G-C3N4α-Fe2O3/C Photocatalysts: Synergistically Intensified Charge Generation and Charge Transfer for NADH Regeneration

Article abstract of DOI:10.1021/acscatal.8b00070

Graphitic carbon nitride (g-C₃N₄) is an emergent metal-free photocatalyst because of its band position, natural abundance, and facile preparation. Synergetic intensification of charge generation and charge transfer of g-C₃N₄ to increase solar-to-chemical efficiency remains a hot yet challenging issue. Herein, a nanoshell with two moieties of α-Fe₂O₃ and carbon (C) is in situ formed on the surface of a g-C₃N₄ core through calcination of Fe³⁺/polyphenol-coated melamine, thus acquiring g-C₃N₄α-Fe₂O₃/C coreshell photocatalysts. The α-Fe₂O₃ moiety acts as an additional photosensitizer, offering more photogenerated electrons, whereas the C moiety bridges a "highway" to facilitate the electron transfer either from α-Fe₂O₃ moiety to g-C₃N₄ or from g-C₃N₄ to C moiety. By tuning the proportion of these two moieties in the nanoshell, a photocurrent density of 3.26 times higher than pristine g-C₃N₄ is obtained. When utilized for photocatalytic regeneration of reduced nicotinamide adenine dinucleotide (NADH, a dominant cofactor in biohydrogenation reaction), g-C₃N₄α-Fe₂O₃/C exhibits an equilibrium NADH yield of 76.3% with an initial reaction rate (r) of 7.7 mmol h^-1 g^-1, among the highest r for photocatalytic NADH regeneration ever reported. Manipulating the coupling between charge generation and charge transfer may offer a facile, generic strategy to improve the catalytic efficiency of a broad range of photocatalysts other than g-C₃N₄.

Full text of DOI:10.1021/acscatal.8b00070

Subscriber access provided by UNIV OF DURHAM
Article
g-C3N4@#-Fe2O3/C Photocatalysts: Synergistically Intensified
Charge Generation and Charge Transfer for NADH Regeneration
Yizhou Wu, Jesse Ward-Bond, Donglin Li, Jiafu Shi, Shaohua Zhang, and Zhongyi Jiang
ACS Catal., Just Accepted Manuscript • DOI: 10.1021/acscatal.8b00070 • Publication Date (Web): 15 May 2018
Downloaded from http://pubs.acs.org on May 15, 2018
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 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 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.
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 41
ACS Catalysis
1
2
3
4
5
6
7
8
9
gꢀC3N4@αꢀFe2O3/C Photocatalysts:
Synergistically Intensified Charge Generation
and Charge Transfer for NADH Regeneration
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
Yizhou Wua,c, Jesse WardꢀBonda,d, Donglin Lia, Jiafu Shib,c,e,*, Shaohua Zhanga,c,
Zhongyi Jianga,c,e,*
a Key Laboratory for Green Chemical Technology of Ministry of Education, School of
Chemical Engineering and Technology, Tianjin University, Tianjin 300072, P. R.
China.
b
Tianjin Engineering Center of Biomassꢀderived Gas and Oil, School of
Environmental Science & Engineering, Tianjin University, Tianjin 300072, P. R.
China.
c Collaborative Innovation Center of Chemical Science and Engineering (Tianjin),
Tianjin 300072, P. R. China.
d
Department of Chemical and Biological Engineering, University of Ottawa, 161
Louis Pasteur Private, Ottawa, ON K1N 6N5, Canada.
e State Key Laboratory of Biochemical Engineering, Institute of Process Engineering,
Chinese Academy of Sciences, Beijing, 100190, P. R. China.
*Corresponding author: Jiafu Shi (shijiafu@tju.edu.cn); Zhongyi Jiang
(zhyjiang@tju.edu.cn)
1
ACS Paragon Plus Environment

ACS Catalysis
Page 2 of 41
1
2
3
4
ABSTRACT
5
6
7
8
9
Graphitic carbon nitride (gꢀC3N4) is an emergent metalꢀfree photocatalyst due to its
band position, nature abundance and facile preparation. Synergetic intensification of
charge generation and charge transfer of gꢀC3N4 to increase solarꢀtoꢀchemical
efficiency remains a hot yet challenging issue. Herein, a nanoꢀshell with two moieties
of αꢀFe2O3 and carbon (C) is in situ formed on the surface of a gꢀC3N4 core through
calcination of Fe3+/polyphenolꢀcoated melamine, thus acquiring gꢀC3N4@αꢀFe2O3/C
core@shell photocatalysts. The αꢀFe2O3 moiety acts as an additional photosensitizer,
offering more photoꢀgenerated electrons, whereas the C moiety bridges a “highway”
to facilitate the electron transfer either from αꢀFe2O3 moiety to gꢀC3N4 or from gꢀC3N4
to C moiety. By tuning the proportion of these two moieties in the nanoꢀshell, a
photocurrent density of 3.26 times higher than pristine gꢀC3N4 is obtained. When
utilized for photocatalytic regeneration of reduced nicotinamide adenine dinucleotide
(NADH, a dominant cofactor in biohydrogenation reaction), gꢀC3N4@αꢀFe2O3/C
exhibits an equilibrium NADH yield of 76.3% with an initial reaction rate (r) of 7.7
mmol hꢀ1 gꢀ1, among the highest r for photocatalytic NADH regeneration ever
reported. Manipulating the coupling between charge generation and charge transfer
may offer a facile, generic strategy to improve the catalytic efficiency of a broad
range of photocatalysts other than gꢀC3N4.
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
KEYWORDS: Charge Generation; Charge Transfer; Graphitic Carbon Nitride;
Photocatalysis; NADH Regeneration
2
ACS Paragon Plus Environment

