of KOH to the glass cell. The absence of
contamination effects from the glass cell under these conditions is demonstrated by the
fact that the ORR mass activity at 0.9 V vs. RHE of Pt/C obtained in our measurements
(0.15 A mgpt-1 , see Figure 3.7) is essentially identical to that reported for a Teflon cell
(0. 12 A mget- Y0"0.
3.3 Results from Thin-film Rotating Disk Electrode
LaNiO 3 has been known as highly active ORR catalyst for some time24,
76
and was
recently applied as cathode catalyst in a sodium-borohydride/air fuel cell' 01 . Figure 3.2
shows the cyclic voltammogram of a thin-film LaNiO 3 electrode in N 2 saturated 0.1 M
KOH at 10 mV s-1 , indicating the absence of any noticeable oxidation/reduction feature in
the range of 0.7 to 1.0 V RHE, consistent with previous reports 76 . It should be noted that
most ORR activity studies in the past use a Hg/HgO reference scale, whereby 0 V RHE
corresponds to approximately -0.92 V Hg/Hg0
8
5.
In 02 saturated 0.1 M KOH, the onset
43
of ORR currents is observed below ~0.95 V vs. RHE (see Figure 3.2), confirming the
high ORR activity of LaNiO 3 . Figure 3.3 compares the ORR activity of the three
different perovskites examined in this work, showing the capacity-corrected net ORR
current densities of the negative-going scan at a rotation rate of 1600 rpm. The ORR
activity of the glassy carbon substrate and of the acetylene black (AB) added to the oxide
electrodes is also shown in Figure 3.3, indicating that background contributions from the
AB and the GC disk are negligible above ~0.7 V vs. RHE. Therefore, even for the oxide
with the lowest measured ORR activity (Lao.7 5 Cao.2 5 FeO 3 ), the background contribution
from the carbon materials is negligible and the measured ORR activity is that of the oxide
itself.
The more than 0.22 V lower ORR activity of Lao. 7 5 Ca 0 .2 5 FeO 3 compared to
LaNiO 3 (Figure 3.3) is somewhat surprising, since similar exchange current densities
were reported previously for LaNiO 3, Lao. 7 Sro. 3FeO 3 , and Lao. 5Sro. 5 FeO 3 82 and since one
would not expect that the exchange of A-site atoms of different sizes (i.e., Sr for Ca)
would produce such a large shift in ORR activity (e.g., the ORR activity of AMnO 3
compounds is within 50 mV for A=La, Pr, Nd, Sm, Gd, Dy, or Y78). This discrepancy
may be related to the strong 02 mass transport limitations in thick high-porosity pellet
electrodes used previously8 2, rendering the quantification of the ORR activity ambiguous,
if not impossible, and highlighting the need for a simple ORR activity measurement
method with well-controlled 02 mass transport properties.
44
400
200
V
Cam
E
0
-200
1OM
%V
%Vu
M
-400~
0.7
0.9
0.8
1.0
Applied potential (V) vs. RHE
Cyclic voltammograms of LaNiO 3 at 10 mV s-1 in 0.1 M KOH at room
temperature in N2-saturated electrolyte at 0 rpm (blue) or in 0 2 -saturated electrolyte at
Figure 3.2.
1600 rpm. Sweep directions are shown by the arrow.
0
E
-
-500
t: -1000
0.4
0.6
0.8
1.0
Applied potential (V) vs. RHE
Figure 3.3.
ORR current densities (capacity-corrected negative-going scans) of GC-
supported and Nafion*-bonded (Nat ion exchanged) thin-film LaNiO 3, LaCuo.5Mno.O3,
Lao.75Cao.25FeO3 electrodes at 1600 rpm in 0 2-saturated 0.1 M KOH at 10 mV s-1. The
45
background ORR activity deriving from either the GC substrate or a thin-film Nafion*bonded acetylene black (AB) thin-film electrode is shown for reference.
Figure 3.4 compares the ORR activity of LaCuo. 5Mno. 50 3 with that of Pt/C at 1600 rpm
and 10 mV s-, showing that the same diffusion limited current density of ~5.7 mA cm-2
was obtained for both catalysts. This value is consistent with the 4-electron limiting
current for 02 reduction to water which has been reported previously for platinum singleand poly-crystalline electrodes in 0.1 M KOH at room temperature 98 , 102,
103.
Therefore,
one can conclude that 02 is reduced to water in an overall 4-electron process on the thinfilm LaCuo.5Mno. 50 3 electrode, which was also confirmed by the Koutecky-Levich
analysis shown in Figure 3.5, where the slope of the lines is related to the number of
electrons, n, exchanged in the overall 02 reduction reaction:
slope = (0. 62nFDo2 13 V-' 6 Coj'
where F is the Faraday's constant, Do is the diffusivity of the 02 molecule in the
electrolyte, o> is the rotation speed, v is the kinematic viscosity, and Co is the
concentration of 02 in the electrolyte. The fact that the same slope is obtained for both
LaCuo. 5Mno. 50 3 and Pt/C confirms the complete 4-electron reduction of 02 to water.
Figure 3.5 also shows the Koutecky-Levich analysis for LaNiO 3 and the essentially
identical slope demonstrates the 4-electron reduction to water on LaNiO 3, in agreement
with previously reported rotating ring disk electrode (RRDE) measurements on LaNiO 383 .
Unfortunately, owing to its low ORR activity (~0.2 mA cmd
isk
at 0.7 V, see Figure 3.3),
no Koutechky-Levich analysis could be conducted with the Lao. 7 5 Cao.2 5 FeO 3 sample.
Even though in general the final product of the reduction (i.e. number of electron
transferred) must be known to quantify the kinetic ORR current (see equation 3), the
transport correction term is negligible (10%) if the measured current is less than 10% of
the minimum limiting ORR current (i.e., that for assuming a 2-electron reduction), which
is the case for the Lao. 7 5 Cao.2 5FeO 3 electrode, so that sufficiently precise kinetic ORR
current densities can be obtained even in this case where the final reaction product (H2 0 2
or H2 0) is unknown.
46
I
0
0 2-sat.
I
I
0.6
0.8
0.1 M KOH
Rotating: 1600rpm
V -2
E
E.
Rate: 10 mV s
LaCu,.,Mn,0.O0
4
CP,
-6 0.0
0.2
0.4
1.0
Applied potential (V) vs. RHE
Figure 3.4.
Polarization curves of GC supported thin-film LaCuo.5Mno. 50
3
and Pt/C
electrodes using a Na+ ion-exchanged Nafion* binder: 1600 rpm in 02-saturated 0.1 M
KOH at 10 mV s-.
A cautionary note on the interpretation of the above described Koutechky-Levich
analysis or of RRDE measurements is suggested here as it was shown that the apparent
number of electron transferred in the ORR deduced from (R)RDE measurements can
depend on electrode thickness
72 ,104,105
. For example, Fe-N-C based catalysts were shown
to quantitatively reduce 02 to H2 0 2 (2-electron process), but owing to the facile nonelectrochemical H2 0 2 decomposition (H 20 2 -+ H 20 + 0.502), RRDE measurements on
thick electrodes (>5 [Lm based on the known packing density of high-surface area
carbons' 06) indicated an apparent mostly 4-electron reduction process on these
catalysts72 . Since our electrode thickness is only -2 lim, we would not expect this artifact
and believe that LaCuo. 5Mno. 50
water.
3
and LaNiO 3 indeed reduce 02 in a 4-electron process to
Nevertheless, to unambiguously identify the ORR product (H 20 vs. H 20 2),
detailed RRDE measurements of electrodes with varying thickness would be necessary.
47
I
I
0.0006
v 0.0004
E
0.0002
3
LaNiO (0.8 V)
o LaCu,,Mn, 05
A PtC (0.8 V)
0.0000
0.00
0.05
1
0.10
3
(0.65 V)
1
0.15
-1/2 (s112
Figure 3.5.
Koutecky-Levich analyses for GC-supported thin-film LaCuO. 5Mno. 5 0 3 ,
LaNiO 3, and Pt/C electrodes (02 saturated 0.1 M KOH at 10 mV s-').
To extract the actual ORR kinetics of these oxides and of Pt/C, ohmic drop and mass
transport corrections were applied to the ORR measurement using the Levich equation
for the convective transport in an RDE configuration (see experimental section). The
resulting kinetic ORR activities were then normalized by the measured specific surface
areas of the oxides (Table 3.1) or of Pt/C (see experimental section) to obtain the specific
ORR activity, is, plotted in Figure 3.6. In terms of specific activity, LaNiO 3 is the most
active of the three studied oxide materials, with a value of -40 iA cm-2oxide compared to
-320 pA cm- 2pt for Pt/C at the typical benchmark condition of 0.9V vs. RHE (Figure 3.6).
Despite the fact that there are many previous articles on the use of LaNiO 3 as ORR
catalyst, specific activities can only be extracted from the study of Bockris and Otagawa8 2
with a value of -0. 1 pA cm 2 oxide (~100 pA cm- 2 eiectrode at a reported roughness factor
-1000 cm2oxide cm-2electrode). The difference between this previously reported low specific
activity for LaNiO 3 and our measurement can be attributed to undefined 02 mass
transport overpotentials in the thick and porous pellet electrodes used 2 , whereby this
artifact could be avoided in our thin-film RDE measurements. The Tafel slopes of the
three different oxides and of Pt/C (Figure 3.6) are all within -60±10 mV decade-'. While
this is consistent with the Tafel slope reported for poly-crystalline Pt at the same
48
conditions 99, reported Tafel slopes for LaNiO 3 scatter over a wide range (45 mV decade~'
60 mV decade 1 82, 120 mV decade- 76), which again is most likely related to the poorly
controlled 02 mass transport in these measurments based on thick porous pellets or gas
diffusion electrodes.
24,
1.0
LU
0.9
U5
1 v
LN1
S LaNiO 3
gilI
v
gi
Pt/C
i
uJ 0.8 -
LaCu Mn
0
0
Mn,,Os
00.7
La,, Ca. 2 FeO
10
100
i, (pA cm-2
1000
or pA cm 2 )
Figure 3.6. ORR specific activities, is, of of LaNiO 3, LaCuo.5 Mno 50 3, Lao. 7 5 Cao.2 5FeO3
(using specific surface areas listed in table 3.1) and of Pt/C versus potential, obtained
after iR and 02 mass-transport correction (from capacity-corrected negative-going scans
at 10 mV s~1 in 0.1 M KOH for oxides, from positive-going scans at 10 mV s-' in 0.1M
KOH for Pt/C). Error bars represent standard deviations from at least three independent
repeat measurements.
Figure 3.7 shows the ORR actvity mass activity, i., of the various perovskites and Pt/C.
At the typical benchmark voltage of 0.9 V vs. RHE, the measured mass activity of Pt/C
(0.15 A mgp- 1) is in good agreement with the literature (0.12 A mgp~1 100). Unfortunately,
most studies on the ORR activity of LaNiO 3 provide no information on catalyst mass in
the electrodes 2 4 , 8 2 , 8 3 or lack detailed experimental conditions' 0 1. Only two studies allow
the evaluation of defined LaNiO 3 mass activities (pure 02, room temperature, 0.9 V vs.
RHE corresponding to -0.02 V vs. Hg/HgO), with values of -2 A goxide~1 (6 M KOH) and
-0.7 A goxide~1 (5 M KOH)1 07, which are reasonably consistent with the value of
49
The mass activities of
(0.1 M KOH) obtained from Figure 3.7.
-1.6 A goxide~
LaCuo. 5Mno. 5 0
3
and Lao.75 Cao.2 5FeO 3 are significantly lower than LaNiO 3 and no
comparison data are available in the literature.
1.0
I
Ui
. LaNiO 3
PtIC
0.9 -
uW 0.8 -
LaCu
0
5
Mn 0O 3
0.7
Lao0.76
2 FeO 3
7 Ca0.26FeO
0.1
1
10
100
1000
im(A g-)
Figure 3.7.
ORR mass activities, im, of of LaNiO 3, LaCuo.5Mno. 5 0 3 , Lao.75 Cao.2 5 FeO3
(using specific surface areas listed in table 1) and of Pt/C versus potential, obtained after
iR and 02 mass-transport correction ((from capacity-corrected negative-going scans at
1
10 mV s~1 in 0.1 M KOH for oxides, from positive-going scans at 10 mV s~ in 0.1M
KOH for Pt/C). Error bars represent standard deviations from at least three independent
repeat measurements.
In the following, we seek to estimate the kinetically limited cathode performance of
LaNiO 3 and Pt/C if used in a fuel cell based on mass activities shown in Figure 3.7. A
recent report compared the cathode performance of Lao. 8Sro. 2Mn0 3.is with that of Pt/C at
equal loading (1 mgoxide or 1 mgpt per Cm2 cathode), showing an only -0.1 V lower
performance for the perovskite catalyst 6 . Another study showed essentially identical
performance of Lao. 6 Sro.4Mno.8Feo.2O3 with that of a Pt/C cathode of, however, undefined
Pt loading"1 . In general, it is not straightforward to compare the potential cathode activity
of "cheap" oxide catalysts with that of expensive precious metal catalysts, since the
cathode loading of non-precious metal oxide catalysts is limited only by electrode
50
thickness constraints, while the loading of a Pt catalyst is constrained by cost
considerations.
Here we assume that the maximum electrode thickness for a "cheap"
oxide-based cathode electrode is on the order of 100 gm as suggested previously 7 and
will compare it to the conventionally used Pt loading of 0.4 mgpt cm-2 cathode. We
emphasize this projection is based on an idealized scenario in the oxide electrode, where
full utilization of the catalyst with minimum mass transport loss is assumed. Given the
density of 7.2 goxide Cm- 3 of LaNiO 3 and assuming 50% porosity in the catalyst layer, the
maximum LaNiO 3 cathode
loading would be
-36 mgoxide cm-2cathode
(obtained by
multiplying cathode thickness, density, and porosity). The kinetically limited cathode
performance of a LaNiO 3 catalyst can then be estimated by multiplying the maximum
LaNiO 3 loading with its mass activity values shown in Figure 3.7, which then can be
compared to the cathode performance of a Pt/C catalyst obtained by multiplying its mass
activity data in figure 7 with a typical loading of 0.4 mgpt Cm-2cathode.
This is shown in
Figure 3.8, indicating that essentially identical cathode performance for Pt/C and LaNiO 3
would be projected under these assumptions. Another strategy to match the activity of
the LaNiO 3 -based cathode to the Pt-based without constructing 100 pim electrode is to
synthesize smaller particles to reduce particle size from 200 nm to 40 nm to allow for 5
times larger surface area per mass. Many synthesis routes have been reported to prepare
oxide particles in the range of 30-100 nm4 . At this particle size, the oxide-based
electrode performance can be as competitive as that of Pt-based at -7 mgoxide cm-2cathode
(with an estimated electrode thickness of -20 pm.) At this loading, the assumption of full
utilization with minimum mass transport loss is likely more realistic.
From this analysis, the oxide-based cathode performance could indeed be competitive
with that of Pt/C but, at the same time, it is clear that oxides with higher ORR activity
than LaNiO 3 are desired with, most importantly, long-term stability over a wider potential
range. The establishment of a reliable and fast ORR activity screening method as
presented in this work can accelerate the development of oxide-based ORR catalysts with
enhanced activity and durability. Identifying more active and more durable oxide-based
ORR electrocatalysts is subject of the next chapter, where we will examine the ORR
activity of a large number of model perovskites using the thin-film RDE technique with
51
alkali ion exchanged Nafion* binder presented here, seeking to develop fundamental
ORR activity descriptors which would allow to more effectively search the large
parameter space of partially substituted AA'BB'0
3
perovskites, analogous to what was
developed for noble metal catalysts2 1 .
Projected kinetically limited
cathode performance
0.95
0.901-
l /C
-
10.85 -
uJ
0.801
10
100
ICathode
Figure 3.8.
1000
(mA cm-2)
Projected kinetically limited cathode performance under an idealized
scenario of full utilization of the catalyst with minimum mass transport loss for electrodes
with different ORR catalysts and catalyst loadings: LaNiO 3 at 36 mgoxide cm-2 electrode or
commercial Pt/C at 0.4 mgpt cm-2 electrode. Projections are based on the mass activity data
in figure 7 (100 kPaabs 02 at room temperature).
3.4 Chapter 3 Conclusion
We report a methodology using the thin-film RDE technique to quantify the mass and
specific ORR activities of sub-micron-sized oxide catalysts, as poor mass transport in
traditional porous electrodes leads to ambiguity in comparing the intrinsic ORR activity
of different oxides. Using this method, we show that Lao. 75Ca 0 .2 5FeO 3 has the lowest mass
and the specific activities, followed by LaCuo.5Mno. 50 3 , and by LaNiO 3, which is the
most active.
The Koutecky-Levich
analysis indicates that, for LaNiO 3 and
LaCuo.5 Mno. 5 0 3 , the reaction proceeds through the 4-electron reduction where water is
52
the final product. Compared to state-of-the-art platinum nanoparticle catalysts, LaNiO 3
has only about an order of magnitude lower specific activity. The performance projection
of LaNiO 3 catalysts in actual fuel cell cathodes, where the catalyst loading of noble-metal
free catalysts is limited by electrode thickness constraints, suggests comparable
performance to platinum nanoparticles supported on carbon, where catalyst loading is
constrained by cost instead. This chapter establishes a fast, reliable technique to screen
sub-micron-sized catalysts, which can aid in the development of highly active and lowcost ORR catalysts.
53
Chapter 4 - Activity Descriptor for ORR
4.1 Introduction
Transition metal oxides have shown reasonably high activity for the ORR in fuel cells" 3
and the OER in water electrolysis'
pH.
35
and direct solar water splitting 08 at neutral and high
However, lack of fundamental understanding of the ORR mechanism and the
material properties that govern the catalytic activity hampers the development of highly
active oxide catalysts.
In this chapter, we report a volcano relationship between a
material property that serves as the activity descriptor and the intrinsic ORR activity of
perovskite-based oxides. Such information has predictive power and provides insights
into the design of new catalysts with enhanced ORR activity similar to those reported for
platinum-based metals18 , 19 , 2 1 .
