Oxygen Vacancy Structure Associated Photocatalytic Water Oxidation of BiOCl

Article abstract of DOI:10.1021/acscatal.6b02613

A central issue in understanding photocatalytic water splitting on a stoichiometric or defective nanostructured oxide surface is its adsorption mode and related reactivity. More than just improving the adsorption of water on oxide surfaces, we demonstrate in this work that surface oxygen vacancies (OVs) also offer a possibility of activating water toward thermodynamically enhanced photocatalytic water oxidation, while the water activation state, as reflected by its capability to trap holes, strongly depends on the structures of OVs. Utilizing well-ordered BiOCl single-crystalline surfaces, we reveal that dissociatively adsorbed water on the OV of the (010) surface exhibits higher tendency to be oxidized than the molecularly adsorbed water on the OV of the (001) surface. Analysis of the geometric atom arrangement shows that the OV of the BiOCl (010) surface can facilitate barrierless O-H bond breaking in the first proton removal reaction, which is sterically hindered on the OV of the BiOCl (001) surface, and also allow more localized electrons transfer from the OV to the dissociatively adsorbed water, leading to its higher water activation level for hole trapping. These findings highlight the indispensable role of crystalline surface structure on water oxidation and may open up avenues for the rational design of highly efficient photocatalysts via surface engineering. (Chemical Equation Presented).

Full text of DOI:10.1021/acscatal.6b02613

Subscriber access provided by UNIV TORONTO
Article
Oxygen Vacancy Structure Associated
Photocatalytic Water Oxidation of BiOCl
Hao Li, Jian Shang, Huijun Zhu, Zhiping Yang, Zhihui Ai, and Lizhi Zhang
ACS Catal., Just Accepted Manuscript • DOI: 10.1021/acscatal.6b02613 • Publication Date (Web): 04 Nov 2016
Downloaded from http://pubs.acs.org on November 4, 2016
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
ACS Catalysis is published by the American Chemical Society. 1155 Sixteenth Street
N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 34
ACS Catalysis
1
2
3
4
5
6
7
8
9
Oxygen Vacancy Structure Associated Photocatalytic Water Oxidation of BiOCl
*
Hao Li, Jian Shang, Huijun Zhu, Zhiping Yang, Zhihui Ai, and Lizhi Zhang
Key Laboratory of Pesticide & Chemical Biology of Ministry of Education, Institute of
Environmental & Applied Chemistry, College of Chemistry, Central China Normal University,
Wuhan 430079, P. R. China
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
*
Eꢀmail: zhanglz@mail.ccnu.edu.cn. Phone/Fax: +86ꢀ27ꢀ6786 7535
1
ACS Paragon Plus Environment

ACS Catalysis
Page 2 of 34
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Abstract: A central issue in understanding the photocatalytic water splitting on stoichiometric or
defective nanostructured oxide surface is its adsorption mode and related reactivity. More than just
improving the adsorption of water on oxide surface, we demonstrate in this work that surface oxygen
vacancies (OVs) also offer a possibility to activate water toward thermodynamically enhanced
photocatalytic water oxidation, while the water activation state, as reflected by its capability to trap
holes, strongly depends on the structures of OVs. Utilizing wellꢀordered BiOCl singleꢀcrystalline
surfaces, we reveal that dissociatively adsorbed water on the OV of (010) surface exhibits higher
tendency to be oxidized than the molecularly adsorbed one on the OV of (001) surface. Analysis of
the geometric atom arrangement shows that the OV of BiOCl (010) surface can facilitate the
barrierless OꢀH bond breaking in the first proton removal reaction, which is sterically hindered on
the OV of BiOCl (001) surface, and also allow more localized electrons transfer from the OV to the
dissociatively adsorbed water, leading to its higher water activation level for hole trapping. These
findings highlight the indispensable role of crystalline surface structure on water oxidation, and may
open up avenues for the rational design of highly efficient photocatalysts via surface engineering.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Keywords: Water oxidation; Photocatalysis; BiOCl; Oxygen vacancies; Facet dependency
Introduction
Water chemistry on oxide surfaces is of great significance for both scientific and fundamental
applications including gas sensors, biomedical implants, batteries, fuel cells and heterogeneous
catalysis. Of particular interest is the photocatalytic water oxidation on semiconducting oxide
surfaces due to the urgent need for efficient environmental control and solar energy conversion,
whose relevant chemical processes taking place on the surfaces involve water acting as both the
2
ACS Paragon Plus Environment

Page 3 of 34
ACS Catalysis
1
ꢀ4
1
2
3
4
5
6
7
8
9
“spectator” (reaction medium) and the key participants. Typically, photocatalytic water oxidation
*
+
*
+
is initiated by the first proton removal reaction (H
2
O + hole ꢀ •OH + H ; the superscript *
denotes the adsorbed state), which involves both the OꢀH bond breaking and the trapping of a
5ꢀ7
*
photogenerated hole on the valence band (VB). Diffusion of the •OH species away from the
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2 2
surfaces allows the generation of free •OH radicals, which can be further coupled to generate H O .
1
*
Both •OH radicals and H O are the key components for the organic pollutants removal. Also, •OH
2
2
species can be further oxidized to O
2
via multihole transfer process, leaving electrons on the
+
conduction band (CB) capable of reducing H or CO
2
for solar fuel production, or fixing N
2
for
2
,8ꢀ10
ammonia synthesis.
Although the position of VB maximum, which determines the oxidizing capability of
photogenerated holes, has been long believed to govern the thermodynamics of photocatalytic water
oxidation, this point of view is not supported by recent theoretical and experimental results yet. For
2
example, photocatalytic water oxidation on many prototypical stoichiometric TiO surfaces is still
thermodynamically unfavorable even though the position of VB is adequately matched with the
water’s oxidation potential. From the surface catalysis standpoint, such unfavorableness is mainly
ascribed to the undesirable electronic state of surfaceꢀadsorbed water. Typically, surface
coordinating unsaturated (5ꢀcoordinated) Ti cations provide the most reliable sites for water
adsorption via the LewisꢀAcid interaction. Such an interaction usually results in the donation of lone
paired electrons from the O atom of adsorbed water to metal cations, decreasing the water’s overall
charge density. Consequently, this interfacial charge transfer behavior will drives the occupied states
2
of adsorbed water lower than the upper edge of TiO VB, impeding the transfer of photoinduced
11ꢀ15
holes to adsorbed water thermodynamically.
In view of the strong adsorption mode dependent
3
ACS Paragon Plus Environment