Page 3 of 41
ACS Catalysis
1
2
3
4
1. INTRODUCTION
5
6
7
8
9
Photocatalysis, as a green and sustainable process for the conversion of solar
energy into chemicals, has gained tremendous interests.1,2 Efficient solarꢀtoꢀchemical
conversion requires rapid energyꢀtoꢀmass transfer in photocatalysis. Two critical steps
in photocatalytic process, i.e., charge generation and charge transfer, determine the
utilization efficiency of solar energy that is finally stored in the form of chemicals
through surface redox reaction.3,4 Design and preparation of photocatalysts capable of
synergistically intensified charge generation and charge transfer with wellꢀmatched
behaviors is the prerequisite of a number of photocatalytic processes, including water
splitting,5ꢀ7 carbon dioxide (CO2) reduction,8ꢀ11 biohydrogenation,12ꢀ14 and so on.
Among various photocatalysts, graphitic carbon nitride (gꢀC3N4) is being
boomingly explored in recent years due to its unique band position, nature abundant
source and facile preparation.1,15 Nevertheless, lower charge generation efficiency and
slower charge transfer rate remarkably inhibit the practical application of pristine
gꢀC3N4. Therefore, tremendous efforts have been devoted to intensify either of the two
steps16, 17 for elevating the solarꢀtoꢀchemical efficiency of gꢀC3N4.
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
Owing to their excellent charge mobility and good compatibility with gꢀC3N4,
carbon materials have been incorporated into gꢀC3N4 to accelerate the charge transfer.
Photoꢀgenerated electrons could be easily extracted from gꢀC3N4 to carbon via πꢀπ
stacking electric field. Then, the charge transfer could be greatly reinforced and,
thereby, the electronꢀhole recombination could be inhibited.18ꢀ21 However, the intrinsic
light absorption of carbon materials limits the efficient utilization of incident light by
3
ACS Paragon Plus Environment

ACS Catalysis
Page 4 of 41
1
2
3
4
5
6
7
8
9
gꢀC3N4. The unmatched charge behaviors, i.e., attenuated charge generation and
reinforced charge transfer, often result in a limited increase of available charge
carriers or even a negative effect in the presence of excessive amount of carbon.1,22,23
Alternatively, several methods have also been developed to intensify the charge
generation of gꢀC3N4, e.g., constructing hierarchically structured gꢀC3N4 to achieve
superior lightꢀharvesting ability,16,24,25 doping heteroatoms into gꢀC3N4 to broaden the
wavelength for visibleꢀlight absorption,26ꢀ28 attaching secondary photosensitizers onto
gꢀC3N4 to offer additional photoꢀgenerated electrons,29ꢀ31 etc. The charge generation
efficiencies are enhanced by 2.4~3.7 times compared to pristine gꢀC3N4, which benefit
to improving the overall solarꢀtoꢀchemical efficiency.16,25,27,29 Albeit some significant
progress in the intensification of either charge generation step or charge transfer step,
the synergistic manipulation of both steps is rarely reported.31,32 Particularly, it is still
challenging but highly desirable to acquire wellꢀmatched charge transfer and charge
generation in gꢀC3N4ꢀbased photocatalysts.
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
In this study, for the first time, gꢀC3N4@αꢀFe2O3/C core@shell photocatalysts with
a gꢀC3N4 core and an αꢀFe2O3/C nanoꢀshell are prepared to achieve highly efficient
regeneration of reduced nicotinamide adenine dinucleotide (NADH). Redox reactions
afford the essential driving force for all varieties of life. NADH is utilized as a
hydride donor in majority of redox reactions performed by oxidoreductases, and in
many energyꢀrelevant catalytic hydrogenation reactions spanning from alcohol
deracemization, oxygen reduction to hydrogen production.33,34 The expensive but
indispensable features of NADH requires facile and effective regeneration methods.
4
ACS Paragon Plus Environment

Page 5 of 41
ACS Catalysis
1
2
3
4
5
6
7
8
9
The gꢀC3N4@αꢀFe2O3/C is formed through calcination of Fe3+/polyphenolꢀcoated
melamine, which exhibits a core@shell structure. The optical and photoelectric
properties of gꢀC3N4@αꢀFe2O3/C confirm the simultaneous intensification of charge
generation and charge transfer. Excitation of the αꢀFe2O3 moiety endows the
nanoꢀshell with charge generation ability, minimizing the inefficient light absorption
from the C moiety. Meanwhile, the C moiety facilitates the electron transfer either
from αꢀFe2O3 moiety to gꢀC3N4 or from gꢀC3N4 to the surface of C moiety, thus
suppressing the initial charge recombination in gꢀC3N4. By altering the ratios of
αꢀFe2O3 moiety to C moiety in the nanoꢀshell, the two steps of charge generation and
charge transfer could be wellꢀmatched and synergistically intensified, leading to
significantly enhanced efficiency for NADH regeneration. Notably, the preparation
process integrates the dual functions of polyphenols,35ꢀ39 i.e., substrateꢀindependent
surface adhesion and transition metal coordination, which offers a facile and generic
method to form αꢀFe2O3/C nanoꢀshell onto a broad range of substrates. The strategy of
synergistically intensifying charge generation and charge transfer based on αꢀFe2O3/C
nanoꢀshell in this study is applicable to other photocatalysts other than gꢀC3N4.
2. MATERIALS AND METHODS
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
2.1. Materials. Melamine (99%), sodium chloride (NaCl, 99.8% GR), bisꢀtris
hydrochloride (99%), ferric trichloride hexahydrate (99% AR), triethanolamine
(TEOA, ≥99% GC), disodium hydrogen phosphate (99 AR), sodium dihydrogen
phosphate (AR), pentamethylcyclopentadienylrhodium (III) chloride dimer (99%),
2,2'ꢀdipyridyl (99% AR) were purchased from Aladdin Industrial Corporation
(Shanghai, China). Tannic acid (95%) was obtained from J&K Chemical (Beijing,
5
ACS Paragon Plus Environment

ACS Catalysis
Page 6 of 41
1
2
3
4
5
6
7
8
9
China). βꢀnicotinamide adenine dinucleotide hydrate (≥98%), amorphous silicon
dioxide (powder, an average size of 0.2ꢀ0.3 ꢁm) and methylviologen dichloride (98%)
were received from SigmaꢀAldrich (St. Louis, USA). Hydrochloric acid (HCl,
36.0ꢀ38.0 wt%) was obtained from Rionion (Tianjin, China). Oolong tea and red wine
were commercial products and purchased from convenience store. All other chemicals
or reagents were used without further purification.
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. Preparation of gꢀC3N4@αꢀFe2O3/C Core@Shell Photocatalysts. 1 g melamine
was dispersed in 15 mL 0.133 M bisꢀtris, 0.8 M NaCl solution. 5 mL tannic acid (10
mg mLꢀ1) was added into the solution. The solution was kept stirring at 25 oC in water
bath for 24 hours. The powder was collected by centrifugation and washed with
deionized (DI) water. To prepare gꢀC3N4@C, the freezeꢀdried powder was sealed in a
o
crucible and calcined with the following process: 300 C for 1 h, 400 oC for 1 h, then
o
o
550 C for 4 h with the heating rate of 5 C minꢀ1. For gꢀC3N4@αꢀFe2O3/C, the
powder was reꢀdispersed in 15 mL DI water. 0.1 mmol FeCl3·6H2O (dissolved in 5
mL DI water) was instantly added into the suspension. The solution was kept stirring
for 5 minutes and centrifuged to obtain the precipitate. The asꢀobtained precipitate
was washed with DI water, freezeꢀdried overnight, and then calcined with the same
procedure. After calcination, the samples were ground for further experiments.
To prepare C·M, 1.68 mmol melamine and 1.68 mmol cyanuric acid were dissolved
in 70 mL DI water at 42 oC in water bath, respectively. The solution of cyanuric acid
was slowly added into the solution of melamine. White precipitate was formed
o
immediately and the suspension was kept stirring at 42 C for 20 minutes. The
precipitate was collected by centrifugation and freezeꢀdried. C·MꢀgꢀC3N4@αꢀFe2O3/C
and C·MꢀgꢀC3N4@C were prepared by replacing melamine with C·M.
6
ACS Paragon Plus Environment

