Overwhelming the Performance of Single Atoms with Atomic Clusters

Subscriber access provided by UNIV OF SOUTHERN INDIANA

Article

Overwhelming the performance of single atoms with atomic clusters for platinum-catalyzed hydrogen evolution Minghao Sun, Jiapeng Ji, Mingyu Hu, Mouyi Weng, Yaping Zhang, Haisheng Yu, JiaJun Tang, Jun-chao Zheng, Zheng Jiang, Feng Pan, Chengdu Liang, and Zhan Lin ACS Catal., Just Accepted Manuscript • DOI: 10.1021/acscatal.9b02305 • Publication Date (Web): 26 Jul 2019 Downloaded from pubs.acs.org on July 26, 2019

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 36 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

ACS Catalysis

1

Overwhelming the performance of single atoms with atomic

2

clusters for platinum-catalyzed hydrogen evolution

3

Minghao Sun1,‡, Jiapeng Ji1,‡, Mingyu Hu2,‡, Mouyi Weng2,‡, Yaping Zhang3, Haisheng Yu4,

4

Jiajun Tang5, Junchao Zheng6,*, Zheng Jiang4,*, Feng Pan2,*, Chengdu Liang1 and Zhan Lin1,5*

5

1

6

Technology, College of Chemical and Biological Engineering, Zhejiang University, Hangzhou

7

310027, P. R. China.

8

2

9

518055, P. R. China.

Zhejiang Provincial Key Laboratory of Advanced Chemical Engineering Manufacture

School of Advanced Materials, Peking University, Shenzhen Graduate School, Shenzhen

10

3

11

and Engineering, Southwest University of Science and Technology, Mianyang 621010, China.

12

4

13

Academy of Sciences, Shanghai 201204, P. R. China.

14

5

15

Guangzhou 510006, P. R. China.

16

6

17

China.

18 19 20 21 22 23

Corresponding Authors * [email protected] (Z. L.) * [email protected] (F. P.) * [email protected] (J. Z.) * [email protected] (Z. J.)

State Key Laboratory of Environment-friendly Energy Materials, School of Materials Science

Shanghai Synchrotron Radiation Facility, Shanghai Institute of Applied Physics, Chinese

School of Chemical Engineering and Light Industry, Guangdong University of Technology,

School of Metallurgy and Environment, Central South University, Changsha 410083, P. R.

ACS Paragon Plus Environment

1

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

Page 2 of 36

1

ABSTRACT

2

As one of dominant configurations, platinum (Pt) single atomic catalysts (SACs) have pushed the

3

performance of hydrogen evolution reaction (HER) to an unprecedented level due to maximized

4

atomic utilization efficiency of Pt atoms. However, the contribution of atomic clusters, which

5

exist in SACs as well, to the overall catalytic performance is always overlooked, thus limiting

6

further enhancing the performance of Pt-catalyzed HER. Herein, we report anchoring Pt atomic

7

clusters on N-doped graphene for ultrahigh performance of HER. Benefiting from optimized

8

electron transfer and larger binding energy between active centers and the substrate, Pt atomic

9

cluster catalysts (ACCs) exhibit higher catalytic activity than their single atomic counterparts

10

after 4000 cycles and more than 16 hours for HER in an acid solution. These findings reveal vast

11

opportunity of enhancing the catalytic performance of chemical reactions with noble metal-based

12

ACCs in the near future.

13

KEYWORDS single atomic catalyst, atomic cluster, platinum, graphene, hydrogen evolution

14

TOC GRAPHICS

15


2

Page 3 of 36 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

ACS Catalysis

1

INTRODUCTION

2

Hydrogen evolution through the electrocatalytic reduction of water is one of the most promising

3

solution to the problem of fossil fuel depletion. Due to outstanding catalytic activity and stability,

4

platinum (Pt) is currently the best metal catalyst for electrocatalytic hydrogen evolution reaction

5

(HER).1-2 However, Pt is intrinsically noble,3-4 it’s necessary to reduce its loading in catalysts

6

while maintaining high catalytic efficiency.5-6 Numerous efforts have been devoted to optimize

7

the configurations of active centers in Pt-based catalysts, commonly used strategies include

8

alloying Pt with transition-metals,7-9 shape-control synthesis,10-11 and down-sizing of active

9

centers.12-13 Among all catalyst configurations, single-atom catalyst (SAC) represents the lowest

10

limit of particle size while maintaining maximum atomic efficiency.14 In particular, extremely

11

high catalytic activity and selectivity were explored for noble metal-based SACs.15-16 Therefore,

12

the fabrication of Pt-based SACs has become the frontier of Pt-catalyzed HER, which has

13

boosted the performance of electrocatalytic HER to an unprecedented level.2, 17-18

14

Currently, recognitions on the catalytic performance of Pt-based SACs are mainly based on Pt

15

single-atoms, which is considered to be the dominant configuration of active centers in Pt-based

16

SACs.2 In fact, single atoms are not the only configuration of active centers in many SACs,

17

where atomic clusters exist as well.19 Owing to extremely high surface free energy, single atoms

18

tend to agglomerate into atomic clusters during synthesis or under working conditions.20-22

19

Therefore, the formation of atomic clusters is more energetically favored and thus more stable

20

under working conditions than single atoms. Moreover, single atoms do not exhibit their

21

overwhelming advantages in every circumstance, e.g., the catalytic activity of Rh-based atomic

22

clusters toward CO oxidation is 5 times higher than their single atomic counterparts.23 As such,

23

it’s reasonable to speculate that atomic clusters in Pt-based SACs also contribute to the


3

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

Page 4 of 36

1

electrocatalytic performance of HER, while catalysts compose of Pt-based atomic clusters could

2

be a promising catalyst configuration to replace state-of-the-art commercial Pt/C catalyst and

3

even push the performance of electrocatalytic HER to a higher level. However, owing to the

4

incapability of current synthesis methods, the comparison of intrinsic catalytic capability of Pt-

5

based atomic clusters with their single atomic counterparts is rarely reported so far.

