Mechanistic Understanding of Size-Dependent Oxygen Reduction Activity and Selectivity over Pt/CNT Nanocatalysts

Article abstract of DOI:10.1002/ejic.201801521

Identifying the underlying nature of the structure sensitivity of oxygen reduction reaction (ORR) over carbon supported Pt catalysts is an important but challenging issue in electrochemical system. In this work, we combine experiments, density functional theory calculations with model calculations to clarify the size-dependent ORR activity and selectivity over differently sized Pt/CNT catalysts. HAADF-STEM, HRTEM and XPS measurements show that the Pt nanoparticles supported on CNT possess a well-defined truncated octahedron shape in most cases and similar electronic properties. The observed size-insensitive TOF_{active site} based on the number of Pt(111) atoms suggests the Pt(111) surface as the dominant active site. Moreover, the Pt(111) surface is also suggested as the dominant active sites for the formation of H₂O₂, and the catalyst with the higher Pt binding energy facilitates the oxygen reduction to H₂O. The insights revealed here could shed new light on the design and optimization of Pt-based ORR catalysts.

Full text of DOI:10.1002/ejic.201801521

A Journal of
Accepted Article
Title: Mechanistic Understanding of Size-dependent Oxygen Reduction
Activity and Selectivity over Pt/CNT Nanocatalysts
Authors: Jie Gan, Wei Luo, Wenyao Chen, Jianing Guo, Zhonghua
Xiang, Bingxu Chen, Fan Yang, Yunjun Cao, Fei Song,
Xuezhi Duan, and Xinggui Zhou
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Eur. J. Inorg. Chem. 10.1002/ejic.201801521
Link to VoR: http://dx.doi.org/10.1002/ejic.201801521

