Supporting Information

Li-N2 battery for ammonia synthesis and

computational insight

Xingyu Ma1, Jiang Li1, Hongjun Zhou1, Jianwei Zhao2, Hui Sun1 *

1
State Key Laboratory of Heavy Oil Processing, Beijing Key Laboratory of Biogas

Upgrading Utilization, College of New Energy and Materials, China University of Petroleum-

Beijing, Fuxue Road No. 18, Changping District, Beijing 102249, P.R. China

2
Shenzhen HUASUAN Technology Co., Ltd., Shenzhen 518055, P.R. China

*
E-mail: sunhui@cup.edu.cn

ABSTRACT: Electrochemical synthesis of ammonia is deemed as an alternative to the fossil-

fuel-driven Haber-Bosch process, in which Li-mediated nitrogen reduction (LiNR) is the most

promising scheme. Continuous lithium-mediated nitrogen reduction for ammonia synthesis (C-

LiNR) has recently been reported in high-level journals, while with many foggy internal

reactions. Synthesizing ammonia in a separate way may be profitable for understanding the

mechanism of LiNR. Herein, a intermittent lithium-mediated nitrogen reduction for ammonia

synthesis (I-LiNR) was proposed, three steps required for I-LiNR were achieved in the

cathode chamber of a Li-N2 battery. Discharge, stand, charge in Li-N2 battery is corresponding

to N2 lithification, protonation and lithium regeneration, respectively. It can also realize quasi-
continuous process with practical significance because it could be carried out through the

identical battery. Products such as Li3N, LiOH, NH3 are detected experimentally, which

demonstrates a definite reaction pathway. The mechanism of Li-N2 battery, Li-mediated

synthesis of ammonia and LiOH decomposition are explored through density functional theory

calculations. The role of Li in dinitrogen activation is highlighted. It expands the range of

LiOH-based Li-air batteries and may guide the study from Li-air to Li-N2, it will lead attention

to the reaction mechanism of Li-mediated nitrogen reduction. Challenges and opportunities of

the procedure are discussed in the end.

RuCl3 (Aladdin Reagent, 40 %, 50 mg) and Super P (Lizhiyuan Battery Materials, 80 mg)

was weighted into 30 ml ethylene glycol (Tianjin Guangfu Fine Chemical Research Institute),

and was stirred until well mixed. Then it was poured in a 50 ml hydrothermal autoclave and

was tighten. After reacted 3 h at 180 °C, the product was washed with ethanol for 3 times by

centrifugal machine and quickly placed in a vacuum environment to dry at 80 °C for 10 h,

named as Ru@C. The prepared catalyst was stored in an inert gas glove box. Each step of the

preparation process needs to be fast to avoid exposure to air for too long. The characterization

of the Ru@C material is shown from Figure S1 to Figure S5 in the supporting information.

2. Preparation of electrode

The electrode was prepared by kneading a slurry made from 90 wt% Ru@C and 10 wt%

PVDF binder (Arkema 5130) in N-methyl-2-pyrrolidone (NMP) (Aladdin Reagent, 99.9 %)

until a homogeneous solution was obtained after breaking. The slurry was knife-coated on a

sheet of carbon cloth, and the coated sheet was dried at 80 °C for 12 h in a vacuum oven to

volatilize NMP, after which round disks were punched out as cathode and stored in an inert gas

glove box.

3. Electrochemical Methods

All cells were assembled inside an Ar-filled glovebox and embedded into a homemade

vented mold. The Li−N2 battery, comprising the cathode above (Ø14 mm), a glass microfiber

1
separator (Whatman, Ø16 mm) soaked with the electrolyte, and a lithium metal anode (Ø16

mm), was assembled in a CR2032-type coin-cell with pores on the top of cathode. The battery

in sealed mold was taken out from glovebox and subsequently connected to raw gas pipeline in

a N2 shielding gas. The N2 with a purity of 99.999% was purified by passing through the gas

washing bottle containing 0.1M KOH, 0.05M H2SO4, molecular sieve and electrolyte, in

sequence. Before the battery performed the discharge procedure, let it stand for 8h with N2

circulating. The flow rate of nitrogen gas and water vapor was the same with the continuous

ammonia synthesis in Li - N2 & H2O battery. The time of rest with H2O(N2) (N2 as a carrier

gas) was the same as that of discharge with N2. The picture of vented mold is shown in Figure

S6. The schematic diagram of raw gas pipeline is shown in Figure S7. The procedure of

discharge, rest and charge was carried out in the LAND multi-channel test system. The cyclic

voltammetry (CV) and Potentiostatic electrochemical impedance spectroscopy (PEIS) were

tested on PMC DC-2000 produced by Princeton Applied Research with frequencies ranging

from 106Hz to 0.01Hz after the voltage was stabilized under a N2 protective atmosphere.