ACS Catalysis
Page 4 of 34
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
water oxidation on semiconductor surfaces, a key strategy to enhance the thermodynamics of water
oxidation is to regulate the waterꢀsurface interaction mode toward water activation for a better hole
15
trapping.
To realize this goal, we propose the introduction of native surface oxygen vacancies (OVs),
because OVs can significantly influence the geometric structure and electronic structure of oxide
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
6,17
surfaces, thus altering their chemical reactivity.
and CO , OVs can act as the direct coordinating sites for enhanced substrates adsorption, and also as
the charge donor sites capable of activating these molecules due to the abundant electrons localized
For the adsorption of inert molecules like O , N
2 2
2
18ꢀ22
at OVs.
To confirm whether water can be activated on the OVs with enhanced thermodynamics
toward oxidation, BiOCl is selected as the model catalyst because its photoreactivity is closely
related to OVs, which can be in situ generated under UV or solar light because the surface BiꢀO
2
3,24
bonds of long length and low energy are weak.
Besides, the successful preparation of two
wellꢀordered, single crystalline surfaces allows the comparative study of water chemistry on OVs
25
with different structures possible.
4
ACS Paragon Plus Environment

Page 5 of 34
ACS Catalysis
1
2
3
4
5
6
7
8
9
Results and Discussion
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 1. Surface structure of different BiOCl surfaces with OVs. Schematic illustration of the (a)
inꢀplane and (b) outꢀplane relaxation over the atoms around the OV on the BiOCl (001) and (010)
2
surfaces, respectively. (c) Bi Lꢀedge extended XAFS oscillation function k χ(k) and (d) the
corresponding Fourier transforms for the BiOCl with and without OVs.
The prototypical surfaces of BiOCl are the (001) and (010) surface, whose surface atom
arrangements are different from each other. The (001) surface of BiOCl is of a closeꢀpacked
structure with exposure of high density O atoms. As the protons in acid solution (pH = 1) during the
synthesis of this surface are abundant, the adsorption of protons onto BiOCl (001) surface leads to
5
ACS Paragon Plus Environment

ACS Catalysis
Page 6 of 34
2
6, 27
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
the formation of a layer of hydroxyl groups (Figure S1a).
The (010) surface of BiOCl
synthesized in near neutral solution possesses an open channel structure with O, Bi and Cl atoms
exposed, in which the coordination of Bi atoms is unsaturated (Figure S1b). After the introduction of
an OV on these two surfaces, local atoms around the OV suffered from a distinct geometrical
distortion according to the density functional theory (DFT) calculation. Parallel to the (001) surface,
two nearest Bi (Bi , Bi ) atoms around the OV in the sublayer were relaxed toward the OV,
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
1 1 2 2
lengthening the corresponding Bi ꢀO (or Bi ꢀO ) bond from 2.28 Å to 2.36 Å (Figures 1a and S2a).
Differently, atoms relaxation perpendicular to the (001) surface was not observed and the average
bond length of BiꢀCl around the OV did not change significantly (Figure S2b). We called this atoms
relaxation behavior as inꢀplane relaxation. On the (010) surface, no atoms relaxed parallel to the
(
010) surface after the introduction of an OV (Figure S2c). However, perpendicular to the surface,
the nearest Bi atom relaxed downward, while a neighboring Cl relaxed upward (Figure 1b and S2d).
3
1
This different relaxation behavior, as characterized by the atoms moving along the open channel
perpendicular to the (010) surface, was denoted as outꢀofꢀplane relaxation. Interestingly, different
from the inꢀplane relaxation of Bi atoms around the OV on (001) surface, the outꢀofꢀplane relaxation
happened on the (010) surface shortened the Bi ꢀO (or Bi ꢀO ) bond length from 2.39 Å to 2.33 Å,
3
1
3
2
and the Bi
characterize these different atom relaxation behaviors, we performed Xꢀray absorption fine structure
XAFS) spectroscopy, which is a powerful elementꢀspecific tool to study local atomic arrangements.
1 1 2 1
ꢀCl (or Bi ꢀCl ) bond length from 2.93 Å to 2.82 Å. In order to experimentally
(
Single crystalline BiOCl nanosheets exposed with (001) or (010) surface (denoted as BOCꢀ001 or
BOCꢀ010) were prepared via a simple hydrothermal method (Figure S3). Their surface properties
were characterized by zeta potential measurement and electrochemical impedance spectroscopy
6
ACS Paragon Plus Environment

Page 7 of 34
ACS Catalysis
1
2
3
4
5
6
7
8
9
(Figure S4). UV light illumination was used to create OVs on the surfaces to obtain their
oxygenꢀdeficient counterparts (denoted as BOCꢀ001ꢀOV or BOCꢀ010ꢀOV) (Figure S5). The Bi
2
Lꢀedge extended XAFS k χ(k) oscillation curves revealed that the formation of OVs did not
significantly change the frequencies, but remarkably reduced the amplitude because of the surface
distortion (Figure 1c). This could be further explicitly manifested by corresponding Fourier
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
transform (FT) of the k χ(k) functions in Rꢀspace (Figure 1d). The FT curve of BOCꢀ001 was
characterized by two main peaks at 1.82 Å and 2.81 Å associated with the BiꢀO and BiꢀCl
coordination, respectively. For BOCꢀ001 with OVs, the BiꢀO peak position shifted to the higher R
direction by 0.04 Å, while the BiꢀCl peak position remained unchanged (Figure 1d). To obtain
quantitative structural parameters around the Bi atoms within BOCꢀ001ꢀOV, leastꢀsquare curve
parameter fitting of the BiꢀO and BiꢀCl peaks was performed (Table S1). The coordination number
for the surface interatomic BiꢀO pair (NBiꢀO) decreased from 4.0 to 3.1, while the surface interatomic
BiꢀCl pair (NBiꢀCl) value slightly changed from 4.0 to 3.9, suggesting the generation of OVs rather
than Cl vacancies on the (001) surface. Moreover, the increased surface BiꢀO bond length
corresponding to the highꢀR shift of BiꢀO peak and the almost unchanged BiꢀCl bond length were
consistent with the inꢀplane relaxation behavior of atoms around the OV of BiOCl (001) surface
(Table S1). Similar with the case of BOCꢀ001ꢀOV, surface NBiꢀO of BOCꢀ010ꢀOV reduced from 3.7
to 2.8, while the NBiꢀCl slightly decreased, indicating the presence of OVs. However, the peak
positions of BiꢀO and BiꢀCl of BOCꢀ010ꢀOV shifted to lower R direction by 0.07 Å and 0.09 Å,
respectively. These shifts indicated the shortening of both BiꢀO and BiꢀCl bond lengths, agreeing
well with the outꢀofꢀplane relaxation behavior of atoms around the OV of (010) surface (Table S1).
7
ACS Paragon Plus Environment