Page 7 of 41
ACS Catalysis
1
2
3
4
5
6
7
8
9
To prepare samples with different ratio of αꢀFe2O3 moiety and C moiety,
FeCl3·6H2O were altered from 0.01 mmol to 0.2 mmol before adding into the
suspension. The acquired samples were denoted as gꢀC3N4@αꢀFe2O3/CꢀX (X was the
millimole of FeCl3·6H2O initially added). Samples prepared by tea and red wine
followed the same procedure. In detail, 1 g oolong tea was dispersed in 60 mL hot
water for 5 minutes to obtain tea water. 5 mL tea or red wine was used as polyphenols
to replace tannic acid.
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
For the samples corroded by hydrochloric acid (HCl), 50 mg samples was dispersed
o
in 10 mL concentrated HCl. The suspension was kept stirring at 80 C in water bath
for 6 hours. After the first 3ꢀhour reaction, another 10 mL concentrated HCl was
added in the suspension. The powder was collected by centrifuged and washed by DI
water until the pH of supernatant turned into neutral.
2.3. Synthesis of [Cp*Rh(bpy)H2O]2+. Pentamethylcyclopentadienylrhodium (III)
chloride dimer (61.8 mg) was dispersed in 5 mL methanol to form a red suspension.
31.2 mg 2,2'ꢀdipyridyl was added into the suspension. The suspension was kept
stirring until it turned into bright orange solution. The solution was evaporated to 1mL
at ambient condition. 10 mL diethyl ether was added into the solution to acquire
o
orange precipitate. The precipitate was collected by centrifugation and dry at 50 C
overnight to obtain [Cp*Rh(bpy)Cl]Cl. [Cp*Rh(bpy)H2O]2+ was formed after
dissolved in DI water. The stock solution (20 mM) was prepared and kept at 4 oC
avoiding direct light irradiation.
2.4. Photocatalytic Regeneration of NADH. Typically, the reaction solution was
composed of 1 mM NAD+, 0.2 mM Rh complex, 400 mM TEOA, 1 mg photocatalyst
and 100 mM phosphate buffer (pH 8). The total volume of the solution was 2 mL. The
solution was placed in a quartz cuvette and kept stirring. A blue LED (405 nm, 100W)
7
ACS Paragon Plus Environment

ACS Catalysis
Page 8 of 41
1
2
3
4
5
6
7
8
9
was used as the light source and the distance was kept at 5 cm. Before turning on the
light, the reaction system was kept stirring in dark for 5 minutes to reach
adsorption/desorption equilibrium, the rationality of which was confirmed and shown
in Figure S15 and S16. The concentration of NADH was determined by measuring
the absorbance of the solution at 340 nm by a UVꢀvis spectrophotometer (Uꢀ3010,
Hitachi). The initial reaction rate was calculated based on the first 6ꢀmin reaction with
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
[Eq.1]:
V
r =
(ct − c0 )
(1)
mpct
where r was the initial reaction rate, V was the volume of reaction solution, c was the
concentration of NADH, mpc was the mass of photocatalyst and t was the reaction
time.
2.5. Enzymatic Reduction of Formaldehyde with in situ NADH Regeneration. The
reaction solution (4 mL) was composed of 1 mM NAD+, 0.2 mM Rh complex, 400
mM TEOA, 2 mg photocatalyst, 100 mM phosphate buffer (pH 8), 0.62 mg ADH
(200 U) and 40 mM formaldehyde (from 37 % formaldehyde solution with 10~15 %
methanol as stabilizer). The produced methanol was detected by gas chromatography
(GCꢀ2010 plus, Shimadzu). The total turnover number of NADH during the methanol
production was calculated with [Eq.2]:
cMeOH ⋅V
TONNADH
=
(2)
mpc
8
ACS Paragon Plus Environment

Page 9 of 41
ACS Catalysis
1
2
3
4
where TONNADH was the turnover number of NADH, cMeOH was the concentration of
5
6
7
methanol,
V was the volume of reaction solution and mpc was the mass of
8
9
photocatalyst. According to the enzymatic reaction,
YADH
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
HCHO + NADH →CH3OH+NAD+
NADH was stoichiometric consumed during the reaction of methanol production. The
mole of regenerated NADH was equal to the methanol obtained during the reaction.
The times of individual NADH used in methanol production was calculated with
[Eq.3]:
cMeOH
NNADH
=
(3)
cNADH
where NNADH was the times of individual NADH utilized for methanol production,
cMeOH was the concentration of methanol and cNADH was the concentration of NADH.
The equilibrium yield of NADH was 76.3% in 16 minutes. The overall reaction rate
was 0.0477 mM minꢀ1. For 40ꢀmin reaction, the theoretical regeneration of NADH
could be obtained was 1.91 mM, which was close to the experimental data, 1.87 mM.
Therefore, we chose the equilibrium yield of NADH as cNADH for calculation.
2.6. Photoelectrochemical Evaluation. The sample was fabricated into electrode on a
FTO glass. In general, 10 mg of sample was dispersed in 1 mL water solution
containing 0.1 mL Nafion and sonicated for 10 minutes. Then, the solution was
o
dropped on the 2×1.5 cm2 FTO glass and dried at 50 C overnight. The film was cut
into 1×1.5 cm2 for evaluation. All the photoelectrochemical measurements were
performed in a threeꢀelectrode cell with a saturateꢀpotassiumꢀchloride sliver electrode
(Ag/AgCl) as a reference electrode, a platinum foil (0.5×0.5 cm2) as a counter
9
ACS Paragon Plus Environment