6

In this work, we report Pt-based atomic cluster catalysts (ACCs) loaded on N-doped

7

graphene/graphitic-C3N4 (AC Pt-NG/C) via a photodeposition route, as a novel configuration of

8

catalytic active centers for HER. The reason why we apply the phrase “atomic clusters” is that

9

we focus on studying the clusters that might exist in traditional SACs. Meanwhile, the word

10

“atomic” best describes the dimension of such clusters. Under visible-light irradiation, graphitic-

11

C3N4 (g-C3N4) produces electrons,24 which are separated by N-doped graphene (NG) and

12

transferred on the π-π conjugation network of NG, allowing for the reduction/anchoring of Pt

13

precursors and the formation of Pt atomic clusters on NG. Comparing with previous works in

14

which only semiconductors are used as substrates to construct SACs via photodeposition,25-26 a

15

heterogeneous-structured substrate (i.e., NG/g-C3N4 composite) is applied in our work, which

16

not only lead to improved photo-electric transformation efficiency of g-C3N4 under irradiation,

17

but also enables optimizing the electronic structure of active catalytic sites in AC Pt-NG/C via

18

tuning the interaction between Pt atomic clusters and substrates (i.e., NG). As proved by DFT

19

calculations, the N content in NG plays a key role in the formation and anchoring of Pt-based

20

atomic cluster, which also contribute to a more desirable adsorption property of hydrogen on Pt-

21

based ACCs than on Pt single atoms or commercial Pt/C catalyst during the catalytic process. As

22

a result, AC Pt-NG/C exhibits high catalytic activity towards HER, exceeding that of commercial

23

20 wt% Pt/C catalysts by a factor of 21. Meanwhile, AC Pt-NG/C retains extraordinary stability


4

Page 5 of 36 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

ACS Catalysis

1

in the chronopotentiometry test for over 16 hours and only 5.1 mV loses after 4000 cycles, which

2

is among the best reported, to the best of our knowledge. This work not only provides a novel

3

strategy for constructing a highly efficient and stable catalyst configuration for HER, but also

4

promotes large-scale application of Pt-based catalysts in the near future.

5

RESULTS AND DISCUSSION

6

Synthesis and characterizations of AC Pt-NG/C. Owing to different chemical potentials,

7

charge transfers between single-atoms and the substrate brings vast opportunity for stabilizing

8

metallic single-atoms by tuning the structure of the substrate.27 Among various elements that

9

were applied to modify carbon substrates, N is the most researched, which not only provides

10

strong anchoring sites for metallic single-atoms but also changes their electronic properties as

11

well.28 Therefore, in this work we selected NG as the substrate for loading Pt atomic clusters,

12

which was achieved by the annealing of a GO/urea mixture (Step I of Figure 1a).29 Because the

13

pyrolysis of urea leads to formation of g-C3N4,24 the pyrolysis product is a NG/g-C3N4 (NG/C)

14

composite.

15

N-doping in NG was first proved by X-ray photoelectron spectroscopy (XPS) measurement.

16

As shown in Figure 1b and 1c, the N content in NG/C increased violently from 1.2 wt% to 12.05

17

wt% (Table S1) after annealing, most of which was attributed to the increment of pyridinic N.

18

The existence of N in GO is attributed to the oxidation treatment with HNO3 in the Hummer’s

19

method.30 Meanwhile, the decrease of O content and increase of C content in the XPS results

20

(Figure S1 and Table S1) are consistent with the digital images of the products (Figure S2), in

21

which the change of color (from dark brown to black) suggests the restoration of graphene-like

22

properties in NG,31 which is favorable for achieving better performance in electrocatalytic HER.


5

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

Page 6 of 36

1

Considering that g-C3N4 also contributes to the increased N content in the XPS analysis (Figure

2

1c), we further carried out Raman spectroscopy to confirm N-doping in NG. As shown in Figure

3

S3, due to the introduction of new structural defects (i.e., N-doping sites) onto carbon skeleton,32

4

the D (ca. 1350 cm-1) band to G (ca. 1570 cm-1) band peak-intensity ratios of the products

5

increased from 0.92 to 1.13 after annealing, while the absence of 2D peak further demonstrates

6

the defective structure of NG.33-34 Next, electron microscopy techniques were applied to confirm

7

the existence of g-C3N4 in NG/C. Due to the similar layered-structure of g-C3N4 and graphene,

8

they tend to become overlapped during annealing.32 For this reason, NG/C appears darker in

9

certain areas of the transmission electron microscopy (TEM, Figure 1d) image comparing with

10

that of GO (Figure S4a), while the typical wrinkled structure (Figure S4b and S4c) and large

11

specific surface area of GO were largely retained in NG/C (353 and 282 m2 g-1, respectively, for

12

GO and NG/C, Figure S5), which is crucial for achieving high loading content and homogeneous

13

anchoring of Pt atomic clusters.35 The hysteresis in the range of P/P0=0.4-0.9 in the isotherms of

14

NG/C suggests existence of mesoporous contents, which might be attributed to the formation of

15

g-C3N4. Meanwhile, the formation of g-C3N4 was further proved by Fourier transform infrared

16

spectroscopy (FTIR). In Figure S6a, the emergence of new peaks at ca. 800 and 1250 cm-1 in the

17

spectra of NG/C are attributed to the breathing vibration of triazine units and stretching mode of

18

C-N in g-C3N4.30 While in the powder x-ray diffraction (XRD) patterns (Figure S6b), the typical

19

interlayer-stacking of aromatic segments peak at 27.5° is indexed as (002) of g-C3N4.36 Based on

20

the above, we conclude that NG/C is a highly N-doped material, while the g-C3N4 on NG/C is

21

favorable for the subsequent synthesis of AC Pt-NG/C via the photodeposition.

22


6

Page 7 of 36 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

ACS Catalysis

1 2

Figure 1. Preparation and characterization of AC Pt-NG/C. (a) Schematic illustration of the

3

main steps involved for the preparation of AC Pt-NG/C. XPS N 1s spectra of (b) GO and (c)

4

NG/C. (d) TEM image of NG/C. (e) ADF-STEM image of AC Pt-NG/C (bright dots in red

5

circles). (f) Normalized XANES spectra at the Pt L3-edge and (g) k2-weighted R-space FT

6

spectra from EXAFS of AC Pt-NG/C, Pt foil and PtO2. Scale bars: d 5 μm and e 5 nm.