ACS Catalysis
Page 8 of 34
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
Figure 2. Theoretical study of water adsorption on the BiOCl (001) surface. (a) Adsorption of H O
2
on defectꢀfree BiOCl (001) surface. (b) Adsorption of H O on the BiOCl (001) surface with an OV
and (c) the corresponding charge density difference. The yellow and blue isosurfaces represent
charge accumulation and depletion in the space, respectively. For clarity, the symmetric parts of the
optimized slabs at the bottom are not shown. The isovalue is 0.005 au. Red dashed lines in the
schematic plots of the water adsorption structures show the ordinary HBs. ꢁE is the water adsorption
energy and the negative adsorption energy indicates the total energy decreases when water is
adsorbed. ꢁρ is the Bader charge change of the adsorbed water. For clarity, the calculated Bader
charge is the sum of Bader charge of three atoms of one water molecule and the positive values of
Bader charge change mean the transfer of electrons from the BiOCl surface to the adsorbed H O. (d)
2
The PDOS of water on (001) surface of BiOCl with different adsorption structures. For the sake of
presentation, the local contributions have been scaled by a factor of ten.
8
ACS Paragon Plus Environment

Page 9 of 34
ACS Catalysis
1
2
3
4
5
6
7
8
9
Successful elucidation of the specific structures of OVs on the two prototypical BiOCl surfaces
allowed us to testify our hypothesis whether OVs might serve as active sites for water activation
toward enhanced thermodynamics of oxidation. On the defectꢀfree BiOCl (001) surface, H
inner layer was mainly adsorbed via a hydrogenꢀbond (HB) network constructed by the interactions
between hydroxyl groups of adsorbed H O and those on the surface (Figure 2a). The representative
HB lengths (2.05 Å and 3.02 Å) were much longer than that (1.82 Å) of bulk water. This indicated
the weak interaction between adsorbed H O with (001) surface, being further confirmed by the small
2
O in the
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
2
adsorption energy of 0.11 eV and the absence of significant interfacial charge transfer according to
Bader charge calculation (Figure 2a). Different from the isolated hydroxyl groups on many TiO2
surfaces, hydroxyl groups on BiOCl (001) surface are rather immobile whose O atoms are strictly
confined to the surface lattice that hinders their efficient threeꢀdimensional relaxation. This
difference might explain the weak interaction between water and the BiOCl (001) surface. After the
introduction of an OV, H
replacing the missing bridging oxygen and chemically bonding to a 3ꢀcoordinated Bi atom (Bi3c
around the OV (Figure 2b). Compared with H O adsorbed on defectꢀfree (001) surface via pure HBs,
H O adsorbed on the OV was more exothermic by 0.42 eV. Interestingly, the molecular plane of
2 w
O was molecularly adsorbed at the OV site with its oxygen (O ) roughly
)
2
2
H
2
O, which was initially perpendicular to the (001) surface, performed reorientation around its
figure axis with the most stable orientation being parallel to the surface, as two relatively strong HBs
(
1.75 Å and 1.73 Å) were formed between the ꢀO H (ꢀO H ) group of adsorbed H O and adjacent
w a w b 2
bridging O atom (O
b
) (Figure 2b). We then adopted charge density difference to trace the origin of
O, as
these two HBs, and first found a distinct charge donation from OV to the adsorbed H
2
characterized by the depletion of electrons localized on the Bi3c near the OV and the accumulation of
9
ACS Paragon Plus Environment

ACS Catalysis
Page 10 of 34
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
electrons on O
w
w
(Figure 2c). Meanwhile, the accumulation of electrons on O enhanced the
hydrogenꢀdonor nature of ꢀO
(ꢀO ) and O provided a clear picture of these two HBs’ formation process (Figure 2c).
Bader charge calculation revealed a net charge transfer of 0.42 e from the defective BiOCl (001)
surface to the adsorbed H O, making the adsorbed H O negatively charged. To estimate the
energetics of H O with different adsorption modes to trap photogenerated holes, we examined the
w
H, and the appearance of charge localization in the region between
28
w
ꢀO H
a
w
H
b
b
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
2
2
partial density of states (PDOS) of the adsorbed H
surface via HB interaction (H 001ꢀHB) was embedded in the VB, approximately 1.76 eV more
negative than the VB edge (also VB maximum) (Figure 2d). However, after the introduction of an
OV, PDOS of OVꢀadsorbed H O (H 001ꢀOV) was shifted to the VB edge by 1.07 eV. This shift
narrowed the distance between PDOS of adsorbed H O and the VB edge from 1.76 eV to 0.69 eV.
2 2
O. PDOS of H O bound to the defectꢀfree (001)
2
O
2
2
O
2
We therefore proposed that OVs on BOCꢀ001 surface could provide the new sites to enhance the
water adsorption, and also activate the adsorbed water, enhancing the hole trapping by the adsorbed
water thermodynamically.
1
0
ACS Paragon Plus Environment

Page 11 of 34
ACS Catalysis
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
Figure 3. Theoretical study of water adsorption on the BiOCl (010) surface. (a) Adsorption of H O
on the defectꢀfree BiOCl (010) surface. (b) Adsorption of H O on the BiOCl (010) surface with an
2
2
OV and (c) the corresponding charge density difference. (d) The PDOS of H O on the (010) surface
of BiOCl with different adsorption structures. For the sake of presentation, the local contributions
have been scaled by a factor of ten.
As for the defectꢀfree (010) surface of BiOCl, both Bi and O atoms of unsaturated coordination
29ꢀ32
are supposed to act as reliable Lewis acid and base sites for the inner layer water adsorption.
Within expectation, H O was adsorbed molecularly with the O chemically bound to a (010) surface
3ꢀcoordinated Bi atom (Bi3c), which was 0.34 eV more exothermic than the H O adsorption on the
perfect (001) surface (Figure 3a). Meanwhile, H O exhibited reorientation with the molecular plane
2
w
2
2
being almost parallel to the (010) surface because of the HB interaction between ꢀOH
surface 3ꢀcoordinated O atom (O3c) with a moderate HB of 1.89 Å (Figure 3a). The formation of this
HB via sharing the H atom of adsorbed H O to electron donors of O3c slightly elongated the OꢀH
b 2
of H O and a
b
2
b
1
1
ACS Paragon Plus Environment