European Journal of Inorganic Chemistry
10.1002/ejic.201801521
FULL PAPER
Mechanistic Understanding of Size-dependent Oxygen Reduction
Activity and Selectivity over Pt/CNT Nanocatalysts
[
a]
[a]
[a]
[b]
[b]
[a]
[c]
Jie Gan, Wei Luo, Wenyao Chen, Jianing Guo, Zhonghua Xiang, Bingxu Chen, Fan Yang,
[
c]
[d]
[a]
[a]
Yunjun Cao, Fei Song, Xuezhi Duan,* Xinggui Zhou
Abstract: Identifying the underlying nature of the structure sensitivity
of oxygen reduction reaction (ORR) over carbon supported Pt
catalysts is an important but challenging issue in electrochemical
system. In this work, we combine experiments, density functional
theory calculations with model calculations to clarify the size-
dependent ORR activity and selectivity over differently sized Pt/CNT
catalysts. HAADF-STEM, HRTEM and XPS measurements show
that the Pt nanoparticles supported on CNT possess a well-defined
truncated octahedron shape in most cases and similar electronic
properties. The observed size-insensitive TOFactive site based on the
number of Pt(111) atoms suggests the Pt(111) surface as the
dominant active site. Moreover, the Pt(111) surface is also
the Pt ORR catalysts. The atomically dispersed Pt species were
reported to possess a high selectivity for the H
2
O
2
formation
) rather than H
O),[24] while the Pt
2
−
+
−
through a 2e pathway (O
2
+ 2H + 2e → H
2
O
2
2
O
−
+
−
through a 4e pathway (O
2
+ 4H + 4e → 2H
particle size effects on the ORR selectivity have not been
understood very much. Notably, the Pt nanoparticles with
differently geometric structure usually exhibit different Pt
electronic properties that would influence the ORR
performance.[19,25-29] However, limited work has been dedicated
to the Pt particle size effects under excluding the Pt electronic
effects. Considering the characteristic of such catalytic system, it
is highly favorable to disentangle the Pt size effects from the
electronic effects and identify the underlying nature of the Pt size
effects from the perspective of active site.
suggested as the dominant active sites for the formation of H
and the catalyst with the higher Pt binding energy facilitates the
oxygen reduction to H O. The insights revealed here could shed new
2 2
O ,
2
Herein, we focus our attention on mechanistic understanding
of the structure sensitivity of carbon nanotubes (CNT) supported
Pt catalyzed ORR. The CNT supported Pt catalysts with different
particle sizes but similar electronic properties are employed to
light on the design and optimization of Pt-based ORR catalysts.
examine the ORR performance in HClO
4
solution.
A
Introduction
methodology by combining experiments, density functional
theory (DFT) calculations with model calculations is proposed to
clarify the size-dependent ORR activity and selectivity over the
Pt/CNT catalysts. The insights here might open up a new
avenue for the design of Pt-based ORR catalysts.
Green electrochemical technologies like fuel cells hold a crucial
place in renewable energy conversion systems, and their large-
scale commercialization is predominantly limited by the cathodic
oxygen reduction reaction (ORR) due to its sluggish kinetics.[1-4]
Carbon supported Pt-based catalysts have been recognized as
[5-
the state-of-the-art ORR catalysts especially in acidic medium.
Results and Discussion
1
3]
In previous studies, strongly adsorbed acidic electrolytes such
as H SO and H PO are found to be unfavorable for the reaction,
while HClO solution is believed to be a promising non-
2
4
3
4
Mechanistic understanding of size-dependent ORR activity
4
adsorbing electrolyte.[14-19] In addition, tremendous efforts have
been devoted to fine-tuning Pt geometric and electronic
properties toward enhanced ORR performance.[1,4,11-14]
The end-close carbon nanotubes support with the highly
crystalline and mesoporous characteristics reflected in Figure S1
by scanning electron microscope (SEM), transmission electron
The ORR over Pt nanoparticles supported on carbon
microscope (TEM) and N
2
physisorption measurements was
12,13,19-23]
materials is
a
typical structure-sensitive reaction.[
used to load differently sized Pt nanoparticles by incipient
wetness impregnation. Figure 1a-1c give high-angle annular
dark-field scanning transmission electron microscope (HAADF-
STEM) images and the corresponding particle size distributions
of 2, 8 and 10 wt% Pt/CNT catalysts. Their average particle
sizes are determined to be 1.7, 2.2 and 2.7 nm, respectively.
Optimal Pt particle sizes were observed in the range of 2-3
nm.[18-21] To date, understanding the nature of the Pt particle size
effects has been a longstanding scientific question. On the other
hand, the ORR selectivity is another important factor to evaluate
Core-level XPS measurements in Figure 1d show that these
three catalysts exhibit similar Pt0 4f7/2 binding energy of ca.
[
a]
J. Gan, W. Luo, Dr. W. Y. Chen, B. X. Chen, Prof. Dr. X. Z. Duan,
Prof. Dr. X. G. Zhou
State Key Laboratory of Chemical Engineering, East China
University of Science and Technology, Shanghai 200237, China
E-mail: xzduan@ecust.edu.cn
http://hgxy.ecust.edu.cn/2016/0419/c1270a9497/page.htm
J. N. Guo, Prof. Dr. Z. H. Xiang
State Key Lab of Organic-Inorganic Composites, Beijing University
of Chemical Technology, Beijing 100029, China
Prof. Dr. F. Yang, Dr. Y. J. Cao
State Key Laboratory of Catalysis, Dalian Institute of Chemical
Physics, Chinese Academy of Sciences, Dalian 116023, China
Prof. Dr. F. Song
Shanghai Institute of Applied Physics, Chinese Academy of
Sciences, Shanghai 201204, China
71.70 eV, indicating similar Pt electronic properties. Valence-
band XPS measurements in Figure 1e are also carried out,
because they can provide the information about Fermi levels
and density of state.[30] Clearly, the densities of Pt 5d states
gradually increase with the Pt loading, being in consistent with
previous observations of metal loading effects,[31] while the
