The procedures and materials for nitrogen reduction outlined below have been optimized for reliable ammonia production. However, certain deviations from the procedure and material vendors, which are specifically called out in-text and in prior work, 1 may lead to poorer ammonia production. We recommend closely following the procedure for reproducibility and high yields. MaterialsTetrahydrofuran (THF, 99+%, stabilized with BHT) and molecular sieves (3Å, 4-8 mesh) were purchased from Acros Organics. Lithium tetrafluoroborate (LiBF4, 98%), dimethyl carbonate (DMC, ReagentPlus, 99%), propylene carbonate (PC, anhydrous, 99.7%), maleic acid (ReagentPlus, ≥99%), sodium salicylate (ReagentPlus®, ≥99.5%), sodium hypochlorite (NaOCl, 10-15%), nickel chloride hexahydrate (NiCl2•6H2O), hydrochloric acid (HCl, 37%), and ferrocenium hexafluorophosphate (FcPF6, 97%) were purchased from Sigma-Aldrich. Hexanes (C6H14), diethyl ether (Et2O), and sulfuric acid (H2SO4, 93-98%, trace metal grade) were purchased from Fisher Scientific. Platinum foil (Pt, 0.025 mm thick, 99.99%, trace metals basis) and sodium citrate (anhydrous) were purchased from Beantown Chemical. Ethyl alcohol (C2H5OH, Koptec, anhydrous, 200 proof), sodium hydroxide (NaOH, Macron Fine Chemicals, pellet form), and acetone (ACS, BDH Chemical) were purchased from VWR International. Sodium nitroprusside (99-102%), ferrocene (Fc, 99%, powder), boric acid (H3BO3, 99.99%, trace metals grade), nickel foil (annealed, 99+%, 0.05 mm), ammonium chloride (NH4Cl, anhydrous, 99.99%), and ammonium hexachloroplatinate (IV) ((NH4)2PtCl6, 43.4% min Pt) were purchased from Alfa Aesar. Isotope labelled nitrogen ( 15 N2, 98%+) was purchased from Cambridge Isotope Laboratories, Inc. Argon gas (UHP, 5.0 grade) and hydrogen gas (UHP, 5.0 grade) were purchased from Airgas. Nitrogen gas was available in-house; it is generated by boil-off of liquid nitrogen from Airgas. Milli-Q water was obtained by filtering deionized (DI) water through a Milli-Q purification system (Merck, Millipore Corporation). Steel cloth (304 stainless steel, 400x400 mesh) and steel foil (cold-worked 304 stainless steel, 0.002" thick) were purchased from McMaster-Carr. Platinum-coated carbon paper (0.5 mg cm -2 60% platinum on Vulcan, carbon paper) was purchased from FuelCellStore. Polyporous Daramic 175 separators were received as a sample from Daramic (Charlotte, NC). Electrolyte solution preparationDry THF was used as the solvent in THF-based experiments described below. It was obtained by drying as-purchased THF over 20% v/v of freshly dried molecular sieves for at least 48 hours in a round-bottom flask sealed with a rubber septum stopper. The sieves were prepared by washing with acetone and heating at 300 °C for 5 hours in a muffle furnace. The water content of dry THF was found to be 7.1±0.3 ppm (n = 3) via Karl-Fischer titration. As-purchased LiBF4, stored in an Ar glovebox, was dissolved in dry THF to obtain electrolyte solutions containing 1 M LiBF4. As discussed in previous work, 1 it is imperative for the LiB...

Alternative approaches for producing ammonia are necessary to reduce the environmental impact of its production. The lithium-mediated electrochemical nitrogen reduction reaction (LM-NRR) is one attractive alternative method for producing ammonia at small scales in a distributed process. This process requires a proton donor in the electrolyte to produce ammonia from nitrogen, but the role of the proton donor in selective ammonia production is not well understood. In this work, we experimentally tested several classes of proton donors for the ability to promote LM-NRR. We found that a wide array of alcohols can promote nitrogen reduction and that n-butanol leads to the highest ammonia Faradaic efficiencies. Among the tested proton donors, even slight changes in the proton donor structure can significantly affect the yield of ammonia. In addition, most active proton donors exhibit a thresholding behavior as a function of their concentration, where the selectivity toward ammonia increases dramatically above a certain concentration of the proton donor. We found evidence to imply that these effects could be due to the proton-donor-induced changes in the properties of the solid electrolyte interphase (SEI), which lead to changes in the diffusion of relevant species through the SEI to the reactive electrode. By selectively allowing for diffusion of nitrogen over the proton donor to the electrode, the SEI can promote selective nitrogen reduction to ammonia. A coupled kinetic transport model of the process was proposed to explain the observed trends and to predict ammonia production as a function of operating conditions.