7

For the synthesis of AC Pt-NG/C, an aqueous dispersion of NG/C containing target amount of

8

H2PtCl6 was exposed to visible-light irradiation (Step II of Figure 1a). The absorption edge of the

9

aqueous mixture is ca. 460 nm as determined by the UV–vis diffuse reflectance spectra (DRS,

10

Figure S7a), which is identical to that of g-C3N4.37 After irradiation, the peak-intensity ratio of


7

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

Page 8 of 36

1

AC Pt-NG/C further decreased to 0.98 comparing with that of NG in the Raman spectra (Figure

2

S7b), this is attributed to the elimination of oxygen-functional groups in NG by the photo-driven

3

electrons as proved by the XPS results (Figure S8a).31 The carbon and nitrogen contents are

4

almost consistent (Figure S8b and S8c), which confirms that NG could separate electrons from

5

g-C3N4 under irradiation and is consistent with the established role of g-C3N4 as a photo-electric

6

transformation mediator. The similar FTIR spectra of AC Pt-NG/C and NG/C (see Figure S6a)

7

suggest that the major fraction of the photo-generated electrons is consumed in the reduction of

8

Pt precursors or dissipated as heat rather than in the elimination of oxygen-functional groups.

9

Due to the Pt content in AC Pt-NG/C is low, no Pt nanoparticles (NPs) could be observed in

10

the SEM (Figure S9a) and TEM images of AC Pt-NG/C (Figure S9b). Particularly, it should be

11

noted the absence of Pt NPs even in the high-resolution transmission electron microscopy

12

(HRTEM, Figure S9c) images of AC Pt-NG/C and no characteristic peaks of Pt crystals could be

13

observed in the XRD patterns of AC Pt-NG/C (Figure S10), which suggests an extremely fine

14

dispersion of Pt in AC Pt-NG/C. Therefore, we further resorted to aberration-corrected annular

15

dark-field scanning transmission electron microscope (ADF-STEM) to investigate the dispersion

16

state of Pt in AC Pt-NG/C. As shown in Figure 1e, taking into consideration the sharp Z-contrast

17

between metallic materials and carbon matrix, we attribute uniformly distributed bright dots

18

(with diameters of ca. 0.5 nm) to small Pt clusters in AC Pt-NG/C,2 which exhibits atomically-

19

dispersion features and amounts to 2.0 wt% in AC Pt-NG/C according to the inductively coupled

20

plasma-optical emission spectrometer (ICP-OES) analysis. Energy dispersive spectroscopy (EDS)

21

mapping (Figure S11a-S11d) proves that only trace amount of Pt (Figure S11b) could be

22

detected on the surface of AC Pt-NG/C, which is consistent with the ICP-OES results.

23

Meanwhile, nitrogen (Figure S11c) and carbon (Figure S11d) elements are distributed


8

Page 9 of 36 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

ACS Catalysis

1

homogeneously all over the surface of AC Pt-NG/C, further proving that NG/C is a highly N-

2

doped substrate.

3

Owing to the surface free energy of single atoms is extremely high which tend to agglomerate

4

during processing, a decreased loading content of metallic active materials in SACs will lead to

5

higher atomic ratio of single atoms in the active centers.38 In this work, further increasing the

6

loading content of Pt (e.g., 3 wt%) in AC Pt-NG/C generates Pt nanoparticles in the products

7

(Figure S12a). Therefore, to compare the efficacy of atomic clusters with single atoms, we

8

halved the content of Pt in AC Pt-NG/C to prepare another catalyst with other synthesis

9

conditions being kept equal, the product is denoted as 1-Pt-NG/C. As shown in Figure S12b, no

10

Pt nanoparticles are detected in the ADF-STEM image of 1-Pt-NG/C. Theoretically, the atomic

11

ratio of single atoms in the active centers of 1-Pt-NG/C should be higher than that of AC Pt-

12

NG/C. In the BET tests (Figure S13), AC Pt-NG/C (272 m2 g-1) and 1-Pt-NG/C (263 m2 g-1)

13

demonstrate similar specific surface areas and the mesoporous feature of NG/C is retained in

14

both catalysts. Size distribution of Pt particles in AC Pt-NG/C and 1-Pt-NG/C are calculated

15

based on the ADF-STEM images (see Figure 1e and Figure S12b), the results (Figure S14)

16

reveal that the average sizes of Pt particles in AC Pt-NG/C (ca. 0.61 nm) is larger than that of 1-

17

Pt-NG/C (ca. 0.54 nm), suggesting that the latter contains a higher ratio of single atoms in active

18

centers. Comparing with previous reports of Pt-based SACs,38 the average sizes of Pt particles in

19

this work is relatively smaller, this could be attributed to the mild synthesis conditions of

20

photodeposition,39 which favors a homogeneous distribution of active centers in the products.

21

To achieve deeper insights into the particle size distribution of the catalysts, CO chemisorption

22

test is applied to investigate the ratio of surface Pt atoms in AC Pt-NG/C and 1-Pt-NG/C.

23

Generally, as the size of nanoparticle decreases, more atoms will be exposed to surface, which


9

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

Page 10 of 36

1

leads to higher CO-adsorption capacity of the nanoparticles. In this work, the CO/Metal mole

2

ratio (RCO/Metal) of AC Pt-NG/C and 1-Pt-NG/C is 0.71 and 0.49, respectively, suggesting that the

3

average size of Pt particles in AC Pt-NG/C is larger, which is consistent with the ADF-STEM

4

results. However, contrary from the expectation that most of the Pt atoms in the catalysts should

5

be exposed to surface, for the average sizes of Pt particles are extremely small,38 it’s surprising to

6

find that a large fraction of Pt atoms in the catalysts do not adsorb CO, especially for 1-Pt-NG/C.

7

According to the Okhlopkova’s work, CO adsorbs strongly on the surface of Pt-group metals.