ACS Catalysis
Page 12 of 34
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
bond to 0.99 Å (Figure 3a). Similar with previous reports over H
of TiO surfaces, these Lewis acidꢀbase interactions remarkably decreased the electron density on
adsorbed H O, as reflected by a net charge transfer of 0.31 e from the adsorbed H O to BiOCl (010)
surface according to the Bader charge calculation. After the generation of an OV on the (010)
surface, the Bi3cꢀbound H O would diffuse to the OV, where spontaneous dissociation took place
with the ꢀO H group being pinned at the OV and the proton fragment ꢀH being transferred to a
2
O adsorption on 5ꢀcoordinated Ti
2
2
2
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
w
a
b
2
2
ꢀfold coordination unsaturated Bi atoms (Bi2c) around the OV (Figure 3b). Apparently, H O on the
OV of BiOCl (010) surface preferred to dissociate in comparison with that of (001) surface, which
might be explained as follows. First, the water dissociation on the OVs of oxide surfaces is highly
related to the sufficient charge transfer from the OV to the adsorbed water. This is because the OꢀH
3
3ꢀ36
bonds of water will not break without sufficient interfacial charge transfer.
isodensity contour plot of the corresponding charge density difference of H O adsorbed on the OV
of BiOCl (010) surface, being characterized by the significant charge depletion on Bi2c atoms around
the OV along with the charge accumulation on the dissociated ꢀO and ꢀH fragments, suggested
that the donation of electrons from the OV of (010) surface to H O was definitely more favorable
than that on the defective (001) surface (Figure 3c). Consistently, Bader charge change (0.89 e) of
dissociatively adsorbed H O was much more than that (0.42 e) of the molecularly adsorbed one,
As expected,
2
w
H
a
b
2
2
confirming the OV on the BiOCl (010) surface allowed more electrons donated to the dissociatively
adsorbed H O. Second, the water dissociation on the OV requires an available site to accept the
2
3
7ꢀ39
dissociated ꢀH fragment for the well stabilization of dissociated state.
surface, even though adsorbed H O could form strong HBs with bridging O atoms that significantly
elongated the OꢀH (or OꢀH ) bond from 0.97 to 1.00 Å, spontaneous dissociation was still
On the OV of BiOCl (001)
2
a
b
1
2
ACS Paragon Plus Environment

Page 13 of 34
ACS Catalysis
1
2
3
4
5
6
7
8
9
prohibited because each bridging O atom had already been bound to a proton, impeding its
40
acceptance of another proton because of steric hindrance (Figure 2b). As for the OV of BiOCl (010)
surface, Bi2c atom near the OV provided the suitable site to bind the ꢀH
a
fragment. We have to point
out here that the binding of dissociated ꢀH fragment to surface Bi2c atom is actually an unusual
b
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
water dissociation mode because the expected site for the dissociated proton transfer is the surface
O atom toward the formation of another ꢀOH group. This unexpected binding behavior is believed
3
c
b
to be a thermodynamic requirement for the surface energy minimization. Our previous work
revealed that the introduction of OVs on the BiOCl surface typically produced new states in the
bandgap, which were fully occupied with electrons along with the formation of dangling bonds on
26
the two Bi2c atoms near the OVs, making the defective surface highly unstable. Similar dangling
3 5
bonds on the coordination unsaturated cations were found on the Ta N (100) surface with nitrogen
vacancies, the Si terminated SiC (001) surface and the FeO (111) surface with OVs, all of which can
be effectively stabilized by the dissociative water adsorption with the ꢀH fragment being bound to
3
5, 41, 42
these unsaturated cations.
2
The corresponding H O adsorption energy of ꢀ1.24 eV on the OV of
BiOCl (010) surface, which was much more exothermic than that (ꢀ0.53 eV) of defective (001)
surface, confirmed the (010) surface with an OV was greatly stabilized by this kind of dissociative
water adsorption behavior. PDOS of molecularly adsorbed H
2
O on the BiOCl (010) surface via
O to (010)
Lewis acidꢀbase interactions (H 010ꢀBi3c) accompanied by the charge transfer from H
2
O
2
surface was deeply embedded in VB by about 2.07 eV (Figure 3d). However, the introduction of an
OV, which reversed the interfacial charge flow direction, remarkably upshifted the PDOS of
2 2
dissociatively adsorbed H O (H O010ꢀOV) to almost overlap the VB edge of BiOCl. This remarkable
1
3
ACS Paragon Plus Environment

ACS Catalysis
Page 14 of 34
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
change suggested the water activation towards oxidation on the OV of BiOCl (010) surface might be
more favorable than that on the defective (001) surface (Figure 3d).
Although the above theoretical calculations revealed that both the (001) and (010) surfaces of
BiOCl could realize water activation after the OV introduction, it remained unknown which
adsorption mode (molecular or dissociative) more favored the photocatalytic water oxidation,
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
considering that both the adsorption and electronic structures of OVꢀadsorbed H O were different on
various BiOCl surfaces. To answer this question, we would like to focus on the first proton removal
*
+
*
+
2
reaction (H O + hole ꢀ •OH + H ), which is considered as the key kinetic step of water
5
ꢀ7
oxidation. As aforementioned, this reaction involves two elementary steps, the OꢀH bond breaking
step that is typically regarded as a surfaceꢀcatalyzed process dependent on the surface local bonding
geometry, and the hole trapping step that is a lightꢀdriving process related to the electronic state of
6
,7
2
adsorbed water. We thus assumed the oxidation of dissociatively adsorbed H O on the OV of
BiOCl (010) surface was more favorable than the molecular adsorbed one on the OV of (001)
surface because of the following two reasons. First, the OꢀH bond breaking on the OV of BiOCl
(010) surface was a barrierless process for the dissociative H
molecular adsorption of H O on the OV of (001) surface, it was merely an activated process and
extra energy was then required to break one of the OꢀH bonds. Second, compared with the PDOS of
molecularly adsorbed H O on the OV of BiOCl (001) surface with a distance of 0.69 eV to the VB
edge, PDOS of dissociatively adsorbed H O on the OV of (010) surface almost overlapped with the
2
O adsorption. However, as for the
2
2
2
VB edge, suggesting its higher activation level and higher tendency to trap holes (Figure 2d and 3d).
The donation of electrons from the OV to the adsorbed water was undoubtedly responsible for the
water activation as we discussed above, while the activation level was highly dependent on the
1
4
ACS Paragon Plus Environment

Page 15 of 34
ACS Catalysis
1
2
3
4
5
6
7
8
9
amount of donated electrons. Bader charge calculations revealed that the dissociatively adsorbed
H
2
2
O accepted more electrons than the molecularly adsorbed H O on the OV of BiOCl surfaces,
which was basically consistent with the fact that dissociated hydroxyl was a better electron acceptor
1
3
than molecular water. Therefore, we believe that OVs on BiOCl surface offer a possibility to
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
energize the sluggish water oxidation through water activation, while the activation states related to
both the geometrical and electronic structure of adsorbed H O are highly dependent on the structures
2
of OVs.
Figure 4. Experimental study of the water adsorption on the BiOCl surfaces. (a) Designed reaction
cell for the FTIR signal recording. BiOCl samples, after being pressed into cylindrical pellets, were
put in the center of the reaction cell. The cover dome was equipped with three windows. Two of the
1
5
ACS Paragon Plus Environment