ACS Catalysis
Page 10 of 41
1
2
3
4
5
6
7
8
9
electrode, and the sample of interest as a working electrode. The electrolyte was 0.1
M Na2SO4. A 300 W xenon lamp couple with an AM 1.5G global filter (100mW cmꢀ2)
were used as radiation source. Linear sweeps, transient photocurrent and
electrochemical impedance spectroscope (EIS) Nyquist plot were measured by a CHI
660D electrochemical workstation. The initial voltage of transient photocurrent and
EIS was kept at ꢀ0.1 V.
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
2.7. Characterizations. Scanning electron microscope (SEM) was carried out on an
FEI Nova XL430 and Hitachi Limited instrument. Transmission electron microscope
(TEM) images were taken on FEI Tecnai G2 F20 equipment. Powder Xꢀray
diffraction (XRD) measurements were performed on a Rigaku D/max 2500V/PC
Xꢀray diffractometer with the graphite filtered Cu Kα radiation (λ = 1.54056 Å), and
the data was acquired in the range of 10ꢀ60o (2θ) at a rate of 6o minꢀ1. Fourier
transform infrared spectroscope (FTꢀIR) was measured on a Nicoletꢀ560 spectrometer,
where 64 scans were accumulated with a resolution of 4 cmꢀ1 for each spectrum. The
57Fe Mössbauer spectrum was carried out on a MFDꢀ500AVꢀ02 Mössbauer
spectrometer made by Topologic Systems, Inc. with a 57Co radioactive source. The
Xꢀray photoelectron spectroscope (XPS) was performed on a PerkinꢀElmer PHI 1600
ESCA Xꢀray photoelectron spectroscope with monochromatic Mg Kα radiation
(1253.6 eV). The UVꢀvis absorption spectra were recorded by a UVꢀvis
spectrophotometer (Uꢀ3010, Hitachi) equipped with an integrating sphere for the
diffuseꢀreflectance
spectroscope
(DRS),
using
BaSO4
as
reference.
Photoluminescence spectra were recorded on Horiba Jobin Yvon Fluorolog3
10
ACS Paragon Plus Environment

Page 11 of 41
ACS Catalysis
1
2
3
4
5
6
fluorescence spectrometer at room temperature with excitation at 350 nm.
3. RESULTS AND DISCUSSION
7
8
9
3.1. Preparation and Characterizations of gꢀC3N4@αꢀFe2O3/C Core@Shell
Photocatalysts. The preparation of gꢀC3N4@αꢀFe2O3/C core@shell photocatalysts are
presented in Scheme 1. Briefly, the solutions of polyphenols (tannic acid, tea or red
wine) were mixed with wellꢀdispersed melamine suspension in bisꢀtris/NaCl buffer.
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
o
The solution were kept stirring at 25 C in water bath for 24 hours to form a
polyphenol coating on melamine. Then, Fe3+ were added to the suspended solution of
melamine@polyphenol mixtures, the color of melamine@polyphenol changed
instantly due to the rapid coordination of phenolic hydroxyl groups with Fe3+,35 which
also indicated the formation of Fe3+/polyphenol coating on melamine. The asꢀformed
melamine@Fe3+/polyphenol was sealed in crucible and then calcined in muffle
furnace. Finally, the core@shell photocatalysts with a gꢀC3N4 core and an αꢀFe2O3/C
nanoꢀshell, denoted as gꢀC3N4@αꢀFe2O3/C, were acquired. For control, gꢀC3N4@C
were also prepared though calcination of melamine@polyphenol.
Scheme 1. Preparation of gꢀC3N4@αꢀFe2O3/C core@shell photocatalysts.
11
ACS Paragon Plus Environment

ACS Catalysis
Page 12 of 41
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
The Xꢀray diffraction (XRD) patterns of gꢀC3N4@αꢀFe2O3/C and gꢀC3N4@C
(Figure 1a) demonstrated that both αꢀFe2O3/C and C nanoꢀshells had negligible
influence on the structure of gꢀC3N4. The peaks at around 12.8o (100) corresponded to
the inꢀplane structure of triꢀ
s
ꢀtriazine. The small shift from 13.1o (PDF No. 87ꢀ1526,
ꢀtriazine during the
JCPDS) to 12.8o was caused by the partial oxidation of triꢀ
s
calcination process in air.40 The peak at around 27.2o was ascribed to the interlayer
periodic stacking.32 Fourier transform infrared (FTꢀIR) spectra (Figure S1) also
confirmed the formation of gꢀC3N4. The bands at 810 cmꢀ1, 1200ꢀ1600 cmꢀ1 and 3100
cmꢀ1 corresponded to triazine units, aromatic CꢀN heterocycles and primary/secondary
amines, respectively.41 Interestingly, no signals of αꢀFe2O3 and C were found in either
gꢀC3N4@αꢀFe2O3/C or gꢀC3N4@C as shown in both XRD patterns and FTꢀIR spectra.
12
ACS Paragon Plus Environment

Page 13 of 41
ACS Catalysis
1
2
3
4
5
6
7
8
9
Integrating C species with gꢀC3N4 caused little changes in XRD patterns and FTꢀIR
spectra due to the similar structure between carbon ring and gꢀC3N4 (Figure 1a and
S2f).19,42 As for gꢀC3N4@αꢀFe2O3/C, considering the trace amount of tannic acid and
Fe3+ used in the experiment and the much thinner coating of Fe3+/polyphenol
compared with melamine, the signals emitted from αꢀFe2O3/C nanoꢀshells may be
shielded by the strong signal of gꢀC3N4. To identify the phase of αꢀFe2O3/C
(particularly the phase of αꢀFe2O3), a supramolecular assembly formed by melamine
and cyanuric acid (C·M) was applied as a downsized precursor to prepare
gꢀC3N4@αꢀFe2O3/C (denoted as C·MꢀgꢀC3N4@αꢀFe2O3/C, and the details can be seen
in 2.2. Preparation of gꢀC3N4@αꢀFe2O3/C Core@Shell Photocatalysts and Figure S2).
As shown in Figure 1b, the XRD pattern of C·MꢀgꢀC3N4@αꢀFe2O3/C exhibited both
signals of gꢀC3N4 and αꢀFe2O3. The signals of αꢀFe2O3 was consistent with cubic
Fe21.34O32 (PDF No. 83ꢀ0112, JCPDS), which could be considered as αꢀFe2O3.
Moreover, Mössbauer spectrum was carried out to probe the chemical environment of
Fe in the gꢀC3N4@αꢀFe2O3/C sample. As shown in Figure S3, the main signal with an
isomer shift of 0.364 mm sꢀ1 was corresponded to αꢀFe2O3.43 TEM images (Figure 1c
and 1d) revealed the sheetꢀlike structure of gꢀC3N4@αꢀFe2O3/C. In the red area
(Figure 1c, inset), the lattices of αꢀFe2O3 were observed with the (110) plane. XRD,
Mössbauer spectrum and HRTEM results confirmed the formation of αꢀFe2O3 in the
nanoꢀshell.
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
To explore the microstructure of the nanoꢀshell, electron energy loss spectroscope
(EELS) was carried out in the edge area to explore the element distribution on the
13
ACS Paragon Plus Environment