We examine the perovskite family, where A sites with rare-earth metal ions, and B sites
with transition metal ions can allow partial substitution to form AA'BB'0 3 to
experimentally examine a large number of oxides (fifteen total) in order to establish a
1 0 and Bockris and Otagawa8 2 have
catalytic descriptor for ORR. Matsumoto et al., ' "'"
reported the geometric currents of perovskites in thick, porous electrodes as a function of
potential but the intrinsic specific ORR activity (kinetic current densities normalized to
catalyst surface area) necessary for ORR mechanistic discussion is not available. In the
last chapter, we describe a methodology using a thin-film rotating-disk electrode with
well-defined oxygen transport13 to allow a precise comparison of the ORR activity of
different transition metal oxides. This method yields a more accurate determination of
the intrinsic ORR activity than that estimated from the data reported by Bockris and
Otagawa
3
as oxygen mass transport resistances in thin-film electrodes are much better
compensated than thick electrodes with very high internal surface area. In this work, we
apply this methodology to assess the ORR activity on fifteen perovskite-based oxides
with various A-site (Laj xCaxBO 3, La+xBO3 .x) and B-site (LaB xB',0 3) substitutions,
which is used to identify material properties (descriptors) that govern the intrinsic ORR
activity. Here we utilize the molecular orbital approach to identify descriptors for the
54
ORR activity of oxides, in contrast to the hypothesis proposed previously by Matsumoto
et al. 24,109,10 which is based on the conjecture that the ORR activity of oxide electrodes
can be greatly influenced by the formation and filling of a o-* band between the eg
orbital of bulk transition metal ions and a molecular orbital of surface oxygen. The use
of the molecular orbital approach to these perovskites is supported by a large number of
reports that the surface of transition metal oxides favors electron localization over the
bulk-itinerant electron state2 9 , 111,112. Herein, we show that the primary descriptor that
governs the ORR activity of the fifteen perovskites examined in this work is the extent of
o * antibonding ("eg") orbital filling of surface transition metal ions. This hypothesis is
further supported by an increasing ORR activity with more hybridization of the B-O (Bsite-metal and oxygen) bond as revealed by 0 K-edge X-ray absorption spectroscopy
(XAS) analysis.
4.2 Experimental Procedure
Materialsynthesis. The perovskites samples used in this study were synthesized with a
co-precipitation method. Rare and alkaline earth nitrate and transition metal nitrate (both
99.98% Alfa Aesar) in a 1:1 mole ratio were mixed in Milli-Q water (18 MQ-cm) at 0.2
M metal concentration. The solution containing rare and alkaline earth, and transition
metals was titrated to 1.2 M tetramethylammonium hydroxide (100% Alfa Aesar). The
precipitate was filtered, collected, and dried. The powder samples were subjected to heat
treatment at 10000 C under Ar atmosphere for the La1 ,CaxMnO 3 (x = 0, 0.5) and
La0 5Ca0 5CrO 3 , LaNi0 5 Mno.O
3
and LaCu0 5 MnO.O
3
samples, at 1000*C under dry air
atmosphere for LaCrO 3, La xCaxFeO 3 (x = 0, 0.25, 0.5), LaIsxCaxCoO 3 (x = 0, 0.5),
La 3 Ni 20 7 ,
La4 Ni 3 0
10
samples, and at 800*C under 02 atmosphere for the LaNiO 3 sample.
LaMnO 3, was prepared from an 800*C heat treatment of LaMnO 3 in air. La2 NiO 4 was
synthesized in 2 steps: 1000'C under Air atmosphere and then heated at 800'C under Ar
atmosphere.
All gases had ultra-high-grade purity (Airgas).
All samples were
characterized by X-ray diffraction to determine phase purity with a Rigaku high-power
rotating anode X-ray powder diffractometer at a scan rate of 0.8 degree/min. All X-ray
diffraction data and refined lattice parameters are shown in Table 4.1 respectively. All
55
the oxides used in this study had a single phase except LaNiO 3, where <2% of NiO was
present.
Average particle sizes with standard deviations and estimated surface area
values were quantified via Philips XL30 FEG ESEM (FEI-Philips), which are listed in
Table 4.2.
0
0
0
Space group
a(A)
b(A)
c(A)
LaCrO3
Pnma
5.48
7.76
5.51
LaMnO3
Pnma
5.66
7.72
5.53
LaFeO3
Pnma
5.57
7.85
5.56
LaCoO 3
R -3 c
5.44
5.44
13.09
LaNiO 3
R -3 c
5.46
5.46
13.14
La2 NiO 4
Fmmm
5.47
5.46
12.68
La 4 Ni3 O10
Fm mm
5.41
5.46
27.99
LaNi0 .Mno.5O
3
Pn m a
5.46
7.74
5.51
LaCuO.5 Mn.O
3
Pn m a
5.48
7.77
5.52
LaO. 5CaO.5CrO3.,
Pn m a
5.42
7.68
5.45
LaO. CaO.5MnO3.
Pn m a
5.42
7.65
5.43
LaO.5CaO.5FeO 3.b
Pn m a
5.55
7.84
5.55
LaO.75Cao.23FeO 3.b
Pn m a
5.52
7.80
5.52
LaO.5Cao.5CoO 3.,
Pnma
5.41
7.59
5.36
LaMnO36
R -3 c
5.52
5.52
13.35
Table 4.1. The derived lattice parameters of the oxide model compounds in this work
56
dumber (pUm)
dia (pm)
LaCrO3
0.64(±0.25)
0.83
1.1
LaMnO3
1.05 (±0.52)
1.51
0.6
LaFeO3
0.71 (±0.34)
1.01
0.9
LaCoO3
0.78 (±0.40)
1.10
0.7
LaNiO 3
0.20 (±0.06)
0.24
3.5
La 2NiO 4
0.49 (±0.25)
0.70
1.2
La4Ni3 ,1,
0.45 (±0.15)
0.53
1.6
LaNio.,Mn,.O 3
0.34 (±0.11)
0.81
1.1
LaCuO.5 Mn.50 O 3
0.58 (±0.28)
0.80
1.1
LaO.5CaO.5CrO3 .a
0.19 (±0.06)
0.22
4.9
LaO.5CaO.5 MnO3 -
0.92 (±0.44)
0.62
2.1
LaO. CaO.,FeO 3.b
0.62 (±0.31)
0.89
1.1
LaO.75CaO.25FeO 3.
0.36 (±0.22)
0.59
1.8
LaO.SCaO.SCoO 3 -
0.43 (±0.23)
0.63
1.6
LaMnO 3.,
1.39 (±0.58)
1.81
0.5
A, (m2
.'1)
Table 4.2. Characterizations of the oxides studied in this chapter. The number averaged
diameter, dnumber, the volume-area averaged diameter, dvia, and the specific surface area, A.
were obtained from particle size distribution measurements. Methodology for calculating
each variable is given elsewhere".
57
Electrode preparation. Glassy-carbon electrodes (5 mm in diameter, Pine Instrument)
were polished with 0.05 Vm Alumina slurry (Buehler) to obtain a mirror finish.
Perovskite oxides were mixed with acid-treated acetylene black (Chevron) at a mass ratio
of 5:1. Catalyst ink was prepared by mixing perovskite oxides and acetylene black with
tetrahydrofuran (THF, 99.9+% Sigma-Aldrich) and an appropriate amount of neutralized
Nafion (5% weight, Ion power) as described elsewhere 3 . The electrode had a final
composition of 250
[goxide cm 2 disk,
50
[gAB Cm2 disk,
and 50
[gNafion cm 2 disk-
Electrochemicalcharacterization. Electrochemical measurements were conducted with a
rotating-disk electrode (Pine Instrument). All measurements were collected in a 100 mL
solution of 0.1 M KOH, prepared from Milli-Q water (18 MQecm) and KOH pellets
(99.99% weight, Sigma-Aldrich) electrolyte, with a VoltaLab PST050 potentiostat. In
this work, error bars represent standard deviations from at least three independent repeat
measurements.
The analysis of ORR kinetic currents were outlined in our previous
work 3 , and the specific ORR activity was obtained from normalizing the kinetic current
by the surface area of each oxide estimated from SEM images (Table 4.2). Because the
errors associated with the surface area estimation (Table 4.2) are smaller than the
experimental uncertainty in the intrinsic ORR activity of oxides from our RDE
measurements (-30 mV standard deviation, translating to a change of -3.2 times in the
ORR current density normalized to oxide surface considering a Tafel slope of 60
mV/decade), the ORR trend are not affected significantly by the errors in the oxide
surface area. In addition, BET measurements of the surface area of select oxides showed
reasonable agreements with those estimated from SEM images (e.g. LaNiO 3 , 5.6 m2/g
from BET measurements vs. 3.5 m2/g estimated from SEM images). Moreover, it should
be noted that the specific ORR current densities for LaCrO 3 and LaFeO 3 were comparable
to the ORR activity of acetylene black (background ORR activity) considering
experimental uncertainty 3 and therefore, the ORR activity values reported here can only
present as upper limits for these two oxides. Furthermore, the ORR activity results
obtained from the thin-film RDE measurement can be well translated to the actual
performance of the air electrode in fuel cells and metal-air batteries as shown previously
for PEMFCs7 . Lastly, it should be mentioned that preliminary stability testing of the
58
most active oxide, LaMnO 3.b, showed very small ORR activity changes (smaller than the
experimental uncertainty in the RDE measurements) during 3.5 hours of continuous
cycling between 0.7 and 1.0 V vs. RHE, which suggests that oxide materials can serve as
reasonably stable ORR catalysts in fuel cells and metal-air batteries.
Hard X-ray Absorption characterization. Hard X-ray Absorption Spctroscopy (XAS)
was collected at beamline X1 1A of the National Synchrotron Light Source at the
Brookhaven National Laboratory with the electron storage ring of 2.8 GeV and a current
in the range of 150-300 mA. The Mn, Ni, Co, and Fe K-edge electron yield modes were
collected from oxide powders mounted on carbon conductive tape with a Lytle detector at
room temperature using Si(1 11) double-crystal monochromator detuned to -70% of the
maximum intensity. The spectra were calibrated to the reference metal foils by setting
the maximum inflection points (E0 ) to their respective reference energies.
X-ray
absorption near edge structure (XANES) was extracted from the absorbance with the
IFEFFIT package by subtracting from the pre-edge region with a linear fit and
normalizing to a per atomic basis with the average of the absorption cross section over
113
the post-edge region'3
Determination of eg filling of non-stoichiometric LaMnO3, 6 . Based on the published
thermodynamics, subjecting LaMnO 3 to a heat treatment in air at 800'C will result in a
phase transformation to an oxygen over-stoichiometric LaMnO 3.6 compound with 6 ~
0.11 (±0.05)"4. To ensure that the synthesis of LaMnO 3+6 was successful, XRD was
collected to confirm the rhombohedral phase of LaMnO 3.8 (Table 4.1).
published 6 = 0.11, LaMnO 3.8 has a high spin
3
t2
From the
. configuration and eg filling of 0.79
e, 079
(±0.05) =0.8"5.
Determination of eg filling of Ca-substituted La 0o5Ca05MnO 3.
X-ray emission spectroscopy
15
Previous studies using
, thermogravimetric analysis (TGA)' 16 , X-ray absorption
near-edge spectroscopy (XANES)' 17 , and magnetic measurement
18
have shown that
charge compensation following Ca-substitution in LaMnO 3 occurs via Mn* formation.
We, therefore, assume that our La0.5Ca 05MnO 3 compound charge-compensates by
forming Mn**. As a result, La0. 5Ca0.5MnO 3 has an average Mn oxidation state of 3.5+ and
therefore an electron configuration of t 2 ,3 eo. 5 . We had conducted Mn K-edge XANES to
59
confirm that the Mn in LaO. 5CaO5MnO 3 shifts to higher oxidation state in relative to
LaMnO 3 (Figure 4.1).
D
CLaMnO3
La Ca Mn0a
6540
6560
6580
Figure 4.1. X-ray Absorption Spectra LaMnO 3 vs. Lao. 5CaO. 5MnO 3
LaFeO3
U)LaO,CaO,FeO
7110 7120 7130 7140 7150
Energy (eV)
Figure 4.2. X-ray Absorption Spectra LaFeO 3 vs. Lao. 5Cao.5FeO 3
Determination of eg filling of Ca-substituted LaO 75Cao.25FeO3 and Lao
0 5CaO
5FeO3 .
The
charge compensation following Ca-substitution in LaFeO 3 is assumed to occur via Fe'
60
formation in high spin configuration. This is in agreement with the observation of Fe*
formation in Ca-substituted LaFeO3 using Mossbauer spectroscopy
9 ,120 .
Our own Fe
K-edge XANES also confirms that the Fe oxidation shifts to higher oxidation state with
increasing Ca content (Figure 4.2).
LaCoO3,
La 0Ca CoOa7675
7700
7725
7750
7775
Energy (eV)
Figure 4.3. X-ray Absorption Spectra LaCoO 3 vs. Lao. 5 Cao.5 CoO 3 .3
Determination of eg filling of Ca-substituted Lao5 Cao5 CoO3 6 . Because the exact spinstate of La 0 5 Ca0 5 CoO3 6 has not been determined yet in the literature, partly due to the
fact that the majority of Co remains 3+ despite the substitution of La * by Ca2 ion 2 , we
approximate the spin state of this compound as a mixture of LaCoO 3 and SrCoO 2 .5 .
Using a mixture of intermediate spin LaCoO 3 (t2g5 eg ) 1
e2)122, we arrive at an approximation of eg
=
2
and high spin SrCoO 2 . (t2g'
1.5 for Lao5 Ca0 5 CoO 3 -8. The concept of
high-spin Co stabilization with Ca-substitution has been proposed in the literature121,13,
14,
so we believe our approximation is within reason, especially for the surface Co. We
had also conducted TGA to confirm the formation of oxygen vacancy (6 TGA= 0.21±0.01)
that had been reported in the literature following Ca-substitution.
Our Co K-edge
XANES additionally demonstrated that the Co edge remained unchanged following
61
Ca-substitution, which provides further evidence for the stabilization of Co 3 * in
Lao 5Cao 5CoO 3M(Figure 4.3).
Determination of eg filling of LaNi0 5 Mno.Os3 and LaCuo5 Mno.O 3 . The eg filling for the
mixed B-site compounds (LaNio5 Mno.5O 3 and LaCu0eMno.O 3 ) is complicated by the
presence of two inequivalent B atoms and thus two different eg fillings. It is also widely
known that a charge-ordered Mn/Ni undergoes charge-disproportionation into Mn" and
Ni2 * in LaNiXMnjxO 31 5 .
Thus, to determine properly the eg filling for the mixed
compounds, the oxidation state of both B atoms was measured with XANES.
Using a
known-relationship where E0 (maximum inflection point of the absorption edge) scales
with oxidation state 126, we estimate the valence state of Mn in both LaNi05 Mn.0 O 3 and
LaCu0 5 Mn0 5 O3 to be 3.7 (high spin configuration, Figure 4.4). A high-spin configuration
of Mn leads to an eg filling of 0.3 for the Mn site. Similar analysis was applied to Ni in
LaNio5 Mno.O 3 , and the oxidation state of 2.3 was found (Figure 4.4). Based on this
information, the eg filling for Ni is thus determined to be 1.7 (low-spin configuration).
While we did not perform this analysis on Cu in LaCue 5 MnO.O 3 , we believe the Cu
oxidation state should be close to that of Ni if not lower due to the higher
electronegativity of Cu atom.
between LaNio5 Mno.O
3
Furthermore, the similarity in the Mn oxidation state
and LaCu0 5 Mno.O
3
also indicates that the Cu and the Ni
oxidation state should be very close to each other (~2.3+). Thereby, we conclude that the
eg filling for Cu is 2.7 or higher. When we apply the eg filling from either Mn or Ni (Cu)
to the volcano plot, we have found that the use of eg filling from Mn is most consistent
with the observed activity trend (vide infra). Had we used the eg filling of average B
atoms, or Ni or Cu, the eg would have resulted in an underestimation of the ORR activity.
The consistency of selecting Mn eg filling with our volcano plot also leads us to propose
that the active site on the mixed compounds is the Mn atom, where both Ni and Cu atoms
had too many eg electrons, rendering their catalytic properties inactive. The proposed Mn
active site is consistent with our observation with the Ruddlesden-Popper compounds that
Ni(+2) and Ni(+2.5) were not active for the ORR. Hence it is unlikely that Ni(2.3+),
which was observed in LaNi0 .5 Mno0 O3, would be the active site.
62
6556
La(NiCu)05 Mn0 5 O3
~6554
La 0.5 Ca 0.5 MnO,3
0)
C
6552
a
LaMnO 3
3.0
3.2
3.8
3.6
3.4
4.0
LaNiO 3
8348
LaNi 5 Mn0 5 O3
8347
mU
8346
a
2.0
Figure 4.4.
LaCuo. 5 Mno. 5 03.
Determination
La N104
2.2 2.4 2.6 2.8
Oxidation state
of B-site
(Top) Mn oxidation
oxidation
3.0
state of LaNio.5 Mno. 50 3 and
state for LaNio.5 Mno. 5 0
3
(red circle)
and
LaCuo.5 Mno.50 3 (blue triangle) is estimated to be ~3.7+. (Bottom) Ni oxidation state in
LaNio5Mno. 5 0 3 (red circle) is estimated to be ~2.3+.
4.3 Method for ORR Activity Comparison
The ORR currents of a LaCu 5 MnO50 3 thin-film electrode at various rotation speeds are
shown in Figure 4.5 as an example. The onset of the ORR occurs at ca. 0.8 V vs. RHE
(reversible hydrogen electrode), and an oxygen-transport-limited current appears at
63
potentials below ca. 0.5 V vs. RHE. The slope of the Koutecky-Levich plot in the masstransport-limited region (0.4 V vs. RHE, Figure 4.5 inset) has a value of ~1-1
(mA-'cm'disk-1), which indicates 4-e~ reduction of oxygen to water",
transport correction to the oxygen diffusion overpotential
1.
The mass-
was applied to all
measurements to obtain the ORR kinetic currents as outlined in our previous work13 . The
surface-area-normalized kinetic current densities of four representative oxides, referred to
as specific activity (i,),
3
are plotted as a function of voltage in Figure 4.6. Because all
fifteen oxide catalysts had comparable Tafel slopes around 60 mV per decade (Figure
4.7), the dashed line in Figure ic shows the intrinsic activity for each catalyst can be
assessed by the potential to achieve a given specific ORR current (25 pA/cm
2
0 x).
For
LaCuO5 Mno.O 3 , LaMnO 3, LaCoO 3 , and LaNiO 3 , this specific activity current can be
reached at potentials of 781(±15), 834(±24), 847(±3), and 908(±8) mV vs. RHE,
respectively. Higher potential indicates higher electrocatalytic activity for a given oxide.
It is interesting to note that oxides such as LaMnO 3+6 and LaNiO 3 have intrinsic ORR
activity comparable to the state-of-the-art Pt/C (Figure 4.8).
0
E
E
-4 -0-6.
.4
<02.
E
-6
7
0.4
Figure 4.5.