positions of Fermi levels are similar, further supporting the three
catalysts with similar Pt electronic properties.
[
[
[
b]
c]
d]
ORR measurements of the three Pt/CNT catalysts and pure
CNT support were performed in O
2 4
-saturated 0.1 M HClO
solutions, and the cyclic voltammetry (CV) curves and the linear
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201801521
FULL PAPER
Figure 1. (a-c) Typical HAADF-STEM images, (d) core-level Pt 4f XPS spectra and (e) valence-band XPS spectra, (f) CV curves at a scan rate of 100 mV s-1 and
g) mass activities at 0.7 V of 2, 8 and 10 wt% Pt/CNT catalysts. The insets in (a), (b) and (c) indicate the corresponding Pt particle size distributions of the three
(
catalysts. The CV curve of pure CNT is also shown in (f). (h) Free energy profiles of ORR via the dissociative mechanism on the Pt(111), Pt(100) and Pt(211)
surfaces. The inset in (h) indicates a comparison for the reaction free energies at 1.23 V of the protonation of O* and OH* on the three Pt surfaces.
sweep voltammetry (LSV) curves are shown in Figure 1f and
Figure S2, respectively. Compared to the pure CNT support,
these CNT supported Pt catalysts exhibit distinct reduction
peaks in the range of 0.6-0.7 V, indicating that the Pt dominates
the ORR process. Moreover, the kinetic control is mostly
observed in Figure S2 at potentials above 0.7 V over the three
catalysts by rotating electrode techniques. According to the
Koutecky-Levich equation presented in the experimental section,
the kinetic current densities of the three catalysts at 0.7 V at the
typical rotating speed of 1600 rpm were obtained to calculate
their mass activities. Figure 1g shows the mass activity of the
three catalysts as a function of the Pt particle size. Obviously,
there are size-dependent ORR activities over the three catalysts.
To gain mechanistic insights into the size-dependent ORR
activity, we employ DFT calculations to understand the ORR
mechanism on three typical Pt surfaces, i.e., Pt (111), Pt (100)
and Pt (211) surfaces. To model the interaction with the solvent,
we placed 1/3 ML adsorbates and 1/3 ML water molecules layer
on the three Pt surfaces as suggested by previous studies.[
As shown in Scheme 1, the ORR process to form water in acidic
medium might proceed by means of a dissociative mechanism
(I), associative mechanism (II) and/or hydrogen peroxide
32,33]
mechanism (III) with
a four-electron pathway. With DFT
structural optimization calculations, the calculated energy
profiles of ORR on the three Pt surfaces for the three reaction
mechanisms were plotted, and the results are shown in Figure
1h and Figure S3. At the ORR equilibrium potential, the
formation of O* on the three Pt surfaces for the Mechanism I is
observed to be exothermic and thus thermodynamically favorable,
2
while the protonation of O * and OOH* on the three Pt surfaces
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201801521
FULL PAPER
is highly endothermic and thus thermodynamically unfavorable
for the Mechanism II and Mechanism III. These indicate that
under the acidic reaction conditions, the ORR mainly proceeds
via the Mechanism I, i.e., the dissociative mechanism, which is
consistent with previous studies.[4,33,,34]
to the other two surfaces is lowly exothermic, which can be
reflected by the trend of the ΔGads values of O* on the three Pt
surfaces in Table 1. It is also observed that the protonation of O*
and OH* are endothermic, suggesting non-spontaneous
occurrence of these steps; the protonation of O* is more
endothermic than that of OH*. All of these results suggest that
the protonation of O* is most likely the rate-determining step.
Since the onset potential (i.e., Uonset) is an important measure
of the ORR activity, the theoretical Uonset values of the three Pt
surfaces were obtained from the correlated free energy profiles
at equilibrium potential, i.e., 0.78, 0.72 and 0.41 eV for the
Pt(111), Pt(100) and Pt(211) surfaces, respectively, and the
corresponding free energy profiles at the calculated Uonset are
shown in Figure 1h. Based on the calculated Uonset values, the
Pt(111) and Pt(100) surfaces have much higher Uonset values
than the Pt(211) surface. This indicates that the Pt(211) surface
provides a negligible contribution to the ORR activity, which is in
coincident with some previous studies.[22] It is noted that in our
DFT studies, the Uonset value on the Pt(111) surface is slightly
higher than that on the Pt(100) surface, while in the DFT studies
of Norskov and coworkers, the Uonset value on the Pt(100)
surface (i.e., 0.80 eV) is slightly higher than that on the Pt(111)
surface (i.e., 0.78 eV).[22] In other words, it is difficult to identify
which one is the dominant active sites for the ORR on the
Pt(111) and Pt(100) surfaces only based on the thermodynamic
stability of reaction intermediates without the consideration of
Scheme 1. Three kinds of ORR mechanisms: (I) dissociative mechanism, (II)
associative mechanism and (III) hydrogen peroxide mechanism.
reaction barrier. However, it remains
a big challenge to
accurately calculate the reaction barriers due to the ORR
reaction characteristics, such as the occurrence of ORR in the
liquid phase, the presence of charge on the catalyst surfaces,
the electric double layers on the electrode surfaces.[35] In
addition, some recent theoretical studies on the Pt(111) surface
showed that the ORR activity and mechanism also highly
depend on the reaction conditions, e.g., potential.[36] All of these
theoretical studies suggest that the identification of clear ORR
catalyst active sites and reaction mechanism is limited by the
current theoretical methods and studies. More advanced and
detailed theoretical calculations would be needed in the future
studies.
Furthermore, Table 1 presents the most stable adsorption
configurations and the corresponding Pt-O bond lengths as well
as the adsorption free energies (ΔGads) of the key intermediates