ACS Catalysis
Page 16 of 34
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
windows were made of ZnSe and permitted entry and exit of the detection infrared beam, whereas
the third (quartz) view window was for the observation as well as the transmission of a UV light.
The IR signal was in situ collected through a HgCdTe (MCT) detector along with the photoreactions.
(b) Representative FTIR spectra of the pristine, vacuumꢀtreated and subsequent UVꢀilluminated
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
BiOCl surfaces and the typical lowꢀtemperature EPR signal corresponding to in situꢀcreated OVs
(inset). (c) Dynamic change of the OV EPR signal intensity (solid lines) and normalized ꢀOH FTIR
signal intensity (dashed lines) of the BiOCl (001) and (010) surface along with the UV light
illumination. Recording of the surface FTIR spectra under certain reaction conditions over
BOCꢀ001ꢀOV (d) and BOCꢀ010ꢀOV (e). High resolution of the O1s (f) and Bi4f (g) spectra of the
BiOCl surfaces containing OVs before and after exposure to water vapor.
To verify these theoretical calculation results of the OVꢀstructureꢀdependent water activation, it
is first necessary to experimentally determine the different water adsorption structures, which is
extremely challenging especially for the study of dissociative adsorption mode. Common methods to
characterize the water adsorption structure on oxide surface are based on scanning tunneling
30,34,39,43,44
microscope and temperatureꢀprogrammed desorption.
In this study, we for the first time
adopted FTIR spectroscopy to study the water adsorption behavior. Since surface hydroxyl groups
are often interfered by the trace amount of adsorbed water via their strong interactions, FTIR is used
to work in tandem with vacuum surface science in a designed reaction cell equipped with facilities
for sample heating, cooling and for the reaction atmosphere controlling (Figure 4a). The FTIR
spectra of untreated BiOCl (BOCꢀ001 or BOCꢀ010) displayed a rising and broad band spanning
ꢀ1
3
250ꢀ3750 cm arisen from the water hydrogenꢀbonded hydroxyl stretching (Figure 4b). However,
ꢀ6
after the thermal treatment (120 min at 380 K) under high vacuum (~10 mbar) to remove most of
1
6
ACS Paragon Plus Environment

Page 17 of 34
ACS Catalysis
1
2
3
4
5
6
7
8
9
the initially surfaceꢀadsorbed molecular water, this feature largely disappeared, accompanying with
ꢀ1
the appearance of a discrete and narrow vibration around 3500 cm corresponding to the surface
hydroxyl groups (Figure 4b). When the whole reaction cell was cooled to room temperature, a UV
light (λ = 254 nm) was then applied to shine the sample through the quartz view window to in situ
create OVs on the BiOCl surfaces. After irradiation, both BOCꢀ001 and BOCꢀ010 exhibited a typical
OV signal according to the low temperature electron paramagnetic resonance (EPR) spectra (inset of
Figure 4b). Meanwhile, the amount of surface hydroxyl groups (dashed lines) over BOCꢀ001 and
BOCꢀ010 gradually decreased along with the OVs’ concentration increase (solid lines) (Figure 4c).
After the in situ formation of OVs, waterꢀsaturated Ar gas was then introduced to the reaction cell to
maximize adsorption of water onto the BiOCl surfaces. After a sufficient static reaction between the
adsorbed water and OVs of BiOCl in the dark, the reaction cell was reꢀevacuated via thermal and
vacuum treatment. This treatment was supposed to desorb the subsequentlyꢀpumped gaseous water,
but left behind the possible newlyꢀformed surface hydroxyl groups arisen from water dissociation at
the OVs, which allowed the direct visualization of the hydroxyl groups’ change. After the reaction
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
between OVs and H O, both the intensity and the position of surface hydroxyl absorption band on
BOCꢀ001ꢀOV exhibited no obvious change, but the absorption band intensity was significantly
enhanced in the case of BOCꢀ010ꢀOV (Figure 4d and 4e). This difference suggested that water was
molecularly adsorbed on the OVs of BOCꢀ001, which could be easily removed by the thermal and
vacuum treatment, but irreversibly dissociated on the OVs of BOCꢀ010 to generate new hydroxyls
covering OVs, resulting in the increase of overall surface hydroxyls density. Moreover, the band
ꢀ1
position of surface hydroxyls on BOCꢀ010ꢀOV slightly shifted from 3572 cm to a higher
ꢀ1
wavenumber of 3599 cm (Figure 4d). This shift was attributed to the increased electron density on
1
7
ACS Paragon Plus Environment

ACS Catalysis
Page 18 of 34
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
the O
w
atom caused by charge transfer from the OV of BiOCl (010) surface to newlyꢀformed
45
dissociated ꢀO
When D O vapor was used instead of H
stretching emerged at 2655 cm on BOCꢀ010ꢀOV, which was stably observed up to 550 K (Figure
e). This clearly demonstrated the dissociative adsorption of D O on the OVs of BiOCl (010) surface.
w a w a
H fragment, leading to a higher O ꢀH vibrational frequency (Figure 3b and 3c).
2
2
O vapor, new absorption band corresponded to the ꢀOD
ꢀ1
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
4
2
The increased density of surface hydroxyl groups arisen from the water dissociation on the OVs of
BOCꢀ010 was also confirmed by the Xꢀray photoelectron spectroscopy (XPS). The O1s spectra of
BOCꢀ001ꢀOV showed a broader shoulder with a chemical shift of 1.5 eV relative to the oxide O 1s
peak, being attributed to surface hydroxyl groups. After the water vapor exposure, the intensity of
this shoulder changed slightly, but was remarkably enhanced in the case of BOCꢀ010ꢀOV (Figure 4f).
This enhancement corresponded to the increase of surface hydroxyl groups originated from the water
dissociation on OVs of BiOCl (010) surface. Along with the water dissociation, we also observed the
electronic states change associated with the OVs oxidation of BOCꢀ010. Bi4f spectra of the
BOCꢀ010ꢀOV showed two additional shoulder peaks at 163.3 and 157.5 eV of lower binding
(
3−x)+
energies, which were corresponded the reduced Bi states of Bi
around the OVs (Figure 4g).
After exposure to water vapor, these shoulder peaks on BOCꢀ010ꢀOV were largely depopulated,
which were slightly attenuated in the case of BOCꢀ001ꢀOV (Figure 4g). Depopulation of the defect
states in the Bi 4f spectra, reflecting a decreased concentration of OVs, supported our DFT
calculation results that interaction of water with the OV of BiOCl (010) surface involved the simple
(
3−x)+
adsorption, and also the surface redox reaction, where Bi
were reoxidized by the adsorbed water
via the charge donation from OVs to the newlyꢀformed hydroxyl groups (Figure 3c).
1
8
ACS Paragon Plus Environment