8

However, back donation of metal d-electrons to the * antibonding orbital of CO is greatly

9

dampened in the case of metal ions, which weakens the strength of CO-Metal bond. As a result,

10

the adsorption energy of CO on positively charged Pt particles is much weaker compared to a

11

negatively charged or neutral particle,40 which deteriorates the adsorptive capability of some Pt

12

atoms toward CO. Based on the above, we speculate that some Pt atoms in AC Pt-NG/C and 1-

13

Pt-NG/C carry positive charges, which leads to the deviation of surface Pt atomic ratios in the

14

catalysts from expectation and is proved by the XPS tests. As shown in the Pt 4f spectrum

15

(Figure S15a and S15b), both AC Pt-NG/C and 1-Pt-NG/C contains considerable amount of Pt2+

16

in the active centers. Meanwhile, the Pt0/Pt2+ ratio of AC Pt-NG/C (ca. 1.25) is higher than that

17

of 1-Pt-NG/C (ca. 1.08), suggesting that a larger fraction of Pt atoms in 1-Pt-NG/C have a direct

18

contact with the substrates comparing with AC Pt-NG/C, which carry positive charges due to the

19

electronegativity of Pt is much smaller than the light atoms (e.g., C, N, O) in NG/C substrates

20

and do not adsorb CO strongly as suggested by the results of CO chemisorption tests. The small

21

peaks located at ca. 74 and 77.5 eV of the fitted spectra might be attributed to the Pt4+

22

coordinated with O atoms in the graphene substrates or the PtO2 contents,41 since the

23

electronegativity of O (3.5) is larger than those of C (2.5) and N (3.0).42 However, it’s unlikely


10

Page 11 of 36 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

ACS Catalysis

1

that the Pt4+ state is related to the (PtCl6)2- ions in the precursors, for we have washed the

2

catalysts thoroughly before using, and the corresponding peaks of Cl are extremely weak in the

3

full XPS spectra of both specimens (Figure S15c and S15d). To sum up, the CO chemisorption

4

tests and XPS results prove that the interaction between the Pt active centers (e.g., Pt atomic

5

clusters and Pt single atoms) and NG/C substrates is strong, moreover, AC Pt-NG/C possesses a

6

higher ratio of Pt atomic clusters in active centers than 1-Pt-NG/C.

7

X-ray absorption of fine structure (XAFS) measurement is a powerful technique that gives

8

structural information of the catalysts at atomic scale, which could be divided into X-ray

9

absorption near-edge structure (XANES) region and extended X-ray absorption fine structure

10

(EXAFS) region. In this work, the white-line (WL) peak of AC Pt-NG/C is located between

11

those of Pt foil and PtO2 (Figure 1f), suggesting some Pt atoms carry positive charges, which is

12

consistent with the XPS results (see Figure S15a). In the Fourier transforms of EXAFS (i.e., R-

13

space, Figure 1g),43 the only prominent peak at ca. 2.0 Å in the spectra of AC Pt-NG/C is related

14

to Pt-N/O/C contributions, suggesting no Pt NPs were formed in AC Pt-NG/C.44 Albeit the

15

difference in Pt contents, 1-Pt-NG/C possesses similar threshold energy (E0) and maximum

16

energy (Epeak) to AC Pt-NG/C in the XANES measurement (Figure S16a),45 which is consistent

17

with previous works yet the reason is still unknown.2 Meanwhile, the peak at ca. 2.0 Å in the

18

Fourier transforms of EXAFS of 1-Pt-NG/C is sharper comparing with that of AC Pt-NG/C

19

(Figure S16b), this might be attributed to the ratio of Pt single atoms is higher in the active

20

centers of 1-Pt-NG/C, which leads to more prominent “single-atomic” features of the catalyst.

21

It’s interesting to note that the strong peak at 2.6 Å in the FT-EXAFS spectra of Pt foil is absent

22

in those of AC Pt-NG/C and 1-Pt-NG/C. As proved by previous works,2 the significant Pt-Pt

23

peak in FT-EXAFS spectra will become dampened and shifted to higher positions in the case of


11

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

Page 12 of 36

1

Pt clusters, meanwhile, the intensity of the Pt-Pt peak become even weaker as the size of Pt

2

cluster reduces, which indicates that FT-EXAFS is sensitive to the size of atomic clusters. In our

3

work, the diameters of Pt atomic clusters are extremely small as proved by ADF-STEM images

4

(see Figure 1e and Figure S12b) and DFT calculations (see details below in Figure 3a), both

5

contribute to the very small peak at ca. 2.6 Å of FT-EXAFS spectra of AC Pt-NG/C. Therefore,

6

it’s reasonable to conclude that a large fraction of the Pt species in AC Pt-NG/C exhibit the

7

typical features of ACCs.

8 9

Figure 2. Electrocatalytic performance of the catalysts. (a) Electrochemical performances of the

10

catalysts in LSV measurements. (b) Tafel plots obtained from the polarization curves in (a). (c)


12

Page 13 of 36 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

ACS Catalysis

1

Mass activity of different catalysts at the potential of -50 mV (vs. RHE) derived from the LSV

2

curves in (a). (d) Chronopotentiometry curves (without IR compensation) of different catalysts.

3

Electrocatalytic measurements. We first carried out the electrochemical impedance

4

spectroscopy (EIS) measurements to gain more information about the electrolytes. In Figure S17,

5

the intersection of the semicircle and real axis in the high frequency region of Nyquist plots

6

represents the resistance of electrolyte solution (Rs).46-47 In this work, Rs kept consistent (ca. 8 Ω)

7

covering both HER and non-HER conditions at room temperature, thus making the results of

8

subsequent tests more convincing.

9

To make the electrochemical results of AC Pt-NG/C and 1-Pt-NG/C more convincing,

10

commercial 20 wt% Pt/C catalysts is applied as a control sample for carrying out electrochemical

11

tests (for more details of the 20 wt% Pt/C catalysts, see EXPERIMENTAL SECTION). In the

12

LSV tests, AC Pt-NG/C exhibits smaller overpotential (, 35.28 mV, Table S2) than that of 1-Pt-

13

NG/C (47.20 mV), commercial 20 wt% Pt/C catalysts (42.03 mV) and state-of-the-art Pt-based

14

SACs for HER (Table S3) at a current density of 10 mA cm–2 (Figure 2a), suggesting the high

15

catalytic activity of AC Pt-NG/C toward HER. Considering that a higher Pt content of may

16

enhance the HER catalytic performance,2 Tafel and Turnover frequency (TOF) analysis was

17

performed to gain further insight into the intrinsic catalytic activity of each Pt atom in the

18

catalysts. As shown in Figure 2b, the Tafel slope value of 20 wt% Pt/C is identical to that of 1-

19

Pt-NG/C (31 mV dec-1) and consistent with those values in published reports,2 while AC Pt-

20

NG/C exhibits smaller Tafel slopes of 27 mV dec-1 under the same conditions, suggesting that

21

the reaction was catalyzed through a Volmer–Tafel process.48-49 Exchange current densities (j0)

22

were obtained from the Tafel plots by an extrapolation method.50 The j0 value of AC Pt-NG/C

23

was calculated to be 0.520 mA cm-2, which is larger than that of 20 wt% Pt/C (0.513 mA cm-2)


13

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

Page 14 of 36

1

and 1-Pt-NG/C (0.240 mA cm-2), suggesting superior catalytic activity of Pt atoms in AC Pt-

2

NG/C.

