Ammonia for fertilizer comes from large industrial plants, while remote farms can face a very different problem: getting agricultural inputs to the field. The Haber-Bosch process, which drives most industrial ammonia production, needs high temperatures, elevated pressures and centralized facilities. It also accounts for around 2% of the world’s annual carbon emissions. Roughly half the nitrogen in the human body has passed through a Haber-Bosch reactor.
On 20 June 2025, chemists and engineers at the University of Glasgow announced a £20 prototype that uses ultrasound, air and water to produce nitrate, a plant nutrient. The university’s account describes air bubbled through deionised water, followed by focused ultrasound that drives tiny bubbles to extreme temperatures. The team’s Cell Reports Physical Science paper, published on 16 July 2025, reports nitrate and nitrogen oxide production in minutes. A STAR Protocols paper, dated 20 April 2026 in the available record, sets out how other labs can design and assemble the sonoreactor.
The £20 figure describes a prototype, not a proven farm product.
Glasgow’s £20 sonoreactor targets fertilizer supply, not energy parity
A Times of India report says the prototype remains experimental and currently uses proportionately more energy than Haber-Bosch. Glasgow’s announcement presents the idea as a possible way to change agricultural supply chains in remote areas. The paper’s claim of a “step change” in energy requirements per mole is relative to previous aqueous-sonication reports, not to industrial ammonia production.
That distinction sets the test for the idea. The researchers are exploring whether nitrogen fixation can be made locally, in small equipment, rather than only in a centralized commodity plant. The logistics case is plausible, but the published results do not establish a cost advantage. They do not show the delivered cost of usable fertilizer from the device, and the team says the process would need to be scaled before individual farmers could make fertilizer on demand.
The price of a prototype says little about what a manufactured, field-ready system would cost. The more consequential step is publishing a method that other groups can try. If those groups can reproduce the output and measure energy use consistently, they can test whether local production makes sense where delivery is difficult. If they cannot, the £20 headline will have outrun the engineering.
Pulsed ultrasound drives bubbles to 5,000°C
The process begins by bubbling air through deionised water for a few minutes, allowing nitrogen and oxygen to dissolve. Focused ultrasound creates small bubbles that oscillate and collapse. This acoustic cavitation produces repeated implosions; gases inside the bubbles can reach 5,000°C, according to the University of Glasgow. The proposed chemistry is that these conditions break nitrogen molecules apart and allow nitrogen and oxygen to react, ultimately forming nitrate.
A high-speed camera helped the team watch bubbles oscillate and implode while it tested different ultrasound methods. The Glasgow account says four-millisecond bursts fired every 80 milliseconds produced the best results. Those timings describe the pulse pattern; the available evidence does not establish that every reported concentration was measured under identical operating conditions. Pulsing is central to the method, but it does not by itself settle the device’s total energy cost or practical output.
The use of sound waves to convert nitrogen from air into nitrate dates to a discovery first identified in the 1930s, according to the Times of India report. Glasgow’s work applies the approach using focused ultrasound. The mechanism explains how nitrate can form; it does not establish that a farm can get enough useful nutrient from the process.
The nitrate measurements show lab output, not crop performance
The paper reports that pulsed 200 kHz ultrasound applied to an air-and-water mixture at room temperature produced about 15 micromolar nitrate after 60 seconds. Its highlights also report 60 micromolar NOx within 480 seconds in a simple, low-cost device. The authors call this a record rate for nitrogen oxide production by sonication in aqueous solution and a step change in energy requirements per mole compared with previous reports.
Glasgow’s laboratory account gives another result: around 40 micromolar nitrate in 20 millilitres of water over eight minutes. The reported values refer to different durations and product measurements, so they are not interchangeable. Concentration alone does not tell a reader how much fertilizer a working system could make.
The research-data collection lists a broad Glasgow team, while the protocol record names support from the Engineering and Physical Sciences Research Council’s “Decarbonising Nitrogen Fixation for Sustainable Net-Zero Agriculture” program and the Royal Society. Those details show a research effort extending beyond a single announcement; they do not establish readiness for deployment. A secondary report says researchers still need to determine how effectively the nitrate supports plant growth. Crop performance remains an open test.
Reproducibility will determine whether the protocol has value
The next useful result is not a higher headline concentration. It is an independent build that reports nitrate output, NOx, energy consumption and operating conditions in enough detail to compare with the Glasgow work. Sonochemical results can depend on reactor geometry and operating conditions. The protocol gives other researchers a way to test reproducibility, but its publication is not evidence that separate builds will perform alike.
That distinction has consequences for the field. If independent labs can follow the protocol and obtain comparable results, the work gains a common baseline for assessing output and energy use. Other decentralized approaches, including plasma and electrochemical nitrogen fixation, could then be compared against a better-defined test. If independent groups cannot reproduce the reported output under comparable conditions, the first task will be to identify which design or operating variables explain the difference. Either outcome is more informative than treating the prototype price as a product quote.
My expectation is that logistics, rather than energy parity with Haber-Bosch, will remain the strongest argument for this research. That argument only holds where the cost and difficulty of bringing fertilizer in exceed the full cost of making and using nitrate locally. The ledger contains no delivered-cost comparison and no evidence of commitments from NGOs, defense logistics organizations, mining operators or fertilizer buyers. Those are possible users to investigate, not customers already in the pipeline.
A field case also depends on the product and the inputs. The prototype uses deionised water, and the available evidence does not show how it performs with other water sources. That leaves a practical deployment question for future testing, especially in places where water treatment itself is difficult. The same standard applies to crop trials: a measurable nitrate concentration is not yet proof of an agronomically useful fertilizer supply.
The prediction is testable. If independent groups reproduce the output and a crop trial shows useful plant response, the logistics case becomes more credible, even without energy parity. If either result fails, the argument for remote deployment weakens. The protocol matters because it makes those tests possible; the £20 estimate cannot substitute for them.
The Glasgow device has not displaced the large Haber-Bosch plant on energy use or scale. It has demonstrated nitrate production from air and water in a laboratory sonoreactor, with results published for other researchers to examine and attempt to reproduce. Roughly half the nitrogen in the human body has passed through Haber-Bosch. The unresolved question is whether some fertilizer can be made reliably nearer to the farms that need it.