Page 19 of 34
ACS Catalysis
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 5. In situ water contact angle measurements on the BiOCl surfaces. (a) Schematic
presentation of water contact angle measurement. A UV light source was set above the BiOCl
cylindrical pellets and water contact angle was in situ monitored via a microscope connected to a
computer. (b) Photos of the water droplets on the BiOCl surfaces under UV light illumination. The
final contact angle was an average of six measurements on the different locations on the surfaces. (c)
Dynamic change of the water contact angles and (d) water contact angle reciprocals on the BiOCl
surfaces along with UV light illumination. Theoretical simulation of water adsorption structures in
the outer layer on the (e) BOCꢀ001 and (d) BOCꢀ010 with an OV.
Enlightened by the above FTIR and XPS results, we further evidenced the different microscopic
waterꢀsurface interaction modes via the macroscopic water adsorption on the BiOCl surfaces
monitored by the in situ water contact angle measurements (Figure 5a). Defectꢀfree BOCꢀ001 was
o
rather hydrophobic with an initial water contact angle of 137.9 , confirming the weak interaction
between water and this surface. Under UV light irradiation for 120 min, the water contact angle on
1
9
ACS Paragon Plus Environment

ACS Catalysis
Page 20 of 34
o
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
BOCꢀ001 gradually decreased to 86 , indicative of slight hydrophilic character (Figure 5b). As
compared with BOCꢀ001, BOCꢀ010 showed a quite different water wetting behavior. First, initial
o
water contact angle on BOCꢀ010 was around 96.4 , significantly smaller than that on BOCꢀ001,
indicating the stronger interaction between water molecules and BOCꢀ010. Second, the water contact
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
o
angle on BOCꢀ010 surface quickly decreased to 21.7 within 120 min of UV light illumination,
showing a more rapid conversion from hydrophobicity to hydrophilicity (Figure 5b). For
comparison, dark control experiments were performed, which revealed that the water contact angles
did not change along with time in the dark (Figure S6). Photoinduced hydrophilic conversion on
both the BiOCl surfaces could be undoubtedly related to the presence of OVs, as the hydrophilic
surface did not appear when control experiments were conducted in the O
2
atmosphere to quickly
4
6ꢀ50
quench the in situꢀformed OVs.
Plotting the water contact angle against UV light illumination
gave trends of the surface hydrophilicity evolution, but could not provide specific kinetic
information on the surface hydrophilicity change because the initial values of water contact angles
on BiOCl surfaces were different (Figure 5c). So we plotted the reciprocal of the contact angle
against the UV light irradiation time, which gave nearly straight lines whose corresponding slopes
46
could be defined as the hydrophilic conversion rates (k ) (Figure 5d). The estimated k values of
h
h
ꢀ5
ꢀ5
ꢀ1
BOCꢀ001 and BOCꢀ010 were 3.59 × 10 and 25.13 × 10 (degree • min) , respectively. Therefore,
the photoinduced hydrophilic conversion rate of BOCꢀ010 was about 7 times that of BOCꢀ001. The
photoinduced hydrophilic conversion rate difference between BOCꢀ001 and BOCꢀ010 under the
same conditions was actually attributed to their different interaction strengths between the inner
layer water and the outer layer water. Molecular water adsorption on the OV of BiOCl (001) surface
w
in the inner layer made O of the adsorbed water a credible HB acceptor site for the enhanced outer
2
0
ACS Paragon Plus Environment

Page 21 of 34
ACS Catalysis
1
2
3
4
5
6
7
8
9
layer water adsorption with a moderate HB bond of 1.80 Å (Figure 5e). However, in the case of
BiOCl (010) surface, water dissociation along with the charge transfer from OV to water offered the
dissociated ꢀO
w
H
a
fragment a perfect HB donor site toward the formation of an intimately
O complex with a much stronger HB of 1.70 Å (Figure 5f). Therefore, both the
surfaceꢀbound OHꢀH
2
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
higher binding energy for water adsorption in the inner layer and the stronger HB water formed in
the outer layer determined the hydrophilicity character of BOCꢀ010 under UV light.
Figure 6. Experimental study of water oxidation on the BiOCl surfaces. (a) EPR spectra of the
2
spinꢀreactive •OH radicals photogenerated by BiOCl in H O. (b) Qualitative determination of the
•
OH radicals in H O or D O photogenerated by BiOCl along with UV light illumination. (c) O
2
2
2
2
1
ACS Paragon Plus Environment

ACS Catalysis
Page 22 of 34
1
8
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2 2
evolution upon UV light irradiation over BiOCl in H O. (d) Time course for the generation of O
18
2
from H
O by BiOCl detected via mass spectra.
The above results and discussions provided the solid evidences that water molecules were
molecularly adsorbed on the OVs of BiOCl (001) surface, but dissociatively adsorbed on the OVs of
BiOCl (010) surface. However, we still needed check our speculation whether the dissociatively
adsorbed water molecules with more suitable geometrical structure and higherꢀlevel of activation
would indeed exhibit higher tendency toward oxidation than the molecularly adsorbed ones. To
answer this question, EPR was first used to determine the primary photocatalytic water oxidation
active species of •OH radicals either adsorbed on BiOCl surfaces or dissolved in the water solution,
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
*
+
*
+
which were generated via the first proton removal reaction (H
2
O + hole ꢀ •OH + H ). To exclude
the possibility of •OH radicals formation via the O reduction pathway, Ar was continuously purged
2
2
to completely remove the waterꢀdissolved O . No EPR signal could be observed in the dark. As
expected, we observed strong fourꢀline EPR spectra with the relative intensities of 1:2:2:1
corresponding to •OH radicals for both BOCꢀ001 and BOCꢀ010 under UV light irradiation (Figure
6a). These EPR signals could be largely inhibited by adding either tertꢀbutyl alcohol (TBA) or
sodium acetate (SA) as the •OH radicals or holes scavenger, indicating the generation of •OH
radicals was directly related to water oxidation by photogenerated holes (Figure 6a). Plotting the
intensity of the EPR signal intensity against the irradiation time gave nearly two straight lines,
whose slopes indirectly gave the •OH radicals’ generation rates and were estimated to be 157.28 and
ꢀ1
742.13 min for BOCꢀ001 and BOCꢀ010, respectively (Figure 6b). Therefore, the water oxidation
on BOCꢀ010 toward the formation of •OH radicals was about 4.7 times that on BOCꢀ001. This
higher photoreactivity of BOCꢀ010 was not attributed to its specific surface area, which was only 2.8
2
2
ACS Paragon Plus Environment