CE
o/2
0)-1/2 (s)
0.6
0.8
EN vs. RHE
Oxygen reduction activity of LaCuo.5 Mn. 5 0
o
1.0
3
electrode in 0 2-saturated
0.1 M KOH at 10 mV/s scan rate at rotation rates of 100, 400, 900, and 1600 rpm
64
The Koutecky-Levich analysis (inset) of the limiting currents (0.4 V)
respectively.
indicates a 4e~ transfer reaction.
-25
00M
C4m
-30
-35
0.80
0.90
0.85
E(V vs. RHE)
Figure 4.6. Specific activities of LaCuo. 5 Mno.5 0 3 , LaMnO 3, LaCoO 3, and LaNiO 3. The
potential at 25 pA/cm 2ox is used as a benchmark for comparison (shown as the
intersection between the activity and a horizontal gray dashed line).
65
1.0
. .
..
- .4 . ..
..
.
.
. ,
LaNio3
Ui0.9
La 4NiaO 1
>0(0.8
3
-LaCoO
-
La NiO
0.7
LhV
La0*Ca 05CoO
LIJ.ao,0
0.61
1.0
LaCu Mn O0
'
10
100
1000
. . ....
....
~LaNio
I
LU0.9
X
w
>0
La 4Nia3O 1
;0.8 -LaCoO
3
La 2NiO4
>
0.7
La Ca CoO
0.6..............
10
LaCuo Mn
5
OS03
100
/S
(pIA CM-
1000
)
OX
Figure 4.7. Tafel plot of the ORR specific activity of oxides studied in this chapter using
thin film RDE.
66
I
-I
I
0.95
6100fthPt/C
0.90
S0.85> 0.85LaMnO
4 0.80
100
10
(A
1000
Cm ox O pAC
Pt
Figure 4.8. Tafel plot of the ORR specific activity of LaMnO 3+8vs. Pt/C from the thin
film RDE experiments. These two catalysts have very similar ORR activities (-50 mV
shift, corresponding to less than an order of activity difference).
Error bars represent
standard deviations of at least three independent measurements.
4.4 ORR activity descriptor identification
To identify an ORR activity descriptor, we begin by examining the relationship between
the 3d-electron number of B-site ions and the ORR activity with the hypothesis that the
3d-electron number, which represents the antibonding electron occupation of the B-O
bond, can influence the B-0
2
interaction strength3 5 (see Scheme 4.1). This hypothesis is
supported by the recent observation that the adsorption energy trend of B-0
2
can be
approximated by that of B-O'8. Interestingly, the comparison between the ORR activity
of the perovskites and the d-electron number per B cation (Figure 4.9) reveals an
M-shaped relationship with the maximum activity attained near d' and d7 , which
resembles the trend reported for oxidation activity of gas-phase CO and hydrocarbon on
perovskites 36 ,129.
67
OH-
02m.2-O
mO2
2-O
myO
12-
2-0~
N02-
B
02-
2-00
02Mn 3* (di)
d-manifold with
oxygen ligands (HS)
d-manifold
e
02Surface d-manifold
(Tetragonal, HS)
dX2 ?
dz,
-4f4--4-4
4-4tg
t2g
Scheme 4.1.
-
Molecular orbital of B0 6 interaction.
d
-$- -t- dxz idyz
The energy levels of the
d-electrons in free B ion, octahedral B0 6 , and surface B0 6 configurations. The initial d
manifold degeneracy is split into the antibonding eg and t2g levels. Mn 3 (d'4) in high spin
configuration is shown as an example of a cation with an antibonding eg-orbital
degeneracy with a single eg electron. At a surface cation, the degeneracy is removed by a
tetragonal site symmetry (C4v) in which the occupied eg electron occupying a z 2 orbital
directed toward a surface OH~ is lowered in energy relatively to the eg electron occupying
the x2_y2 orbital.
We further show that the intrinsic ORR activity of all the oxides exhibits a volcano shape
as a function of the eg-filling of B ions. Assuming that the 02 molecule likely adsorbs on
the surface B sites end-on, the eg orbital directed toward an 02 molecule overlaps the 02p, orbital more strongly than the overlap between the t2g and the
O-2p,
orbitals, and
therefore suggests that the filling of eg rather than t2g of the B-ion should more accurately
determine both the energy gained by adsorption/desorption of oxygen on B ions and the
ORR activity if such an adsorption/desorption process was involved in the rate-limiting
step of the ORR. This conjecture is supported by the correlation between increasing eg
filling and decreasing onset temperature of oxygen release measured via 02 temperatureprogrammed desorption in the order of LaCrO 3 (eg = 0) > LaMnO (eg = 1) ~ LaCoO (eg
3
3
1) LaNiO 3 (eg ~ 1) > LaFeO 3 (eg = 2) , which is indicative of the strength of bonding
68
between B-site cations and oxygen. We have taken into account whether partial
substitution on the A or B sites undergoes charge compensation via oxygen vacancy or
changing of B ion valency in the eg calculation of the surface B ions, which was
determined from a combination of literature data, thermogravimetry and XAS (see Table
4.3). Plotting ORR activity as a function of the eg filling produces a definitive volcano
plot (Figure 4.10) with a voltage span of 0.25 V corresponds to ~ 4 orders of magnitude
in the intrinsic ORR activity (using ~60 mV/decade Tafel slope.)
@2
Alcm
0"
1.0 -C-
OU LaNiO3
LaM3NiO
M
y
0.8 -
44j
La 2 NiO
La Ca CrO
1
2
3 4
5
6
7
8
9
d-electron
Figure 4.9. The potentials at 25 gA/cm 2 ox of the perovskite oxides have M-shaped
relationship with the d-electron number. The data symbols vary with the type of B ions
(square red for Cr, orange circle for Mn, beige upside up triangle for Fe, green upside
down triangle for Co, blue diamond for Ni, and purple left triangle for mixed compounds),
where x = 0 and 0.5 for Cr, and 0, 0.25, and 0.5 for Fe. Error bars represents standard
deviations of at least 3 measurements.
69
o.9
La Ca MnO
17-~~~~3
05
LaCoO
LaMn 0.5 Ni 0 5
-X-
50. 8
LaMn C
LaMnO
Y T La 37Ni3
La0 5Ca ,,Co0 3 .
3
La2 NiO
4
0.7-La Ca CrO
0
Figure 4.10.
La Ca FeO
1
e 9 electron
2
The potentials at 25 gA/cm 2 .x as a function of eg-orbital in perovskite-
based oxides. The data symbols vary with the type of B ions (square red for Cr, orange
circle for Mn, beige upside up triangle for Fe, green upside down triangle for Co, blue
diamond for Ni, and purple left triangle for mixed compounds), where x = 0 and 0.5 for
Cr, and 0, 0.25, and 0.5 for Fe. Error bar represents standard deviations of at least 3
measurements.
70
Valence
Spin state
Assignment
Example of reference
LaCrO3
Cr3 *
n/a
t2g
LS and HS identical
LaMnO3
Mn 3 *
H.S.
t2g
e
LaFeO3
Fe3+
H.S.
t2
e
LaCoO3
Co 3 +
I.S.
LaNiO 3
Ni3+
L.S.
La 2NiO 4
Ni 2'
n/a
La 4Ni 3O 10
Ni2.7.
L.S.
LaNiO..MnO._O
Magnetization
Mossbauer
spectroscopy 20
e
t2g
Magnetization13 1
Extrapolation from
RNiO 3 magnetization13 2'
Extrapolation from
t2g
Ni2.3+
7
Mn - *
L.S. (Ni)
H.S. (Mn)
tag6e . (Ni)
t2g e 3(Mn)
Method
See Supplementary
Cu2.3*
L.S. (Cu)
e
13
Method
e(1n)
LS and HS identical
(Mn)
e ~03
6
t2g
LaO.5CaO.SCrO 3 -a
Cr2.*
n/a
t2geg
(Assume b= 0)
LaO.5CaO.5MnO3 a
Mn 35
H.S.
t2g
Fe 35
3
t2g
Magnetization1",
eg
X-ray Emission 1 5
Mossbauer
H.S.
(Assume 8= 0)
LaO.75CaO.25FeO3.b
5
Fe3.2
*
H.S.
( Mn
t2g e
(Assume b= 0)
LaO.5CaO.5CoO 3-4
Co 3*
I.S./H.S.
t2g
e
(OTGA= 0.21±0.01)
LaMnO3.16
(Assume b = 0.1)
LaNiO 3 and La 2NiO 4
See Supplementary
H.S. (Mn)
3
(Assume b= 0)
LaO.5CaO.5FeO3.b
LS and HS identical
Mn3.7*
3
LaCuo.5 Mno.0 O 3
tag 692
H.S.
t2g eg
5
spectroscopy11,120
Mossbauer
spectroscopy11,120
La1 ,CaCoO3-8
extrapolation 12 1
X-ray Emission
Spectroscopy
15
Table 4.3. Summary of literature information on the spin state of the perovskite oxides
71
4.5 Proposed ORR mechanism
We explain the volcano trend of oxide ORR activity with the following. Too little of eg
filling in LajCaCrO3 (t2g3 e 0 for x = 0, t2g 2-ego for x = 0.5) can result in too strong B-02
bonding while too much eg filling in LalCaxFeO3 (t2g3 e
t2g3eg
2
for
x =
0, t2g3 e1- 5 for x = 0.25,
for x = 0.5) and LalxNiO 3.x [(LaNiO 3)1,jLa 2NiO 4),, having (1-x) t2 g6eg'
+ x
t2g 6eg 2 )]
can lead to too weak 02 interaction, neither of which are optimum for the ORR activity.
On the other hand, a moderate amount of eg-filling (~1) in Laj xCaxMnO 3 (t2g 3eg' for x = 0,
t 2 g3 e 0. 5 for x =
0.5), LaCoO 3 (t2g 5 eg'), and LaNiO 3 (t2g 6 eg'), yields the highest activity. The
striking correlation between the eg filling and 4 orders of magnitude of the ORR activity
suggests that the eg filling represents a unifying and primary descriptor for the ORR
activity of these oxides. The importance of eg filling also explains the origin of the peaks
at high-spin d4 and low-spin d'in the M plot of ORR activity versus the number of d
electrons (Figure 1d): both possess eg filling of ~1. The M shape thus arises from the
spin-state transition on the B ions from high spin (d <6) to low spin (d > 6).
The importance of a single eg electron for ORR catalysis can be molecularly explained by
the previously proposed competition between the Of/OH- displacement (Step 1) and OH-
regeneration (Step 4) on the surface of the transition metal ions as rate-limiting for ORR
in an alkaline solution 3 4 (see Figure 4.11). In this scheme, B is a transition metal cation
with an eg electron ordered into an orbital directed toward the surface OH- ion. As a first
order of approximation, the kinetics of the 0 22 /OH- exchange can be rationalized in terms
of the energy gained by breaking the B-OH- bond to form B-0 2 2 . The presence of a
single eg represents a a * electron that can serve to destabilize the B-OH- bond and
promote the
02 /OH-
exchange reaction.
The displacement of the OH- ion that
experiences a a * electron in the B-OH- bond by the 02+ e- (or O2,ads ) species removes the
highly energetic a * electron from the B-OH- bond to give a more stable B-0
2
configuration. If the eg electron is more than 1 (the right branch in Figure 4.10), the 022
/OH- exchange does not gain sufficient energy during the displacement and the ORR
kinetics can be limited by the rate of the 0 2 /OH exchange (Step 1). On the other hand, if
there is less than one eg electron on the B cation (the left branch in Figure 4.10), the B-0 2-
72
is not sufficiently de-stabilized and the ORR kinetics can be instead limited by the rate of
surface OH~ regeneration (Step 4).
OHH2 0+e
OH-
+e-
Im+
O-B-O
OH-
4
02-
002-
I(m+1)+
I(m+1)+
O-B-O
OH-
O-B-O
3
OOH-
H 2 0+e
2
O-B-O
eFigure 4.11.
OH-
Proposed ORR mechanism on perovskite oxide catalysts 1 .
The ORR
proceeds via 4 steps: 1 surface hydroxide displacement, 2 surface peroxide formation, 3
surface oxide formation, and 4 surface hydroxide regeneration.
While our observation on the importance of o * electron transfer to 02 in ORR activity is
in qualitative agreement with the previous hypothesis of Matsumoto et al. 24,109, 110 our
work offers a quantitative correlation between the eg filling of surface transition metal
ions and intrinsic ORR activity up to 4 orders of magnitude and predicts eg
i1 to be a
key criterion for developing the highest activity. In addition, an important distinction of
our analysis is the assumption of a localized eg electron in an orbital directed toward an
02 molecule from the surface B cations instead of the traditional band theory" 1"'". We
here discuss many works that have provided evidence to this assumption to date. In the
case of LaNiO 3, which has itinerant bulk a * electrons, the site symmetry of the bulk
rhombohedral structure in LaNiO 3 does not permit a localization of an eg electron in the
bulk; but the symmetry of a surface B site changes to tetragonal, which allows ordering
of a localized eg electron directed toward the surface anion species. This phenomenon is
also displayed in low-temperature LaCoO 3 , where eg localization gives rise to the
73
ferromagnetic surface intermediate-spin Co(III):t2g5 eg' despite the low-spin Co(III):t 2g 6 ego
bulk phase' 12 . A similar argument applies to the metallic system La 1 _SrMnO3, which
exhibits a transition with increasing x from localized eg electrons to a narrow a * band on
the high-spin Mn atoms in the bulk 29 . Here, the localization of the surface eg electrons is
stabilized by a reduction of the number of Mn near neighbors as well as by site
symmetry.
4.6 Influence of B-O covalency on ORR activity
To further support that the eg filling of surface B ions governs the ORR activity, we
examine the influence of the covalency of the B-O bond of the perovskites on the ORR
activity.
Stronger covalency of the B-O bond should increase the driving force and
thereby facilitate the 0 2 /OH exchange on the surface B ions, which can be considered as
the rate-limiting step of ORR for oxides (Figure 4.11).
To quantify the B-O bond
covalency of the perovskites, we performed the 0 K-edge XAS, which revealed
excitations from the 0 Is orbital to the unoccupied states above the Fermi level from
regions within ~10 nm from the surface (50% of the signal comes from top 1 nm6 1). The
0 K-edge spectra collected from perovskites consist of two major excitations: 0 Is -+ B
3d - O 2p band from 527 - 533 eV, and 0 1s -- La 5d - 0 2p band from 533 - 537 eV6 .
As discussed in earlier chapter, the specificity of Is -- 2p excitation in the former
permits the quantification of the covalency or hybridization effect of the perovskites in
this study. To probe the effect of hybridization on the ORR activity without interference
from the eg filling effect, we plot the influence of the hybridization parameter on the ORR
activity of perovskites at constant eg filling in Figure 4.12.
Increasing hybridization
positively affects the ORR activity (Figure 4.12), which is in agreement with the
molecular picture of the proposed ORR mechanism.
74
0.95
0.90
,
09 @ 25 pAlcmo
*
e =14
e
LaNiO3
0.85
0.80
LaCoO
0
LaMnO,
0.6
0.5
0.4
0.3
Abs./(holeg+1/4 hole g)
Figure 4.12. The role of B-O covalency on the ORR activity of perovskite oxides. The
potentials at 25 pA/cm 2
as a function of the B-O covalency at eg filling = 1. Error bars
represents standard deviations of at least 3 measurements.
4.7 Chapter 4 Conclusion
This chapter shows that the design principle for enhancing the ORR activity of transition
metal oxide perovskites is an eg filling having a value of -1 for maximum activity, which
can be further improved by increasing the covalency between the metal 3d and oxygen 2p
orbitals. Our findings can be explained by a competition between a rate-limiting 0 2 2-/OH
exchange (Step 1 in Figure 4.11) and the regeneration of OH on the surface (Step 4 in
Figure 4.11).
This exchange depends on an energy gained by transferring a single
o*-antibonding eg electron of the B-OH- bond to the O2~ adsorbates, thereby stabilizing
the displacement. The greater the covalent contribution to the B-0 22 - bond, the greater
the energy realized by the exchange to stabilize the adsorbate and therefore the faster
ORR kinetics. This work shows that tuning surface electronic structure features such as
transition metal eg filling and covalency is a promising strategy to develop highly active
non-precious-metal-containing oxide catalysts for oxygen reduction in electrochemical
conversion and storage devices. Considering the synergism between the OER and the
75
ORR expected from the pioneering ab initio works of Rossmeisl et al.,
these
descriptors could also influence the OER activity of oxides, and such a study would test
the validity of the proposed influence of eg filling of transition metal ions on 02
electrocatalysis.
76
Chapter 5 - Activity Descriptor for OER
5.1 Introduction
The ability to design cost-effective, highly active catalysts for energy storage applications
is a critical element in the pursuit of sustainable energy. The OER in particular is an
enabling process for many energy storage options such as direct-solar'
electricity-driven water splitting (H20
Mx
(MA0
2 ->
+ 02)136,137.
->
3
and
H2 + MO2)5, and rechargeable metal-air batteries
The OER kinetics is sluggish, however, even when facilitated
by the state-of-the-art, precious-metal-containing catalysts.
This limitation imposes a
significant overpotential requirement' 3 5 , 138, similar to the case of the ORR8 '
18
for fuel
cell applications 7 . First-row transition-metal oxides such as NiCo 20413 9 and LaNiO3 5 are
cost-effective alternatives, but they underperform relative to IrO2, which is considered to
be state-of-the-art. To assist the discovery of highly active catalysts based on abundant
elements, substantial effort has been devoted to understanding the mechanism and
parameters that govern the OER activity (namely activity descriptor). Pioneering works
of Trasatti
40
, Bockris and Otagawa35 , and Rossmeisl et al.13
have reported the enthalpy
of a lower to higher oxide transition, the 3d electron number of bulk transition metal ions,
or the surface oxygen binding energy as an OER activity descriptor, respectively.
However, it is not straightforward to predict new transition-metal oxides with high OER
activity with these proposed descriptors"' " 144. In this chapter, we develop an OER
activity design principle, where a near-unity occupancy of the eg orbital of surface
transition-metal ions and high covalency with oxygen can enhance the intrinsic OER
activity of perovskite transition-metal oxides in alkaline solution (40H~ -> 02 + 2H 2 0 +
4e-). We further show that this design principle has predictive power, from which an
oxide catalyst with much higher activity than the state-of-the-art IrO 2 catalyst was found.
5.2 Experimental Procedure
Material synthesis. The perovskites oxides were synthesized with a co-precipitation
method as described in the previous chapter with the exception of Ba 0 5 Sr 0 5 Co0 8 Fe0 .2 O3
77
(BSCF), which was synthesized with a nitrate combustion method.