in the Mechanism I, e.g., O* and OH* species. Clearly, the most
stable adsorption site of the O* species is the fcc site on the
Pt(111) surface, while the bridge site on the Pt(100) and Pt(211)
surfaces; that of the OH* species is the top site on the Pt(111)
and Pt(100) surfaces, while the bridge site on the Pt(211)
surface. As also seen in Figure 1h, at the ORR equilibrium
potential, the formation of O* on the Pt(111) surface compared
Table 1. Structural and energetic parameters of O* and OH* species on
Pt(111), Pt(100) and Pt(211) surfaces.
Recently, we have developed a simple yet effective method of
model calculations to discriminate the dominant active sites for
the Pt/CNT and Ru/CNT catalyzed hydrolytic dehydrogenation of
ammonia borane [
37,38]
as well as Pd/CNT catalyzed selective
hydrogenation of acetylene.[
39]
As a consecutive effort, we
further employed this method to discriminate the dominant active
sites for the ORR over Pt/CNT catalysts. For this method, high
resolution transmission electron microscopy (HRTEM)
measurements were first carried out to analyze the Pt particle
shape and the type of the surfaces. As observed in Figure 2a-2c,
the Pt nanoparticles in most cases for the 2-10 wt% Pt/CNT
catalysts are prone to exist as the top slice of truncated
cuboctahedron, which is consistent with the previous
studies.[37,40] To determine the type of the top surface of the
above shape, we further carried out the fast Fourier transform
(
FFT) to find a mixture of (100) and (111) surfaces with the top
o
surface of (100) based on the 70.5 angle between (111)
surfaces as well as 54.8o angle between (111) and (100)
surfaces. As schematically shown in Figure 2d, a representative
shape of CNT supported Pt nanoparticle consists of (111), (100),
edge and corner atoms. Moreover, the numbers of (111), (100),
*
The higher is the ΔGads value, the weaker is the adsorption of the
intermediates on the Pt surfaces.
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201801521
FULL PAPER
Figure 2. (a-c) Typical HRTEM images and the corresponding FFT analyses of 2-10 wt% Pt/CNT nanocatalysts. (d) Schematic diagram of truncated
cuboctahedron for CNT supported Pt nanoparticle. (e) Number of atoms per mole of Pt for the specific sites and normalized TOFs of the catalysts as a function of
Pt particle size.
edge and corner atoms over the differently sized Pt/CNT catalysts
per mole of Pt could be estimated by the number of specific sites
of each particle times the number of particles. The
corresponding results are presented in Figure 2e.
indicates that the ORR on the last two catalysts predominately
produces water via the four-electron pathway, while that on the
first one partially proceeds via the two-electron pathway.
Correspondingly, the 2 wt% Pt/CNT ORR catalyst gives rise to
It is further assumed that the activity of each specific type of
active site is the same regardless of the Pt particle size,
because the three Pt/CNT catalysts exhibit similar Pt electronic
properties mentioned above. In other words, the difference in the
Pt catalyzed ORR activity mainly originates from the difference
in the number and type of the active sites. If one type of Pt
active site is responsible for the ORR activity, the kinetic current
of catalysts would increase linearly with the number of the
specific Pt active sites. In this regard, turnover frequency (TOF)
and normalized one of each type of active site, calculated from
the data in Figure 1g and 2e, should be constant. It can be seen
in Figure S4 and 2e that only when the Pt(111) surface acts as
the active sites, the corresponding TOFactive site and normalized
one appear to be almost constant. This suggests that the
Pt(111) surface is the dominating active sites for the ORR over
the Pt/CNT catalysts.
2 2
much higher H O yield (i.e., 18.13%) than the other two
catalysts, which are 4.44% and 3.01%, respectively. Combining
these results with the three catalysts with different Pt particle
sizes, it is suggested that there is size-dependent ORR
selectivity over the three catalysts.
It is well-known that the equilibrium potential for the ORR via a
four-electron pathway is 1.23 V, while that for the formation of
hydrogen peroxide via a two-electron pathway is 0.70 V.[
Considering that the OOH* has been demonstrated to be the
key decisive intermediate for the formation of hydrogen peroxide
via the two-electron pathway,[41] DFT calculations were
employed to understand the size-dependent ORR selectivity
over the three catalysts. Figure 3c shows the most stable
adsorption configurations of OOH* as well as the corresponding
lengths of the Pt-O and O-O bonds on the Pt(111), Pt(100) and
Pt(211) surfaces. Specifically, the most stable adsorption site of
the OOH* species is the top site on the Pt(111) and Pt(100)
surfaces, while the bridge site on the Pt(211) surface. As seen in
Figure 3d, the adsorption free energies of the OOH* (ΔGOOH*) on
the Pt(111) surface is observed to be higher than those on the
Pt(100) and Pt(211) surfaces, which suggests that the formation
of OOH* is lowly exothermic on the Pt(111) surface compared to
the other two surfaces.
41]
Mechanistic understanding of size-dependent ORR
selectivity
In addition to the electrocatalytic ORR activity, the ORR
selectivity is another important factor in the evaluation of the
Pt/CNT ORR electrocatalysts. Figure 3a shows rotating ring-disk
electrode (RRDE) polarization curves of the three Pt/CNT
catalysts by the rotating ring and disk electrode testing. Based
on the data in Figure 3a, the average number of electron
Furthermore, Figure 3e presents the calculated energy
profiles on the Pt(111), Pt(100) and Pt(211) surfaces at 0.7 V.
Clearly, the Pt(111) surface is more close to the ideal
catalytically active sites for the oxygen reduction to hydrogen
peroxide than the Pt(100) and Pt(211) surfaces. This suggests
2 2 2
transferred (n per O ) and the average yield of H O (%) for the
