Electrosynthesis of ammonia with high selectivity and high rates via engineering of the solid-electrolyte interphase

Abstract

Ammonia is a large-scale commodity essential to fertilizer production, but the Haber-Bosch process leads to massive emissions of carbon dioxide. Electrochemical ammonia synthesis is an attractive alternative pathway, but the process is still limited by low ammonia production rate and faradaic efficiency. Herein, guided by our theoretical model, we present a highly efficient lithium-mediated process enabled by using different lithium salts, leading to the formation of a uniform solid-electrolyte interphase (SEI) layer on a porous copper electrode. The uniform lithium-fluoride-enriched SEI layer provides an ammonia production rate of 2.5 ± 0.1 μmol s^-1 cm_geo ^-2 at a current density of -1 A cm_geo ^-2 with 71% ± 3% faradaic efficiency under 20 bar nitrogen. Experimental X-ray analysis reveals that the lithium tetrafluoroborate electrolyte induces the formation of a compact and uniform SEI layer, which facilitates homogeneous lithium plating, suppresses the undesired hydrogen evolution as well as electrolyte decomposition, and enhances the nitrogen reduction.

Keywords: electrochemical ammonia synthesis; high rates; high selectivity; lithium-mediated nitrogen reduction; solid-electrolyte interphase.

Graphical abstract

Figure 1

Theoretical investigation of the SEI layer by using different lithium salts (A) Schematic of the mechanism for Li-mediated ammonia synthesis. Although the accurate mechanisms are still not entirely understood, it is broadly believed that this LiNR process relies on the metallic lithium reduced from Li⁺ to dissociate N₂ followed by a sequence of electron and proton transfers to form NH₃ with suitable proton donors. (B) Calculated Gibbs formation free energy of Li-containing compounds as a function of voltage (versus Li/Li⁺). (C) The Li⁺ conductivity and energy barrier of Li surface mobility for the Li₂CO₃, LiOH, LiHF₂, and LiF at the operating voltage (U_Li/Li+ = 0 V). The error bars represent the uncertainty of calculated Li⁺ conductivity. (D) Schematic illustration of proposed Li diffusion model for a LiF-enriched SEI layer during the LiNR process. (E) The Gibbs adsorption free energy of NH₃ on different Li-containing compounds.

Figure 2

Fabrication of porous Cu electrodes for Li-mediated ammonia synthesis (A–C) Representative SEM images of the stainless steel (SS) mesh (A) and porous Cu electrode (B and C). (D) Cross-section SEM images of the porous Cu electrode. (E) Cyclic voltammetry of different porous Cu electrodes at scan rate of 30 mV s⁻¹. (F) Current density change versus scan rate of different porous Cu electrodes and the calculated specific capacitances. The change in current density was determined at −0.5 V versus reference electrode.

Figure 3

Electrochemical performance of Li-mediated ammonia synthesis (A and B) Digital photos of the setup for working, counter, and reference electrodes (A) and the glass cell sitting in the autoclave (B). The distance between WE and RE was fixed around 0.5 cm for all the experiments. (C and D) Digital photos of the autoclave sitting in the fume hood (C) and Ar glovebox (D). (E) LSV of the porous Cu electrode using different lithium salts. Inset in (E) is a digital photo of the porous Cu electrode (0.2 cm_geo²). (F and G) Chronopotentiometry (CP) of the porous Cu electrode at current densities of −0.1, −0.2, and −0.5 A cm_geo⁻² (F) and −1.0 A cm_geo⁻² (G) with different lithium salts. Inset in (G) is the cross-section SEM image of the porous Cu electrode without porous Cu on the backside. The black lines represent the data of the porous Cu electrode without porous Cu on the backside. All the experiments here were using the porous Cu electrodes that were synthesized at the same condition, and 2 M lithium salt in tetrahydrofuran solutions containing 1 vol % ethanol under 20 bar N₂. A total charge of 240 C was passed for the CP measurements at varied current densities.

Figure 4

Efficiency of the Li-mediated ammonia synthesis (A and B) Faradaic efficiencies (A) and NH₃ production rates (B) of the porous Cu electrode using different lithium salts at current densities ranging from −0.075 to −1.0 A cm_geo⁻². The shadows in (B) are guides to the eye. (C) A comparison of NH₃ production metrics at ambient temperature between our work and reported highest rates in non-aqueous electrolytes in the literature. (D) Accumulated NH₃ in the electrolyte, deposited layer including SEI, and gas phases using different lithium salts at −1.0 A cm_geo⁻². The left and right y axis in (D) represents the weight and corresponding percentage of NH₃, respectively. The calculated faradaic efficiencies and NH₃ production rates at current densities ranging from −0.1 to −1.0 A cm⁻² are based on the experiments shown in Figure 3. The error bars represent the standard deviation of independent repeats of the same experiment (n ≥ 3).

Figure 5

XPS investigation on the SEI layers without degassing damage and air exposure (A and B) Depth-profiling XPS spectra of F 1s (A) and C 1s (B) for the SEI-LiBF₄. (C and D) Depth-profiling XPS spectra of F 1s (C) and C 1s (D) for the SEI-LiPF₆. (E and F) Depth-profiling XPS spectra of Cl 2p (E) and C 1s (F) for the SEI-LiClO₄. The commercial LiBF₄, LiPF₆, and LiCl powder were used as reference samples.