In the BSCF
synthesis, alkaline earth and transition-metal nitrates (all 99.998+% Sigma-Aldrich) were
prepared at 0.2 M concentration, to which glycine (>99% Sigma-Aldrich) was added at
0.1 M concentration. The mixture was heated until vigorously evaporating, at which
point, sparks are emitted as a result of combustion; it was then calcined at 11 00 C under
air atmosphere for 24 hours.
Figure 5.1 shows the diffraction pattern of the as-
synthesized BSCF, where the crystal was found to be single phase (P m -3 m) with lattice
parameter of 3.99 A. The particle size distributions and estimated surface areas of BSCF
are listed in Table 5.1.
Electrochemical characterization. The oxide electrode consisted of perovskite oxides,
acetylene black (Chevron), and neutralized Nafion* (Ion power); they were prepared at a
loading of 0.25 mgox cm -2disk, 0.05 mgA cm-2 disk, and 0.05 mgNafion cm-2 disk, similar to the
case of the ORR. Electrochemical measurements were conducted with a rotating-disk
electrode (Pine) using a VoltaLab PST050 potentiostat. The electrolyte was prepared
from Milli-Q water (18 MQ-cm) and KOH pellets (99.99% weight, Sigma-Aldrich). All
measurements were collected under 02 saturation (ultra-high-grade purity, Airgas) to
ensure the 0 2/H2 0 equilibrium at 1.23V vs. RHE at a rotation rate of 1600 rpm where no
mass transport limitation was observed.
The analysis of OER kinetic currents was
capacitance-corrected by taking the average between positive and negative-going scans,
and then iR-corrected (Figure 5.2).
The specific OER activity was obtained from
normalizing the kinetic current by the surface area of each oxide estimated from particle
size distribution. All the potentials are iR-corrected potentials unless stated as applied
potentials. The rotating ring disk electrode (Pine) with Pt-ring was calibrated with a
potassium ferricyanide couple as described elsewhere7 0 . Generated 02 was collected at a
ring potential of 0.4 V vs. RHE.
78
dnu,,er (ym)
dvia (Pm)
A, (m2 g 1)
Bao.5 Sro.sCoo
0 FeO.20 3.6
0.84 (±1.27)
7.01
0.2
BaO5 SrO5Co 0 8Fe 2 0 3.8
0.23 (±0.09)
0.30
3.9
(ball-milled)
Table 5.1. The number averaged diameter,
dnumber,
the volume-area averaged diameter,
dvia, and the specific surface area, As, were obtained from particle size distribution
measurements as described in the previous chapter. Note that we also measured the BET
area for BSCF, which were found to be within a factor of 2 of the specific area
determined from SEM.
U)
4,
30
Figure 5.1.
40
50 60 70 80
20 (Cr source)
90
X-ray Diffraction Pattern of Bao.5 Sro.5 Coo. 8Feo. 2 O3 .. The experiment was
conducted at room temperature using Cr source. The pattern revealed a space group of P
m -3 m (simple cubic) with a lattice parameter of 3.99 A.
79
Synthesis of rutile Ir0
2
nanoparticle. IrO 2 were synthesized in a 2-step reaction as
described elsewhere". Briefly, a solution of either IrCl4-xH2O in tetralin and oleylamine
was prepared at room temperature and then heated to 60 *C under Ar atmosphere for 20
minutes, followed by an injection of tetrabutylammonium borohydride in chloroform.
The reaction mixture was heated up to 200 *C for 19 hours. Afterward, the nanoparticles
were precipitated by ethanol addition and collected by centrifugation at room temperature
before re-dispersing in hexane, which was followed by a 500 *C heat-treatment for 20 hrs
under 02 atmosphere.
structure (P4 2 /mnm).
Both electron and X-ray diffraction reveal the rutile crystal
Catalyst inks were prepared by horn sonication of oxides in 70
wt% 2-propanol (99.9+% Sigma-Aldrich) as solvent. 10 pL of catalyst ink was drop-cast
onto a glassy carbon electrode (0.196 cm 2 area, Pine Instrument) polished to a mirror
finish with 0.05 gm alumina slurry (Buehler). The catalyst thin-film was dried at ambient
conditions overnight. We find that the OER specific activity of rutile IrO 2 was insensitive
to the loading, which suggests full utilization of the catalyst layer in our electrode.
6
-- -
Raw data (10 mV/s scan)
- Average forward-backward
iR correction
I
E4
E
2
1.0
1.2
1.4
1.6
E(V vs. RHE)
Figure 5.2.
Ohmic and capacitive corrections of the as-measured OER activity of
example perovskite oxide catalyst. The as-measured OER activity of the Lao.5 Cao.5Co0 3.
6
thin-film composite catalyst ("raw", solid red line) is capacity-corrected by taking an
average
of forward and backward
(positive
and negative-going)
scans.
The
capacity-corrected OER current (short-dash blue line) is then ohmically corrected with
80
the measured ionic resistance (~~45 0) to yield the final electrode OER activity (long-dash
green line).
Determination of eg filling of LaNio5 MnosO3 and LaCu05 Mn0 5 O3 .
LaNiO5 MnO.O
3 and
The eg fillings of
LaCu 5 Mn 5 O 3 are complicated by the fact that these compounds
undergo charge-disproportionation between Mn and Ni/Cu atoms. To estimate the eg
filling, we use the strategy from previous chapter, which relies on tracking the oxidation
states of each metal with X-ray Absorption Spectroscopy (XANES.)
From a known-
relationship where EO (maximum inflection point of the absorption edge) scales with
oxidation state 12 6 , we estimate the Mn electronic configuration to be t2g3 ego~3 for both
LaNio5 Mn.0 O3 and LaCu 5 MnO.O 3 , and Ni to be
t2g6 eg'~ 7
for LaNi0 5MnoO
5 0 3 and Cu to be
for LaNio5 MnoO 3. When we apply the eg filling from each B atom to the OER
volcano plot, we have found that the use of eg filling from Ni is most consistent with the
t2 e
7
observed activity for LaNi0 5 MnO.O
and Mn for LaCuO5 Mno.O 3 . Had we used the eg
3
filling of average B atoms, or Mn for LaNiO5MnO50 3 or Cu for LaCu0 5MnO50 3 , the eg
would have resulted in an underestimation of the OER activity. The consistency eg filling
selection with our volcano plot also leads us to propose that the active site for the OER is
the Ni atom for LaNiO5 Mno.O
in LaNi0 5 MnoO5 0
3
3 and
the Mn atom for LaCue 5 Mno.O 3 , where the Mn atom
has too little eg electron filling and Cu atoms in LaCue5 Mn. 0 O 3 has too
many eg electrons, rendering their catalytic properties inactive.
Determination of e, filling of Bao5Sro5 Co08Feo.20 3 4 compound. The eg filling for the
mixed A-site and B-site BSCF is highly complicated by the presence of four distinct
We
cations and a large concentration of oxygen vacancies inside the host structure.
measure the oxygen vacancy concentration with thermogravimetry (TGA), where we
found
6
TGA
to be -0.4.
While this
6
TGA
is in agreement with the literature value14 1 it
alone still cannot determine the oxidation state of Co and Fe atoms. Therefore we resort
to the use of XANES to estimate the Co oxidation state. Using CoO (Sigma-Aldrich) and
LaCoO 3 as model compounds, we estimate the oxidation state of the Co to be -2.8+
(Figure 5.3). Considering the previous works of Harvey et al.142 "
43
and Arnold et al."4,
whose XANES and X-ray Photoemission Spectroscopy studies have shown that the Co
oxidation state in BSCF is partially reduced below 3+ and that the Fe oxidation state is
81
partially oxidized above 3+, our estimated oxidation state is therefore not unexpected.
Taking into account that the surface Co state in BSCF has been reported to be in an
intermediate spin state", similar to the case of LaCoO 3, the electronic configuration of
the Co in BSCF is likely t2g eg.
7650
7700
7750
7800
7850
Energy (eV)
16
LaCoOs3
14
0
0.
BSCF
12
Coo
10L
1. 5
2.0
2.5
3.0
3.5
Oxidation state
Figure 5.3.
Co K-edge XANES of CoO, LaCoO 3 , and BSCF. (A) The Co K-edge
spectra. (B) The edge position (Eo, defined to be the energy at the highest first derivative
of the absorbance) of BSCF fits right in between CoO and LaCoO 3 . By using the known
metal oxidation vs. E0 relationship, we estimate the Co oxidation state to be -2.8+.
82
Ball-milling of Ba0 5 Sr0 5 Co0 8 Fe 0 .203 6 compound. The as-synthesized BSCF powder was
ball-milled with a Fritsch pulverisette 6 planetary mill.
A 5:2 mass ratio of 1 mm
zirconia milling beads to BSCF was prepared with 10 mL of N-methyl-2-pyrrolidinone
(NMP) solvent. The materials were milled in a zirconia crucible at 500 rpm for 8 hours,
with a 30-minute rest period every one-hour. The BSCF-NMP mixture was then dried in
a furnace at 195 *C in air, and then at 240 *C in vacuum overnight to evaporate off any
remaining solvent. The powder was then finely ground by a mortar and pestle before
electrochemical characterization.
We caution, however, that this ball-milling method
does not result in a sufficient miniaturization of the BSCF particle; a strategy based on
nanoscale oxide synthesis will likely produce more uniform, high-surface-area oxides.
We studied the ball-milling sample at both 0.25 mgox cm-2 disk, and 0.05 mgox cm
disk
to
ensure that there is no transport limitation in our measurement. The mass and specific
activities were found to be the same regardless of the loading.
5.3 OER Activity Descriptor Identification
As discussed in the previous chapter, our approach is based on a concept of molecular
orbital bonding. As the eg-orbital of surface transition-metal ions a-bonds with a
surface-anion adsorbate 4 5 ,
146,
its occupancy can greatly influence the binding of
oxygen-related intermediate species on B sites and thus the OER activity. We selected
the perovskite structure, Ai-xA'xByB' 1-y0 3, where A or A' is a rare-earth or alkaline-earth
metal, and B or B' is a transition-metal as model system to verify this hypothesis.
Representative OER currents of oxides collected in this study are shown in Figure 5.4,
which were obtained by using a methodology based on thin-film rotating-disk electrodes
with well-defined oxygen transport1 3 . The background contributions for OER from a
carbon additive (acetylene black, AB) and Nafion* in the oxide films and glassy carbon
electrode (GCE) are negligible below ~1.65 V vs. Reversible Hydrogen Electrode (RHE),
as shown in Figure la. To separate the surface area effect from the intrinsic catalytic
activity, we analyzed the surface-area-normalized kinetic current density (referred to as
specific activity, is) as a function of voltage vs. RHE to assess the intrinsic OER activities
of the perovskite oxides. Our measurements consistently yield a comparable or higher
specific OER activity of oxides than those found in previous reports 35 , 147-150 (Table 5.2).
83
This can be attributed to a higher catalyst utilization in thin-films (<10 cm 20x/cm
GCE
13, 96
( 104 cm
2
2
isk) on
than traditional porous electrodes having a very high internal surface area
x/cm
2
isk)
used in earlier work, thus yielding more accurate determination of
the specific activity of powder oxide catalysts.
2
E
E 1
0
1.2
1.3
1.4
1.5
1.6
1.7
1.8
Eapplied (V vs. RHE)
Figure 5.4. Example OER currents of Lai.xCaxCoO 3 and LaCoO 3 thin-films on GCE in
0 2 -saturated 0.1 M KOH at 10 mV s-
scan rate at 1600 r.p.m, which were
capacitance-corrected by taking an average of the positive and negative scans. The
contributions from AB and binder (Nafion*) and GCE are shown for comparison.
84
Compounds
Electrode i
Loading
Mass i
Specific i
(reference)
(mA cm.2 disk)
(mg cm-2isk)
(A g-1 )
(ptA cm-20 )
LaNiO 3
~20-2,000
n.a.
n.a.
(Bockris & Otagawas 2 )
(extrapolated)
-2-40
n.a
n.a
-13-50
~2-10
n.a
n.a
3-30
~0.1-1
n.a
n.a
0.2-1
~3-10
n.a
n.a
n.a.
~0.1-2.5
n.a
n.a
n.a.
~0.2-0.6
n.a
n.a
n.a.
~3.5
0.25
-14
~370
~0.5-3
0.25
-2-15
-250-800
~0.2-4
0.25
-0.8-16
~90-1,500
~0.03-0.4
0.25
~0.12-1.6
La xSrxCoO 3
30-360
(extrapolated)
(Bockris & Otagawas 2 )
La xSrxFeO 3
(Bockris & Otagawa8 2 )
La xSrxMnO 3
(Bockris & Otagawas 2 )
La xSrxCoO 3
(Matsumoto et al.'8 )
La1 xSrxFeO 3
(Wattiaux et al."')
La1 xSrxMnO 3
(Matsumoto & Sato"")
LaNiO3
(this work)
La xCaxCoO 3
(this work)
La xCaxFeO 3
(this work)
La, CaMnO 3
~20-100
(this work)
Table 5.2.
Benchmark OER activities of various state-of-the-art perovskite oxide
catalysts. All the OER activities were ohmic-corrected and normalized to the actual
surface area at 0.4 V overpotential in basic electrolyte (>13 pH).
85
Our result shows that the specific OER activities of the perovskite oxides exhibit a
volcano shape as a function of the eg-filling of surface B cations (Figure 5.5), where the
activities are compared in terms of the overpotential required to provide a specific current
of 50 pA/cm 20 x. In this analysis, the eg filling of B ions was estimated from a
combination of literature data, thermogravimetry and X-ray adsorption results15 2 ; partial
substitution on the A or B sites resulting in oxygen vacancy or B-ion valency changes
were taken into consideration. The assignment of the eg-orbital occupancy for oxides
with two different B ions such as LaMno. 5 CuO.50
3
was based on the more active surface
transition-metal cation under an assumption that the less active cation contributes
negligibly. Using a Tafel slope of ca. 60 mV/decade, the eg occupancy correlates with
the specific OER activities of oxides across four orders of magnitude, over a voltage span
of 0.3 V, an observation that allows identification of eg-filling of surface transition-metal
ions as a universal activity descriptor for these oxides.
It should be noted that the
eg-filling descriptor proposed in this study is fundamentally different from the number of
3d-electrons (both eg and t2g electrons) of bulk B cation proposed by Otagawa and
Bockris35 or the conduction band described by Matsumoto148 . The filling of the surface
transition-metal antibonding states of eg-symmetry parentage proposed by us is more
appropriate for catalysis since the a-bonding eg orbital has a stronger overlap with the
oxygen-related adsorbate than the a-bonding t2g orbital and thus can more directly
promote electron transfer between surface cation and adsorbed reaction intermediates.
Moreover, lowering of the site symmetry at the surface can induce orbital localization
and modify the surface electronic state to differ from that of the bulk
112,
and the usage of
the surface eg filling as opposed to the bulk value is crucial to prescribe accurately the
catalytic trend for ORR on oxides51 .
86
1.4
@/ = 50 pA cm
LaNlU
A Ba,,Sr.,CooffFe 2O3
LaCoO3
LaO5Cao MnO
0
3
y
1.6
>an
Li4 1.7
La0 .,Ca.,Co03
LaMn,,NIe0O
La-SCaxFeO 3
LaMn.Cu, 505
LaMnO3+,
LaMnO
3
LaCrOs
1.8 -
0.0
-
0.5
-
1.0
-
1.5
-
2.0
2.5
e electron
Figure 5.5.
The relation between the OER catalytic activity, defined by the
overpotentials at 50 pA cm-x of OER current, and the occupancy of the eg-symmetry
electron of surface B transition metal. Data symbols vary with type of B ions (Cr, red;,
Mn, orange; Fe, beige; Co, green; Ni, blue; mixed compounds, purple), where x = 0,
0.25, and 0.5 for Fe.
Error bars represent standard deviations of at least three
independent measurements. The dash volcano lines are shown for guidance only.
5.4 Proposed OER mechanism
We explain the OER activity trend in Figure 5.5 as follows. Considering the ab initio
finding of Rossmeisl et al. 1, whose work suggests a universal dependence of the OER
activity on the oxygen binding strength, we propose that the eg filling of surface
transition-metal cations can greatly influence the binding of OER intermediates to the
oxide surface and thus the OER activity. For the right branch of the volcano, the ratedetermining step (RDS) is the formation of the 0-0 bond in OOH adsorbate on B ions
(B(m+l)+02- + OH~ -+ B(m-OOH~ + e~) while for the left branch of the volcano, the
deprotonation of the oxy-hydroxide group to form peroxide ions (B(m)-OOH~ + OH- -+
B"'*-02 2- + H 20 + e~) is rate-limiting (Scheme 5.1)135. Having eg close to unity can
allow the oxide catalyst to facilitate these RDSs efficiently and thus lead to the highest
OER activity. Our molecular orbital model is further supported by the observation that
87
the OER activity can be enhanced with increasing covalent mixing between B and 0
atoms of oxides at a constant eg filling (Figure 5.6). Greater B-O covalency, where the
active redox couple of B ions is becoming pinned at the top of the
O-2p level,
promotes
the charge transfer between surface cations and adsorbates such as 022- and 02. in the
RDSs of OER, which can result in higher OER activity.
1.50
@ 50 pA cm2
1.55
LaNiO 3
iU
1.60
LaCoO3
1.65
LaMnO 3
I.IU
0.2
0.4
e,=
0.6
0.8
Abs.I(hole +1/4hole )
eg
t2g
Figure 5.6. The role of B-O covalency on the OER activity of perovskite oxides. The
potentials at 50 pA/cm2ox as a function of the B-0 covalency, which is estimated by the
normalized soft X-ray absorbance at eg filling = 1. Error bars represent standard
deviations of at least 3 measurements. The B-0 covalency data was from chapter 2.
88
OH-
H2 0 + e-.
IM+
0-B-0
I
r
I(m+1)+
02-'
0
2
2
+
eOH-
4
'002-
OER
RDS
0-B-C
OH-
OH-
OOH-
I(m+1)
0-B-C
3
H2 0+ e-
0-BC
OHOH-
OH-
I M+
0 2 + e-
O-Bm_.H20 + e-.
1
02-'
002-
ORR
RDS
(mn+1)+
O-B-O
OH-
OH-
4
2
e~
OOH-
|(m+1)+
0-B-C
3
IM+
0-B-0
H20 + e-
OH-
Scheme 5.1. Proposed OER mechanism on perovskite transition-metal oxide catalysts.
Both (top) the OER and (bottom) the ORR proceeds via 4 electron transfer steps. In the
OER case, the RDS are the 0-0 bond formation (2 in OER) and the proton extraction of
the oxy-hydroxide group (3 in OER).
For the ORR, the RDS are either the surface
hydroxide displacement (4 in OER) or the surface hydroxide regeneration (1 in OER).