4. Characterization Methods

The X-ray Photoelectron Spectroscopy (XPS) was taken on a K-Alpha machine made from

American Thermo Scientific Company. Field-emission Scanning electron microscope (SEM)

imaging and Energy Dispersive Spectroscopy (EDS) mapping were conducted on a

GeminiSEM300 model produced by the German Carl Zeiss Company, and the acceleration

voltage is 20kV. The transmission electron microscope (TEM) and high-resolution TEM

images were carried out with a FEI Tecnai G2 F20 microscope. X-Ray Powder Diffraction

(XRD) data was collected on an Empyrean X-ray diffractometer produced by the Netherlands

Panalytical Company. Cu-Kα rays were used, the tube voltage was 40 kV and the current was

40 mA. The characterization of the electrode after the electrochemical test was an ex-situ test.
2
The battery was disassembled in a Ar-filled glove box and the cathode was extracted and placed

into a self-made vacuum sealing box to transfer. Exposure to air for less than 20s before

Characterization.

5. Ammonia detection and calculation

NH4Cl (0.3146g) was dried in an oven at 60 °C for 8 hours, and it was dissolved into

100ml ultrapure water as 1000ppm ammonia solution. It was then diluted to configure 0.1 0.2,

0.4, 0.6, 0.8, 1ppm ammonia standard solution using volumetric flask gradually. The above

standard solution was tested by UV Spectrophotometric Analysis spectrophotometry produced

from Shanghai Yuanxi Instrument Co., Ltd., and the standard curve of absorbance-ammonia

yield using indophenol blue reagent method was drawn according to Lambert Beer's law, shown

in Fig. S8.

Chromogenic reagent (A): 4 g of KOH (Tianjin Guangfu Fine Chemical Research

Institute), 5 g of salicylic acid (Tianjin Guangfu Fine Chemical Research Institute), and 5 g of

sodium citrate (Tianjin Guangfu Fine Chemical Research Institute) were dissolved in 100 ml

of ultrapure water by ultrasonic for 1 h.

Oxidation solution (B): 3.04 ml of sodium hypochlorite (Aladdin Reagent, available

chlorine 10-15 %) was added into 100 ml of ultrapure water by ultrasonic for 30 min.

Catalytic reagent (C): 0.25 g of sodium nitroferricyanide dihydrate (Aladdin Reagent) was

dissolved in 20 ml of ultrapure water by ultrasonic for 1 h.

The above reagents are stored with the amber reagent bottle. 2 ml of the standard solutions

or sample electrolyte was pipetted into 10ml centrifuge tube, to which 2 mL of reagent (A), 1

mL of reagent (B) and 0.2 mL of reagent (C) were then added. After 2 h, the concentration of

3
the produced indophenol blue was measured using UV-vis spectrophotometer.

30ml of ultrapure water was used to collect the outgoing gas passing through the cathode

of the battery, and 2ml of the collected liquid was taken out as the bottom liquid for ammonia

detection. The calculation formula was as follows. The electric power during discharge was

required for nitrogen reduction, and it was used in the calculation of faradic efficiency (FE).

The FE calculation during the cycle of discharge and charge also used the electricity during

discharge.

The NH3 production and formation rate was determined using the following equation:

r(NH3) = c(NH3) × V / t

c(NH3) is the measured ammonia concentration, V is the volume of ultrapure water, t is

the duration of discharge and charge.

The Faradaic efficiency (FE) was calculated as follows:

FE = 3 × m(NH3) × F / 17 × Q

F is the Faraday constant and Q is the capacity required during discharge

6. Computational methods

Spin-polarized DFT calculations were performed using the Vienna ab initio simulation

package (VASP).1,2 The revised Perdew-Burke-Ernzerhof (RPBE) was selected for the

exchange-correlation potential.3,4 The pseudo-potential was described by the projector-

augmented-wave (PAW) method.5 The energy cutoff for the plane wave basis set used was 400

eV. K-mesh resolution was set to 0.06 2π/Å. The geometry optimization is performed until the

Hellmann–Feynman force on each atom is smaller than 0.02 eV·Å-1. The SCF energy criterion
4
is set to 10−6 eV in iterative solution of the Kohn-Sham equation. SIGMA was set to 0.2. The

Ru(101) slab model was created by a 2a*3b supercell with 12 Å vacuum and 3 atomic layers.

The Ru(101)-Li slab model was built by placing one layer of Li on top of Ru(101) model. Only

the top one layer was relaxed and all others were fixed. The adsorption energy is calculated by

Eads=EAB-EA-EB, in which EAB is the total energy of adsorbed state, EA is the energy of adsorbate

and EB is the energy of adsorbent. The free energy is corrected by ΔG=ΔE+ΔZPE-TΔS, where

ΔE is adsorption energy, ΔZPE is zero point energy, ΔS is entropy.

5
Figure S1. The SEM of Ru@C.

6
Figure S2. (A) The TEM and (B) particle size analysis of Ru@C.

7
Figure S3. The XRD of Ru@C.

8
Figure S4. (A) The High magnification TEM and (B) crystal plane spacing analysis of

Ru@C.