Page 23 of 34
ACS Catalysis
1
2
3
4
5
6
7
8
9
2 2
times that of BOCꢀ001 (Figure S4c). Meanwhile, when D O was used instead of H O, BOCꢀ001
showed a deuterium kinetic isotope effect (KIE = kH2O/kD2O) of 2.92, a characteristic value indicating
5
ꢀ7
breakage of the OꢀH bond is the rateꢀlimiting step of the first proton removal reaction. However,
as for BOCꢀ010, whose in situ formed OVs could directly catalyze the OꢀH bond breaking prior to
oxidation due to water dissociation, showed a much smaller KIE of 1.28 (Figure 6b). We also
qualitatively measured the amount of free •OH radicals as well as their coupling product of H O via
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
2
photoluminescence. In accordance with the EPR results, BOCꢀ010 possessed much faster reactive
oxygen species generation rate than BOCꢀ001 (Figure S7b and S7c). Moreover, the higher
concentration of •OH and H O generated by BOCꢀ010 surface endowed its higher photoreactivity
2
2
than BOCꢀ001 for the anaerobic removal of pentachlorophenol with good recyclability and stability
Figure S7dꢀf and S8). The multihole water oxidation products of O over the BiOCl surfaces were
also investigated. Although the adsorption of O was more preferential than that of H O on the OV
of BiOCl surfaces (Figure S9), as compared with TiO ꢀP25 which cannot in situ generate OVs under
UV light, BiOCl still exhibited remarkable capability to oxidize water into O (Figure 6c). Since no
electron scavenger was added during O evolution reaction, these results clearly demonstrated the
(
2
2
2
2
2
2
indispensable role of the in situ formed OVs on the BiOCl surfaces toward enhanced
thermodynamics for water oxidation, while the higher photoreactivity of BOCꢀ010 was attributed to
the higherꢀlevel activation of dissociatively adsorbed H
2
O. In order to exclude the possible
16
18
2
interference of O in air, water oxidation with BiOCl photocatalysts were also carried in H O
2
1
8
18
under UV light. Photogenerated O
2
was detected by mass spectrometry and the signal for O
2
(m/z
=36) gradually increased for the BiOCl catalysts, while was not detected in case of P25 (Figure 6d).
1
8
2
The much higher signal intensity for O of BOCꢀ010 confirmed its higher photoreactivity toward
2
3
ACS Paragon Plus Environment

ACS Catalysis
Page 24 of 34
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
water oxidation than that of BOCꢀ001. Because only traces of H
2
was detected, the twinborn
for the OVs
quenching, being reflected by equilibrium over the generation and consumption of OVs, and/or
photoinduced electrons might be either scavenged by some of the surfaceꢀadsorbed O
2
2
4,51
migrated into the aqueous solution to became solvated electrons (Figure S10).
Similar water
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
activation scheme could also be observed over BiOBr and BiOI, the homologous photocatalysts of
BiOCl. Both stoichiometric BiOBr and BiOI could not oxidize water to generate •OH radicals under
visible light, but after the introduction of OVs on the catalysts surface, distinct EPR signals
correspond to •OH radicals were promptly observed, indicating the indispensable role of OVs for
water activation (Figure S11).
Conclusions
In summary, DFT calculations together with XAFS spectroscopy were first adopted to determine the
specific structures of OVs on the (001) and (010) surfaces of BiOCl. Further theoretical and
experimental results demonstrated OVs could activate adsorbed water via their localized electrons,
offering a possibility to enhance the thermodynamics toward water oxidation, while the water
activation state was supposed to be highly dependent on the structures of OVs. In our case, the OVs
of BiOCl (010) surface with dissociatively adsorbed water exhibited much higher photoreactivity
toward oxidation than those of BiOCl (010) surface with the molecularly adsorbed water, generating
2 2 2
more waterꢀoxidized species like •OH, H O and O . It was proposed that OVs of BiOCl (010)
surface could facilitate the barrierless OꢀH bond breaking of the first proton removal reaction, which
was sterically hindered on the OVs of BiOCl (001) surface, and also allow more electron transfer
from the OVs to dissociatively adsorbed water, leading to a higher water activation level. These
2
4
ACS Paragon Plus Environment

Page 25 of 34
ACS Catalysis
1
2
3
4
5
6
7
8
9
findings highlight the indispensable role of crystalline surface structure on water oxidation, and may
open up new avenues for the rational design of new photocatalysts via surface engineering.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
5
ACS Paragon Plus Environment

ACS Catalysis
Page 26 of 34
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Experimental Section
Chemicals. All chemicals used were purchased from Sinopharm Chemical Reagent Co., Ltd.
Shanghai, China) and were of analytical grade and used without purification.
Preparation of BiOCl single-crystalline nanosheets. Wellꢀdefined BiOCl singleꢀcrystalline
(
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
5
2
nanosheets were prepared by our previous method. Typically, Bi(NO
mL distilled water with stoichiometric amount of KCl. The solution with pH adjusted be either 1 or
was then poured into an autoclave and was heated at 220 °C for 24 h. Resulting precipitates were
3 3 2
) •5H O was added into 18
6
washed with deionized water and ethanol for 6 times and dried at 80 °C for 24 h in air. We denoted
the BiOCl nanosheets obtained under pH = 1 with (001) surface exposed as BOCꢀ001 and the BiOCl
nanosheets obtained under pH = 6 with (010) surface exposed as BOCꢀ010. 0.2 g BOCꢀ001 or
BOCꢀ010 was then immersed into 50 mL distilled water and illuminated with a UV light (λ = 254
nm, 30 W) under continuous Ar purging for 4 h to obtain its oxygenꢀdeficient counterpart
BOCꢀ001ꢀOV or BOCꢀ010ꢀOV.
Characterization. In situ diffuse reflectance FTIR spectra were recorded by Nicolet iS50FTꢀIR
spectrometer. XPS was obtained with PerkinꢀElmer PHI 5000C and all binding energies were
calibrated by using the contaminant carbon (C1S = 284.6 eV) as a reference. Water contact angles of
BiOCl surfaces were measured by a sessile drop method at room temperature with a contact angle
goniometer (OCAꢀ20, DataPhysics, Germany) equipped with video capture. For every measurement,
2 ꢂL of deionized water droplet was used. Electron paramagnetic resonance (EPR) spectra were
conducted on a Bruker EMX EPR Spectrometer (Billerica, MA). Bismuth Lꢀedge extended Xꢀray
absorption fine structure (EXAFS) spectra were performed at the beamline 1W1B of Beijing
Synchrotron Radiation Facility, Institute of High Energy Physics, Chinese Academy of Sciences.
2
6
ACS Paragon Plus Environment