89
5.5 Application of the Activity Descriptor
The OER design principle shown in Figure 5.5 was applied to obtain a highly active OER
catalyst, BSCF, which was best known for its high performance as cathode in solid oxide
fuel cells 5 3 and oxygen permeation membranes.
BSCF are in the intermediate spin state 4 2'
143
Because the surface Co cations of
the electronic configuration of Co cation
(estimated oxidation state of~2.8, see Section 5.2, Experimental Procedure) in BSCF can
be assigned as t2g 5eg~". We found that the BSCF had the highest OER activity among all
oxides studied as predicted by the eg activity descriptor (Figure 5.5). To ensure that the
observed oxidation current is from oxygen evolution, we applied a rotating-ring-disk
electrode measurement to detect 02 gas generated from the catalyst by electrochemically
reducing the generated gas at the Pt ring electrode that was held at the ORR potential of
0.4 V vs. RHE (Figure 5.7).
The reduction current detected on the ring electrode
suggests that the oxidation current from BSCF can be accounted for fully by the OER. In
comparison to the state-of-the-art catalyst, the intrinsic OER activity of BSCF is two
orders of magnitude higher than that of IrO2 nanoparticles (davg ~6 nm) in Figure 5.8.
The exceptional OER activity of BSCF was further confirmed by galvanostatic
measurements, where no degradation in the OER activity of BSCF was observed in this
study (Figure 5.9). Once the sizes of BSCF particles were reduced to sub-micron range
via ball-milling, the mass activity of ball-milled BSCF (BM-BSCF) was found to
compare favorably with IrO2, and the overpotential of BM-BSCF was shown to be nearly
identical to that of Ir0 2 at a mass activity of 10 A g-,x (Figure 5.9).
90
20
0
-20
00M%
41-40
C
, .,rIM
-60
-8 0 t
0
--.
25
50
75
I
100
Time (second)
Figure 5.7.
Evidence of 02 generated from BSCF using rotating ring disk electrode
measurements. The 02 gas generated from BSCF on GCE disk (OER current given as
idisk)
is reduced at the Pt ring at a constant potential of 0.4 V vs. RHE.
1.65
1.60
1.55
LU
1.50
1.45 "
1E-3
0.01
0.1
1
10
i (mA cm-X
Figure 5.8. The OER specific activity of BSCF (square) and IrO 2 (circle) as extracted
from the polarization.
Error bars represent standard deviations of at least three
independent measurements.
91
1.56
,
.
,
.
*
BSCF (0.2 m2 g-1)
1.54
>
BM-BSCF (3.9 m2 g 1 )
1.52 .IrO
2
(71 m2 g1 )
4W
1.50
-
@i =1OAg 1
.
--
50
0
100
I ox
150
Time (second)
Figure 5.9. Galvanostatic (constant current) measurement of mass activities of BSCF,
ball-milled BSCF (BM-BSCF), and IrO2.
5.6 Chapter 5 Conclusion
This work reports a model for the OER electrocatalytic activity of oxides in alkaline
solution using the catalytic trend observed in the case of the ORR. The principle was
used to discover an oxide OER catalyst that is competitive against state-of-the-art
precious-metal containing electrocatalyst. Considering the OER volcano trend of oxides
(peak at eg -1.2) with that of ORR (peak at eg -0.8) reported previously 51, and the fact
that the ORR activity of BSCF is not among the highest 52 , we arrive at an important
implication that no simple oxide can be optimal for both the ORR and the OER. Our
work suggests the importance of developing a transition-metal oxide having a
surface-cation eg occupancy close to unity and high B-O covalence as a promising
strategy to create bi-functional oxide catalysts and electrodes for the development of
efficient, rechargeable metal-air batteries, regenerative fuel cells, and other rechargeable
air-based energy storage devices.
92
Chapter 6 - Influence of Cation Species
6.1 Introduction
Understanding electrochemical splitting and forming of oxygen bond via the ORR and
the OER is central to many sustainable energy storage and conversion technologies2,4,7,
17, 154.
Occurring at the solid-liquid interface, both ORR and OER are multi-electron
transfer reactions that have been intensively studied as model prototypical multiintermediate electrochemical reactions.
In the past decade, fundamental works by
Rossmeisl and Norskov have applied ab initio calculations, in combination with
Bronsted-Evans-Polanyi Relations to describe the ORR and the OER mechanisms
20
,13 5 .
Using this strategy, they have developed the activity descriptor approach, which has led
to discoveries of novel catalytic surfaces for the ORR2 1, 155 . This Rossmeisl-Norskov
model, however, assumes that the ions in the electrolyte do not specifically influence the
electrocatalytic behavior of the ORR/OER surfaces
56
. This is in contrast to the recent
findings by Stamenkovic and Markovic, whose works report that a presence of a small
concentration of Li+ in K*-containing alkaline electrolyte can specifically influence the
ORR electrokinetics, depending on the nature of the catalytic surfaces
157, 158
In this
chapter, we seek to understand the effect of Li* ion on the ORR/OER electrocatalysis,
especially on transition metal oxide surfaces, which we have previously reported as
promising candidates for alkaline ORR and OER electrocatalysts5 1 ,
52, 159.
As the
presence of the cationic species such as Li+ is unavoidable in aqueous Li-air and other
metal-air batteries 4 1 ,
137,
understanding the role of the cation on the ORR/OER
electrokinetics on transition metal oxides is critical to developing a more efficient, costeffective aqueous metal-air devices.
The model for ion-surface interaction originates from the early works on the ORR
electrokinetics of Pt(l 11) in strongly absorbing anion-containing electrolytes. In H2 SO 4 ,
for example, the sulfate anion was found to interact strongly with Pt( 111) surface,
consequently poisoning and rendering the surface ineffective for catalyzing the ORR 160
This effect has been observed for phosphate anion, which also interacts with Pt(l 11)
93
surface in a similar fashion' 61 . Consequently, the ORR of Pt(1 11) in weakly absorbing
electrolyte such as HClO 4 displays faster kinetics than strongly absorbing H 2 SO 4 and
H3 PO 4 due to less surface poisoning and thus more available active site1 62 .
The
perspective on the cation-surface interaction, on the other hand, has received substantially
less attention. Due to the deep energetic levels of the cationic species, the cation's role in
electrocatalysis is believed to serve primarily as ions to neutralize surface in the double
layer space charge region. However, recent report has shown that the effect of these
interface cations on electrocatalysis are non-negligible and can significantly influence the
catalytic rate by more than an order of magnitude for Pt surfaces157 . The explanation thus
far concluded the stronger non-covalent interaction between surface Pt oxide and smaller
Li+ than bigger K* to be the main factor behind the activity difference. Strmcnik and coworkers thereby proposed that smaller Li+ can interact more strongly with the Pt oxide
and consequently has a longer resident time at the surface than larger K+. While there,
Li+ also has a more "rigid" hydration layer than larger K+ and thus, in combination with
its long surface resident time, can interfere with the availability of the nearby active site.
Recently, the cation effect has shown to prevail on Au surfaces, where the presence of the
oxide species on Au was proposed to underpin the observed cation influence on the
ORR158'163.
While previous studies of cation influences on Pt and Au are the first demonstrations of
the role of the cation on the ORR electrokinetics, the proposed mechanism of cationoxide influences is convoluted by the presence of both metallic and oxide surfaces at the
ORR turnover condition.
Consequently the mechanistic investigation of cation's
influence on Pt and Au can be challenging to interpret, as the possibility of the selective
cation interaction with either phase cannot be excluded. In this aspect, transition metal
oxide catalyst offers an all-inclusive oxide-phase surface to build the understanding of the
cation effect from. To further demonstrate the nature of the cation influence and its
interaction with the oxide species, we examine the mechanism of the cation influence on
the ORR kinetics of LaMnO 3+6 and the OER kinetics of rutile IrO 2 (r-IrO 2 ) and and
Bao. 5 Sro.5 Coo.sFeo.203-6 (BSCF) catalysts. These compounds were selected as they were
recently reported to be among the most active ORR/OER electrocatalytic compounds
among the class of transition metal oxide materials; LaMnO 3+6 was found to be one of the
94
most active ORR compounds5 1'164,165
while r-IrO 2 has long enjoyed its long reputation
as the OER benchmark catalyst'13 5, 140
BSCF was recently reported to have an intrinsic
OER activity even higher than r-Ir025 2 . In this work, we examine the influence of the
cation on these compounds' kinetics behaviors, from which the mechanism for cation
effect will be postulated.
6.2 Experimental Procedure
Transition metal oxide synthesis. The Perovskites, LaMnO 3+6 and BSCF, and the rutile rIrO2 were synthesized as reported elsewhere
52, 159, 166.
In the LaMnO 3+8 synthesis, rare
earth nitrate and manganese nitrate (99.98% Alfa Aesar) at the respective stoichiometric
ratio were mixed in Milli-Q water (18 MQ cm) at metal concentrations on the order of
0.1 M. This solution was then titrated with 1.2 M tetramethylammonium hydroxide (Alfa
Aesar) until a solid precipitate was formed. The precipitate was filtered, collected, and
dried at 200 'C for 12 hours, and then subjected to heat-treatment at 1000 'C under argon
(Airgas 99.999%) for 12 hours. The oxide was subsequently heat-treated at 800 'C under
dry air (Airgas) to obtain oxygen over-stoichiometric phase.
In the BSCF synthesis,
alkaline earth and transition-metal nitrates (99.998+% Sigma-Aldrich) were prepared at
0.2 M concentration in Milli-Q water, to which glycine (>99% Sigma-Aldrich) was added
at 0.1 M concentration. The mixture was heated until vigorously evaporating, at which
point, sparks are emitted as a result of combustion; it was then calcined at 11 00 C under
air atmosphere for 24 hours.
In the r-IrO2 synthesis, a solution of 0.6 mmol of
IrCl 4 xH 2 O, 10 mL of tetralin and 10 mL of oleylamine was prepared at room
temperature (20 C) in a four-neck flask. The mixture was heated to 60 'C under Ar
atmosphere (Airgas 99.999%) and kept at that temperature for 20 min to dissolve the
precursor homogeneously into the solution with continuous stirring. 2.4 mmol of TBABH
in 2 mL of chloroform was quickly injected into the above solution and left for 20 min.
The reaction mixture was further heated up to 200 'C and kept at that temperature for 19
hrs. The solution was cooled down, and Ir NPs were precipitated by ethanol addition and
collected by centrifugation. The product was re-dispersed in hexane and then annealed in
500 'C for 20 hrs in pure 02 (Airgas 99.999%).
X-ray Diffraction revealed all the
studied oxide materials to be single-phase.
95
Oxide electrodepreparation. The procedure for preparing oxide ORR/OER electrode is
described in details in previous chapter' ,166.
Briefly, the perovskites were mixed with
acetylene black carbon (AB) at a 5/1 mass ratio of oxide/carbon. Catalyst inks were
prepared by horn sonication (Bransonic 3510, Branson) of the oxide, AB, and an
appropriate amount of the Nat-exchanged Nafion* solution with tetrahydrofuran (THF,
99.9+% Sigma-Aldrich), yielding inks with final concentrations of 5 mgoxide mlink-1,
1 mgAB mlink,
and I
mgNafion mlink-.
Next, 10 pl of catalyst ink were drop-cast onto a
glassy carbon (GC) electrode (0.196 cm 2 area, Pine, USA) polished to a mirror finish
with 0.05 mm alumina slurry (Buehler).
composition of 250 ptgoxide
IrO 2 ,
cm-2disk,
The perovskite electrode had a final
50 pgAB cm
disk,
and 50 pgNafion
CM-2 disk-
For the r-
the electrode was prepared from an ink containing r-IrO 2 solubilized in 70% 2-
propanol (Sigma, 99.9+% Sigma-Aldrich). The rutile electrode had a final composition
of 50 pgoxide cm-2 disk.
Platinum electrodepreparation. Glassy carbon disk electrodes coated with a thin film of
high surface area carbon-supported platinum nanoparticle ("Pt/C") catalyst (46 %wt. Pt,
Tanaka Kinzoku, Japan) were prepared as described elsewhere ' 151. Briefly, 15 pl of an
aqueous Pt/C ink of a composition of 0.3 mgcatalyst ml- and 0.08 mgNafion mE 1 was dropcast onto a GC disk elecktrode and allowed to dry overnight, yielding a Pt loading of
54 gpt cm-2 disk. and 3 pgNafion cm-2 disk.
The Pt/C surface area was estimated from the
electrochemical H2 underpotential deposition and found to be consistent with the
electrochemical surface area reported elsewhere 0 .
Electrochemicalcharacterization. All electrochemical measurements were conducted in
a three-electrode glass cell (Pine Instrument) and using a rotator (Pine) to which working
electrode in a thin-film rotating disk electrode (RDE) configuration was attached; the
potential was controlled using a VoltaLab PST050 potentiostat. The 0.1 M concentration
of alkaline electrolyte was prepared from Milli-Q water (18 MW cm) and either LiOH,
NaOH, or KOH pellets (99.99% purity, Sigma-Aldrich). All measurements were
conducted at 10 mV s-1 in either N2 or 02 (ultra-high purity grade, Airgas) at room
temperature. A saturated calomel electrode (Pine Instrument) was used as a reference.
Error bars represent standard deviations from at least three independent repeat
96
measurements. The mass transport correction was performed for the ORR using the wellknown RDE correction (Levich equation). The OER activity was analyzed without masstransport correction.
6.3 The Influence of Li* Cation on the ORR
A typical ORR result of Pt/C RDE in 0.1 M KOH, NaOH, and LiOH at various rotation
rates is shown in Figure 6.1.
The electrocatalytic activity of Pt/C was highest in K*
cation containing electrolyte, followed by Na*, and Li+, respectively, in agreement with
the previous work''
158
Interestingly, the relative difference between the specific
activities in each electrolyte we observed (-
order of magnitude) was not as high as the
difference reported by Strmcnik et al (-2 order of magnitude' 5 7 ).
We attribute this
observation to the size difference between the two Pt/C studied with our smaller 2 nm
(vs. 5 nm) size having less sensitivity to the cation effect, which should be expected as
different Pt planes have different sensitivities to the cation' 58 and that smaller particle
forms more under-coordinated facets
167
. We note that the presence of Nafion* ionomic
binder in our samples also influenced the catalytic difference as a result of an additional
interaction of the sulfone group with the cation on the platinum surface' 6 8, 169 (Figure
6.2). The ionomer influence, however, is at a smaller factor and is not enough explain the
difference between our result and the literature value.
To measure the cation influence on the transition metal oxide surfaces, we utilize the
thin-film RDE technique to quantify the ORR activity of model transition metal oxide
surface. Figure 6.3 show exemplary ORR currents of LaMnO 3+6 in 0.1 M LiOH, 0.1 M
NaOH, 0.1 M KOH.
These ORR currents have been background-corrected and are
shown at a scan rate of 10 mV/s as examples. The onset for the ORR activity in the K*
electrolyte was found near 1 V vs. RHE with lower potentials found for smaller cations
such as Na* and Li+. The trend that the ORR activity decreases with decreasing cation
size (Li+ < Na* < K) is similar to the case found for the ORR kinetic activities of Pt/C in
the same set of electrolytes (Figure 6.1) and is also similar to the earlier report of the
ORR trend for Pt(1 11), poly-Pt, and Pt/C157 . The fact that both perovskite and Pt/C carry
the same cation trend is peculiar when considering that the non-covalent cation effect
97
mechanism proposed by Strmcnik et al. is based on having the hydration layer of the
cation that is non-covalently bound on the surface oxide species blocking nearby open,
active metal sites (Scheme 6.1). This mechanism presents a great deal of difficult to
understand the ORR trend in the context of the perovskite oxide catalyst, where the
surface is uniformly terminated with either a hydroxide or an oxide species as active
sites 35 ,
82, 170,
and therefore does not have any open "metal" site to be blocked.
Furthermore, when considering the fact that the earlier cation mechanism predicts a
strong non-covalent cation "poisoning" of the surface oxide/hydroxide species, the fact
that the ORR current on an oxide-terminated surface can be observed in strongly
absorbing cation electrolyte such as LiOH forces us to re-consider how the cation
influences the ORR electrokinetics on the oxide surfaces.
'I I
'
1 '
'
Pt/C (10 gp CM
-O.IM KOH
-1
-
E-2
0.1M NaOH
0.1M LiOH
%nw
W
-5
0.0
0.2
0.4
0.6
0.8
1.0
Applied potential (V vs. RHE)
Figure 6.1. The ORR activity of thin-film Pt/C RDE in the positive potential-going scan
direction at 10 mV/s in various alkaline electrolytes at 900 r.p.m.
98
1.00
,
,--w
.o
... .
.O .w
-e-o.0. 1M KOH w. Nafion
0.95
1
0.90
-4-.1
t
0.85
u
0.80
-e-O.1MO. Nafio
n~
-M&40.1M KOH w.o. Nafion
0.75
0.1
1
iOR
(mA cm2)
Figure 6.2. The influence of Nafion* on the ORR specific activity of thin-film Pt/C
RDE in KOH and LiOH.
6.3 The Influence of Li+ Cation on the OER
To better understand how the cation influences the surface oxide species during the
turnover condition for oxygen electrocatalysis, we conduct the OER measurement of two
benchmark OER oxide catalysts in the same set of alkaline electrolytes as the ORR.
Figure 4 shows typical findings for the OER kinetic currents of r-IrO2 and BSCF in the
same set of electrolytes (i.e. 0.1 M LiOH, 0.1 M NaOH, and 0.1 M KOH). Both the rIrO 2 and BSCF have onset potentials of the OER activities at about 1.4 V vs. RHE, with
higher onset potentials with smaller cations. Interestingly, the same cation trend, where
the OER activity decreases with decreasing cation size (Lie < Na* < K*) for both rutile
and perovskite surfaces is observed, which indicates that the trends for the cation on
oxygen electrocatalysis are the same for both r-IrO 2 and BSCF surfaces. Because the
rate-limiting species on the transition metal oxide surfaces during the OER turnover are
likely a hydroxo/oxyhydroxo/oxo species5 2,
135, 171,
the fact that both surfaces have the
same trend in the cation effect is not surprising. However, when considering that the
non-covalent cation effect should poison all of the oxide surfaces and block them from
being active, it is interesting to observe that, like in the case of the ORR, the OER current
in LiOH can be observed at all (in fact, very close to that measured in KOH).
99
0
000
,LaMn03+,
-1
E
(0.25 mg cm , )
E
@ 1600rpm
-2
-
-0.M KOH
%W.