ACS Catalysis
Page 14 of 41
1
2
3
4
5
6
7
8
9
surface of the photocatalyst (Figure S4). C and N elements were uniformly
distributed in gꢀC3N4@αꢀFe2O3/C and gꢀC3N4@/C. As shown in Figure 1d (inset) and
S4, the boundary of C EELS (red color) extended further than the boundary of N
EELS (green color) in gꢀC3N4@αꢀFe2O3/C, indicating the formation of the nanoꢀshell
on gꢀC3N4. The average thickness of the boundary was ~8 nm. The elemental
mapping of Fe (Figure S5e) confirmed the uniform distribution of Fe on the surface
of gꢀC3N4@αꢀFe2O3/C. Moreover, the structure of melamine@Fe3+/polyphenol was
also investigated by TEM. As shown in Figure S6, a continuous coating shell was
formed on the core of melamine. The EELS, elemental mapping and TEM results
verified the formation of αꢀFe2O3/C nanoꢀshell and core@shell structure of
gꢀC3N4@αꢀFe2O3/C.
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
14
ACS Paragon Plus Environment

Page 15 of 41
ACS Catalysis
1
2
3
4
5
6
7
8
9
Figure 1. (a) XRD patterns of gꢀC3N4, gꢀC3N4@C and gꢀC3N4@αꢀFe2O3/C. (b) XRD
pattern of C·MꢀgꢀC3N4@αꢀFe2O3/C. (c) TEM and (inset) HRTEM images of
gꢀC3N4@αꢀFe2O3/C. (d) TEM and (inset) overlapped EELS images of
gꢀC3N4@αꢀFe2O3/C. N EELS in green color was covered by C EELS in red color and
the extra boundary of C of ~8 nm was marked.
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
The chemical environment for C, N and Fe were further probed by Xꢀray
photoelectron spectroscope (XPS). The existence of Fe element in
gꢀC3N4@αꢀFe2O3/C was also verified by full scan XPS spectra with a ratio of 4.5 wt
%
(Figure 2a), which was consistent with the result of TEM energy dispersive
spectrometer (EDS, 4.25 wt%, Figure S5f). Inductively Coupled Plasma Mass
Spectrometry (ICPꢀMS) was also carried out to determine the amount of Fe in bulk
gꢀC3N4@αꢀFe2O3/C. The content of Fe element was 3.5 wt%. By contrast with the
XPS/TEM results, the slight lower content of Fe measured by ICPꢀMS further
confirmed the enrichment of αꢀFe2O3 moiety on the surface of gꢀC3N4. In the
highꢀresolution C1s spectra (Figure 2b), two main peaks at 284.7 eV and 288.1 eV
were observed, which were assigned to sp2 CꢀC bonds and sp2 NꢀC=N bonds,
respectively. The sp2 NꢀC=N bonds corresponded to the triꢀ
sꢀtriazine unit, whereas the
sp2 CꢀC bonds mainly belonged to the C moiety/nanoꢀshell on the surface.25 Two
small peaks at 285.9 eV and 288.9 eV corresponded to the residual C=O and CꢀO
species in the nanoꢀshell after calcination.44 The N 1s spectra in Figure S7 again
evidenced that the nanoꢀshell had negligible influence on the primary structure of
15
ACS Paragon Plus Environment

ACS Catalysis
Page 16 of 41
1
2
3
4
5
6
7
8
9
gꢀC3N4. The O 1s spectra (Figure 2c) exhibited two peaks at 533.2 eV and 531.8 eV
for both gꢀC3N4@C and gꢀC3N4@αꢀFe2O3/C, which corresponded to C=O and CꢀO.44
Two extra peaks in O 1s spectrum of gꢀC3N4@αꢀFe2O3/C were observed at 529.8 eV
and 530.8 eV, which were originated from αꢀFe2O3 and FeꢀOꢀC, respectively.44 The
existence of FeꢀOꢀC indicated the strong interaction between αꢀFe2O3 moiety and C
moiety, which may serve as an electron pathway to facilitate the electron transfer via
FeꢀOꢀC.45 In Fe 2p spectrum (Figure 2d), the binding energy of 711.3 eV and 724.8
eV belonged to 2p3/2 and 2p1/2 of αꢀFe2O3, respectively. The peak at 718.6 eV was the
satellite peak of 2p3/2 for αꢀFe2O3.46 Both O 1s and Fe 2p spectra confirmed the
formation of αꢀFe2O3 in the gꢀC3N4@αꢀFe2O3/C sample.
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 2. (a) Full scan XPS spectra and highꢀresolution (b) C 1s and (c) O 1s spectra
of gꢀC3N4@C and gꢀC3N4@αꢀFe2O3/C. (d) Highꢀresolution Fe 2p spectrum of
16
ACS Paragon Plus Environment

Page 17 of 41
ACS Catalysis
1
2
3
4
gꢀC3N4@αꢀFe2O3/C.
5
6
7
8
9
3.2. Optical and Photoelectric Properties of gꢀC3N4@αꢀFe2O3/C Core@Shell
Photocatalysts. The charge generation and charge transfer behaviors were
characterized by a series of optical and photoelectric methods. UVꢀvis diffuse
reflectance spectroscope (DRS) results revealed that gꢀC3N4 had an absorption edge at
464.6 nm (Figure 3a). After the introduction of αꢀFe2O3/C or C nanoꢀshell onto
gꢀC3N4, the visible light absorption was expanded to 800 nm. It was worth noting that
the expanded visible light absorption was primarily attributed to the C moiety or the C
nanoꢀshell, which had no function in generating available charge carriers. The C
moiety would even weaken the incident light absorbed by gꢀC3N4, leading to a
decrease of effective light absorption. The corresponding bandgap energies of
gꢀC3N4@C and gꢀC3N4@αꢀFe2O3/C were 2.48 eV and 2.51 eV (Figure 3b), which
was smaller than gꢀC3N4 (2.60 eV). The band position was calculated using XPS
valence spectra (Figure 3c) and bandgap energies. gꢀC3N4 exhibited a band position
of valence band (VB) at 1.78 eV and conduction band (CB) at ꢀ0.82 eV. As for
gꢀC3N4@C, VB stayed in almost the same positon (1.77 eV) while CB became less
negative (ꢀ0.712 eV). According to previous literature,15 CB of gꢀC3N4 could be
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
connected to the LUMO levels of triꢀ
carbon pz orbitals. Accordingly, VB of gꢀC3N4 was a combination of the HOMO
levels of triꢀ ꢀtriazine, which was derived from the nitrogen pz orbitals. More carbon
sꢀtriazine, which predominantly consisted of
s
pz orbitals introduced by C nanoꢀshell in gꢀC3N4@C may stabilize the excited
17
ACS Paragon Plus Environment

