Solar Ammonia for Shipping

By Miguel R.L. (aka Lambda), ETH Zürich engineer

In part 1 of the ammonia vs biofuels discussion, we looked at the large-scale destruction that would ensue if we scaled biofuel production to satisfy global demand. But while we have established that biofuels lose out handily to PV in the area metric, it might be reasonable to assume that biofuels cost less per unit of energy and produce less CO2 compared to manufactured systems like PV farms. After all, it’s nature doing the work of splitting the hydrogen from water, and then combining it into hydrocarbons using the carbon from CO2.

I suspect that our reckless extraction and use of fossil fuels produced by photosynthetic organisms millions of years ago may have habituated us to the idea that using nature practically comes for free. Be as it may, we should remind ourselves that fossil fuels are anything but free. The fossil fuel climate debt is large and growing, and it will be paid no matter what. Similarly, we cannot conveniently take the cost of biofuels at face value and ignore hidden costs and unpaid externalities. The same goes for ammonia synthesis, or any human activity for that matter. We must therefore preface a full technical feasibility and costing analysis in parts 3 and 4 of this discussion with an examination of the lifecycle emissions of the alternative fuels under consideration.

Current lifecycle emissions of biofuels and ammonia

For this section, we’ll be referring to recent studies, one from 2019 and one from 2021 for biofuels, and one detailed one from 2021 for solar PV.

By performing a lifecycle GHG emissions analysis on corn ethanol using the GREET model, scientists from Argonne National Laboratory found that corn ethanol has a carbon intensity (CI) of 53.2 gCO2e/MJ, which includes land-use change (LUC) GHG emissions of 7.4 gCO2e/MJ. In 14 years, the reduction in CI was a fairly minimal 23%. Similar research carried out in 2019 found that California’s Low-Carbon Fuel Standard certified “ethanol production pathways had an average CI of 66 gCO2e/MJ (including LUC GHG emissions), which is just 35% lower than that of the California gasoline blendstock (100.82 gCO2e/MJ).”

The US average petroleum blendstock had a CI of 93 gCO2e/MJ. Compared to the 2019 figures, we are looking at a 43% emissions reduction of corn ethanol compared to fossil fuels. Meaning that, not only are US taxpayers and consumers forking out considerable subsidies for ethanol and biodiesel, but they are doing so for a fuel that still emits over half of gasoline’s lifecycle emissions.

Other studies present more hopeful numbers, especially pertaining to sugarcane ethanol, which has the lowest life-cycle GHG emissions of all biofuels and thus represents the best case scenario for current production. To allow for a better comparison with the 2019 study, I’ve also used the GREET model’s numbers. Not only is the GREET model one of the most recent of the 4 models examined in the study, but at 24.0 gCO2e/MJ, it also nearly matches BioGrace’s numbers (23.0 gCO2e/MJ). GREET’s numbers also include emissions resulting from transportation beyond Brazil’s borders. However, due to significant variability in modeling assumptions, none of the sugarcane models account for LUC emissions. These need not be negligible, as exemplified by the case of palm oil biofuels.

Now, onto solar PV’s life-cycle analysis (LCA). It’s important to note that there are multiple solar PV LCA review papers out there, but most of them refer to data from 2012 or not long thereafter, which just wouldn’t be fair to PV given the significant improvements that have occurred in the last decade or so. Summarizing the numbers for the lifecycle emissions, we obtain the following Sciencedirect.com:

For the glass-glass (G-G) solar module numbers, the low efficiency of 19.40% should be stressed. Multiple modules coming onto the market should achieve over 22.5% shortly. And about half of the per kWp GHG emissions associated with module production, transport, and end-of-life disposal (but excluding BOS, installation, and operation) are solely related to the electricity needs of Poly-Si and monocrystalline Czochralski silicon (Cz-Si) — i.e., the ingot silicon used in wafer production. In light of the high energy intensity of Cz-Si, large emissions reductions are not unrealistic. Case in point, Nexwafe, who is targeting commercial deployment in 2022, claims a 70% emissions decrease for its novel wafer manufacturing process that dispenses with many of the conventional processing steps.

Higher module efficiencies, solar panel recycling, and a greening grid will all further lower PV’s carbon intensity. And some modules being installed today will likely continue to generate electricity half a century into the future, a far cry from the 30-year lifetime used in the calculations by Müller et al. All in all, I wouldn’t be surprised if solar ammonia approached 3 gCO2e/MJ carbon intensity by 2030 in the best solar locations.

Estimating electrolyzer lifecycle emissions is not straightforward, but Fig. 4 of a 2021 study puts them at about half of solar BOS (solar-related equipment and components other than the modules), with both being dwarfed by PV module contributions. Since until very recently electrolyzer manufacturing could best be described as artisanal, significant LCE reductions can also be expected for electrolyzers. I’ve also assumed that the construction emissions of the Haber-Bosch reactor used to synthesize ammonia from nitrogen and hydrogen gas are compensated by a reduction in the required electronics equipment (both electrolyzers and solar PV work on DC power) as well as the avoidance of a grid connection (assumed islanded operation).

Applying an additional 40% electricity-to-ammonia energy consumption penalty (to be justified in part 3), we conclude that ammonia produced with today’s lowest-carbon solar energy has an approximate CI of about 8.1 gCO2e/MJ as seen in the above table. This is almost exactly a third of the LCE of today’s best biofuels, and 15% of the LCE of corn ethanol, which, as mentioned in the first part of this series, accounts for 64% of current worldwide ethanol production.

So, yes, all things considered, most current biofuels produce fewer emissions than fossil fuels, but it’s not a close contest if solar ammonia is allowed to compete. 