3

TOF value is a direct criterion of the catalytic activity of a single active site.51 In this work,

4

the TOF values of the specimens were obtained at thermodynamic potential (e.g. 0 V vs RHE)

5

according to the exchange current densities, among which the TOF value of AC Pt-NG/C was

6

calculated to be 0.0927 S-1, which was an order of magnitude higher than that of 20 wt% Pt/C

7

(0.00915 S-1) and much higher than that of NP Pt-NG/C (0.0854 S-1). In most cases, a decreased

8

loading content of active materials will lead to higher TOF value of a catalyst,52 therefore, we

9

were surprised to find that AC Pt-NG/C with higher loading content of Pt than that of 1-Pt-NG/C

10

exhibited higher TOF value, which suggests that Pt atomic cluster possesses better HER catalytic

11

capability than Pt single atoms. For a better understanding of the HER catalytic activity, areal

12

current density of AC Pt-NG/C is compared with commercial Pt/C catalysts 1-Pt-NG/C and

13

state-of-the-art Pt-based SACs for HER (Figure S18, Table S3 and S4) at =50 mV versus RHE,

14

among which AC Pt-NG/C presents the highest value (36.9 mA cm–2). The results are more

15

striking when current densities are normalized to Pt mass. As shown in Table S4, the mass

16

current density of AC Pt-NG/C reached 6.508 A mg-1 at =0.05V (vs RHE), which is 1.46 and

17

21.06 times higher than that of 1-Pt-NG/C (4.456 A mg-1) and 20 wt% Pt/C (0.309 A mg-1),

18

respectively (Figure 2c). These results confirm the excellent intrinsic catalytic activity of Pt

19

atomic clusters in AC Pt-NG/C.

20

To achieve further insight into the catalytic activity of Pt atomic clusters and Pt single

21

atoms, we construct two more specimens by further decreasing the loading content of Pt in the

22

catalysts, namely, 0.5-Pt-NG/C and 0.25-Pt-NG/C. As shown in Figure S19a and S19b, no Pt

23

particles up to 1 nm can be observed in the ADF-STEM images of 0.5-Pt-NG/C and 0.25-Pt-


14

Page 15 of 36 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

ACS Catalysis

1

NG/C, while their Pt0/Pt2+ ratios (0.29 and 0.17, respectively, for 0.5-Pt-NG/C and 0.25-Pt-NG/C,

2

see Figure S19c-S19d) decrease consecutively comparing with that of 1-Pt-NG/C (1.08),

3

suggesting that the ratios of Pt atomic clusters in the catalysts decrease in the order of 1-Pt-

4

NG/C0.5-Pt-NG/C0.25-Pt-NG/C. In the LSV measurements (Figure S20a), the 0.5-Pt-NG/C

5

(66.10 mV) and 0.25-Pt-NG/C (95.75 mV) catalysts exhibit larger  than that of 1-Pt-NG/C

6

(47.20 mV) at a current density of 10 mA cm–2, suggesting the catalytic activity of 1-Pt-NG/C is

7

higher toward HER. Meanwhile, the areal current densities and mass activities of the catalysts

8

increase in the order of 0.25-Pt-NG/C0.5-Pt-NG/C1-Pt-NG/C (Figure S20b and S20c),

9

confirming that the Pt atomic clusters are more active than their single atomic counterparts. It

10

should be noted that the differences in mass activity of 1-, 0.5- and 0.25-Pt-NG/C are small

11

(Figure S20c). Generally, for Pt-based SACs with carbonaceous substrates, Pt single atom is the

12

major configuration of active centers when the loading content of Pt is below 1 wt%,2, 38 which

13

should be the same case for 1-, 0.5- and 0.25-Pt-NG/C in this work, for the mild synthesis

14

conditions of photodeposition is favorable for achieving uniform distribution of active centers.39

15

Therefore, albeit the large differences in Pt loading content and areal current densities, it’s

16

reasonable that the Pt single atoms in 1-, 0.5- and 0.25-Pt-NG/C exhibit similar mass activities.

17

To conclude, the above results are enough to back up the conclusion that Pt atomic clusters are

18

more active than their single atomic counterparts toward HER.

19

Furthermore, long-term cycling tests were carried out to evaluate the stability of AC Pt-

20

NG/C under working conditions. As shown in Figure S21a, after 4000 cyclic voltammetry (CV)

21

cycles, the  of 20 wt% Pt/C shifted negatively by more than 56.4 mV at a current density of 10

22

mA cm–2, suggesting a significant decrease in catalytic activity, which may be attributed to the

23

weak interaction between Pt NPs and carbon substrate, thus leading to serious detachment and


15

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

Page 16 of 36

1

agglomeration of Pt NPs during the tests.53 In contrast, the polarization curve of AC Pt-NG/C

2

exhibited negligible potential decay after 4000 cycles (Figure S21b) and only lost 5.1 mV after

3

4000 cycles, which is an order of magnitude smaller than that of 20 wt% Pt/C and much lower

4

than that of 1-Pt-NG/C (8.1 mV, Figure S21c). Moreover, AC Pt-NG/C exhibit overwhelming

5

advantage over 20 wt% Pt/C in the chronopotentiometry test. As shown in Figure 2d, the

6

catalytic activity of 20 wt% Pt/C and 1-Pt-NG/C kept fading throughout the test, which lost 73

7

and 34 mV, respectively, after 16 hours. Meanwhile, AC Pt-NG/C kept highly active with rare

8

attenuation under the same condition (5 mV). The above results confirmed that the stability of

9

AC Pt-NG/C is among the best reported for Pt-based SACs (Table S5). It’s important to note that

10

the higher catalytic stability of AC Pt-NG/C comparing with 1-Pt-NG/C proves that Pt atomic

11

clusters are more stable than Pt single atomic under the working conditions of HER. Therefore,

12

more work should be devoted to construct Pt-based ACCs featured by higher ratios and more

13

uniform distribution of Pt atomic clusters.