ACS Catalysis
Page 18 of 41
1
2
3
4
5
6
7
8
9
electrons at lower energy. Another explanation was that the work function of C was
less negative than CB of gꢀC3N4, leading to a band bending of CB towards less
negative after the formation of gꢀC3N4@C.7 Considering the preparation process of in
situ formation of C nanoꢀshell on gꢀC3N4, the interaction between gꢀC3N4 and C
nanoꢀshell would turn into covalent bonds during calcination. The energy level
hybridization were likely to happen at molecular level, where the orbitals played a
circuital role in the formation of energy band. Notably, the detected VB of
gꢀC3N4@αꢀFe2O3/C became less positive (1.48 eV) than gꢀC3N4 and gꢀC3N4@C,
which may be arisen from the rearrangement of energy levels among CB of αꢀFe2O3,
the work function of C and/or VB of gꢀC3N4. Since the signal of αꢀFe2O3 moiety in
DRS was covered by the strong absorption of C moiety, the calculated CB of
gꢀC3N4@αꢀFe2O3/C (ꢀ1.032 eV) from XPS valence spectra and DRS results cannot
reflect the real position of CB. We deduced that the CB of gꢀC3N4@αꢀFe2O3/C was
located at the same position with that of gꢀC3N4@C, considering the little impact of
relative low CB of αꢀFe2O3 to that of gꢀC3N4.
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
The charge recombination was measured through photoluminescence (PL) spectra
(Figure 3d). Commonly, carbon materials could afford outstanding electron mobility
to semiconductors. Electrons and holes were separated at the interface, which then
decreased the charge recombination and prolonged the life time of excited charge
carriers.19 As shown in Figure 3d, gꢀC3N4@αꢀFe2O3/C and gꢀC3N4@C showed much
weaker peak intensity compared with gꢀC3N4, indicating the positive role of C
moiety/nanoꢀshell
in
suppressing
charge
recombination.
Moreover,
18
ACS Paragon Plus Environment

Page 19 of 41
ACS Catalysis
1
2
3
4
5
6
gꢀC3N4@αꢀFe2O3/C showed a little higher luminescence intensity than gꢀC3N4@C,
mainly a result of the enhanced charge generation from αꢀFe2O3 moiety.
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 3. (a) UVꢀvis diffuse reflectance spectra. (b) Calculated band gap patterns
based on UVꢀvis diffuse reflectance spectra. (c) Valance band XPS spectra and (d) PL
spectra of gꢀC3N4, gꢀC3N4@C and gꢀC3N4@αꢀFe2O3/C.
Photoelectronic properties of gꢀC3N4, gꢀC3N4@C and gꢀC3N4@αꢀFe2O3/C were
further investigated to elucidate the charge generation and charge transfer behaviors
over gꢀC3N4@αꢀFe2O3/C. As shown in Figure 4a, gꢀC3N4@αꢀFe2O3/C exhibited the
highest photocurrent density of 8.783 ꢁA/cm2, which was 3.26 and 2.10 times that of
gꢀC3N4 and gꢀC3N4@C, respectively. The photocurrent density reflected the available
charge carriers migrating in the circuit, which was determined by charge generation
19
ACS Paragon Plus Environment

ACS Catalysis
Page 20 of 41
1
2
3
4
5
6
7
8
9
and charge transfer. Therefore, to interpret the prominent photocurrent density of
gꢀC3N4@αꢀFe2O3/C, we favored to firstly investigate the charge generation ability.
Methylviologen dichloride (MVCl2) was adopted as a fast electron scavenger in
electrolyte to measure the carrier density.47 In the presence of MVCl2, the electron
transfer efficiency was considered to be 100%, in which case the charge
recombination of photocatalysts would be minimized. The asꢀobserved carrier density
could be ascribed to the contribution of charge generation. Meanwhile, the
photocurrent onset potential in a voltammogram was equal to the quasi Fermi level of
majority carriers (electrons). The carrier density ratio between different samples could
be calculated in the quasi Fermi level by Nernst equation [Eq.4]:47
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
Ef 1 − Ef 2 = kT ln(N f 1 / N f 2 ) / e
(4)
where Ef was the quasi Fermi level,
k
was the Boltzmann’s Constant,
was the elementary charge.
T
was the
temperature, Nf was the carrier density and
e
As shown in Figure 4b, gꢀC3N4@C (0.220 V vs. RHE) had a less positive onset
potential than gꢀC3N4 (0.239 V vs. RHE), whereas gꢀC3N4@αꢀFe2O3/C (0.250V vs.
RHE) had a more positive onset potential than gꢀC3N4. The carrier density of
gꢀC3N4@C was only a half as that of gꢀC3N4. However, gꢀC3N4@αꢀFe2O3/C exhibited
higher carrier density than both gꢀC3N4 and gꢀC3N4@C, with the ratio of 1.53 and
3.21, respectively. The decreased carrier density of gꢀC3N4@/C was the result of the
undesirable light absorption from C nanoꢀshell that attenuated the charge generation
ability. As shown in Figure S8 and Table S1, the charge generation ability decreased
sharply with the increased thickness of C nanoꢀshell. As for gꢀC3N4@αꢀFe2O3/C,
20
ACS Paragon Plus Environment

