(14c) Spatial Variation in Cost of Renewable Electricity-Driven Continuous Ammonia Production across the United States

Ammonia production contributes more than 1% of the global greenhouse gas emissions (GHG) while being used to serve a majority of the demand for nitrogen-containing fertilizer for agricultural use. The predominant route for ammonia production relies on natural gas as a source of energy and hydrogen for thermochemical Haber-Bosch (HB) synthesis, and is estimated to result in about 2.3 tonnes of CO₂ per tonne NH₃ ammonia produced.[1] Declining costs of variable renewable energy (VRE) based electricity and electrolyzers have raised interest in producing low-carbon H₂ via electrolysis, as well as its use in decarbonization of industrial ammonia production. This electrically-driven ammonia route could not only serve existing uses for fertilizer production, but also be deployed to service energy needs for other end-use sectors where ammonia is being contemplated (e.g. marine transport). Here, we study the spatial variations in cost of producing ammonia across the U.S. predominantly based on the temporal variations in using renewable electricity input in conjunction with thermochemical HB synthesis. The plant configuration targets continuous ammonia supply to an end-use customer throughout the year while rely on energy input as electricity, supplied from a combination of on-site VRE generation as well as grid imports. Production of feed H₂ and N₂ for HB synthesis is modeled using pressure swing adsorption (PSA) and low temperature proton exchange membrane (PEM) electrolysis. The energy inputs and capital cost for the HB synthesis loop, including compressors, heat recovery and reactor system, are parameterized based on offline Aspen simulations.

Our approach goes beyond prior techno-economic assessments of electricity-driven ammonia production by explicitly accounting for variability in electricity supply and its implications on plant design, cost and emissions. This is achieved by using a least-cost integrated design and operations modeling framework that treats as variables the relative sizing of various units (e.g. electrolyzer, PSA, renewables capacity), including possible on-site deployment of alternative forms of on-site storage (battery energy storage, gaseous H₂ and liquid N₂). The model accounts for intra-annual variability in inputs (VRE resource, grid electricity marginal price and emissions) at an hourly resolution and includes various constraints such as: a) such as inter-temporal constraints governing energy storage (H₂, battery) and electrolysis operation, b)inflexible ammonia production because of the relatively harsh operating conditions (500^oC, 250 bar) , with possibility to go offline for maintenance with minimum up and downtime requirements, c) constraints enforcing hourly (constant) and annual production (>95% availability) requirements as well as any emissions intensity constraints. The overall mixed-integer linear programming (MILP) model is able to optimize for the minimum annualized cost of providing round-the-clock ammonia under the required system emission and flexibility constraints.

We used the model to evaluate the levelized cost of ammonia for multiple electricity source and location scenarios in the US. Our analysis shows that the levelized cost of a grid-connected ammonia is nearly 50% higher than current natural gas, HB-produced ammonia in the absence of carbon tariffs. Relying on a grid electricity mix with higher VRE penetration (e.g. New York) allows for producing ammonia with lower overall emission intensity (e.g. 1.6 CO₂ tonne/tonne) compared to conventional natural gas-based routes. We also evaluated dedicated VRE-based ammonia production for locations in close proximity to existing NH₃ production facilities and agricultural hubs in the US, to identify the cost-optimal VRE mix and storage requirements. Our analysis shows that a standalone renewable ammonia production facility makes use of storage of intermediate products (N₂, H₂) in the production process so as to be able to dispatch them during non-availability of renewable electricity. To meet the minimum power input necessary to operate the thermochemical HB process, electrochemical storage (e.g. Li-ion) is also needed. However, if the thermochemical HB process can be operated at less than nameplate feed flow rates, the need for Li-ion battery storage is minimized, allowing for more cost-effective production options.

[1] IEA (2019), The Future of Hydrogen, IEA, Paris https://www.iea.org/reports/the-future-of-hydrogen