three Pt/CNT catalysts were obtained. As shown in Figure 3b,
the average numbers of electron transferred of the 2, 8 and 10
wt% Pt/CNT catalysts are 3.64, 3.91 and 3.94, respectively. This
that the Pt(111) surface is more likely for the formation of H
2 2
O
compared to the Pt(100) and Pt(211) surfaces.
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201801521
FULL PAPER
Figure 3. (a) RRDE polarization curves, (b) the average n per O
2
and the average yield of H
O
2 2
(%) of the three Pt/CNT catalysts. (c) Stable adsorption
configurations of OOH*, (d) the corresponding ΔGOOH* and (e) the free energy profiles for hydrogen peroxide mechanism at 0.7 V on Pt(111), Pt(100) and Pt(211)
surfaces.
Based on all the above analyses, the size-dependent ORR
activity and selectivity over the Pt/CNT catalysts were
understood in detail by a combination of the experiments, model
calculations and DFT calculations. The Pt(111) surface is
suggested as the dominant active sites not only for the oxygen
reduction to water, but also for the oxygen reduction to hydrogen
peroxide. It is noted that for the 2 wt% Pt/CNT catalyst, it
while the lower Pt binding energy for the formation of OOH* and
thus the formation of hydrogen peroxide. These results could
support the above deduction that the Pt catalyst with the lower
Pt binding energy would facilitate the formation of hydrogen
peroxide.
2 2
exhibits much higher H O yield than the other two catalysts with
the high Pt loadings in Figure 3. Although the three catalysts
show similar Pt electronic properties, they have different Pt
particle size distributions. Especially for the catalyst with the low
Pt loading, there exist relatively more small-sized Pt
nanoparticles immobilized on the carbon nanotubes support in
Figure 1. Our previous studies have demonstrated that when
using Pt loading to change the Pt particle size for the Pt/CNT
catalysts, the smaller-sized Pt nanoparticles exhibit much lower
Pt binding energy.[42] These Pt nanoparticles with the lower Pt
binding energy might be favorable for the formation of hydrogen
peroxide.
Along this line, we further probed the effects of the Pt
electronic properties in Figure S5 over the three catalysts on the
average number of electron transferred and the average yield of
H
2
O
2
of the ORR, where these three catalysts exhibit similar Pt
particles sizes in Figure S6 (i.e., 2.1 ± 0.3, 2.2 ± 0.4 and 2.0 ±
.4 nm) and their more preparation and characterization details
Figure 4. The average n per O
2 2 2
and average yield of H O (%) of the three
catalysts with different Pt0 4f7/2 binding energy.
0
are presented in Supporting Information. As clearly seen in
Figure 4 and S7, increasing the Pt0 4f7/2 binding energy gives
rise to higher average number of electron transferred and lower
Notably, although in our work the Pt(111) surface is found to
act as the dominant active sites for both the oxygen reduction to
water (main reaction) and that to hydrogen peroxide (side
reaction), their reaction rates are remarkably different. According
H
2
O
2
yield. This could be because the higher Pt binding energy
is favorable for the oxygen adsorption and thus its dissociation,
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201801521
FULL PAPER
to our above DFT and experimental studies, on the Pt catalysts,
the ORR process to form water in acidic medium have been
shown to mainly proceed by means of the dissociative
mechanism. In other words, it is much easier to dissociate the
oxygen molecule to O* species rather than to form the OOH*
species.
inductively coupled plasma atomic emission spectrometer (Agilent 725-
ES ICP-AES).
Catalytic measurements: The as-obtained Pt/CNT catalysts were
evaluated in a single-compartment conventional three-electrode cell at
room temperature and ambient pressure using electrochemical
measurements (RRDE-3A, Japan) on
a
computer controlled
bipotentiostat CHI 760E (Shanghai Chenhua Co., China). A rotating
2
glassy carbon disk electrode of 5 nm diameter (0.196 cm ) coated by a
Conclusions
catalyst film was used as the working electrode. Pt wire and saturated
calomel electrode (AGCL) were used as the counter electrode and
4
reference electrode respectively, with 0.1 M HClO solution aqueous
In summary, we have clarified the underlying nature of the size-
dependent ORR activity and selectivity over the Pt/CNT
catalysts from the catalyst active site point of view by combining
the experiments, DFT calculations with model calculations. With
employing the model of truncated octahedron for Pt nanoparticle,
the observed size-insensitive TOFactive site based on the number
of Pt(111) atoms suggests Pt(111) as the dominant active sites
for the ORR. The origin of the size-dependent ORR activity is
mainly attributed to the difference in the Pt geometric properties
rather than the Pt electronic properties. Moreover, the Pt(111)
surface is also suggested as the dominant active sites for the
solution used as electrolyte. The catalyst ink was prepared by
ultrasonically dispersing 5 mg fine catalyst powder in 1 mL absolute
ethanol and 50 µL 5 wt% Nafion solution for 20 min. 10 µL of this ink was
deposited onto the surface of glassy carbon disk electrode followed by air
drying to gain a Pt/CNT catalyst loading of ca. 240 µg cm-2.
The ORR activity was evaluated by CV at a scan rate of 100 mV s-1 and
LSV at a scan rate of 5 mV s-1 in O
2 4
-saturated 0.1 M HClO . The
potential scan range was set between -0.25 and 0.95 V referred to that of
the Ag/AgCl potential, and the potentials reported in this study were all
normalized to that of the reversible hydrogen electrode (RHE) according
to the following equation:
2 2 2 2
formation of H O . The difference in the H O selectivity mainly
originates from the difference in the Pt binding energy. These
results could be valuable for the design and optimization of Pt-
based electrocatalysts for the ORR.
1
1
1
1
1
(
1)