0.IM NaOH
/F-
-0.1M
0.75
0.80
0.85
0.90
LIOH
0.95
Applied potential (V vs. RHE)
0.96
0.92
0.88
Li
0.84
0.80
0.76
0.1
1
RDE
CMd.sk)
Figure 6.3. The ORR activity of thin-film LaMnO 3+6(supported on acetylene black and
neutralized Nafion*) RDE in the negative potential-going scan direction at 10 mV/s in
various alkaline electrolytes at 1600 r.p.m.
100
Block
Scheme 6.1. Graphic illustration of Strmcnik and co-workers' model for cation's noncovalent interaction, at which the cation species interacts with
OHads
group to interfere
the open active Pt site by "blocking" mechanism 5 , m
6.4 Proposed Cation Mechanism
To explain the observed cation effect on the transition metal oxide surfaces, we take hints
from the observation that the difference between the ORR kinetic currents is about an
order of magnitude, whereas the difference between the OER kinetic currents is
significantly less. This observation reveals the cation effect to be potential-dependent;
there is less cation on the surface at the OER potential than the ORR potential.
Combining this with the fact that the cation did not effectively "poison" the whole
transition metal oxide surfaces completely (as there is still measurable ORR/OER
activity), we propose that the cation influences the ORR/OER by ways of modifying the
hydration of the reaction intermediates (Scheme 6.2). As we have previously proposed
the competition between 0 22-/0H- species, and 0 2 /OOH- as rate-limiting for the ORR
and the OER, these negatively charged species draw in cation during the ORR/OER
turnover at the alkaline pH. The potential-dependent factor supports this fact as it has
been observed previously that the absorbing anion on the surface can be de-protonated at
higher potential (equivalent to a cation removal) 173 . This is likely a result of the charge
transfer between surface and the adsorbing species, which helps reducing the net charge
on the adsorbate.
Our proposed mechanism is similar to that proposed earlier by
Strmcnik et al., with one exception: the cation does not virtually "block" the ORR/OER
101
active site, but rather modify their hydration energetics and effectively the reaction
kinetics. We point out here that we cannot differentiate whether the perturbation of the
ORR/OER intermediates from the cation is a result of the cation-direct interaction
(Scheme 6.2 Top) or cation-hydration-mediated interaction (Scheme 6.2 Bottom).
Further spectroscopic studies using for example, sum frequency generation would be
essential to evaluate this proposed mechanism.
1.64
1.60
w
zi
U; 1.56
1.52
4U
1.48
0.01
0.1
'RDE
10
1
(mdsk)
1.64
-
1.60
-
U;
Ba0,Sr 0.,Co 0 Fe0.20
(0.25 mg cmk)
3
1.56
1.52
1.48
WU
1.44
LiOH
NaOH
-KOH
... I
I
. .
0.1
l
. . . . ... I
1
RDE
10
(mA cmk
Figure 6.4. The OER specific activity of thin-film rutile Ir02 and BSCF (the latter is
supported on acetylene black and neutralized Nafion*) RDEs.
102
Scheme 6.2: Proposed cation effect on the ORR/OER on transition metal oxide surfaces.
(Top) cation or (Bottom) its hydration interacts with the hydroxo/superoxo/peroxo
species and its energies during the turnover condition. The kinetics is affected by the
intermediate energy modifications. The B label represents active transition metal atom,
which we previously proposed as the active site for oxygen electrocatalysis on transition
metal oxide surfaces.
6.5 Chapter 6 Conclusion
The influences of cationic species on oxygen electrocatalytic activities on oxide surfaces
are studied in this chapter. We found that both the ORR and the OER activities increases
with increasing cation size (K* > Na* > Li+), following the Hofmeister series, with the
effect of the cation more pronounced in the ORR than the OER. We rationalize this
observation by proposing that the cation serves to modify the solvation energetics of
reaction intermediate, which in turns, affect the catalytic activities. At higher potential
(OER turnover condition), the electrode draws more negative charge away from the
103
adsorbate, which effectively reduces the need for hydration and hence less pronounced
effect in the OER activity.
Our proposed mechanism is congruent with the proposed
competition between 0 22-/0H- species, and 0 2 -/00H~ as rate-limiting for the ORR and
the OER on transition metal oxides. The complexity of the solid-liquid layer however
will need further experimentation to resolve. Our finding demonstrates the importance
and the opportunity to tailor the electrified interface for catalysis and energy conversion
applications.
104
Chapter 7 - Summary and Outlook
7.1 Thesis Summary
We present a model to understand the catalytic trend of transition metal oxide catalysts
for oxygen electrochemical reactions. To develop the model, we conducted a systematic
examination of the electrocatalytic activity of model oxide compounds using thin-film
oxide rotating disk electrode methodology, which de-convolutes the oxygen electrokinetics from the transport and the ohmic losses, and thereby allow quantitative
comparison between different oxide catalysts. We found that the oxygen electrocatalytic
activities (both ORR and OER) for perovskite transition metal oxide catalysts correlate to
the a* orbital ("eg") occupation. We propose that this correlation is a consequence of the
importance of the eg-symmetry parentage electron on the oxygen binding strength. From
our finding, we hypothesize that the rate-limiting step of the oxygen electrochemical
reactions, is the competition between 0 2 2 /OH displacement and OH~ regeneration for the
ORR, and for the OER, is the competition between the formation of the 0-0 bond and
the deprotonation of the oxy-hydroxide species.
This hypothesis is in qualitative
agreement with the model previously proposed for precious metal and precious metal
oxide surfaces - that both the ORR and the OER can be described by the same activity
dmntaeta
135, 138
. We also demonstrate that
descriptor despite following different rate-limiting steps3
transition metal-oxygen covalency carries a significant influence on the oxygen
electrocatalytic activity, which can serve as a secondary activity descriptor. To explain
this observation, we postulate that transition metal-oxygen covalency promotes charge
transferability between transition metal and oxygen, which allows the surface to perform
redox on the adsorbed oxygen.
We further discuss how covalency evolves in the
perovskite oxide using molecular orbital theory, which hints at us to further design nextgeneration perovskite oxide with stronger extent of transition metal-oxygen covalency.
Lastly, we demonstrate the application of our catalytic trend by finding a novel OER
catalyst, which has catalytic activity that rivals state-of-the-art, precious-metal-containing
catalyst.
105
7.2 Outlook
This thesis shows that it is possible to develop a descriptor framework to understand
oxygen electrocatalysis on non-precious-metal surfaces. Our proposed "eg" mechanism
relies heavily on (1) ex-situ characterization of transition metal oxide catalysts, (2) the
ability to distinguish surface state from bulk, and (3) assumption of the octahedral
geometry during turnover condition. These are big assumptions. Therefore, the ability to
do in situ spectroscopic capabilities to evaluate the mode of surface-oxygen interaction
during the turnover condition, from which the role of eg-symmetry parentage electron can
be assessed in a more accurate manner, will substantially improve this work, and also our
understanding of the fundamental mechanism of the oxygen electrochemical reactions.
Infrared and Raman Spectroscopy on well-defined single crystal of transition metal
oxides in operando may be able to address this need. Raman Spectroscopy, a technique
highly sensitive to local symmetry of the vibrational mode, may help reveal, for example,
if BSCF undergoes change from cubic (first-order forbidding Raman) to tetragonal
during turnover condition, although differentiating surface from bulk contribution is not
straightforward.
Similarly Infrared Spectroscopy, which probes molecule's vibrational
behavior, may be able to resolve 0-0 bond during 0-0 bond formation, where the
interaction with surface causes 0-0 bond to develop a non-zero dipole. Information from
these experiments will prove decisive at determining the validity of many electrocatalysis
models, including ours.
An interesting question for the proposed eg model is whether such a correlation would
extend to a molecular transition metal complex catalysis system. It has long been known
that enzymes are made of transition metal ions coordinated to various ligand
chemistries 174 .
In theory, the molecular orbital assumption of the perovskite oxide
system should directly translate to these molecular catalysts.
However, to our
knowledge, such correlation has not been done yet.
The proposed eg model also opens question about whether such a descriptor can be
appropriately applied to design a direct-solar water-splitting system. One approach can
be to directly apply BSCF coating on buried p-n junction, where photo-generated carrier
106
within the buried junction develops a photovoltage to assist with water splitting (Figure
7.1). Another approach would be to integrate the eg catalytic functionality directly in an
oxide semiconductor electrode (Figure 7.2). This approach, albeit more simple, has many
open questions. For example, what should the role of eg be? Two possibilities are the
use of eg band as a Fermi level pinner, where upon illumination of the bulk oxide
semiconductor, the photo-generated hole can be swept to the surface eg state (Figure 7.2,
Top), and the use of eg band as a photo-absorbing band itself (Figure 7.2, Bottom).
Future work in this area will plant the seed for the next-generation design of a device that
can directly capture solar energy and store it in term of chemical fuels.
Catalyst
Photo-absorber
Redox
A
n-Si
p-Si
eg
0 2/H 2 0
Density of state
Figure 7.1.
Schematic
drawing
of an
integrated
catalyst-p-n junction
photoelectrochemical water splitting.
107
for
Oxide semiconductor
Redox
0 2/H 20
eg
Density of state
Redox
Oxide semiconductor
0 2 /H 2 0
eg
Density of state
Figure 6.2. Schematic drawings of a proposed catalytically active oxide semiconductorliquid junction for photoelectrochemical water splitting. The catalytically active eg state
can either be integrated as a Fermi-level pinner at the surface (Top), or as a valence band,
where hole can be generated upon excitation (Bottom).
108
Reference
1.
Lewis, N. S.; Nocera, D. G., Powering the planet: Chemical challenges in solar
energy utilization. Proc.Nat. Acad. Sci. 2006, 103, 15729-15735.
2.
Cook, T. R.; Dogutan, D. K.; Reece, S. Y.; Surendranath, Y.; Teets, T. S.; Nocera,
D. G., Solar Energy Supply and Storage for the Legacy and Non legacy Worlds. Chem.
Rev. 2010, 110, 6474-6502.
3.
Gray, H. B., Powering the planet with solar fuel. Nature Chem. 2009, 1, 7.
4.
Walter, M. G.; Warren, E. L.; McKone, J. R.; Boettcher, S. W.; Mi,
Q. X.;
Santori, E. A.; Lewis, N. S., Solar Water Splitting Cells. Chem. Rev. 2010, 110, 64466473.
5.
Kanan, M. W.; Nocera, D. G., In situ formation of an oxygen-evolving catalyst in
neutral water containing phosphate and Co2 . Science 2008, 321, 1072-1075.
6.
Reece, S. Y.; Hamel, J. A.; Sung, K.; Jarvi, T. D.; Esswein, A. J.; Pijpers, J. J. H.;
Nocera, D. G., Wireless Solar Water Splitting Using Silicon-Based Semiconductors and
Earth-Abundant Catalysts. Science 2011, 334, 645-648.
7.
Gasteiger, H. A.; Kocha, S. S.; Sompalli, B.; Wagner, F. T., Activity benchmarks
and requirements for Pt, Pt-alloy, and non-Pt oxygen reduction catalysts for PEMFCs.
Appl. Catal., B 2005, 56, 9-3 5.
8.
Gasteiger, H. A.; Markovic, N. M., Just a Dream-or Future Reality? Science 2009,
324, 48-49.
9.
Bruce, P. G.; Freunberger, S. A.; Hardwick, L. J.; Tarascon, J. M., Li-0 2 and Li-S
batteries with high energy storage. Nat Mater 2012, 11, 19-29.
10.
Dunn, B.; Kamath, H.; Tarascon, J. M., Electrical Energy Storage for the Grid: A
Battery of Choices. Science 2011, 334, 928-935.
11.
Stephens, I. E. L.; Bondarenko, A. S.; Gronbjerg, U.; Rossmeisl, J.; Chorkendorff,
I., Understanding the electrocatalysis of oxygen reduction on platinum and its alloys.
Energy Environ. Sci. 2012, Advance Article.
12.
Spendelow, J. S.; Wieckowski, A., Electrocatalysis of oxygen reduction and small
alcohol oxidation in alkaline media. Phys Chem Chem Phys 2007, 9, 2654-2675.
109
13.
Suntivich, J.; Gasteiger, H. A.; Yabuuchi, N.; Shao-Horn, Y., Electrocatalytic
measurement methodology of oxide catalysts using a thin-film rotating disk electrode. J.
Electrochem. Soc. 2010, 157, B1263-B1268
14.
Piana, M.; Catanorchi, S.; Gasteiger, H. A., Kinetics of Non-Platinum Group
Metal Catalysts for the Oxygen Reduction Reaction in Alkaline Medium. In Proton
Exchange Membrane Fuel Cells 8, Pts 1 and 2, Fuller, T.; Shinohara, K.; Ramani, V.;
Shirvanian, P.; Uchida, H.; Cleghorn, S.; Inaba, M.; Mitsushima, S.; Strasser, P.;
Nakagawa, H.; Gasteiger, H. A.; Zawodzinski, T.; Lamy, C., Eds. Electrochemical
Society Inc: Pennington, 2008; Vol. 16, pp 2045-2055.
15.
Marcinkoski, J.; James, B. D.; Kalinoski, J. A.; Podolski, W.; Benjamin, T.;
Kopasz, J., Manufacturing process assumptions used in fuel cell system cost analyses. J
PowerSources 2011, 196, 5282-5292.
16.
Rasten, E.; Hagen, G.; Tunold, R., Electrocatalysis in water electrolysis with solid
polymer electrolyte. Electrochim. Acta 2003, 48, 3945-3952.
17.
Lee, Y.; Suntivich, J.; May, K. J.; Perry, E. E.; Shao-Horn, Y., Synthesis and
activities of rutile IrO2 and RuO 2 nanoparticles for oxygen evolution reaction
electrocatalysis. J Phys. Chem. Lett. 2012, 3, 399-404.
18.
Stamenkovic, V. R.; Fowler, B.; Mun, B. S.; Wang, G. F.; Ross, P. N.; Lucas, C.
A.; Markovic, N. M., Improved oxygen reduction activity on Pt 3Ni(1 11) via increased
surface site availability. Science 2007, 315, 493-497.
19.
Stamenkovic, V. R.; Mun, B. S.; Mayrhofer, K. J. J.; Ross, P. N.; Markovic, N.
M.; Rossmeisl, J.; Greeley, J.; Norskov, J. K., Changing the activity of electrocatalysts
for oxygen reduction by tuning the surface electronic structure. Angew Chem Int Edit
2006, 45, 2897-2901.
20.
Norskov, J. K.; Rossmeisl, J.; Logadottir, A.; Lindqvist, L.; Kitchin, J. R.;
Bligaard, T.; Jonsson, H., Origin of the overpotential for oxygen reduction at a fuel-cell
cathode. J Phys. Chem. B 2004, 108, 17886-17892.
21.
Greeley, J.; Stephens, I. E. L.; Bondarenko, A. S.; Johansson, T. P.; Hansen, H.
A.; Jaramillo, T. F.; Rossmeisl, J.; Chorkendorff, I.; Norskov, J. K., Alloys of platinum
and early transition metals as oxygen reduction electrocatalysts. Nature Chem. 2009, 1,
552-556.
110
22.
Lima, F. H. B.; Zhang, J.; Shao, M. H.; Sasaki, K.; Vukmirovic, M. B.; Ticianelli,
E. A.; Adzic, R. R., Catalytic activity d-band center correlation for the 02 reduction
reaction on platinum in alkaline solutions. J. Phys. Chem. C 2007, 111, 404-410.
23.
Meadowcroft, D. B., Low-cost oxygen electrode material. Nature 1970, 226, 847-
848.
24.
Matsumoto, Y.; Yoneyama, H.; Tamura, H., Catalytic activity for electrochemical
reduction of oxygen of lanthanum nickel-oxide and related oxides. J. Electroanal.Chem.
1977, 79, 319-326.
25.
Lu, Y. C.; Gasteiger, H. A.; Parent, M. C.; Chiloyan, V.; Shao-Horn, Y., The
influence of catalysts on discharge and charge voltages of rechargeable Li-oxygen
batteries. Electrochem. Solid-State Lett. 2010, 13, A69-A72.
26.
Bashyam, R.; Zelenay, P., A class of non-precious metal composite catalysts for
fuel cells. Nature 2006, 443, 63-66.
27.
Lefevre, M.; Proietti, E.; Jaouen, F.; Dodelet, J. P., Iron-based catalysts with
improved oxygen reduction activity in polymer electrolyte fuel cells. Science 2009, 324,
71-74.
28.
Bednorz, J. G.; Muller, K. A., Possible high Tc superconductivity in the Ba-La-
Cu-0 system. Z Phys. B-Condens. Mat. 1986, 64, 189-193.
29.
Goodenough, J. B.; Zhou, J. S., Transport properties. In Localized to itinerant
electronic transition in perovskite oxides, Springer-Verlag Berlin: Berlin, 2001; Vol. 98,
pp 17-113.
30.
Wang, J.; Neaton, J. B.; Zheng, H.; Nagarajan, V.; Ogale, S. B.; Liu, B.;
Viehland, D.; Vaithyanathan, V.; Schlom, D. G.; Waghmare, U. V.; Spaldin, N. A.; Rabe,
K. M.; Wuttig, M.; Ramesh, R., Epitaxial BiFeO 3 multiferroic thin film heterostructures.
Science 2003, 299, 1719-1722.
31.
Haeni, J. H.; Irvin, P.; Chang, W.; Uecker, R.; Reiche, P.; Li, Y. L.; Choudhury,
S.; Tian, W.; Hawley, M. E.; Craigo, B.; Tagantsev, A. K.; Pan, X.
Chen, L.
Q.; Kirchoefer,
Q.; Streiffer,
S. K.;
S. W.; Levy, J.; Schlom, D. G., Room-temperature
ferroelectricity in strained SrTiO 3 . Nature 2004, 430, 758-761.
111
32.
Vonhelmolt, R.; Wecker, J.; Holzapfel, B.; Schultz, L.; Samwer, K., Giant
negative magnetoresistance in perovskitelike La 2/3Bal/3 MnOx ferromagnetic films. Phys.
Rev. Lett. 1993, 71, 2331-2333.
33.
Salamon, M. B.; Jaime, M., The physics of manganites: Structure and transport.
Rev. Mod Phys. 2001, 73, 583-628.
34.
Imada, M.; Fujimori, A.; Tokura, Y., Metal-insulator transitions. Rev. Mod Phys.
1998, 70, 1039-1263.
35.
Bockris, J. 0.; Otagawa, T., The Electrocatalysis of Oxygen Evolution on
Perovskites. J Electrochem. Soc. 1984, 131, 290-302.
36.
Tejuca, L. G.; Fierro, J. L. G.; Tascon, J. M. D., Structure and reactivity of
perovskite-type oxides. Adv. Catal. 1989, 36, 237-328.
37.
