Solar Ammonia In A Climate Crisis: Electrolysis Is No Barrier To Low-Cost Renewable Ammonia
By Miguel Rico Luengo (aka Lambda), ETH Zürich engineer.
Published20 hours ago
In parts one and two of this discussion, we have delved into the environmental impacts of biofuels and ammonia. The time has arrived to examine the second and third most important considerations for alternative shipping fuels: economics and technical feasibility. In this third part, we will explain how to synthesize ammonia using electrolysis and continue with a discussion on capacity factors, wrapping up with a review of the main ammonia costs. Part 4 will discuss other implementation considerations of ammonia-powered ships, compare costs with biofuels and bunker fuels, and provide a high-level summary of the entire topic.
Achievable electrolysis efficiencies for solar ammonia
The inquisitive reader will be wondering why I assumed a 40% solar-to-ammonia energy penalty factor in parts 1 and 2. To arrive at this figure, I considered the ammonia synthesis methods by Haldor-Topsøe, a catalytic process technology development company that commands a market-leading position in the catalyst and ammonia industries. This Danish company has been developing a solid oxide electrolyzer cell (SOEC) technology to electrolyze water at high temperatures, and is targeting mass production in 2023 with a 0.5–5 GW factory. In conjunction with a Haber-Bosch reactor, which converts H2 and N2 into NH3, this SOEC electrolyzer produces ammonia at a specific electricity consumption of about 7.22 kWh/kgNH3, or 26 MJ/kg. Of this figure, 6% accounts for gas pressurization, Haber-Bosch, and ammonia refrigeration. Interestingly, no air separation unit is needed to separate the oxygen from the nitrogen in the air; that task is performed by the SOEC electrolyzer. The elegance of this integrated process is shown diagrammatically below:
All in all, Haldor-Topsøe’s ammonia synthesis converts 71.5% of electrical input energy into chemical energy in the form of chilled ammonia. The following table lists other reported water electrolysis efficiencies to show why this electricity penalty number is likely realistic. Note that H2Pro’s electrolyzer is the only non-SOEC system on the list:
I’ll briefly note that when I tried to verify Helmeth’s LHV efficiency using their stated numbers, I did not arrive at 88.8% efficiency as claimed, but the above 82.9%. Bloom Energy’s efficiency also doesn’t account for a steam input. Sunfire’s claimed numbers are the highest at the system level, but testing in an industrial environment in August of 2020 lends some credibility to their technology. Zooming out from the stack to the system level will evidently result in increased power draw, but these auxiliary losses are fairly minimal at an approximate 6–12% of total energy consumption as evidenced across several electrolyzer types by Table 6 of a very exhaustive IRENA report.
Other companies developing SOEC technology include Toshiba, Ceres Power, OxEon Energy, and Nexceris. And with solid-oxide cell technology lending itself to both electrolysis and fuel cell operation (reversible solid-oxide cells, or r-SOCs), it is conceivable that many of the companies now solely developing solid-oxide fuel cells, or SOFCs, might also develop electrolyzers in the future.
Do the above numbers violate the laws of thermodynamics? Far from it. Using hydrogen from water electrolysis and applying Haber-Bosch, the energy input per kg of NH3 can be no less than 21.3 MJ. Direct electrochemical ammonia synthesis methods under development would lower this limit to 19.9 MJ/kgNH3, which is just shy of ammonia’s 18.6 MJ/kg LHV energy content. Haldor-Topsøe’s 26.0 MJ/kgNH3 are therefore well within thermodynamic limits.
Electrolyzer costs not an issue
Hopefully I’ve convinced the reader that the assumed conversion efficiencies are realistic, but how about the costs? After all, efficiency is useless if costs are prohibitive. Well, contrary to my prior assumptions, electrolyzer costs are not negligible. But they are not a deal breaker, either. Let’s take a closer look.
A recent study by the European Technology and Innovation Platform for Photovoltaics (ETIP PV) assumed 2021 electrolyzer costs of €400/kW. This figure exceeds IRENA’s 2020 study numbers of $450/kW, a price level that covers the full system cost, “including the electrolyzer stack, balance of plant (BoP), installation, civil works, grid connection, and utilities.” Taking the ETIP-PV study’s Rajasthan electrolyzer capacity factor of 32.2% (closest to Solar Star PV farm’s proven 32.8% CF), we obtain an approximate electrolyzer cost contribution to the levelized cost of hydrogen (LCOH) of ¢0.57/MJ. Interestingly, the base growth scenario of the ETIP-PV study projects a system electrolyzer CAPEX of approximately $260/kW by 2030. This figure seems conservative if you consider that Stiesdal is targeting serial production of €200/kW alkaline electrolyzers by 2023. That would be about $330/kW at the system level, including BoP when accounting for a 70% electrolyzer stack contribution to the system-level cost, a percentage extrapolated from Fig. 10 of yet another IRENA report. That is, to reach the 2030 ETIP-PV study’s base growth numbers, we will only need to cut $70 in CAPEX costs in 7 years. Further, 2025 cost projections of the ETIP-PV’s fast-growth scenario will be reached in 2023 by Stiesdal. The study’s numbers thus trail 2 years behind near-term industry targets. BNEF even projects $115/kWe prices for alkaline electrolyzers in 2030 in China. Assuming those are stack-level costs, this figure would correspond to $165/kW at the system level (i.e., almost a full $100/kW less than assumed by the ETIP-PV study by the same time).
