Can AI-optimized plasma make green ammonia more cost-competitive?

Applications of AI


[Disclosure: AFN’s parent company, AgFunder, is an investor in Faraday Earth.]

Faraday Earth, a start-up that uses non-thermal plasma to produce green ammonia, claims its system could reach a levelized cost of about $500 per tonne, putting it in close proximity to the fossil fuel-derived gray ammonia that is the backbone of most nitrogen-based fertilizers.

Couple this with a decentralized, container-sized, modular system that reduces transportation costs associated with traditional ammonia supply chains, and the economics suddenly become more attractive, says co-founder Debayan Saha. agfunder news.

“Ammonia is an almost $100 billion market, and fertilizers make up the largest portion of this market. In fact, ammonia feeds half of the world’s population. And our technology essentially solves the bottleneck in traditional ammonia production processes, where nitrogen is very unreactive.”

Unlike the traditional Haber-Bosch process, which uses high heat, high pressure, and catalysts to force nitrogen and hydrogen to react to produce ammonia in a huge centralized plant, Faraday Earth instead uses high-voltage electric fields to produce a non-thermal plasma.

The high-energy gas, also known as the “fourth state of matter,” can “excite” nitrogen molecules, making them more reactive.

Saha, who founded Faraday Earth with Dr. Shashi Ranjan in 2023, says this is important because nitrogen gas is notoriously inert.

Each molecule consists of two nitrogen atoms held together by a strong triple bond. Once nitrogen is activated in a plasma reactor, it can be combined with green hydrogen, and Faraday Earth can be produced by electrolysis (splitting water into hydrogen and oxygen) or sourced from other low-carbon sources, such as naturally occurring geological hydrogen.

Once produced, ammonia can be liquefied and injected directly as fertilizer, dissolved in water to produce ammonia water for industrial use, or used as a feedstock for other fertilizers and chemicals. However, Faraday Earth’s initial focus is on ammonia as the end product.

eureka moment

Saha and Ranjan demonstrated a new activation pathway for nitrogen, but the approach was initially not very efficient, Ranjan says. “When we started, the numbers were 60 to 70 times worse. [than they are today];I tried to optimize it using only normal experimental methods, but the progress was not very good. ”

Saha, who met Ranjan at Stanford University 10 years ago, added that the new moment came when the two began experimenting with machine learning to optimize plasma and maximize yield.

“I knew that Google DeepMind was using AI extensively to optimize plasma reactors for fusion reactors, and it occurred to me that AI could be used to optimize this type of plasma.”

Faraday Earth, which is incorporated in the US and has R&D operations in India, has since created an AI-based digital twin that shows what is happening inside the reaction. Given the number of interdependent variables, it is difficult to optimize through trial and error experimentation.

This allows for “near real-time” adjustment of voltage, current, flow rate, electrode-related parameters, and other reactor conditions, significantly improving performance.

Not all green ammonia technologies are the same…

Although several startups have entered the green ammonia space, Saha argues that alternative approaches such as electrochemical nitrogen reduction (using electricity to produce ammonia in water-based systems) may struggle because nitrogen is less soluble in water systems. This means that much of the nitrogen in the bubbles is not participating in the reaction, he explains.

“What ends up happening is that only the nitrogen in the outer shell of the bubble takes part in the reaction, and all the nitrogen inside the bubble is just wasted.”

He is also skeptical of the lithium-mediated approach (an electrochemical route that uses lithium to activate nitrogen and turn it into ammonia), citing concerns about reactor stability and potential constraints in the lithium supply chain.

On the other hand, plasma-based approaches that first convert nitrogen to nitrates can be less efficient if the goal is ammonia, as additional energy must be spent converting those nitrates back to ammonia, he argues.

In contrast, Faraday Earth seeks to directly activate nitrogen in the plasma to produce ammonia, rather than producing nitrate first. The company still uses a catalyst to help react nitrogen and hydrogen, but says the plasma does the hard part: activating the nitrogen, replacing the need for highly specialized nitrogen-activated catalysts.

Spraying ammonia fertilizer on the field - istock - Smederevac
Ammonia is an important component of most nitrogen fertilizers and is either used directly or converted into products such as urea, ammonium nitrate, and ammonium sulfate. Image courtesy: istock-Smederevac

Progress to date

Faraday Earth built a demo unit to show its technology in action and secured a letter of intent, a pilot contract offer, a purchase order, and at least one paying customer. Ranjan said the company is seeing interest from existing ammonia producers, distributors and suppliers serving markets including fertilizer, cold storage, meat processing and other industrial applications. “We have signed a memorandum of understanding for commercialization with one of the world’s leading companies in this field.”

Rather than building large centralized plants, Faraday Earth envisions containerized units (typically about 40 feet by 10 feet for large systems) that can produce several tons of ammonia per day. Customers with large demands can stack multiple modules, and small users can deploy smaller systems.

This could enable decentralized ammonia production for fertilizer and industrial applications, especially in places where ammonia delivered by logistics is expensive, Saha said.

Target of $500/ton

On cost, he says, “Instead of just looking at production costs, the levelized cost of ammonia is the real parameter we measure (including capital expenditures, operating costs, energy inputs, maintenance, etc.). We expect the levelized cost to be in the range of about $500 per tonne for the coming period. It could potentially go even lower if we can reach the theoretical limit.”

As a comparison, he says: [produced via Haber Bosch] Prewar [with Iran]ranged from $300 to $700 per ton. ”

Faraday Earth’s system is designed to pair well with renewable electricity, as the plasma can be switched on and off quickly, avoiding the long heating and cooling cycles associated with thermal processes, Saha said, noting that the Iran war has put both the price and availability of key fertilizer inputs into sharp focus in recent weeks.

“Ammonia is an almost $100 billion market, and fertilizers make up the largest portion of this market. In fact, ammonia feeds half of the world’s population. And our technology essentially solves the bottleneck in traditional ammonia production processes, where nitrogen is very unreactive.”

Ryan Lee, investment associate at Agfunder, which recently invested in Faraday Earth, said green synthesis of key commodities “represents a fundamental decentralization opportunity, and Faraday Earth is developing technology to address bottlenecks in how the world produces essential inputs. Localization of production in the agricultural economy, particularly in the Global South, has the potential to restructure supply chain dependencies and decarbonize critical industries.”

Read more:

U.S. Senate Fertilizer Hearing: Farmers “cannot take more” due to price hikes

Brazil’s potash insists on local supply as fertilizer market turmoil due to Iran war

Fertilizer prices for U.S. corn increase by up to $35 per acre as Iran crisis worsens

North Carolina farmers weigh in on growing Persian Gulf fertilizer crisis: ‘We will see significant reductions in yields and acres’



Source link