0.5
j
jk jd jk B
2
3
1
6

B  0.2nFC (D ) 
(2)
O
O
where j is the measured current density, j
density and the diffusion current density, respectively. ω is the rotating
k
and j
d
are the kinetic current
Experimental Section
rate of rotating disk electrode (400, 625, 900, 1225, 1600, 2025 rpm). n is
transferred electron number, F is the Faraday constant (96485 C mol ),
Catalyst preparation: Multi-walled carbon nanotubes (CNT, purchased
from Beijing Cnano Technology Limited) were used to support Pt
catalysts with different Pt contents (i.e., 2, 8 and 10 wt%) by incipient
-1
= 1.26×10-3 mol/L), D
= 1.93x10 cm /s), v is the kinematic viscosity of
the electrolyte (v = 1.009x10 cm /s). The constant 0.2 is adopted when
the rotation speed is expressed in rpm.
C
O
is the bulk concentration of O
2
(C
O
O
is the O
2
-5
2
diffusion coefficient (D
O
wetness impregnation with an aqueous solution of H
2
PtCl
6
(Sinopharm
-2
2
Chemical Reagent Co. Ltd). The precursors of Pt/CNT catalysts were
o
dried in stagnant air at room temperature for 12 h, then at 80 C for 12 h,
2 2
/N
at 250 oC for 3 h, and passivating in
followed by reducing in 10% H
% O /Ar at room temperature for 30 min. The as-prepared catalysts
were denoted as 2 wt% Pt/CNT, 8 wt% Pt/CNT and 10 wt% Pt/CNT. The
pristine CNT were treated by acid oxidation of 8 M HNO (65%, Shanghai
Lingfeng Chemicals) mixed with 8 M H SO (98%, Shanghai Lingfeng
The TOF of the active site,[8] e.g., Pt(111) site, is calculated by
normalizing cathode kinetic current to the number of Pt(111) atoms over
differently sized Pt/CNT catalysts, as following:
1
2
3
2
4
1
1
19
Chemicals) in an ultrasonic bath of 60 oC for 2 h, and acid oxidation
followed by high temperature, respectively. The as-obtained samples
were labelled as CNT-O and CNT-D. The Pt/CNT-D catalysts were
prepared using the same procedure mentioned above.
(3)
TOFactive site (esite s )  j  S / (1.610 ) / Nactive site
k
disk
Where j
k
is the kinetic current density at 0.7 V, Sdisk is the area of glassy
carbon disk electrode of 5 nm diameter.
Catalyst characterization: The morphology of CNT was characterized
by Field emission scanning electron microscope (SEM, Nova NanoSEM
The TOFactive site is normalized to the highest TOFmax for each Pt/CNT
catalyst, which can be expressed by
450, US) and transmission electron microscope (TEM, JEOL JEM-1400,
Japan). The size distribution and the mean particle size of all catalysts
were calculated on the basis of the sizes of at least 300 random particles
by HAADF-STEM (Tecnai G2 F20 S-Twin) images. The shape of Pt
particles was observed by HRTEM (JEOL JEM-2100, Japan). Nitrogen
NormalizedTOFactive site  TOFactive site / TOFmax
(4)
For the calculation of electron transfer number (n per O
2
) and yield of
2 2
H O on different catalysts, based on both ring and disk currents from
adsorption/desorption isotherms were measured at 77
K with a
rotating ring-disk electrode (RRDE) of
estimated by the following equations:
4
nm diameter, they were
Quadrasorb SI analyzer. The pore size distributions were derived from
the desorption branch by using the Barrett-Joyner-Halenda (BJH) model.
The High-resolution core-level (CL) and valence-band (VB) X-ray
4
id
(5)
(6)
photoelectron spectroscopy (XPS) data were collected with
a
n 
i  i / N
d
r
ThermoFisher ESCALAB250Xi equipped with an Al Kα X-ray (1486.6 eV,
excitation source working at 15 kV). The analyzer was in the constant
analyzer energy (CAE) mode at a pass energy of 30 eV for all the
valence-band XPS measurements. The binding energies were measured
with an accuracy of 0.05 eV. The C 1s peak at 284.6 eV was taken as an
internal standard to correct the shift in the binding energy caused by
sample charging. The elements of the samples were analyzed by
ir / N
H O %  200
2
2
i  i / N
d r
where i
the current collection efficiency of the Pt ring disk determined using a
solution of Fe(CN) -
(N=0.42).
d r
is the disk current density, i is the ring current density and N is
3
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201801521
FULL PAPER
Computational details
Acknowledgments
Method: The density functional theory (DFT) simulations of the
adsorption and reaction process involved in ORR on Pt surfaces were
performed by using the Vienna ab initio simulation package (VASP).[43-46]
The ion-electron interaction was described with the projector augmented
This work was financially supported by the Natural Science
Foundation of China (21776077), the Shanghai Natural Science
Foundation (17ZR1407300 and 17ZR1407500), the Program for
Professor of Special Appointment (Eastern Scholar) at Shanghai
Institutions of Higher Learning, the Shanghai Rising-Star
Program (17QA1401200), the State Key Laboratory of Organic-
Inorganic Composites (oic-201801007) and the Open Project of
State Key Laboratory of Chemical Engineering (SKLChe-15C03).
We thank the use of XPS Facility within Analytical Centre of
Dalian Institute of Chemical Physics.
method.[
47,48]
wave (PAW)
Electron exchange-correlation was
represented by the functional of Revised Perdew, Burke, and Ernzerhof
RPBE) of generalized gradient approximation (GGA).[49,50] A cutoff
(
energy of 400 eV was used for the plane-wave basis set. The Hellman-
Feynman forces on each ion was minimized to be less than 0.03 eV/Å.
Models: The Pt(111), Pt(100) were respectively modelled using a p(3 x
3) supercell slab model, containing four atomic layers slabs, with a
relaxation of the top two layers. The Brillouin-zone integration has been
performed with a 3 x 3 x 1 Monkhorst-Pack k-point mesh[51,52] for the
above models. For Pt(211), a p(1 x 3) supercell with ten atomic layers
was chosen as the step model, including a relaxation of the top six layers.
The Brillouin-zone integration has been performed with a 4 x 3 x 1
Monkhorst-Pack k-point mesh for Pt(211) model. For all surfaces, a
vacuum thickness of 12 Å was used to avoid the interaction from the top
supercells. All of the surface models were constructed based on bulk fcc
Pt crystal whose lattice constant was 3.98 Å.
Keywords: Sustainable chemistry • Catalysis • Platinum • Active
site • Density functional calculations
[
[
[
[
[
1]
2]
3]
4]
5]
Z. W. Seh, J. Kibsgaard, C. F. Dickens, I. Chorkendorff, J. K. Nørskov, T.
F. Jaramillo, Science 2017, 355, eaad4998.
P. Z. Chen, Y. Tong, C. Z. Wu, Y. Xie, Acc. Chem. Res. 2018, 51, 2857-
2
866.
D. Y. Chung, J. M. Yoo, Y. E. Sung, Adv. Mater. 2018, 30, 1704123-
704143.
Y. Jiao, Y. Zheng, M. Jaroniec, S. Z. Qiao, Chem. Soc. Rev. 2015, 44,
060-2086.
1
Free energy calculations: The overpotential during the ORR process
can be determined by examining the reaction free energies of the
different elementary steps.[53] In order to obtain the reaction free energy
of each elementary step of ORR, we calculated the adsorption free
energy of O*, OH* and OOH*. Since it is difficult to obtain the exact free
energy of O, OH and OOH radicals in the electrolyte solution, the
adsorption free energy ΔGads of O*, OH* and OOH* are relative to the
2
R. Chattotꢀ, O. L. Bacq, V. Beermann, S. Kühl, J. Herranz, S. Henning, L.
Kühn, T. Asset, L. Guétaz, G. Renou, J. Drnec, P. Bordetꢀ, A. Pasturelꢀ,
A. Eychmüller, T. J. Schmidtꢀ, P. Strasser, L. Dubau, F. Maillard, Nat.
Mater. 2018, 17, 827-833.
[
6]
M. F. Li,1 Z. P. Zhao, T. Cheng, A. Fortunelli, C. Y. Chen, R. Yu, Q. H.
Zhang, L. Gu, B. Merinov, Z. Y. Lin, E. Zhu, T. Yu, Q. Y. Jia, J. H. Guo,
L. Zhang, W. A. Goddard III, Y. Huang, X. F. Duan, Science 2016, 354,
1414-1419.
free energy of stoichiometrically appropriate amounts of H
g), defined as follows:
GO*  G(H O(g) *  O*H (g))  O*
2 2
O (g) and H
(


H2

H2O

*
2
2
(7)
[
[
7]
8]
L. Su, W. Jia, C. M. Li, Y. Lei, ChemSusChem 2014, 7, 361-378.
Y. H. Bing, H. S Liu, L. Zhang, D. Ghosh, J. J. Zhang, Chem. Soc. Rev.
2010, 39, 2184-2202.