Stamenkovic, V. R.; Mun, B. S.; Arenz, M.; Mayrhofer, K. J. J.; Lucas, C. A.;
Wang, G. F.; Ross, P. N.; Markovic, N. M., Trends in electrocatalysis on extended and
nanoscale Pt-bimetallic alloy surfaces. Nat. Mater. 2007, 6, 241-247.
38.
Hammer, B.; Norskov, J. K., Theoretical surface science and catalysis -
Calculations and concepts. In Advances in Catalysis, Vol 45: Impact of Surface Science
on Catalysis, Gates, B. C.; Knozinger, H., Eds. Elsevier Academic Press Inc: San Diego,
2000; Vol. 45, pp 71-129.
39.
Ballhausen, C. J.; Gray, H. B., Molecular orbitaltheory: an introductorylecture
note and reprintvolume. W.A. Benjamin: 1965.
40.
Goodenough, J. B., Electronic and ionic transport properties and other physical
aspects of perovskites. Rep. Prog.Phys. 2004, 67, 1915-1993.
41.
Goodenough, J. B.; Kim, Y., Challenges for Rechargeable Li Batteries. Chem.
Mat. 2010, 22, 587-603.
42.
Marianetti, C. A.; Kotliar, G.; Ceder, G., Role of hybridization in NaxCoO 2 and
the effect of hydration. Phys. Rev. Lett. 2004, 92.
43.
Arima, T.; Tokura, Y.; Torrance, J. B., Variation of optical gaps in perovskite-
type 3d transition-metal oxides. Phys. Rev. B 1993, 48, 17006-17009.
44.
Maiti, K., Role of covalency in the ground-state properties of perovskite
ruthenates: A first-principles study using local spin density approximations. Phys. Rev. B
2006, 73.
112
45.
Van Aken, B. B.; Palstra, T. T. M.; Filippetti, A.; Spaldin, N. A., The origin of
ferroelectricity in magnetoelectric YMnO 3 . Nat. Mater. 2004, 3, 164-170.
46.
Chakhalian, J.; Freeland, J. W.; Habermeier, H. U.; Cristiani, G.; Khaliullin, G.;
van Veenendaal, M.; Keimer, B., Orbital reconstruction and covalent bonding at an oxide
interface. Science 2007, 318, 1114-1117.
47.
Chakhalian, J.; Rondinelli, J. M.; Liu, J.; Gray, B. A.; Kareev, M.; Moon, E. J.;
Prasai, N.; Cohn, J. L.; Varela, M.; Tung, I. C.; Bedzyk, M. J.; Altendorf, S. G.; Strigari,
F.; Dabrowski, B.; Tjeng, L. H.; Ryan, P. J.; Freeland, J. W., Asymmetric Orbital-Lattice
Interactions in Ultrathin Correlated Oxide Films. Phys. Rev. Lett. 2011, 107.
48.
Mannhart, J.; Schlom, D. G., Oxide Interfaces-An Opportunity for Electronics.
Science 2010, 327, 1607-1611.
49.
Takaobushi, J.; Ishikawa, M.; Ueda, S.; Ikenaga, E.; Kim, J. J.; Kobata, M.;
Takeda, Y.; Saitoh, Y.; Yabashi, M.; Nishino, Y.; Miwa, D.; Tamasaku, K.; Ishikawa, T.;
Satoh, I.; Tanaka, H.; Kobayashi, K.; Kawai, T., Electronic structures of Fe 3..xMxO4
(M=Mn,Zn) spinel oxide thin films investigated by X-ray photoemission spectroscopy
and X-ray magnetic circular dichroism. Phys. Rev. B 2007, 76.
50.
Shimakawa, Y.; Kubo, Y.; Hamada, N.; Jorgensen, J. D.; Hu, Z.; Short, S.;
Nohara, M.; Takagi, H., Crystal structure, magnetic and transport properties, and
electronic band structure of A2Mn 2O7 pyrochlores (A = Y, In, Lu, and TI). Phys. Rev. B
1999, 59, 1249-1254.
51.
Suntivich, J.; Gasteiger, H. A.; Yabuuchi, N.; Nakanishi, H.; Goodenough, J. B.;
Shao-Hom, Y., Design principles for oxygen reduction activity on perovskite oxide
catalysts for fuel cells and metal-air batteries. Nature Chem. 2011, 3, 546-550.
52.
Suntivich, J.; May, K. J.; Goodenough, J. B.; Gasteiger, H. A.; Shao-Horn, Y., A
perovskite oxide optimized for oxygen evolution catalysis from molecular orbital
principles. Science 2011, 334, 1383-1385.
53.
Fujii, T.; de Groot, F. M. F.; Sawatzky, G. A.; Voogt, F. C.; Hibma, T.; Okada,
K., In situ XPS analysis of various iron oxide films grown by N0 2-assisted molecularbeam epitaxy. Phys. Rev. B 1999, 59, 3195-3202.
113
54.
Guittet, M. J.; Crocombette, J. P.; Gautier-Soyer, M., Bonding and XPS chemical
shifts in ZrSiO 4 versus SiO 2 and ZrO2 : Charge transfer and electrostatic effects. Phys.
Rev. B 2001, 63, art. no.-125117.
55.
Kourkoutis, L. F.; Xin, H. L.; Higuchi, T.; Hotta, Y.; Lee, J. H.; Hikita, Y.;
Schlom, D. G.; Hwang, H. Y.; Muller, D. A., Atomic-resolution spectroscopic imaging of
oxide interfaces. Philos. Mag. 2010, 90, 4731-4749.
56.
Graetz, J.; Hightower, A.; Ahn, C. C.; Yazami, R.; Rez, P.; Fultz, B., Electronic
structure of chemically-delithiated LiCoO 2 studied by electron energy-loss spectrometry.
J Phys. Chem. B 2002, 106, 1286-1289.
57.
de Groot, F. M. F.; Grioni, M.; Fuggle, J. C.; Ghijsen, J.; Sawatzky, G. A.;
Petersen, H., Oxygen 1s X-ray-absorption edges of transition-metal oxides. Phys. Rev. B
1989, 40, 5715-5723.
58.
Tofield, B. C.; Fender, B. E. F., Covalency parameters for Cr3+, Fe3 * and Mn4 in
an oxide environment. J. Phys. Chem. Solids 1970, 31, 2741-&.
59.
Bocquet, A. E.; Mizokawa, T.; Saitoh, T.; Namatame, H.; Fujimori, A., Electronic
structure of 3d-transition-metal compounds by analysis of the 2p core-level
photoemission spectra. Phys. Rev. B 1992, 46, 3771-3784.
60.
Abbate, M.; Degroot, F. M. F.; Fuggle, J. C.; Fujimori, A.; Strebel, 0.; Lopez, F.;
Domke, M.; Kaindl, G.; Sawatzky, G. A.; Takano, M.; Takeda, Y.; Eisaki, H.; Uchida, S.,
Controlled-valence properties of La.xSrxFeO 3 and La-xSrxMnO 3 studied by soft-X-Ray
absorption-spectroscopy. Phys. Rev. B 1992, 46, 4511-4519.
61.
Abbate, M.; Goedkoop, J. B.; Degroot, F. M. F.; Grioni, M.; Fuggle, J. C.;
Hofmann, S.; Petersen, H.; Sacchi, M., Probing depth of surface X-ray absorption
spectroscopy measured in total electron-yield-mode. Surf Interface Anal. 1992, 18, 6569.
62.
Medarde, M.; Dallera, C.; Grioni, M.; Voigt, J.; Podlesnyak, A.; Pomjakushina,
E.; Conder, K.; Neisius, T.; Tjernberg, 0.; Barilo, S. N., Low-temperature spin-state
transition in LaCoO 3 investigated using resonant x-ray absorption at the Co K edge. Phys.
Rev. B 2006, 73, 054424.
63.
Goodenough, J. B., Covalency Criterion for Localized vs Collective Electrons in
Oxides with the Perovskite Structure. J Appl. Phys. 1966, 37, 1415-&.
114
64.
Duffy, J. A., Ionic-covalent character of metal and nonmetal oxides. J. Phys.
Chem. A 2006, 110, 13245-13248.
65.
Sarma, D. D.; Maiti, K.; Vescovo, E.; Carbone, C.; Eberhardt, W.; Rader, 0.;
Gudat, W., Investigation of hole-doped insulating Lai.xSrxCrO 3 by soft-x-ray absorption
spectroscopy. Phys. Rev. B 1996, 53, 13369-13373.
66.
Toulemonde, 0.; N'Guyen, N.; Studer, F.; Traverse, A., Spin state transition in
LaCoO 3 with temperature or strontium doping as seen by XAS. J. Solid State Chem.
2001, 158, 208-217.
67.
Saitoh, T.; Bocquet, A. E.; Mizokawa, T.; Namatame, H.; Fujimori, A.; Abbate,
M.; Takeda, Y.; Takano, M., Electronic structure of Laj.xSrxMnO3 studied by
photoemission and x-ray-absorption spectroscopy. Phys. Rev. B 1995, 51, 13942-13951.
68.
Filpi, A.; Boccia, M.; Gasteiger, H. A., Pt-Free Cathode Catalyst Performance in
H2 /0 2 Anion-Exchange Membrane Fuel Cells (AMFCs). In Proton Exchange Membrane
Fuel Cells 8, Pts 1 and 2, Fuller, T.; Shinohara, K.; Ramani, V.; Shirvanian, P.; Uchida,
H.; Cleghorn, S.; Inaba, M.; Mitsushima, S.; Strasser, P.; Nakagawa, H.; Gasteiger, H.
A.; Zawodzinski, T.; Lamy, C., Eds. Electrochemical Society Inc: Pennington, 2008; Vol.
16, pp 1835-1845.
69.
Neburchilov, V.; Wang, H. J.; Martin, J. J.; Qu, W., A review on air cathodes for
zinc-air fuel cells. J. Power Sources 2010, 195, 1271-1291.
70.
Paulus, U. A.; Schmidt, T. J.; Gasteiger, H. A.; Behm, R. J., Oxygen reduction on
a high-surface area Pt/Vulcan carbon catalyst: a thin-film rotating ring-disk electrode
study. J. Electroanal.Chem. 2001, 495, 134-145.
71.
Lima, F. H. B.; Zhang, J.; Shao, M. H.; Sasaki, K.; Vukmirovic, M. B.; Ticianelli,
E. A.; Adzic, R. R., Catalytic activity-d-band center correlation for the 02 reduction
reaction on platinum in alkaline solutions. J. Phys. Chem. C 2007, 111, 404-410.
72.
Bonakdarpour, A.; Lefevre, M.; Yang, R. Z.; Jaouen, F.; Dahn, T.; Dodelet, J. P.;
Dahn, J. R., Impact of loading in RRDE experiments on Fe-N-C catalysts: Two- or fourelectron oxygen reduction? Electrochem. Solid State Lett. 2008, 11, B105-B108.
73.
Gasteiger, H. A.; Garche, J., Fuel cells. Wiley-VCH: 2008.
115
74.
Koh, S.; Hahn, N.; Yu, C. F.; Strasser, P., Effects of Composition and Annealing
Conditions on Catalytic Activities of Dealloyed Pt-Cu Nanoparticle Electrocatalysts for
PEMFC. J Electrochem. Soc. 2008, 155, B1281-B1288.
75.
Imaizumi, S.; Shimanoe, K.; Teraoka, Y.; Yamazoe, N., Oxygen reduction
property of ultrafine LaMnO3 dispersed on carbon support. Electrochem. Solid State Lett.
2005, 8, A270-A272.
76.
Bursell, M.; Pirjamali, M.; Kiros, Y., Lao. 6Ca 0 .4 CoO 3 , La 0 .Ca 0 .9MnO 3 and LaNiO 3
as bifunctional oxygen electrodes. Electrochim. Acta 2002, 47, 1651-1660.
77.
Lee, C. K.; Striebel, K. A.; McLarnon, F. R.; Cairns, E. J., Thermal treatment of
Lao.6Cao. 4CoO 3 perovskites for bifunctional air electrodes. J Electrochem. Soc. 1997,
144, 3801-3806.
78.
Hyodo, T., Hayashi, M., Miura, N., Yamazoe, N., Catalytic Activities of Rare-
Earth Manganites for Cathodic Reduction of Oxygen in Alkaline Solution. 1996, 143,
266-267.
79.
Yuasa, M.; Shimanoe, K.; Teraoka, Y.; Yamazoe, N., Preparation of carbon-
supported nano-sized LaMnO 3 using reverse micelle method for energy-saving oxygen
reduction cathode. Catal. Today 2007, 126, 313-319.
80.
Shimizu, Y.; Uemura, K.; Matsuda, H.; Miura, N.; Yamazoe, N., Bi-functional
oxygen-electrode using large surface-area Lai-xCaxCoO 3 for rechargable metal-air
battery. J Electrochem. Soc. 1990, 137, 3430-3433.
81.
Yuasa, M.; Sakai, G.; Shimanoe, K.; Teraoka, Y.; Yamazoe, N., Exploration of
reverse micelle synthesis of carbon-supported LaMnO 3 . J Electrochem. Soc. 2004, 151,
A1477-A1482.
82.
Bockris, J. 0.; Otagawa, T., Mechanism of oxygen evolution on perovskites. J
Phys. Chem. 1983, 87, 2960-2971.
83.
Matsumoto, Y., Yoneyama, H., Tamura, H., The Mechanism of Oxygen
Reduction at a LaNiO 3 Electrode. 1978, 51, 1927-30.
84.
Rios, E.; Gautier, J. L.; Poillerat, G.; Chartier, P., Mixed valency spinel oxides of
transition metals and electrocatalysis: case of the MnxCo3.xO 4 system. Electrochim.Acta
1998, 44, 1491-1497.
116
85.
Bard, A. J.; Faulkner, L. R., Electrochemical methods: Fundamentalsand
applications.2nd ed.; John Wiley & Sons: New York, 2001.
86.
Miyazaki, K.; Sugimura, N.; Matsuoka, K.; Iriyama, Y.; Abe, T.; Matsuoka, M.;
Ogumi, Z., Perovskite-type oxides Lai-xSrxMnO 3 for cathode catalysts in direct ethylene
glycol alkaline fuel cells. J. Power Sources 2008, 178, 683-686.
87.
Tulloch, J.; Donne, S. W., Activity of perovskite Lal-xSrxMnO 3 catalysts towards
oxygen reduction in alkaline electrolytes. J. Power Sources 2009, 188, 359-366.
88.
Piana, M.; Catanorchi, S.; Gasteiger, H. A., Kinetics of Non-Platinum Group
Metal Catalysts for the Oxygen Reduction Reaction in Alkaline Medium. 2008, 16, 20452055.
89.
Breger, J.; Meng, Y. S.; Hinuma, Y.; Kumar, S.; Kang, K.; Shao-Horn, Y.; Ceder,
G.; Grey, C. P., Effect of high voltage on the structure and electrochemistry of
LiNio.5 Mno. 50 2 : A joint experimental and theoretical study. Chem. Mat. 2006, 18, 47684781.
90.
Ferreira, P. J.; la 0', G. J.; Shao-Horn, Y.; Morgan, D.; Makharia, R.; Kocha, S.;
Gasteiger, H. A., Instability of Pt/C electrocatalysts in proton exchange membrane fuel
cells - A mechanistic investigation. 2005, 152, A2256-A2271.
91.
Garsuch, R. R.; Le, D. B.; Garsuch, A.; Li, J.; Wang, S.; Farooq, A.; Dahn, J. R.,
Studies of lithium-exchanged nafion as an electrode binder for alloy negatives in lithiumion batteries. J. Electrochem. Soc. 2008, 155, A721-A724.
92.
Hung, M. H.; Rao, M. V. M.; Tsai, D. S., Microstructures and electrical properties
of calcium substituted LaFeO 3 as SOFC cathode. Mater. Chem. Phys. 2007, 101, 297302.
93.
Knizek, K.; Daturi, M.; Busca, G.; Michel, C., Structural, electro-magnetic and
catalytic characterisation of the LaMnlXCux0 3.- system. J. Mater. Chem. 1998, 8, 18151819.
94.
Moriga, T.; Usaka, 0.; Nakabayashi, I.; Kinouchi, T.; Kikkawa, S.; Kanamaru, F.,
Characterization of oxygen-deficient phases appearing in reduction of the perovskite-type
LaNiO 3 to La 2 Ni 2 0 5 . Solid State Ion. 1995, 79, 252-255.
95.
Zook, L. A.; Leddy, J., Density and solubility of nafion: Recast, annealed, and
commercial films. Anal. Chem. 1996, 68, 3793-3796.
117
96.
Schmidt, T. J.; Gasteiger, H. A.; Stab, G. D.; Urban, P. M.; Kolb, D. M.; Behm,
R. J., Characterization of high-surface area electrocatalysts using a rotating disk electrode
configuration. J Electrochem. Soc. 1998, 145, 2354-2358.
97.
Chen, S.; Sheng, W. C.; Yabuuchi, N.; Ferreira, P. J.; Allard, L. F.; Shao-Horn,
Y., Origin of Oxygen Reduction Reaction Activity on "Pt 3Co" Nanoparticles: Atomically
Resolved Chemical Compositions and Structures. J. Phys. Chem. C 2009, 113, 11091125.
98.
Mayrhofer, K. J. J.; Crampton, A. S.; Wiberg, G. K. H.; Arenz, M., Analysis of
the impact of individual glass constituents on electrocatalysis on pt electrodes in alkaline
solution. J Electrochem. Soc. 2008, 155, P78-P8 1.
99.
Mayrhofer, K. J. J.; Wiberg, G. K. H.; Arenz, M., Impact of glass corrosion on the
electrocatalysis on Pt electrodes in alkaline electrolyte. J. Electrochem. Soc. 2008, 155,
P1-P5.
100.
Mayrhofer, K. J. J.; Juhart, V.; Hartl, K.; Hanzlik, M.; Arenz, M., Adsorbate-
Induced Surface Segregation for Core-Shell Nanocatalysts. Angew. Chem. -Int. Edit.
2009, 48, 3529-3531.
101.
Ma, J.; Liu, Y.; Yan, Y.; Zhang, P., A Membraneless Direct Borohydride Fuel
Cell Using LaNiO 3 -Catalysed Cathode. Fuel Cells 2008, 8, 394-398.
102.
Markovic, N. M.; Gasteiger, H. A.; Philip, N., Oxygen reduction on platinum
low-index single-crystal surfaces in alkaline solution: Rotating ring disk(Pt(hkl)) studies.
J Phys. Chem. 1996, 100, 6715-6721.
103.
Schmidt, T. J.; Stamenkovic, V.; Ross, P. N.; Markovic, N. M., Temperature
dependent surface electrochemistry on Pt single crystals in alkaline electrolyte - Part 3.
