Copyright WILEY-VCH Verlag GmbH & Co. KGaA, 69469 Weinheim, Germany, 2018.

Supporting Information
for Adv. Energy Mater., DOI: 10.1002/aenm.201801536

Selective CO2 Reduction on 2D Mesoporous Bi Nanosheets

Hui Yang, Na Han, Jun Deng, Jinghua Wu, Yu Wang,
Yongpan Hu, Pan Ding, Yafei Li, Yanguang Li,* and Jun Lu*
Copyright WILEY-VCH Verlag GmbH & Co. KGaA, 69469 Weinheim, Germany, 2018.

Supporting Information

Selective CO2 Reduction on Two-Dimensional Mesoporous Bi Nanosheets

Hui Yang†, Na Han†, Jun Deng, Jinghua Wu, Yu Wang, Yongpan Hu, Pan Ding, Yafei Li,
Yanguang Li* and Jun Lu*

Miss. H. Yang, Dr. N. Han, Mr. J. Deng, Dr. J. H. Wu, Mr. Y. P. Hu, Mr. P. Ding, and Prof.
Y. G. Li
Institute of Functional Nano & Soft Materials (FUNSOM)

Jiangsu Key Laboratory for Carbon-Based Functional Materials and Devices

Soochow University, Suzhou 215123, China

E-mail: [email protected]

Dr. Y. Wang, Prof. Y. F. Li

College of Chemistry and Materials Science
Nanjing Normal University
Nanjing 210023, China

Dr. J. Lu
Chemical Sciences and Engineering Division
Argonne National Laboratory
9700 Cass Ave, Lemont, IL 60439, USA
E-mail: [email protected]
†
These two authors contribute equally.
Experimental Section

Material characterizations: XRD was collected on a PANalytical X-ray diffractometer using

Cu Kα radiation. SEM was carried out on a Zeiss scanning electron microscope. TEM images

and SAED were taken on an FEI Tecnai transmission electron microscope with an

acceleration voltage of 200 KV. AFM measurements were conducted on a Veeco (MultiMode

V) atomic force microscope.

Preparation of ultrathin BOC nanosheets: In a typical synthesis, 0.2 mmol of Bi(NO3)3·5H2O

was first dissolved in 40 mL of NMP under magnetic stirring, to which was further added 1 g

1
of urea to form a homogeneous solution. This reaction solution was then transferred to a 100

mL Teflon-lined autoclave, and heated at 150 ℃ for 12 h. Solid product was collected by

centrifugation, repetitively washed with ethanol and water, and lyophilized.

Cathodic transformation of BOC nanosheets to mpBi nanosheets: 1 mg of Bi2O2CO3 and 0.5

mg of Ketjenblack carbon were dispersed in 250 μL of ethanol with the addition of 6 μL of 5

wt% Nafion, and vigorously sonicated for 40 min to form a homogenous ink. The ink was

dropcast onto a 1 × 1 cm2 Teflon-treated carbon fiber paper (HCP020P, from HESEN) to

reach catalyst areal loading of 1 mg/cm2. Electrochemical experiments were carried out in a

custom-designed H-shaped cell with two compartments separated by a piece of SELEMION

anion exchange membrane. The working electrode and the reference electrode (saturated

calomel electrode or SCE) were placed in the cathodic compartment. The counter electrode

(graphite rod) was placed in the anodic compartment. Each compartment contained ~30 mL

electrolyte (0.5 M NaHCO3), leaving a headspace of ~25 mL. All potentials were converted to

the RHE scale (ERHE = ESCE + 0.241 V + 0.0591 V × pH). Cyclic voltammetry (CV) was

performed at a scan rate of 10 mV/s controlled by CHI 660E potentiostate. For the cathodic

transformation from BOC nanosheets to mpBi nanosheets, the working electrode was biased

at -0.62 V for 2 h. It was then quickly taken out from the electrolyte, and rinsed with

deionized water for characterizations as described in the main text.