Page 21 of 41
ACS Catalysis
1
2
3
4
5
6
7
8
9
despite the existence of C moiety, the coꢀexcitation of the αꢀFe2O3 moiety and gꢀC3N4
exhibited an even higher carrier density than that of gꢀC3N4. Therefore, αꢀFe2O3
moiety in the nanoꢀshell was not just eliminate the inefficient light absorption of C
moiety but endowed gꢀC3N4@αꢀFe2O3/C with extra charge generation ability
compared to gꢀC3N4. As for the charge transfer behaviors, electrochemical impedance
spectroscope (EIS) was carried out to study the resistance during interfacial charge
transfer (Figure 4c). The arc radius of gꢀC3N4@αꢀFe2O3/C and gꢀC3N4@C on the EIS
Nyquist plot was smaller than that of gꢀC3N4, indicating the smaller resistance of
charge transfer of gꢀC3N4@αꢀFe2O3/C and gꢀC3N4@C than that of gꢀC3N4. The
smallest resistance observed in gꢀC3N4@αꢀFe2O3/C indicated the interaction between
αꢀFe2O3 moiety and C moiety could further promote the charge transfer. Combined
with the results of PL spectra, it can be derived that C nanoꢀshell in gꢀC3N4@C or C
moiety in gꢀC3N4@αꢀFe2O3/C was responsible for the reinforced charge transfer. In
consequence, for gꢀC3N4@C, the enhanced charge transfer and partly light shielding
by the C nanoꢀshell led to the slightly altered photocurrent density. By contrast, the
coꢀexcitation of αꢀFe2O3 and gꢀC3N4 as well as the Cꢀreinforced electron transfer
moiety in gꢀC3N4@αꢀFe2O3/C eventually led to the significantly improved
photocurrent density.
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
21
ACS Paragon Plus Environment

ACS Catalysis
Page 22 of 41
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 4. (a) Photocurrent responses under chopped simulated sunlight in 0.1 M
Na2SO4. (b) Voltammogram under chopped simulated sunlight in 0.1 M Na2SO4 with
1 mM MVCl2. (c) EIS Nyquist plots under dark in 0.1 M Na2SO4.
3.3. Photocatalytic Activities of gꢀC3N4@αꢀFe2O3/C Core@Shell Photocatalysts.
Then, gꢀC3N4@αꢀFe2O3/C photocatalysts were applied for NADH regeneration. As
was well known, redox reactions afford the essential driving force for all varieties of
life, where NADH was evolved as one essential hydride donor. Even in majority of
redox reactions in vitro performed by oxidoreductases, NADH was frequently utilized
for energy conversion and storage.13,48,49 The expensive but indispensable features of
NADH required effective regeneration systems. In our regeneration system (Figure
5a), a rhodium complex, [Cp*Rh(bpy)H2O]2+, was utilized as a mediator to transfer
22
ACS Paragon Plus Environment

Page 23 of 41
ACS Catalysis
1
2
3
4
5
6
7
8
9
excited electrons and hydrogen to NAD+, thereby accomplishing the process of
NADH regeneration.50 The photocurrent density measured in the rhodium complex
solution suggested that the αꢀFe2O3/C nanoꢀshell may possess better affinity to the
rhodium complex (Figure S9). The photocatalytic NADH regeneration was then
conducted under the following conditions by using TEOA as a sacrificial agent, a blue
LED (100W, 405nm) incident light with single wavelength as the light source, the
rhodium complex as the electron mediator. In Figure 5b, gꢀC3N4@αꢀFe2O3/C showed
enhanced photocatalytic NADH regeneration yield compared to gꢀC3N4 and
gꢀC3N4@C. The reaction enabled by gꢀC3N4@αꢀFe2O3/C reached equilibrium in
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
approximately 16 minutes, acquiring a yield of 76.3% and an initial reaction rate (
of 7.7 mmol hꢀ1 gꢀ1 (based on the first 6ꢀmin reaction, calculated with [Eq.1] in 2.5),
which was, to the best of our knowledge, among the highest for photocatalytic
r)
r
NADH regeneration.33 Notably, the slow reaction rate of the first 2ꢀmin was due to the
photoꢀactivation of mediator.50 The kinetic analysis was also performed in the first
6ꢀmin reaction. The ln(C/C0) versus time showed linear curve, identifying the
firstꢀorder kinetics of the reaction (Figure 5c). The reaction kinetics constants of
gꢀC3N4@C and gꢀC3N4@αꢀFe2O3/C were calculated, which exhibited increasing
factors of 1.49, 2.92 compared with gꢀC3N4, respectively. The recycling ability of
gꢀC3N4@αꢀFe2O3/C was shown in Figure S10. The relative activity retained 66.8%
after 8 cycles. We further incubated the photocatalyst in the reaction medium for 2
hours and explored the photocatalytic activity. As shown in Figure S11, the relative
activity decreased about 20% after incubation. The activity loss was attributed to the
23
ACS Paragon Plus Environment

ACS Catalysis
Page 24 of 41
1
2
3
4
5
6
7
8
9
selfꢀreduction of photocatalyst by accumulated excited electrons. The higher loss of
activity during recycling was attributed to the operation, considering the few amount
of photocatalyst (1 mg) we used in experiment.
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
The amount of Fe3+ was then altered to prepare gꢀC3N4@αꢀFe2O3/C with different
contents of αꢀFe2O3. All gꢀC3N4@αꢀFe2O3/C samples exhibited better NADH
regeneration yield compared to gꢀC3N4@C (Figure 5d). The highest yield achieved by
gꢀC3N4@αꢀFe2O3/Cꢀ0.1 was contributed to the wellꢀmatched behaviors of charge
generation and charge transfer. To verify the universality of our method, natural
polyphenols, such as tea and red wine, were also applied to form αꢀFe2O3/C
nanoꢀshell on gꢀC3N4. The NADH regeneration yields of gꢀC3N4@αꢀFe2O3/C derived
from tea and wine were also investigated as a function of reaction time (Figure S12).
Both gꢀC3N4@αꢀFe2O3/C samples showed yield enhancement when compared with
gꢀC3N4 and gꢀC3N4@C (Figure 5e). After reaction equilibrium, the NADH
regeneration yields of gꢀC3N4@αꢀFe2O3/C prepared by tea and wine were 75.6% and
73.3%, respectively. The corresponding initial reaction rates were calculated as 8.9
mmol hꢀ1 gꢀ1 and 8.6 mmol hꢀ1 gꢀ1. Hydrochloric acid (HCl) was then applied to
corrode gꢀC3N4@αꢀFe2O3/C and gꢀC3N4@C (as a reference) to study the chemical
stability of αꢀFe2O3 moiety in the nanoꢀshell. In Figure S13, the NADH regeneration
yield increased after the corrosion of HCl for both gꢀC3N4@αꢀFe2O3/C and
gꢀC3N4@C. The nonꢀdecrease phenomenon indicated that αꢀFe2O3 moiety was quite
stable in the nanoꢀshell, which was ascribed to the strong interaction of FeꢀOꢀC
between αꢀFe2O3 moiety and C moiety. The increased yield was due to the exfoliation
24
ACS Paragon Plus Environment