14

Theoretical calculations and mechanism discussion. Owing to the electrical conductivities

15

of semiconductors are low, most of the electron (e-)/hole (h+) pairs generated under irradiation

16

are recombined immediately after generation, resulting in low photoelectric transformation

17

efficiency (PTE) of the semiconductors.39 Compositing photo-reactive semiconductors with

18

electrical conductors (e.g., graphene or carbon nanotubes) is an efficient strategy that improves

19

the PTE of semiconductors by separating the as-generated e- from h+ as soon as they are

20

generated under irradiation.54 As a result, photo-generated electrons are shuttled and transferred

21

on electrical conductors, which binds and reduces redox mediators at locations on electrical

22

conductors distinct from anchoring sites of semiconductors, which has already been proved by

23

previous works.55 Therefore, in this work, it’s reasonable to conclude that Pt catalytic active


16

Page 17 of 36 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

ACS Catalysis

1

centers locate at locations distinct from g-C3N4 on NG. More importantly, owing to the electrical

2

conductivity of g-C3N4 is extremely low,37 Pt catalytic active centers loaded on g-C3N4 will not

3

contribute to the electrocatalytic performance of catalysts at the working conditions of this work.

4

Based on the above, density functional theory (DFT) calculations were carried out to obtain

5

atomic-scale insight into the catalytic effect of AC Pt-NG/C. First of all, we tried to clarify the

6

most energetically favorable adsorption site for Pt single atoms on NG. All the inequivalent Pt-

7

adsorption configurations (Figure S22a-S22d) around N atoms (namely, pyridinic, graphitic, and

8

pyrrolic N atoms in Figure S22e) and pristine graphene were considered. The adsorption energies

9

(𝐸𝑎 ) were calculated by the following equation:

10

𝐸𝑎 = 𝐸𝑃𝑡/𝑠𝑢𝑏 − 𝐸𝑃𝑡 − 𝐸𝑠𝑢𝑏

11

where 𝐸𝑃𝑡/𝑠𝑢𝑏 is the total energy of the substrate (i.e., all N configurations in NG, or pristine

12

graphene) with an adsorbed Pt single atom, 𝐸𝑃𝑡 is the energy of a Pt single atom referred to its

13

bulk metal, and 𝐸𝑠𝑢𝑏 is the energy of substrate. According to the calculated results (Table S6), Pt

14

single atoms are more prone to adsorbing on pyridinic N comparing with other substrates, with

15

the largest adsorption energy of -5.173 eV (on site Ⅲ, Figure S22a). The Bader charge analysis

16

suggests that 0.252 e- is transferred from the adsorbed Pt single atom to the pyridinic N,

17

confirming site Ⅲ in Figure S22a is the most energy favorable adsorption site for a Pt single

18

atom. We also proved that pyrrolic N tend to transform into pyridinic N (Figure S23), this is

19

consistent with the XPS results in Figure 1c, which suggests vast increment of pyridinic N after

20

annealing. Besides, as shown in Figure S24, the Pt atom adsorbed on site Ⅱ of pyrrole N splits

21

the substrate, which has a negative effect on base structures. Based on the above experimental

(1)


17

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

Page 18 of 36

1

and DFT calculation results, pyridinic N substrate was selected to investigate the adsorption

2

structure of Pt clusters.

3 4

Figure 3. DFT calculation for the energetics of Pt atoms agglomeration to promote the HER

5

activity. Energy favorable adsorption structures of: (a) Pt single atom, (b) Pt6-cluster and (c) Pt9-

6

cluster on pyridinic N graphene substrate were obtained by PSO algorithm. Their corresponding

7

Bader charge were analyzed to distinguish Pt atoms using gradient colors. The profiles of each

8

atom are shown at the bottom. (d) Calculated free energy diagram for hydrogen evolution at

9

equilibrium (applied potential U = 0) with above different catalyst configurations. The red,

10

yellow and blue lines represent the calculated ∆GH* for single Pt single atom, Pt6-cluster and Pt9-

11

cluster respectively, comparing with Pt (111) slab (green line).

12

In order to predict the adsorption structures more efficiently and investigate the catalytic

13

performance of Pt clusters on N-doped graphene, the structure prediction of Pt clusters (n = 4~12)

14

and their adsorption structures on the pyridinic N in NG were performed within the evolutionary

15

scheme using the CALYPSO package based on the particle swarm optimization (PSO)

16

algorithm.56-57 The adsorption structures of clusters consist of 6 and 9 Pt atoms were investigated

17

and compared (denoted as Pt6-cluster and Pt9-cluster, respectively) according to the size range of

18

bright dots in the ADF-STEM image (Figure S14). As shown in Figure 3a-3c, Bader charge


18

Page 19 of 36 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

ACS Catalysis

1

analysis was also conducted to clarify charge transfer and total electronic charge of adsorbed Pt

2

atoms (see details in Table S7), before hydrogen adsorption. The results reveal that Pt atoms

3

present various degrees of electron-deficient and electron-rich phenomenon, which were

4

described by gradient colors. In the case of Pt6-cluster and Pt9-cluster, the Pt atoms (blue atoms

5

in Figure S25 and S26) which are directly bonded to the pyridinic N have the largest charge

6

transfer (0.500 e- and 0.516 e-, respectively, for Pt6-cluster and Pt9-cluster), implying their

7

stronger bonding energy to the substrate than that of Pt single atom (0.252 e-), which contributes

8

to the anchoring of Pt atomic cluster. Meanwhile, the Pt atoms in top layer of Pt 9-cluster

9

structure can be classified by the amount of charge transfer, i.e., the neutral Pt atoms (gray atoms,

10

Pt(4), Pt(5) and Pt(6) in Figure 3c) with almost no charge transfer while the other 3 Pt atoms

11

(light red atoms, Pt(7), Pt(8) and Pt(9) in Figure 3c) are electron-rich within the Bader volume.