Examining advanced biofuels

Thus far, we have omitted alternative biofuel production methods of reduced carbon emissions and environmental impact. 2G lignocellulosic biofuels from waste biomass are one such alternative. We could assume these will become dominant in the near future, but more than a decade (closer to two decades) of unfulfilled promises give pause to such wishful thinking.

No rigorous analysis that I’ve seen predicts a worldwide halt of maize, soy, and sugarcane biofuel cultivation in the next couple of years in favor of 2G feedstocks. Even if we pretended otherwise and assumed an economically viable, large-scale, and timely energy extraction from waste biomass, it would make far more sense to just go with the simplest and cheapest form of waste bioenergy: biogas from anaerobic digestion.

And, ironically, the lowest-carbon method to use biogas involves both hydrogen production and electrolysis. Let me explain. Biomethane makes up around 50% of biogas, the rest being primarily CO2. Methane’s chemical formula is CH4, or 1 carbon and 4 hydrogen atoms. If combusted, methane produces CO2. Even worse than CO2’s GHG warming potential is uncombusted (fugitive) methane, a fitting name for a gas that has every reason to be on the run.

Consequently, it’s the addition of carbon that turns hydrogen and oxygen into potent and long-lived greenhouse gasses. Methane pyrolysis addresses this very issue. It’s able to split methane into solid carbon and hydrogen gas using heat. Some forms of methane pyrolysis, including Hazer’s method, produce high-quality synthetic graphite (credit to FragileFungi for the reference), while lower quality solid carbon unsuitable for electrochemical storage could be permanently sequestered in abandoned mines. In other words, directly using methane produces avoidable GHG emissions, while pyrolyzing it first allows for easy carbon sequestration and even makes it available for the battery industry. Finally, the leftover hydrogen could be used for ammonia synthesis, which lines up with plans from companies such as Monolith Materials and CAC-H2., The former has already built a commercial plant. Other industrial scale biogas hydrogen production projects are in the works.

Where does the aforementioned electrolysis come in? If you recall, CO2 makes up about 50% of biogas. Emitting such a concentrated stream of biogenically captured CO2 back into the atmosphere is wasteful no matter how you look at it. Co-electrolysis of CO2 and water can help. This is a process, being commercialized by several companies (Sunfire, Twelve, Haldor-Topsøe), by which CO2 and water are electrolyzed at the same time to produce syngas for liquid hydrocarbon synfuel production. Such synfuels could be used as drop-in replacements for aircraft. Although, longer term, we would do well to eliminate even these aviation fuels in favor of battery electrification and fuel cell propulsion, both of which can eliminate tailpipe CO2 emissions, NOx formation, and contrails.

In terms of GHG emissions, avoiding biogas CO2 emissions altogether is better than co-electrolysis powered by low-carbon renewable energy, which itself beats directly converting methane into synfuels. Remember, we do not want to put carbon into the atmosphere unless we really think there is no good alternative to doing so. Problem avoidance is better than problem remediation. Bar sequestering all biogenic carbon, using renewable electricity and electrolysis to make carbon-neutral (or even carbon-negative) synfuels and ammonia is likely to be the lowest carbon way to process biogenic energy sources.

In a level economic playing field, the synthesis of such carbon-negative ammonia would see some form of financial support for its contributions to a liveable climate. Battery energy storage monetization would be a welcome plus. But historical precedents suggest it will play out differently: we continue to explicitly and implicitly subsidize entrenched, destructive ways even when cheaper and better ones become available.

In the case of biofuels, much of that support is of questionable origin. Some of its chief supporters happen to be the same actors who continue to flog the dead automotive hydrogen horse, or who foolishly endorse costly and impracticable hydrogen boilers instead of cheap electric heat pumps for space heating. I’m of course referring to the oil and gas industry.

Total, Phillip 66, BP, Eni, ExxonMobil, Marathon, HollyFrontier, and CVR Energy (to name a few) are all pursuing projects in biofuel refining. Some oil companies, such as Marathon, are entering the agricultural processing business. Chevron too, seemingly not content with destroying rich and productive ecosystems and people’s lives by dumping toxic waste in Ecuador, is trying its hand with environmentally destructive agribusiness. It certainly fits its previous pattern of rapacious behavior. But why are oil and gas companies pivoting to biofuels?

Very simply put, their bottom line is at stake. According to the International Energy Agency (IEA), 14% of current refining capacity in advanced economies “faces the risk of lower utilisation or closure” by 2030. Risking stranded assets, European and US refinery operators are switching to processing biofuels. In case their high-GHG blue hydrogen ruse is revealed as the environmental scam it currently is, they will have hedged their bets with biofuels, whose greenwashing is not yet widely questioned by the general public.

Carbon-negative farming would lessen biofuel’s environmental impact (although not its space requirements), but high variability in carbon uptake and dubious measurement methodologies make the extent of carbon sequestration uncertain. Crucially, it should be uncontroversial to assume that intact ecosystems that have been spared from conversion into inefficient open-air energy bioreactors will have superior carbon sequestration abilities, let alone a greater societal value. And that’s without accounting for the initial emissions of direct and indirect LUC emissions. Luckily, solar farms in deserts escape this level of ecosystem destruction.

That concludes today’s examination of lifecycle emissions. We’ve seen that any GHG emissions reductions achievable through hydrocarbon biofuels are bested by either: (a) desert solar farms, which unlike biofuels free up farmland for rich natural ecosystems, (b) methane pyrolysis, which produces hydrogen and solid carbon for the battery industry, or (c) co-electrolysis, which gives biogenically captured CO2 a second life as aircraft synfuels. The next part in this series will analyze the ammonia synthesis costs and address any perceived technical feasibility issues associated with electrolysis. Part 4 will finish with a safety, scale-up, and cost comparison.

Featured image: public domain (CC0) image.