CO2RR measurements: For CO2RR measurements, reduced mpBi nanosheets supported on

carbon fiber paper electrode was directly used as the working electrode in the same two-

compartment cell. Working electrodes of commercial Bi nanopowder (Macklin, 99.99%, 200

mesh, corresponding SEM and XRD are available in Figure S8) were prepared under similar

conditions except for using 1 mg of Bi nanopowder for the catalyst ink preparation. The

electrolyte was pre-saturated with N2 (pH = 8.4) or CO2 (pH = 7.4). A constant CO2 flow of

20 sccm was continuously bubbled into the electrolyte to maintain its saturation during

CO2RR measurements. Cyclic voltammetry and polarization curves were carried out at a scan
2
rate of 10 mV/s. All the potential readings were iR-corrected, and reported against SCE unless

otherwise specified.

In order to analyze the reduction production and calculate their Faradaic efficiency,

electrolysis was performed at a few selected potentials for 1-5 h. During the electrolysis,

gaseous products in the headspace of the cathodic compartment were periodically vented into

a gas chromatograph (GC, Aligent 7890B) equipped with a molecular sieve 5A and two

porapak Q columns. The concentration of H2 was analyzed by a thermal conductivity detector

(TCD), and the concentration of CO was analyzed by a flame ionization detector (FID) with a

methanizer. The concentration of gaseous products was quantified by the integral area ratio of

the reduction products to standards. The Faradaic efficiency of gas product was calculated as

below:

𝑣 𝑥
𝑄𝑐𝑜 ( )×( )×𝑁×𝐹
60 𝑠/𝑚𝑖𝑛 24000 𝑐𝑚3 /𝑚𝑜𝑙
𝐹𝐸(%) = × 100% = × 100%
𝑄𝑡𝑜𝑡 𝑗

where v (=20 sccm) is the flow rate of CO2, x is the measured concentration of product in 1

mL sample loop based on the calibration of the GC with a standard gas, N (=2) is the number

of electrons required to form a molecule of CO or H2, F is the Faraday constant (96500 C mol-
1
), and j is the recorded current.

Liquid products were collected at the conclusion of each electrocatalysis and analyzed by
1
H NMR (Aligent DD2-600) and ion chromatography (IC, Dionex ICS-600). For the NMR,

0.5 ml of the catholyte was mixed with 0.1 ml of D2O and 0.2 ml of dimethylsulfoxide

(DMSO, as the internal standard). The concentration of formate was quantitatively determined

from its NMR peak area relative to that of the internal standard using the calibration curve

from a series of standard HCOONa solutions (Figure S4a,b). For IC, 1 ml of the catholyte was

extracted and diluted 10-50 times with deionized water. The concentration of formate was

quantitatively determined from its IC peak area using the calibration curve from a series of

standard HCOONa solutions (Figure S4c,d). Formate Faradaic efficiency was calculated as
3
follows:

𝑄𝐻𝐶𝑂𝑂− 𝑛𝐻𝐶𝑂𝑂− × 𝑁 × 𝐹
𝐹𝐸𝐻𝐶𝑂𝑂− (%) = × 100% = × 100%
𝑄𝑡𝑜𝑡 𝑗 × 𝑡

where nHCOO- is the measured amount of formate in the cathodic compartment and t is the

reaction time. Formate partial current density at different potentials was calculated by

multiplying the overall geometric current density and its corresponding Faradaic efficiency.