12

Both kinds of Pt atoms construct bridge sites in the Pt9-cluster synergistically, this configuration

13

are expected to promote synergistic catalytic effect, which might be the primary reason for

14

enhanced HER activity of Pt9-cluster.


19

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

Page 20 of 36

1 2

Figure 4. Free energy diagram for hydrogen evolution at equilibrium. The red, yellow and blue

3

lines represent the calculated ∆GH* of inequivalent catalytic sites for single Pt single atom, Pt6-

4

cluster and Pt9-cluster respectively.

5

To gain further insight into the HER catalytic capability of AC Pt-NG/C, the Gibbs free

6

energies (∆𝐺𝐻∗ ) of atomic hydrogen adsorption, which is a reasonable descriptor of the catalytic

7

ability of a material toward HER, were calculated for various Pt based active centers and the

8

majority of inequivalent catalytic sites (Figure S25-S27) by the following equation:

9

𝑆𝐴𝐶 𝑆𝐴𝐶 𝑃𝑡 𝑃𝑡 ∆𝐺𝐻∗ = ∆𝐸𝐻∗ − ∆𝐸𝐻∗ + ∆𝐺𝐻∗

(2)

10

𝑆𝐴𝐶 𝑃𝑡 where ∆𝐸𝐻∗ and ∆𝐸𝐻∗ are the hydrogen chemisorption energies of SAC surface and Pt (111)

11

𝑆𝐴𝐶 𝑃𝑡 respectively, ∆𝐺𝐻∗ and ∆𝐺𝐻∗ are the Gibbs free energies of atomic hydrogen adsorption on


20

Page 21 of 36 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

ACS Catalysis

1

SAC surface and Pt (111) slab. Based on the theory proposed by Nørskov and co-workers, the

2

desirable ∆GH* value of a HER catalyst should be around 0.00 eV and the Gibbs free energy of

3

Pt (111) on the fcc site was calculated previously (-0.1 eV).58 Comparing with other catalytic

4

centers, the calculated ∆GH* in bridge sites between neutral and electron-rich atoms in Pt9-cluster

5

(ie. the adsorption sites in Figure S25a-S25b and Figure S26a-S26c) are closer to 0.00 eV,

6

suggesting its better HER catalytic capability in the view of thermochemistry.58 As illustrated in

7

Figure 3d and Figure 4, the optimal ∆𝐺𝐻∗ values of Pt single atom, Pt6-cluster and Pt9-cluster

8

configuration are -0.176, -0.096 and -0.008 eV, respectively, suggesting the capability of various

9

Pt active centers to catalyze HER is in the order of: Pt9-cluster > Pt6-cluster > Pt slab (111) > Pt

10

single atom. It should be noted that the ∆𝐺𝐻∗ value of a Pt single atom adsorbed on pyridinic N is

11

the lowest, which means that hydrogen atom is bonded too strongly on the substrate thus

12

decreasing the efficiency of hydrogen adsorption and release.59 To sum up, the calculated results

13

are consistent with the experimental results which also proved a higher intrinsic activity of Pt

14

atoms in the atomic clusters comparing with those in their single atomic form.

15

CONCLUSIONS

16

In conclusion, we demonstrate that Pt-based ACCs, as a promising type of catalyst configuration,

17

is highly active toward HER, which not only exhibit higher catalytic activity than their single

18

atomic counterparts and commercial 20 wt% Pt/C catalysts, but also retain highly active after

19

4000 catalytic cycles or after 16 hours in the chronopotentiometry tests. DFT calculation results

20

reveal that the pyridinic N contents in the substrate of the Pt-based ACCs play a key role in

21

stabilizing the Pt atomic clusters, which also promotes charge transfer between the Pt atomic

22

cluster and the substrates, leading to optimized hydrogen adsorption and releasing properties on

23

the catalyst surface. This work not only provides an efficient strategy for constructing highly


21

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

Page 22 of 36

1

active catalyst configurations for HER, but also brings vast opportunity of enhancing the

2

catalytic performance of various chemical reactions, such as photocatalysis and organic synthesis.

3

EXPERIMENTAL SECTION

4

Materials. Analytical grade urea and ethanol were obtained from Shanghai Chemical Reagents,

5

China. Chloroplatinic acid (H2PtCl6, 99.995%) and Nafion (5 wt% in a mixture of lower

6

aliphatic alcohol and water) were acquired from Sigma-Aldrich. 20 wt% Pt/C catalysts with a

7

specific surface area of 200.1 m2 g-2 (Figure S28a) was acquired from Macklin. Meanwhile, the

8

pore volume is 0.427 cm3 g-1 and the average pore diameter is 12.8 nm (Figure S28b). The Pt

9

nanoparticles are uniformly distributed on the catalyst surface (Figure S28c). The average size of

10

the Pt nanoparticles is calculated to be 3.32 nm based on the HRTEM results (Figure S28d).

11

Graphite rod was obtained from Alfa Aesar (99.9995%). All of the chemicals used in this

12

experiment were of analytical grade and used without further purification. Deionized (DI) water

13

(18.2 MΩ at 25 °C) was obtained from Milli-Q System (Millipore, Billerica, MA).

14

Synthesis of NG/C composite. NG/C was prepared according to previous reports,24 in which the

15

mass ratio of NG to g-C3N4 is 3:2. In a typical procedure, graphene oxide (GO) solution (3 mg

16

mL−1) was prepared by a modified Hummer’s method.1 Subsequently, 1.2 g urea (NH2CONH2)

17

was added into 6 ml GO solution under vigorous stirring until fully dissolved. Then, the mixture

18

was dehydrated at 60 °C in a vacuum oven for 12 hr. The as-obtained GO/urea mixture was

19

loaded into the center of a tube furnace (calibrated using a k-type thermocouple probe). After

20

flushing the tube with Ar (150 sccm) for 30 min, the temperature was ramped at 10 °C min −1

21

until 600 °C and maintained for 2 hr under continuous Ar flow, followed by naturally cooling to


22

Page 23 of 36 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

ACS Catalysis

1

room temperature. The final product of NG/C composite was collected and washed three times

2

with DI water under sonication.