Page 27 of 34
ACS Catalysis
1
2
3
4
5
6
7
8
9
The electron energy in the storage ring was about 2.2 GeV with a current of about 100 mA. Incident
Xꢀray energy was scanned across the EXAFS region of the Bi Lꢀedge (13426 eV) using a Si (111)
double crystal monochromator. The EXAFS data were converted to Rꢀspace by taking the Fourier
transform of χ(k) function. Fitting was performed simultaneously on all data sets in Rꢀspace using
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
the program Artemis to optimize the coordination number (N), the distance (R), and disorder (σ ).
DFT theoretical calculation. All the above calculations were performed using the firstꢀprinciples
density of functional theory (DFT) + U calculations with the exchangeꢀcorrelation energy functional,
which were described by generalized gradient approximation with PerdewꢀBurkeꢀErnzerhof (PBE)
5
3
exchangeꢀcorrelation function. In order to test the thickness of the atomic layers of the slabs for all
the BiOCl models, (1 × 1) surface was applied and this part of calculations were implemented by the
CASTEP code in which planeꢀwave pseudopotential approach and ultrasoft pseudopotentials were
54,55
employed for all the atoms with a kinetic energy cutoff of 380 eV.
All the structures were
ꢀ5
relaxed to an energy convergence of 10 eV/atom and a force convergence of 0.03 eV/Å during the
geometrical optimization process. The MonkhorstꢀPack mesh for BiOCl (001) and (010) surfaces
56
were 4 × 4 × 1 and 2 × 4 × 1, respectively. To simulate the water adsorption on the BiOCl surfaces,
a (2 × 2) surpercell was used and enabled by a VASP code in which the projector augmented wave
5
7, 58
(
PAW) method represented the electronꢀion interaction with a kinetic energy cutoff of 520 eV.
ꢀ5
The energy and force converged to 10 eV/atom and 0.02 eV/Å during the geometrical optimization,
respectively. In the energy and DOS calculation parts, the kꢀpoints were increased to 4×4×1 and
3×6×1 for the (001) and (010) surface, respectively. The vacuum in all the models was kept at 20 Å.
The interaction energy of the adsorbate water on BiOCl surface was calculated as: ꢁE = E(BiOCl +
2 2 2
H O) ꢀ E(BiOCl) ꢀ E(H O). The charge density difference was calculated as: ∆ρ = ρ(H O/BiOCl) ꢀ
2
7
ACS Paragon Plus Environment

ACS Catalysis
Page 28 of 34
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
ρ(H
whereas ρ(H
taken in the same geometry of the interacting system. The desorption energy of O
surface was calculated as: ꢁE = E(BiOCl) + E(O ) ꢀ E(BiOCl + O ).
2
O) ꢀρ(BiOCl), where ρ(H
2
O/BiOCl) is the density of the interacting waterꢀsurface system,
2
O) and ρ(BiOCl) are the densities of the two isolated (noninteracting) subsystems,
ꢀadsorbed BiOCl
2
2
2
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Photocatalytic water oxidation. All photocatalytic water oxidation experiments were conducted at
ambient temperature using a UV light (λ = 254 nm, 30 W) or a 500 W Xe arc lamp with a 420 nm
cutoff filter as the visible light source. The photogenerated •OH radicals were detected via EPR
using 5,5ꢀdimethylꢀ1ꢀpyrrolineꢀNꢀoxide as the radical spinꢀtrapped reagent. For photocatalytic
2
oxygen evolution, 0.2 g photocatalyst was added into 100 mL of H O in a Pyrex flask reactor sealed
with a silicone rubber septum. The reactor was also surrounded with water circulation in the outer
jacket in order to maintain at room temperature. The mixture was continuously stirred in the dark
with highꢀpurity Ar bubbled to completely remove the dissolved O
2
for 120 min. Isotopic study was
O in a sealed Pyrex vessel. Generated
O during the photocatalytic water oxidation was monitored by a mass spectrometer (Trace
1
2
8
performed by mixing 0.02 g BiOCl mixed with 1 mL of H
1
8
2
1300ꢀISQ, Thermo).
Supporting Information. Other experimental details, additional SEM and TEM images, Xꢀray
diffraction patterns (XRD) and additional data. This material is available free of charge via the
Internet at http://pubs.acs.org.
ACKNOWLEDGMENT
This work was supported by National Natural Science Funds for Distinguished Young Scholars
(Grant 21425728), National Basic Research Program of China (973 Program) (Grant
2
8
ACS Paragon Plus Environment

Page 29 of 34
ACS Catalysis
1
2
3
4
5
6
7
8
9
2013CB632402), National Key Research and Development Program of China (Grant
2
016YFA0203002), National Science Foundation of China (Grant 51472100), SelfꢀDetermined
Research Funds of CCNU from the Colleges’ Basic Research and Operation of MOE (Grant
CCNU14Z01001 and CCNU16A02029), Excellent Doctorial Dissertation Cultivation Grant from
Central China Normal University (Grant 2015YBZD018 and 2016YBZZ034). We also thank the
National Supercomputer Center in Jinan for providing high performance computation.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
REFERENCES
(1) Hoffmann, M. R.; Martin, S. T.; Choi, W.; Bahnemann, D. W. Chem. Rev. 1995, 95, 69–96.
(2) Ma, Y.; Wang, X.; Jia, Y.; Chen, X.; Han, H.; Li, C. Chem. Rev. 2014, 114, 9987–10043.
(3) Bourikas, K.; Kordulis, C.; Lycourghiotis, A. Chem. Rev. 2014, 114, 9754–9823.
(4) Kumar, S.; Schelling, P. K. J. Chem. Phys. 2006, 125, 204704.
(5) Valdés, A.; Kroes, G.ꢀJ. J. Phys. Chem. C 2010, 114, 1701–1708.
(6) Zhao, W.ꢀN.; Liu, Z.ꢀP. Chem. Sci. 2014, 5, 2256–2264.
(7) Chen, J.; Li, Y.ꢀF.; Sit, P.; Selloni, A. J. Am. Chem. Soc. 2013, 135, 18774–18777.
(8) Youngblood, W. J.; Lee, S.ꢀH. A.; Maeda, K.; Mallouk, T. E. Acc. Chem. Res. 2009, 42,
1966–1973.
(9) Qu, Y.; Duan, X. Chem. Soc. Rev. 2013, 42, 2568–2580.
(10)Li, H.; Shang, J.; Ai, Z.; Zhang, L. J. Am. Chem. Soc. 2015, 137, 6393–6399.
(11)Salvador, P. J. Phys. Chem. C 2007, 111, 17038–17043.
(12)Hahn, K. R.; Tricoli, A.; Santarossa, G.; Vargas, A.; Baiker, A. Langmuir 2012, 28, 1646–1656.
2
9
ACS Paragon Plus Environment