Solar-driven overall CO2/H2O splitting: Commercial 20 wt% Ir/C (from Fuel Cell Store) was

used as the OER electrocatalyst, and similarly loaded onto 1 × 1 cm2 carbon fiber paper

electrode to achieve an areal loading of 1 mg cm-2. Overall CO2/H2O splitting reaction was

carried out inside the same two-compartment cell in a two-electrode configuration and

powered by a commercial 1 cm2 GaInP/GaInAs/Ge multi-junction solar cell. Voltage between

two electrodes was measured by potentiostat and current through the circuit was measured by

2461-digital source-meter (Keithley). All electrochemical results for this part were non-iR-

compensated. The solar-to-formate conversion efficiency was calculated as below:

𝜃
∆𝑟𝐺𝑚 × 𝑛𝐻𝐶𝑂𝑂−
= × 100%
𝑊

𝜃 (298.15𝐾) 𝜃 (𝐻𝐶𝑂𝑂𝐻)
1 𝜃 (𝑂 ) 𝜃 𝜃
∆𝑟 𝐺𝑚 = ∆𝑓 𝐺𝑚 + ∆𝑓 𝐺𝑚 2 − ∆𝑓 𝐺𝑚 (𝐶𝑂2 ) − ∆𝑓 𝐺𝑚 (𝐻2 𝑂)
2
𝜃
where ∆𝑟𝐺𝑚 (=270.138 kJ∙mol-1) was the energy gain during CO2RR, and W was light energy

input (100 mW)

Computation details: All the DFT computations were preformed using the VASP code,[1, 2]

where the exchange-correlation interactions were modeled by GGA-PBE[3] functional, and the

ion-electron interactions were described by the projector-augmented plane wave approach.[4, 5]

A plane-wave cutoff energy of 420 eV with Fermi-level smearing of 0.05 eV for slabs and

0.01 eV for gas-phase species was used in all calculations. The Brillouin zone was sampled

with a 3×3×1 Γ centered k points. The convergence thresholds of energy and forces were set

as 1×10-5 eV and 0.02 eV/Å, respectively. All the surfaces included a vacuum space of at least

4
15 Å to keep the image interaction negligible. The free energy pathways on the constructed

surfaces were explored via the computational hydrogen electrode (CHE) method.[6]

Refernence

[1] G. Kresse, J. Hafner, Phys. Rev. B 1993, 47, 558.

[2] G. Kresse, J. Furthmueller, Phys. Rev. B 1996, 54, 11169.
[3] J. P. Perdew, K. Burke, M. Ernzerhof, Phys. Rev. Lett. 1996, 77, 3865.
[4] P. E. Blöchl, Phys. Rev. B 1994, 50, 17953.
[5] G. Kresse, D. Joubert, Phys. Rev. B 1999, 59, 1758.
[6] J. K. Norskov, J. Rossmeisl, A. Logadottir, L. Lindqvist, J. R. Kitchin, T. Bligaard, H.
Jonsson, J. Phys. Chem. B 2004, 108, 17886.

Figure S1. SEM images of products in (a) H2O, (B) EG, (C) DMSO.

5
Figure S2. High resolution TEM analysis of the exposed surface facets of Bi nanoparticles.

Inserts are the corresponding fast Fourier transform patterns.

6
Figure S3. Total charge integrated from the Bi3+/Bi0 cathodic wave, from which we further

estimated the percentage of surface Bi sites.

7
Figure S4. Identification and quantification of formate by NMR and IC. (a) A typical NMR

spectrum of the catholyte. (b) Standard curve of formate by plotting the formate concentration

with respect to the formate/DMSO NMR peak area ratio. (c) A typical IC spectrum of the

catholyte. (d) Standard curve of formate by plotting the formate concentration with respect to

the absolute IC peak area.

8
Figure S5. Calculated partial current density of (a) formate and (b) CO in (black) N2-

saturated and (red) CO2-saturated 0.5 M NaHCO3.

Figure S6. SEM image of mpBi nanosheets after the long-term stability test.

9
Figure S7. (a-e) Top and side views of optimized structure of (a) (001), (b) (110), (c) (012),

(d) (111) and (e) (101) planes of Bi. (f-h) DFT simulations of (f) CO2 reduction to formate,

(g) CO2 reduction to CO, and (h) H2 evolution on different planes of Bi.

10
Figure S8. (a) SEM and (b) XRD of commercial Bi powders.

11