The oxygen reduction reaction. Phys. Chem. Chem. Phys. 2003, 5, 400-406.
104.
Bonakdarpour, A.; Dahn, T. R.; Atanasoski, R. T.; Debe, M. K.; Dahn, J. R.,
H20 2 release during oxygen reduction reaction on Pt nanoparticles. Electrochem. Solid
State Lett. 2008, 11, B208-B211.
105.
Inaba, M.; Yamada, H.; Tokunaga, J.; Tasaka, A., Effect of agglomeration of Pt/C
catalyst on hydrogen peroxide formation. Electrochem. Solid State Lett. 2004, 7, A474A476.
118
106.
Gu, W.; Baker, D. R.; Liu, Y.; Gasteiger, H. A., Proton Exchange Membrane Fuel
Cell (PEMFC) Down-the-Channel Performance Model. In Handbook ofFuel Cells Fundamentals, Technology andApplications, Vielstich, W.; Gasteiger, H. A.; Yokokawa,
H., Eds. Wiley: Chichester, 2009; Vol. 6, pp 631-657.
107.
Karlsson, G., Reduction of Oxygen on LaNiO3 in Alkaline Solution. 1983, 10,
319-331.
108.
Zou, Z. G.; Ye, J. H.; Sayama, K.; Arakawa, H., Direct splitting of water under
visible light irradiation with an oxide semiconductor photocatalyst. Nature 2001, 414,
625-627.
109.
Matsumoto, Y.; Yoneyama, H.; Tamura, H., Influence of nature of conduction-
band of transition-metal oxides on catalytic activity for oxygen reduction. J. Electroanal.
Chem. 1977, 83, 237-243.
110.
Matsumoto, Y.; Yoneyama, H.; Tamura, H., A new catalyst for cathodic reduction
of oxygen: Lanthanum nickle oxide. Chem. Lett. 1975, 661-662.
111.
Morin, F. J.; Wolfram, T., Surface states and catalysis on d-band perovskites.
Phys. Rev. Lett. 1973, 30, 1214-1217.
112.
Yan, J. Q.; Zhou, J. S.; Goodenough, J. B., Ferromagnetism in LaCoO 3. Phys.
Rev. B 2004, 70, 014402.
113.
Ravel, B.; Newville, M., ATHENA, ARTEMIS, HEPHAESTUS: data analysis
for X-ray absorption spectroscopy using IFEFFIT. J. Synchrot. Radiat. 2005, 12, 537541.
114.
Mizusaki, J.; Mori, N.; Takai, H.; Yonemura, Y.; Minamiue, H.; Tagawa, H.;
Dokiya, M.; Inaba, H.; Naraya, K.; Sasamoto, T.; Hashimoto, T., Oxygen
nonstoichiometry and defect equilibrium in the perovskite-type oxides La1 .xSrxMnO 3 6
Solid State Ionics 2000, 129, 163-177.
115.
Tyson, T. A.; Qian,
Q.; Kao, C.
C.; Rueff, J. P.; de Groot, F. M. F.; Croft, M.;
Cheong, S. W.; Greenblatt, M.; Subramanian, M. A., Valence state of Mn in Ca-doped
LaMnO 3 studied by high-resolution Mn K-beta emission spectroscopy. Phys. Rev. B
1999, 60, 4665-4674.
119
116.
Rormark, L.; Wiik, K.; Stolen, S.; Grande, T., Oxygen stoichiometry and
structural properties of La.xAxMnO 3±6 (A = Ca or Sr and (0 < x < 1). J Mater. Chem.
2002, 12, 1058-1067.
117.
Croft, M.; Sills, D.; Greenblatt, M.; Lee, C.; Cheong, S. W.; Ramanujachary, K.
V.; Tran, D., Systematic Mn d-configuration change in the Lai.xCaxMnO 3 system: A Mn
K-edge XAS study. Phys. Rev. B 1997, 55, 8726-8732.
118.
Ritter, C.; Ibarra, M. R.; DeTeresa, J. M.; Algarabel, P. A.; Marquina, C.; Blasco,
J.; Garcia, J.; Oseroff, S.; Cheong, S. W., Influence of oxygen content on the structural,
magnetotransport, and magnetic properties of LaMnO 3+6. Phys. Rev. B 1997, 56, 89028911.
119.
Li, J., Investigation of orthorhombic perovskite La 1 .xCaxFeO 3-y(0 < x < 0.5).
Phys. Scr. 1992, 45, 62-64.
120.
Russo, U.; Nodari, L.; Faticanti, M.; Kuncser, V.; Filoti, G., Local interactions
and electronic phenomena in substituted LaFeO 3 perovskites. Solid State Ionics 2005,
176, 97-102.
121.
Mastin, J.; Einarsrud, M. A.; Grande, T., Crystal structure and thermal properties
of Lai.xCaxCoO 3. 6 (0 < x < 0.4). Chem. Mat. 2006, 18, 1680-1687.
122.
Takeda, T.; Watanabe, H.; Yamaguch.Y, Magnetic structure of SrCoO 2 .5. J. Phys.
Soc. Jpn. 1972, 33, 970-&.
123.
Samoilov, A. V.; Beach, G.; Fu, C. C.; Yeh, N. C.; Vasquez, R. P., Giant
spontaneous Hall effect and magnetoresistance in Lai.xCaxCoO 3 (0.1
x < 0.5). J. Appl.
Phys. 1998, 83, 6998-7000.
124.
Samoilov, A. V.; Beach, G.; Fu, C. C.; Yeh, N. C.; Vasquez, R. P., Magnetic
percolation and giant spontaneous Hall effect in Lai.xCaxCoO 3 (0.2 < x < 0.5). Phys. Rev.
B 1998, 57, 14032-14035.
125.
Sanchez, M. C.; Garcia, J.; Blasco, J.; Subias, G.; Perez-Cacho, J., Local
electronic and geometrical structure of LaNii.xMnx03+ perovskites determined by x-rayabsorption spectroscopy. Phys. Rev. B 2002, 65, 144409.
126.
Arcon, I.; Mirtic, B.; Kodre, A., Determination of valence states of chromium in
calcium chromates by using X-ray absorption near-edge structure (XANES)
spectroscopy. J Am. Ceram. Soc. 1998, 81, 222-224.
120
127.
Markovic, N. M.; Gasteiger, H. A.; Ross, P. N., Oxygen reduction on platinum
low-index single-crystal surfaces in alkaline solution: Rotating ring disk(Pt(hkl)) studies.
J. Phys. Chem. 1996, 100, 6715-6721.
128.
Fernandez, E. M.; Moses, P. G.; Toftelund, A.; Hansen, H. A.; Martinez, J. I.;
Abild-Pedersen, F.; Kleis, J.; Hinnemann, B.; Rossmeisl, J.; Bligaard, T.; Norskov, J. K.,
Scaling relationships for adsorption energies on transition metal oxide, sulfide, and
nitride surfaces. Angew Chem Int Edit 2008, 47, 4683-4686.
129.
Dowden, D. A., Crystal and ligand field models of solid catalysts. Catal.Rev.
1971, 5, 1-32.
130.
Yokoi, Y.; Uchida, H., Catalytic activity of perovskite-type oxide catalysts for
direct decomposition of NO: Correlation between cluster model calculations and
temperature-programmed desorption experiments. Catal. Today 1998, 42, 167-174.
131.
Yan, J.
Q.; Zhou, J. S.;
Goodenough, J. B., Bond-length fluctuations and the spin-
state transition in LCoO 3 (L=La, Pr, and Nd). Phys. Rev. B 2004, 69, 134409.
132.
Medarde, M. L., Structural, magnetic and electronic properties of RNiO 3
perovskites (R equals rare earth). JPhys-CondensMat 1997, 9, 1679-1707.
133.
Zhou, J. S.; Goodenough, J. B.; Dabrowski, B.; Klamut, P. W.; Bukowski, Z.,
Enhanced susceptibility in LNiO 3 perovskites (L = La, Pr, Nd, Ndo. 5 Smo. 5 ). Phys. Rev.
Lett. 2000, 84, 526-529.
134.
Goodenough, J. B.; Cushing, B. L., Oxide-based ORR catalysts. In Handbook of
fuel cells - Fundamentals,technology and applications,Vielstich, W.; Gasteiger, H. A.;
Yokokawa, H., Eds. Wiley: Chichester, 2003; Vol. 2, pp 520-533.
135.
Rossmeisl, J.; Qu, Z. W.; Zhu, H.; Kroes, G. J.; Norskov, J. K., Electrolysis of
water on oxide surfaces. J. Electroanal.Chem. 2007, 607, 83-89.
136.
Lu, Y. C.; Xu, Z. C.; Gasteiger, H. A.; Chen, S.; Hamad-Schifferli, K.; Shao-
Horn, Y., Platinum-gold nanoparticles: A highly active bifunctional electrocatalyst for
rechargeable lithium-air batteries. J. Am. Chem. Soc. 2010, 132, 12170-12171.
137.
Armand, M.; Tarascon, J. M., Building better batteries. Nature 2008, 451, 652-
657.
138.
Koper, M. T. M., Thermodynamic theory of multi-electron transfer reactions:
Implications for electrocatalysis. J. Electroanal.Chem. 2011, 660, 254-260.
121
139.
Davidson, C. R.; Kissel, G.; Srinivasan, S., Electrode kinetics of the oxygen
evolution reaction at NiCo 2O4 from 30% KOH - Dependence on temperature. J.
Electroanal.Chem. 1982, 132, 129-135.
140.
Trasatti, S., Electrocatalysis by oxides - attempt at a unifying approach. J.
Electroanal.Chem. 1980, 111, 125-131.
141.
Jung, J.; Misture, S. T.; Edwards, D. D., Oxygen stoichiometry, electrical
conductivity, and thermopower measurements of BSCF (Bao. 5 Sro. 5 CoxFe 1..xO 3 . , 0 <x
0.8) in air. Solid State Ionics 2010, 181, 1287-1293.
142.
Harvey, A. S.; Yang, Z.; Infortuna, A.; Beckel, D.; Purton, J. A.; Gauckler, L. J.,
Development of electron holes across the temperature-induced semiconductor-metal
transition in Bai.xSrxCoiyFeyO 3.6 (x, y=0.2-0.8): a soft x-ray absorption spectroscopy
study. J. Phys.-Condes. Matter 2009, 21, 10.
143.
Harvey, A. S.; Litterst, F. J.; Yang, Z.; Rupp, J. L. M.; Infortuna, A.; Gauckler, L.
J., Oxidation states of Co and Fe in BaixSrxCoiyFeyO 3-6 (x, y = 0.2-0.8) and oxygen
desorption in the temperature range 300-1273 K. Phys. Chem. Chem. Phys. 2009, 11,
3090-3098.
144.
Arnold, M.; Xu,
Q.; Tichelaar,
F. D.; Feldhoff, A., Local Charge Disproportion in
a High-Performance Perovskite. Chem. Mat. 2009, 21, 635-640.
145.
Ballhausen, C. J.; Gray, H. B., The Electronic Structure of the Vanadyl Ion. Inorg.
Chem. 1962, 1, 111-122.
146.
Betley, T. A.; Wu, Q.; Van Voorhis, T.; Nocera, D. G., Electronic design criteria
for 0-0 bond formation via metal-oxo complexes. Inorg. Chem. 2008, 47, 1849-1861.
147.
Matsumoto, Y.; Kurimoto, J.; Sato, E., Oxygen evolution on SrFeO 3 electrode. J.
Electroanal.Chem. 1979, 102, 77-83.
148.
Matsumoto, Y.; Manabe, H.; Sato, E., Oxygen evolution on Lai.xSrxCoO 3
electrodes in alkaline solutions. J Electrochem. Soc. 1980, 127, 811-814.
149.
Matsumoto, Y.; Sato, E., Oxygen evolution on Lai.xSrxMnO 3 electrodes in
alkaline solutions. Electrochim. Acta 1979, 24, 421-423.
150.
Wattiaux, A.; Grenier, J. C.; Pouchard, M.; Hagenmuller, P., Electrolytic oxygen
evolution in alkaline medium on La 1 .xSrxFeO 3. perovskite-related ferries. 1.
Electrochemical study. J Electrochem. Soc. 1987, 134, 1714-1718.
122
151.
Matsumoto, Y.; Sato, E., Electrocatalytic properties of transition metal oxides for
oxygen evolution reaction. Mater. Chem. Phys. 1986, 14, 397-426.
152.
Materials and Methods are available as supporting materials on Science online.
153.
Shao, Z. P.; Haile, S. M., A high-performance cathode for the next generation of
solid-oxide fuel cells. Nature 2004, 431, 170-173.
154.
Dotan, H.; Sivula, K.; Gratzel, M.; Rothschild, A.; Warren, S. C., Probing the
photoelectrochemical properties of hematite (a -Fe 2O 3) electrodes using hydrogen
peroxide as a hole scavenger. Energy Environ. Sci. 2011, 4, 958-964.
155.
Stephens, I. E. L.; Bondarenko, A. S.; Perez-Alonso, F. J.; Calle-Vallejo, F.;
Bech, L.; Johansson, T. P.; Jepsen, A. K.; Frydendal, R.; Knudsen, B. P.; Rossmeisl, J.;
Chorkendorff, I., Tuning the Activity of Pt(1 11) for Oxygen Electroreduction by
Subsurface Alloying. J. Am. Chem. Soc. 2011, 133, 5485-5491.
156.
Rossmeisl, J.; Skulason, E.; Bjorketun, M. E.; Tripkovic, V.; Norskov, J. K.,
Modeling the electrified solid-liquid interface. Chem. Phys. Lett. 2008, 466, 68-71.
157.
Strmcnik, D.; Kodama, K.; van der Vliet, D.; Greeley, J.; Stamenkovic, V. R.;
Markovic, N. M., The role of non-covalent interactions in electrocatalytic fuel-cell
reactions on platinum. Nature Chem. 2009, 1, 466-472.
158.
Strmcnik, D.; van der Vliet, D. F.; Chang, K. C.; Komanicky, V.; Kodama, K.;
2
You, H.; Stamenkovic, V. R.; Markovic, N. M., Effects of Li+, K+, and Ba + Cations on
the ORR at Model and High Surface Area Pt and Au Surfaces in Alkaline Solutions. J.
Phys. Chem. Lett. 2011, 2, 2733-2736.
159.
Suntivich, J.; Gasteiger, H. A.; Yabuuchi, N.; Shao-horn, Y., Electrocatalytic
measurement methodology of oxide catalysts using a thin-film rotating disk electrode. J.
Electrochem. Soc. 2010, 157.
160.
Thomas, S.; Sung, Y. E.; Kim, H. S.; Wieckowski, A., Specific adsorption of a
bisulfate anion on a Pt(1 11) electrode. Ultrahigh vacuum spectroscopic and cyclic
voltammetric study. J. Phys. Chem. 1996, 100, 11726-11735.
161.
He,
Q. G.; Yang,
X. F.; Chen, W.; Mukerjee, S.; Koel, B.; Chen, S. W., Influence
of phosphate anion adsorption on the kinetics of oxygen electroreduction on low index
Pt(hkl) single crystals. Phys. Chem. Chem. Phys. 2010, 12, 12544-12555.
123
162.
Wang, J. X.; Markovic, N. M.; Adzic, R. R., Kinetic analysis of oxygen reduction
on Pt( 111) in acid solutions: Intrinsic kinetic parameters and anion adsorption effects. J
Phys. Chem. B 2004, 108, 4127-4133.
163.
Kajii, K.; Ohsaka, T.; Kitamura, F., Crucial effect of rubidium cation on catalytic
activity of a polycrystalline gold electrode toward oxygen reduction reaction in alkaline
media. Electrochem. Commun. 2010, 12, 970-972.
164.
Yuasa, M.; Nishida, M.; Kida, T.; Yamazoe, N.; Shimanoe, K., Bi-Functional
Oxygen Electrodes Using LaMnO 3/LaNiO 3 for Rechargeable Metal-Air Batteries. J.
Electrochem. Soc. 2011, 158, A605-A610.
165.
Yuasa, M.; Yamazoe, N.; Shimanoe, K., Durability of Carbon-Supported La-Mn-
Based Perovskite-Type Oxides as Oxygen Reduction Catalysts in Strong Alkaline
Solution. J Electrochem. Soc. 2011, 158, A41 1-A416.
166.
Lee, Y.; Suntivich, J.; May, K. J.; Perry, E. E.; Shao-Horn, Y., Synthesis and
Activities of Rutile IrO 2 and RuO2 Nanoparticles for Oxygen Evolution in Acid and
Alkaline Solutions. J. Phys. Chem. Lett. 2012, 399-404.
167.
Tritsaris, G. A.; Greeley, J.; Rossmeisl, J.; Norskov, J. K., Atomic-Scale
Modeling of Particle Size Effects for the Oxygen Reduction Reaction on Pt. Catal.Lett.
2011, 141, 909-913.
168.
Subbaraman, R.; Strmcnik, D.; Paulikas, A. P.; Stamenkovic, V. R.; Markovic, N.
M., Oxygen Reduction Reaction at Three-Phase Interfaces. ChemPhysChem 2010, 11,
2825-2833.
169.
Subbaraman, R.; Strmcnik, D.; Stamenkovic, V.; Markovic, N. M., Three Phase
Interfaces at Electrified Metal-Solid Electrolyte Systems 1. Study of the Pt(hkl)-Nafion
Interface. J. Phys. Chem. C 2010, 114, 8414-8422.
170.
Goodenough, J. B.; Manoharan, R.; Paranthaman, M., Surface protonation and
electrochemical activity of oxides in aqueous-solution. J. Am. Chem. Soc. 1990, 112,
2076-2082.
171.
Man, I. C.; Su, H. Y.; Calle-Vallejo, F.; Hansen, H. A.; Martinez, J. I.; Inoglu, N.
G.; Kitchin, J.; Jaramillo, T. F.; Norskov, J. K.; Rossmeisl, J., Universality in oxygen
evolution electrocatalysis on oxide surfaces. ChemCatChem 2011, 3, 1159-1165.
124
172.
Cuesta, A.; Strmcnik, D.; Escudero-Escribano, M.; Kodama, K.; Stamenkovic, V.
R.; Markovic, N. M., Enhanced electrocatalysis of the oxygen reduction reaction based
on patterning of platinum surfaces with cyanide. Nature Chem. 2010, 2, 880-885.
173.
Kolics, A.; Wieckowski, A., Adsorption of bisulfate and sulfate anions on a
Pt(l 11) electrode. JPhys Chem B 2001, 105, 2588-2595.
174.
Armstrong, F. A.; Hirst, J., Reversibility and efficiency in electrocatalytic energy
conversion and lessons from enzymes. P Natl Acad Sci USA 2011, 108, 14049-14054.
125