Novel methods for producing ammonia, a large-scale industrial chemical, are necessary for reducing the environmental impact of its production. Lithium-mediated electrochemical nitrogen reduction is one attractive alternative method for producing ammonia. In this work, we experimentally tested several classes of proton donors for activity in the lithium-mediated approach. From these data, an interpretable data-driven classification model is constructed to distinguish between active and inactive proton donors; solvatochromic Kamlet–Taft parameters emerged to be the key descriptors for predicting nitrogen reduction activity. A deep learning model is trained to predict these parameters using experimental data from the literature. The combination of the classification and deep learning models provides a predictive mapping from proton donor structure to activity for nitrogen reduction. We demonstrate that the two-model approach is superior to a purely mechanistic or a data-driven approach in accuracy and experimental data efficiency.

Methods to produce ammonia from air, water, and renewable electricity are necessary to transition ammonia production away from the CO2-emitting Haber-Bosch process. In this vein, a fully electric process in which water-splitting-derived hydrogen and air-separation-derived nitrogen are reacted in an electrochemical process to produce ammonia is attractive. Such a process has the potential to be highly flexible and utilize intermittent renewable energy well. Here, we evaluated the cost-effectiveness of large-scale fully electric ammonia production relying on renewable electricity sources in conjunction with different types of storage and flexible operation, using a mixed-integer linear programming framework. The approach incorporates a first-principle, chemistry-independent representation of reactor power consumption and its dependence on reactor sizing and electrochemical parameters, the impact of product separation and recycling unconverted reactants, and plant dynamics in response to temporal variability in renewable energy availability. Given the emerging nature of electrochemical ammonia synthesis from nitrogen and hydrogen, we used the model to identify the reaction descriptors and their threshold values that enable cost parity between fully electric ammonia production with commercially viable production using thermochemical synthesis coupled with electrolytic hydrogen powered by renewable energy. We found that ammonia can be produced in an economically competitive manner, i.e. at costs <1 $/kg, at large scales if the electrochemical reactor can produce ammonia at partial currents exceeding 400 mA cm -2 , energy efficiencies exceeding 30%, and process lifetimes of several years. In light of this, novel chemistries that can reduce nitrogen at high rates and moderate (<2.5 V) overpotentials are necessary for economic, large-scale ammonia production.

Cost-effective, low-carbon ammonia production is necessary for decarbonizing its existing uses but could also enable decarbonization of other difficult to electrify end uses such as shipping, where the energy density is a key criterion. Here, we assess the levelized cost of ammonia production (95% availability) at industrial-scale quantities (250 tonnes/day) in 2030 from integrating commercial technologies for renewable electricity generation, electrolysis, ammonia synthesis, and energy storage. Our analysis accounts for the spatial and temporal variability in cost and emissions attributes of the electricity supply from variable renewable energy (VRE) sources and the grid and its implications on plant design, operations, cost, and emissions. On the basis of 2030 technology cost and grid projections, we find that grid-connected ammonia in midcontinental U.S. costs 0.54–0.64 $/kg, in comparison to 0.3–0.4 $/kg for natural-gas-based ammonia and, depending on the generation mix of the grid, may have higher or lower CO2 emissions. Fully VRE based ammonia production, even with simultaneous wind and PV utilization, is more expensive than grid-connected outcomes, due to the need for storage to manage VRE intermittency and continuous ammonia production. Using a combination of VRE and grid electricity at locations of existing ammonia facilities in the midcontinental U.S. can achieve 2–80% CO2 emissions reduction per tonne of ammonia in comparison to natural gas routes and corresponds to a levelized cost range of 0.57–0.85 $/kg NH3. Further cost reductions are shown to be possible if the ammonia synthesis loop can be made more flexible, which reduces the need for a round-the-clock electricity supply and substitutes use of battery storage with ammonia storage.