Fear not if you understandably skimmed through the above number-packed paragraph. All of it is meant to convey that the ETIP-PV study’s projections present a more than adequate assessment of solar hydrogen cost reductions. The summary of the above numbers is that neither capacity factors nor capital costs will be an impediment to the continued cost decreases of renewable hydrogen (and by extension ammonia). IRENA, too, suggests a minimal cost decrease for an increase in the electrolyzer annual full load hours from 3200 h to 4200 h (Fig. ES1 and 1). Agora Energiewende (Fig. 2) backs up this assessment.
Relevant parallels can be drawn with the solar industry. Just a decade ago, solar PV could not compete with conventional energy generation technologies at the utility scale. The capacity factor wasn’t the problem, the high installation costs were. Capacity factors have barely improved since then but costs have cratered. As new nuclear energy installations can wistfully attest, low costs are not achieved with high capacity factors, but with low CAPEX resulting from mass industrialization and deployment. Today, solar is fast becoming the cheapest energy source nearly everywhere. As manufactured systems with similarly low operational capacity factors as solar, electrolyzers will be blessed with the same cost reduction phenomena.
Even if, in a hypothetical scenario, capacity factors were deemed crucial in lowering hydrogen production costs, there would be multiple sites around the world with a combination of both excellent wind and solar resources. These regions include Chile, large parts of Australia, South Africa, Namibia, Kazakhstan, and the Tibetan Plateau. Also relevant would be Mauritania, Morocco, Sudan, Egypt, and the Horn of Africa, all of which lie in the immediate vicinity of some of the world’s busiest shipping lanes. And with bids for 24-hour solar electricity (hybrid PV-CSP) coming in at $39.99/MWh it is not out of the question that near round-the-clock solar electricity might reach $30/MWh (if not $20/MWh eventually). Offshore wind, another energy source that’s enjoying sustained cost reductions, has demonstrated capacity factors of 54% as averaged over 2 years of operation.
What about water usage in arid regions? Topography permitting, the water demands of the electrolysis process could be supplied by ingenious pumped hydro with desalination. This system could simultaneously produce fresh water for industry or agrivoltaic crops and increase the electrolyzer capacity factor. Cooling needs would be met with dry cooling, a technology especially suited for systems like SOECs that shed heat at a high temperature.
A summary of costs
As for transportation and distribution costs, the literature is scarce and the estimates vary widely. I chose the more expensive of the two logistics cost studies for calculation of the full costs. Without further ado, here’s the relevant chart summarizing the main costs, including cost targets by three commercial entities as well as Lazard’s most recent numbers:
Please note that the IEEJ transportation costs above do not only cover the long trip from the Middle East to Japan with a comparatively small fleet of 11 and 19 ships in 2030 and 2035, respectively, but also account for distributing the ammonia to power plants for electricity production. Naturally, with this analysis focusing on ammonia for ships that dock in ports, these last costs should be omitted. And scaling up ammonia consumption to the entire shipping industry would certainly decrease the per unit cost of the ammonia carrier ships. Needless to say, more than 19 ammonia carrier ships are required to service the global market.
Evidently, the need for ammonia carrier ships implies that cargo ships cannot make brief detours to tank directly from ammonia production sites. However, multiple Middle Eastern and North African countries are so close to one of the world’s main shipping routes (Asia-Europe) that transportation costs in this case would virtually disappear. The result is a feasible ¢1.35/MJ,NH3,LHV for $20/MWh electricity.
Solar ammonia, energy is unstoppable
For those that doubt that continued electricity cost reductions can result in more widespread $10/MWh solar LCOE, it should be illuminating to consider that, in August of 2016, Chile announced a then record-low, unsubsidized bid price of $29.1/MWh for a 120 MW utility-scale PV project. Less than 5 years later, in April of 2021, it was announced that the 600 MW Al Shuaiba PV project in Saudi Arabia, a region with a lower solar irradiance than Atacama, would sell power at $10.4/MWh. At almost a third the price, this should be a cautionary tale for betting against PV costs from continuing their inexorable downward march. Projecting the average 1979–2020 US inflation rate of 3.5% until 2040 would raise today’s best PV PPA costs to $19.99/MWh in 2021 dollars. The implicit assumption in this LCOE is that there will be no further solar PV cost reductions. This would be ahistorical; the average solar PV learning rate, the cost decrease associated with each doubling of cumulative capacity, stood at 40% from 2006 to 2020, the historical average being 23.8%. Yet, as of 2020, solar PV produced less than 4% of worldwide electricity. With many industries still left to electrify, and the majority of the world still in the process of developing economically, it would be ill-advised not to expect a few extra doublings in installed capacity.
That concludes the third part of this series. In the final article, we will compare the costs of all relevant shipping fuels, dispel other doubts regarding the viability of ammonia for shipping, and present a future roadmap of where the industry as a whole could be headed.