(EO*  EH2  EH2O  E ) T (S  SH2  SH2O  S* )
*
O*
(EZPE(O*)  EZPE(H2 )  EZPE(H2O） EZPE(*)

)
[
9]
N. Jung, D. Y. Chung, J. Ryu, S. J. Yoo, Y. E. Sung, Nano Today 2014
9, 433-456.

GOH*  G(H O(g) *  OH *1/ 2H (g))  
 0.5 H2  H2O  *
2
2
OH*
(8)

(EOH*  0.5 EH2  EH2O  E ) T  (S  0.5 SH2  SH2O  S* )
[10] V. R. Stamenkovic, B. Fowler, B. S. Mun, G. F. Wang, P. N. Ross, C. A.
*
OH*
Lucas, N. M. Markovic, Science 2007, 315, 493-497.

(EZPE(OH*)  0.5 EZPE(H2 )  EZPE(H2O） EZPE(*) )
[
11] A. Kulkarni, S. Siahrostami, A. Patel, J. K. Nørskov, Chem. Rev. 2018,
18, 2302−2312.
1

GOOH*  G(2H O(g) *  OOH *3 / 2H (g))
2
2
[
12] A. M. G. Marín, R. Rizoa, J. M. Feliu, Catal. Sci. Technol. 2014, 4, 1685-


OOH* 1.5 H2  2 H2O  *
(
9)
1698.
(EOOH* 1.5 EH2  2 EH2O  E* )
T (SOOH* 1.5 SH2  2 SH2O  S* )
(EZPE(OOH*) 1.5 EZPE(H2 )  2 EZPE(H2O） EZPE(*)
[13] S. Sui, X. Y. Wang, X. T. Zhou, Y. H. Su, S. Riffatc, C. J. Liu, J. Mater.
Chem. A 2017, 5, 1808-1825.

[
[
14] Y. Nie, L. Li, Z. D. Wei, Chem. Soc. Rev. 2015, 44, 2168-2201.

)
15] J. X. Wang, N. M. Markovic, R. R. Adzic, J. Phys. Chem. B 2004, 108,
4
127-4133.
16] Y. J. Deng, G. K. H. Wiberg, A. Zana, M. Arenz, Electrochim. Acta 2016,
04, 78-83.
For each elementary step, the reaction free energy was defined as the
difference between free energies of the final and initial states and was
given by the following expression:
[
[
2
17] M. Nesselberger, S. Ashton, J. C. Meier, I. Katsounaros, K. J. J.
Mayrhofer, M. Arenz, J. Am. Chem. Soc. 2011, 133, 17428-17433.
18] K. Kinoshita, J. Electrochem. Soc. 1990, 137, 845-848.