Page 25 of 41
ACS Catalysis
1
2
3
4
5
6
7
8
9
of the samples after the corrosion. As shown in Figure S14, the SEM images of
samples before and after corrosion indicated that the bulk gꢀC3N4@αꢀFe2O3/C and
gꢀC3N4@C were exfoliated into flakes with increased BET surface area (Table S2),
which shortened the distance of electrons transfer from the bulk to the surface and
provided more active sites for surface reaction.51,52 Finally, the photocatalytic NADH
regeneration driven by gꢀC3N4@αꢀFe2O3/C was coupled with enzymatic reduction of
formaldehyde to examine the reactivity of photoꢀregenerated NADH towards a
commercial enzyme, alcohol dehydrogenase (ADH, from Saccharomyces cerevisiae).
In Figure 5f, 1.87 mM methanol was generated in 40 minutes in the presence of 1
mM NAD+, confirming the reactivity of photoꢀregenerated NADH. The total turnover
number of NADH during 40 minutes was 3.74 mmol gꢀ1 and each single NADH
molecule was utilized 2.45 times for methanol production (for details of calculation,
see 2.5). Control experiments were also carried out, which indicated light,
photocatalyst and ADH were all indispensable for the production of methanol in our
system.
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
It should be noted that the sustainability of the photocatalytic NADH regeneration
was restricted by the utilization of
a
noble metalꢀbased mediator
([Cp*Rh(bpy)H2O]2+). The mediator commonly used in the system not only promoted
the electron transfer between photocatalyst and NAD+, but more importantly,
exhibited 100% selectivity towards enzymatic active 1,4ꢀNADH. Great efforts have
been devoted to this field to explore a mediatorꢀfree system. Liu et al.12 reported the
possibility of direct electron transfer between gꢀC3N4 based photocatalyst and NAD+.
25
ACS Paragon Plus Environment

ACS Catalysis
Page 26 of 41
1
2
3
4
5
6
7
8
9
Without the assistant of mediator, the yield of NADH could reach 50% in alkaline
media (pH = 10). The selectivity of 1,4ꢀNADH was only about 34% according to the
subsequent enzymatic reaction, which left us a research space to further improve the
selectivity for acquiring a better and more sustainable system. Wang et al.53 reported a
heterogeneous catalytic NADH regeneration system using a commercial Pt/Al2O3 as
catalyst and H2 as reductant. The yield of NADH reached 50% with a selectivity of
100% towards 1,4ꢀNADH. The mediatorꢀindependent system exhibited a superior
selectivity and comparable activity by contrast with mediatorꢀdependent system
expect that the utilization of noble metal Pt and H2 may lower the sustainability. If the
mechanism of the superior selectivity is revealed and then applied to nonꢀnoble metals,
such as nickel, iron or copper, in the future, it would be a promising strategy to
explore noble metalꢀfree mediator or even highly integrated heterogeneous
photocatalysts for NADH regeneration and enzymatic reactions.
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
26
ACS Paragon Plus Environment

Page 27 of 41
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. (a) Schematic illustration of the coupling system with photocatalytic NADH
regeneration and enzymatic reduction of formaldehyde. (b) Timeꢀresolved NADH
regeneration yields and (c) kinetic analysis of gꢀC3N4, gꢀC3N4@C and
gꢀC3N4@αꢀFe2O3/C. (d) Timeꢀresolved NADH regeneration yields of
gꢀC3N4@αꢀFe2O3/C as a function of Fe3+ amounts initially added for the preparation
of gꢀC3N4@αꢀFe2O3/C. (e) NADH regeneration yields and initial reaction rates of
gꢀC3N4, gꢀC3N4@C and gꢀC3N4@αꢀFe2O3/C derived from tannic acid, tea and red
wine. (f) Methanol production enabled by the coupling system with photocatalytic
27
ACS Paragon Plus Environment

ACS Catalysis
Page 28 of 41
1
2
3
4
5
6
NADH regeneration of gꢀC3N4@αꢀFe2O3/C and enzymatic reduction of formaldehyde
under different conditions.
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
3.4. Mechanism Analysis. The behaviors of charge generation and charge transfer in
photocatalytic process are proposed in Scheme 2a. The charge transfer between
αꢀFe2O3 moiety and gꢀC3N4 follows the Zꢀscheme mechanism. Firstly, αꢀFe2O3
moiety absorbs light and generates excited electronꢀhole pairs. The holes are
quenched by TEOA whereas the electrons in CB of αꢀFe2O3 transfer to VB of gꢀC3N4
through C moiety. Meanwhile, electrons in VB of gꢀC3N4 are excited by incident light
and then transfer to CB of gꢀC3N4. The holes left in VB of gꢀC3N4 are occupied by the
electrons from αꢀFe2O3 moiety. The excited electrons in CB of gꢀC3N4 migrate to C
moiety and participate in NADH regeneration. The complicated electron pathways are
guaranteed by three factors: (1) outstanding electron mobility of C moiety; (2) strong
interactions of FeꢀOꢀC bond between αꢀFe2O3 moiety and C moiety and (3) excellent
interfacial compatibility between C moiety and gꢀC3N4. The energy levels of different
species in our photocatalytic NADH regeneration system are shown in Scheme 2b.
The redox potentials of rhodium complex (ꢀ4.01 eV) and NADH (ꢀ4.59 eV) are higher
than CB of αꢀFe2O3 (ꢀ4.78 eV), in which case only the electrons in CB of gꢀC3N4 have
sufficient energy for NADH regeneration. This theoretically supports the existence of
Zꢀscheme mechanism. The coꢀexcitation of two photosensitizers and Cꢀreinforced
electron transfer increase the amount of available charge carriers for redox reaction.
Briefly, the excitation of αꢀFe2O3 moiety minimizes the inefficient light absorption by
28
ACS Paragon Plus Environment

Page 29 of 41
ACS Catalysis
1
2
3
4
5
6
7
8
9
C moiety and offers more photoꢀgenerated electrons, while the electron transfer
“highway” constructed through C moiety maximizes the efficient utilization of
photoꢀgenerated electrons. The synergistic effect of αꢀFe2O3 moiety and C moiety in
the αꢀFe2O3/C nanoꢀshell leads to a significantly enhanced photocatalytic efficiency.
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
Scheme 2. (a) Illustration of electron generation and transfer pathway in
gꢀC3N4@αꢀFe2O3/C and photocatalytic NADH regeneration coupled with enzymatic
reduction of formaldehyde. (b) Normalized potential energy diagram of photocatalytic
pathway. Red arrows represent the electron pathway in NADH regeneration process.
The redox potentials are obtained from literature.15,19,54ꢀ57
4. CONCLUSIONS
29
ACS Paragon Plus Environment