3

Synthesis of AC Pt-NG/C. For the preparation of all catalysts in this work, UV irradiation was

4

provided by a Xenon-lamp parallel light source system (BEIJING PERFECTLIGHT

5

SCIENCE&TECHNOLOGY Co. Ltd, CHF-XM-500W), operating at a current of 15 A, with the

6

distance between liquid surface and the light source been kept at 5 cm. In a typical synthesis, 30

7

mg NG/C composite was dispersed in 20 ml deionized water under sonication for 5 min. Then, a

8

H2PtCl6 aqueous solution (2 mg mL-1, 0.796 mL) was injected into the above aqueous dispersion

9

under vigorous stirring. The mass ratio of Pt to NG/C is 2 wt% (For the synthesis of 1-Pt-NG/C,

10

0.5-Pt-NG/C and 0.25-Pt-NG/C, the mass ratio of Pt to NG/C is 1 wt%, 0.5 wt% and 0.25 wt%,

11

respectively). Subsequently, the mixture was exposed to visible lights-irradiation under

12

continuous stirring for 6 h. Finally, the products were collected and washed three times with DI

13

water under sonication followed by dehydrating in a vacuum oven.

14

Characterizations. Visible light was provided by a Xenon-lamp parallel light source system

15

(BEIJING PERFECTLIGHT SCIENCE&TECHNOLOGY Co. Ltd, CHF-XM-500W). XRD

16

spectra were obtained at room temperature using an X’Pert PRO (PANalytical, Netherlands)

17

instrument with Cu Ka radiation. Chemisorption of CO was performed on a AutoChem1 II 2920

18

(Micromeritics) in pulse mode, using He as the carrier gas. Raman spectroscopy was conducted

19

using a Laser Confocal Raman Microspectroscopy (LabRAM HR Evolution, Horiba Jobin Yvon).

20

FTIR spectra were recorded on FT-IR TENSOR 27 (Bruker) equipment using a KBr pellet in the

21

range of 800-4000 cm-1. UV-vis spectra were measured by UV3600, Shimadzu. XPS analyses

22

were performed on an Escalab 250Xi XPS system with Al Kα (1486.6 eV) source. ICP-OES

23

measurements were carried out with an Optima 5300DV system (Perkin Elmer). XAFS


23

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

Page 24 of 36

1

measurements were performed on the BL14W1 beamline at the Shanghai Synchrotron Radiation

2

Facility (Shanghai Institute of Applied Physics, China), operated at 3.5 GeV with injection

3

currents of 140–210 mA. A Si (111) double-crystal monochromator was used to reduce the

4

harmonic component of the monochrome beam. Pt foil and PtO2 were used as reference samples

5

and measured in the transmission mode. ADF-STEM images were obtained on a FEI TITAN

6

Chemi STEM equipped with a CEOS (Heidelburg, Germany) probe corrector, operating at 200

7

kV. Transmission electron microscopy (TEM) and high-resolution transmission electron

8

microscopy (HRTEM) were conducted using a JEM-2100F field-emission TEM. Scanning

9

electron microscopy (SEM) was performed with a Hitachi S-4800 field emission scanning

10

electron microscope (FE-SEM) (5 kV) equipped with an energy-dispersive X-ray (EDX)

11

spectroscope. The Brunauer-Emmett-Teller (BET) specific surface area of the samples was

12

measured by nitrogen adsorption at 77 K on a surface area and porosity analyzer (Micrometrics

13

ASAP 2020). Pore size distribution plots were derived from the adsorption branch of the

14

isotherm based on density functional theory (DFT) calculations. Cyclic voltammetry (CV), linear

15

sweep voltammetry (LSV) and electrochemical impedance spectroscopy (EIS) were performed

16

on a CHI660E electrochemical workstation (CH Instruments, China) equipped with a rotating

17

disk electrode (AFMSRCE, PINE ROTATING ELECTRODE).

18

Electrochemical Measurements. Electrochemical measurements were carried out in a three-

19

electrode system in a H2SO4 electrolyte (0.5 M). All the measured potentials were calibrated to

20

reversible hydrogen electrode (RHE) using the following equation: ERHE = EHg/Hg2SO4 + 0.64

21

V+0.0592pH17. A glassy carbon rotating-disk electrode with a diameter of 3 mm (Pine

22

Instruments), Hg/Hg2SO4 electrode (saturated K2SO4 aqueous solution, 0.64 V RHE) and

23

graphite rod (Alfa Aesar, 99.9995%) were applied as the working electrode, reference electrode


24

Page 25 of 36 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

ACS Catalysis

1

and counter electrode, respectively. For the preparation of working electrode, 2 mg of the

2

catalyst and 12 µl Nafion solution was first dispersed in 1 ml water/ethanol (v: v =1:9) solution

3

under sonication. The above dispersion (10 µl) was then loaded onto a glassy carbon electrode

4

and dried naturally at ambient environment, the loading amount of catalyst was ca. 0.283 mg cm–

5

2

6

r.p.m. The polarization curves were corrected with 95% IR-compensation (IR corrected).49

7

Areal/mass current densities were obtained by normalizing the current to the geometric area of

8

the electrode and Pt mass, respectively. Electrochemical impedance spectroscopy (EIS) was

9

conducted in a frequency range of 0.1–100,000 Hz at an overpotential of 200 mV (versus RHE)

10

to obtain the solution resistance (Rs). All data presented was corrected with the solution

11

resistance Rs.50, 60 The number of active sites and turnover frequency (TOF) of the catalysts were

12

calculated according to previous reports.49, 52

13

DFT calculations. All first-principles calculations were performed by the Vienna Ab-initio

14

Simulation Package (VASP).61-62 The exchange-correlation interactions were treated within the

15

generalized gradient approximation of the Perdew-Burke-Ernzerhof (PBE) type.63 Moreover, the

16

electron wave functions were expanded by a plane wave cutoff of 400 eV and a k-mesh of 5×5×1

17

was adopted to sample the Brillouin zone. Atomic relaxation was performed until all forces on

18

each atom were

Overwhelming the Performance of Single Atoms with Atomic Clusters

Recommend Documents