9
Figure S5. The Ru3p and C1s with Ru3d by XPS of Ru@C.

10
Figure S6. The picture of mold with built-in button battery.

11
Figure S7. The schematic diagram of the ventilation line.

12
Figure S8. The V-t curve of 24 h discharge in Li-N2 battery.

13
Figure S9. The cathode N1s spectrum of Li-N2 battery in LiCF3SO3 electrolyte, after (A)

discharge 6 h, (B) discharge 15 h.

14
Figure S10. The SEM surface scan of cathode in Li-N2 battery with LiCF3SO3 electrolyte

after discharge 24h.

15
Figure S11. (A) The V-t curve of discharge to 0 V and corresponding (B) cathode N1s

spectrum of Li-N2 battery in LiTFSI electrolyte.

16
Figure S12. The XRD of cathode after discharge and rest with H2O for 6 h, 15 h, 24 h.

17
Figure S13. The PEIS of initial battery, discharge 6 h and rest with H2O for 6 h.

Table S1. The data of PEIS after fitted from Figure S13.

State R0 Rct

Initial battery 16.59  42.91 

Discharge in N2 35.59  121.3 

Rest in H2O 71.61  185.2 

18
Figure S14. The CV curve of battery under Ar and N2 atmosphere from 1 V ~ 4.5 V with scan

rate of 0.5 mV s-1.

19
Figure S15. (A) The absorbance of the standard sample by indophenol blue method and (B)

the straight line after fitting

20
Figure S16. The yield and FE of discharge, rest, charge in intermittent ammonia synthesis

based on Li - N2.

21
Figure S17. Schematic diagram of the V-t curve of quasi-continuous ammonia synthesis.

22
Figure S18. The structural changes of the discharge phase of a Li-N2 battery.

23
Table S2. The Gibbs free energy and the bond length of N-N distance of Li-N2 battery.

State Free energy (eV) N-N distance (Å)

+Li+N2 0 1.11737

LiN2 -1.35013 1.17124

Li2N2 -2.74279 1.18378

Li3N2 -3.24372 1.2072

Li4N2 -4.2223 1.27243

2Li2N -5.12267 4.49449

Li3N+Li2N -6.28785 3.94193

2Li3N -8.34963 3.85815

24
Figure S19. (A) The total TDOS of N2 adsorption on Ru, inside is that of N. (B) The PDOS

of Ru-5s, Ru-4p, Ru-4d. (C) The PDOS of N-2s, N-2p.

25
 Figure S20. (A) The TDOS of Li4N2 adsorption on Ru, inside is that of N and Li. (B) The

PDOS of Ru-5s, Ru-4p, Ru-4d. (C) The PDOS of N-2s, N-2p. (D) The PDOS of Li-2s, Li-2p.

26
 Figure S21. (A) The TDOS of 2Li2N adsorption on Ru, inside is that of N and Li. (B) The

PDOS of Ru-5s, Ru-4p, Ru-4d. (C) The PDOS of N-2s, N-2p. (D) The PDOS of Li-2s, Li-2p.

27
Figure S22. The structural changes of Li3N react with H2O.

28
Table S3. The Gibbs free energy of Li3N react with H2O.

State Free energy (eV)

Li3N 0

Li2NH+LiOH -1.26983

Li2NH 0.67884

LiNH2+LiOH -0.90065

LiNH2 1.15511

NH3+LiOH -0.16146

NH3 1.92807

29
Figure S23. The structural changes of N2 to NH3 on Ru with one layer of Li.

30
 Figure S24. (A) The TDOS of N2 adsorption on Ru with one layer of Li, inside is that of N

and Li. (B) The PDOS of Ru-5s, Ru-4p, Ru-4d. (C) The PDOS of N-2s, N-2p. (D) The PDOS

of Li-2s, Li-2p.

31
Figure S25. The Gibbs free energy of N2 to NH3 on Ru without Li layer.

32
References

(1) Kresse, G.; Furthmiiller, J. Efficiency of ab-initio total energy calculations for metals

and semiconductors using a plane-wave basis set. Computational Materials Science

1996, 6 (1), 15.

(2) Kresse, G.; Furthmiiller, J. Efficient iterative schemes for ab initio total-energy

calculations using a plane-wave basis set. Physical review B 1996, 54 (16), 11169.

(3) Perdew, J. P.; Burke, K.; Ernzerhof, M. Generalized Gradient Approximation Made

Simple. Physical review letters 1996, 77 (18), 3865.

(4) Hammer, B.; Hansen, L. B.; Nørskov, J. K. Improved adsorption energetics within

density-functional theory using revised Perdew-Burke-Ernzerhof functionals. Physical

review B 1999, 59 (11), 7413.

(5) Blöchl, P. E. Projector augmented-wave method. Physical review B 1994, 50 (24),

17953.

33