G  E  ZPE TS  G +G  Gfield
(10)
U
pH
[
[
19]
I. Katsounaros, S. Cherevko, A. R. Zeradjanin, K. J. J. Mayrhofer,
Where ΔE is the difference of adsorption energy between product and
Angew. Chem. Int. Ed. 2014, 53, 102-121.
reactant molecules adsorbed on catalyst surface, obtained from DFT
[
20] M. H. Shao, A. Peles, K. Shoemaker, Nano Lett. 2011, 11, 3714-3719.
21] A. Anastasopoulos, J. C. Davies, L. Hannah, B. E. Hayden, C. E. Lee, C.
Milhano, C. Mormiche, L. Offin, ChemSusChem 2013, 6, 1973-1982.
22] J. Greeley, J. Rossmeisl, A. Hellman, J. K. Norskov, Z. Phys. Chem.
2007, 221, 1209-1220.
U
calculations. ΔG = eU, where U is the electrode potential relative to the
standard hydrogen electrode. T is the temperature, and e is the charge
transferred. ΔZPE is the change of zero-point energy computed by DFT
[
[
calculations. ΔS is the entropy change from the DFT calculations. ΔGpH is
the correction of the H+ free energy by the concentration dependence of
+
[23] N. Markovic, H. Gasteiger, P. N. Ross, J. Electrochem. Soc. 1997, 144,
591-1597.
the entropy: ∆GpH=-k
B
Tln[H ]. k
B
is Boltzmann constant. ΔGfield is the
1
electric field in the double layer. We neglect this in the work. The free
+
-
Tln[H+].
[24] C. H. Choi, M. Kim, H. C. Kwon, S. J. Cho, S. Yun, H. T. Kim, K. J. J.
energy of H + e is calculated by 1/2 GH2 -eU+k
B
Mayrhofer, H. Kim, M. Choi, Nat. Commun. 2016, 7, 10922-10931.
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201801521
FULL PAPER
[
[
[
25]
Y. H. Chung, D. Y. Chung, N. Jung, Y. E. Sung, J. Phys. Chem. Lett.
013, 4, 1304-1309.
2
26] Y. Peng, B. Z. Lu, N. Wang,a L. G. Li, S. W. Chen, Phys. Chem. Chem.
Phys. 2017, 19, 9336-9348.
27] V. Stamenkovic, B. S. Mun, K. J. J. Mayrhofer, P. N. Ross, N. Markovic,
J. Rossmeisl, J. Greeley, J. K. Nørskov, Angew. Chem. 2006, 118,
2
963-2967.
28] Z. Y. Zhou, X. W. Kang, Y. Song, S. W. Chen, J. Phys. Chem. C 2012,
16, 10592-10598.
29] G. A. Tritsaris, J. Greeley, J. Rossmeisl, J. K. Nørskov, Catal. Lett. 2011,
41, 909-913.
[
[
[
[
1
1
30] L. C. Bai, X. Wang, Q. Chen, Y. F. Ye, H. Q. Zheng, J. H. Guo, Y. D. Yin,
C. B. Gao, Angew. Chem. Int. Ed. 2016, 55, 15656-15661.
31] G. Zhou, L. Barrio, S. Agnoli, S. D. Senanayake, J. Evans, A. Kubacka,
M. Estrella, J. C. Hanson, A. M. Arias, M. F. Garcia, J. A. Rodriguez,
Angew. Chem. Int. Ed. 2010, 49, 9680-9684.
[
[
[
[
32] J. Rossmeisl, A. Logadottir, J. K. Nørskov, Chem. Phys. 2005, 319, 178-
184.
33] J. K. Nørskov, J. Rossmeisl, A. Logadottir, L. Lindqvist, J. Phys. Chem.
B 2004, 108, 17886-17892.
34] L. H. Ou, F. Yang, Y. W. Liu, S. L. Chen, J. Phys. Chem. C 2009, 113,
2
0657-20665.
35] L. Yu, X. L. Pan, X. M. Cao, P. Hu, X. H. Bao, J. Catal. 2011, 282, 183-
90.
36] J. A. Keith, T. Jacob, Angew. Chem. Int. Ed. 2010, 49, 9521-9525.
1
[
[
37]
W. Y. Chen, J. Ji, X. Feng, X.Z. Duan, G. Qian, P. Li, X. G. Zhou, D.
Chen, W. K. Yuan, J. Am. Chem. Soc. 2014, 136, 16736-16739.
[
[
[
[
38] W. Z. Fu, W. Y. Chen, G. Qian, D. Chen, W. K. Yuan, X. G. Zhou, X. Z.
Duan, React. Chem. Eng. 2019, 4, 316-322.
39] Y. Q. Cao, W. Z Fu, Z. J. Sui, X. Z. Duan, D. Chen, X. G. Zhou, Ind. Eng.
Chem. Res. 2019, 58, 1888-1895.
40] B. Lim, M. J. Jiang, P. H. C. Camargo, E. C. Cho, J. Tao, X. M. Lu, Y. M.
Zhu, Y. N. Xia, Science 2009, 324, 1302-1305.
41] S. Siahrostami, A. V. Casadevall, M. Karamad, D. Deiana, P. Malacrida,
B. Wickman, M. E. Escribano, E. A. Paoli, R. Frydendal, T. W. Hansen,
I. Chorkendorff, I. E. L. Stephens, J. Rossmeisl, Nat. Mater. 2013, 12,
1137-1153.
[42] W. Y. Chen, J. Ji, X. Z. Duan, G. Qian, P. Li, X. G. Zhou, D. Chen, W. K.
Yuan, Chem. Commun. 2014, 50, 2142-2144.
[
[
[
[
[
[
[
43] G. Kresse, J. Furthmüller, Phys. Rev. B 1996, 54, 11169.
44] G. Kresse, J. Hafner, Phys. Rev. B 1993, 47, 558.
45] G. Kresse, J. Furthmüller, Comput. Mater. Sci. 1996, 6, 15-50.
46] G. Kresse, J. Hafner, Phys. Rev. B 1994, 49, 14251.
47] P.E. Blöchl, Phys. Rev. B 1994, 50, 17953.
48] G. Kresse, D. Joubert, Phys. Rev. B. 1999, 59, 1758-1775.
49] J. P. Perdew, K. Burke, M. Ernzerhof, Phys. Rev. Lett. 1996, 77, 3865-
3
868.
50] B. Hammer, L. B. Hansen, J. K. Nørskov, Phys. Rev. B 1999, 59, 7413-
421.
[
7
[51] M. Methfessel, A. T. Paxton, Phys. Rev. B. 1989, 40, 3616-3621.
[52] H. J. Monkhorst, J. D. Pack, Phys. Rev. B. 1976, 13, 5188-5192.
[53] H. X. Xu, D. J. Cheng, D. P. Cao, X. C. Zeng, Nat. Catal. 2018, 1, 339-
348.
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201801521
FULL PAPER
Entry for the Table of Contents
FULL PAPER
The size-dependent ORR activity and
selectivity over Pt/CNT catalysts have
been clarified by combining the
experiments, DFT calculations and
model calculations. The Pt(111)
surface, especially for the catalyst with
the higher Pt binding energy, is
favourable for the oxygen reduction to
Oxygen Reduction*
Jie Gan, Wei Luo, Wenyao Chen,
Jianing Guo, Zhonghua Xiang, Bingxu
Chen, Fan Yang, Yunjun Cao, Fei Song,
Xuezhi Duan,* Xinggui Zhou
Page No. 1– Page No. 8
H
2
O.
Mechanistic Understanding of Size-
dependent Oxygen Reduction Activity
and Selectivity over Pt/CNT
Nanocatalysts
*one or two words that highlight the emphasis of the paper or the field of the study
This article is protected by copyright. All rights reserved.