The brine contained 75 milligrams per litre of lithium and 35 milligrams per litre of dissolved silicon, at a pH of 7.2 and an electrical conductivity of about 20 millisiemens. Diatoms were an obvious candidate for the job: these single-celled algae are among the most productive photosynthetic organisms on Earth, and they absolutely require dissolved silicon, which they take up as silicic acid to build their ornate glass-like cell walls, called frustules.

After ten days of cultivation in the brine, the species differences were striking. The diatom Halamphora sp. removed 93 percent of the dissolved silicon, the best performance of any strain tested. Conticribra weissflogii managed 81 percent, while Fistulifera pelliculosa, the strongest biomass producer at 2 grams dry cell weight per litre, removed only 29 percent. Thalassiosira pseudonana managed just 6 percent, and two local isolates from Cornwall, one from Swanpool and one from Kennall Vale, removed 34 and 32 percent respectively. Crucially, lithium concentrations in the brine stayed essentially unchanged throughout the experiments, meaning the metal remained available for downstream extraction.

The researchers took care to rule out abiotic explanations. Control flasks containing brine without any diatoms showed no significant change in silicon concentration over the incubation period, and comparisons between untreated, filter-sterilised and autoclaved brines demonstrated minimal abiotic silicon precipitation. Because the experiments were run in glass vessels, the team also checked for silicon leaching from the glass and found it negligible. The observed silicon depletion was therefore attributed primarily to biological uptake by the diatoms, which incorporate the element into their silica frustules through a tightly regulated biomineralisation pathway involving specialised intracellular organelles called silica deposition vesicles.

Selectivity matters as much as removal. Halamphora sp. achieved a silicon-over-lithium separation factor exceeding 1,000, compared with 473 for C. weissflogii and only 9 for T. pseudonana. Scanning electron microscope images showed that Halamphora sp. cells grown in the brine retained an intact morphology comparable to cells grown in standard algal medium, with no obvious frustule deformation or dissolution, suggesting the organism tolerates the saline, mineral-rich environment. The authors note, however, that microscopy alone cannot demonstrate internal metabolic homeostasis, and a complete silicon mass balance was beyond the scope of the study.

The next challenge was economics. Inoculating brine with large amounts of pre-grown biomass is expensive, so the team tested how the starting cell density affected performance. At a low inoculum of 0.05 grams dry cell weight per litre, only about 24 percent of silicon was removed over ten days, whereas an inoculum of 1 gram per litre achieved up to 97 percent removal. Because the goal was to maximise silicon uptake while minimising the biomass investment, the researchers chose the low inoculum for further work and looked for ways to boost the algae’s growth inside the brine itself.

The answer was nitrogen. The brine contained plentiful phosphorus but appeared nitrogen-limited, so the team supplemented cultures with urea, a cheap nitrogen source. They compared a single dose of 26 milligrams per litre at inoculation with a semi-continuous urea feeding regime, or SCUF, in which the same dose was added every 36 hours. Halamphora sp. consumed the urea at an average rate of 0.18 milligrams per litre per hour, remarkably similar in brine and in standard f/2 medium, indicating the brine supplies sufficient trace elements for basic metabolism. With SCUF, biomass reached 0.45 grams per litre after ten days, roughly nine times the starting inoculum and 1.5 times the yield from a single urea addition. The fed-batch regime also pushed the silicon-lithium separation factor above 1,000, compared with 891 for a single dose.

The effect on silicon removal was dramatic. Within 48 hours, cultures under SCUF removed up to 97 percent of dissolved silicon, dropping the concentration from about 34 milligrams per litre to 9 milligrams per litre. Without added urea, only 8 percent was removed in the same period, and a single initial urea dose achieved 54 percent. Lithium concentrations were unaffected by urea supplementation or by the increased biomass, confirming that the treatment preserves the metal for recovery. The researchers propose that sustained nitrogen flux supports the continuous biosynthesis of the organic matrix molecules, long-chain polyamines and silaffins, that catalyse the polycondensation of silicic acid into amorphous silica inside the cells.

The biomass itself turned out to be a valuable co-product. Fatty acid analysis showed that the highest lipid yield per litre, 71 milligrams per litre over ten days, came from brine cultures under SCUF, containing 16 percent lipid by dry weight. Interestingly, nitrogen-starved cultures without urea accumulated far more lipid per cell, up to 41 percent of dry weight, a well-documented stress response that steers the fatty acid profile toward compositions suited to biodiesel, rich in C16, C16:1 and C18:1 chains. Under nitrogen-replete SCUF conditions, the profile shifted toward long-chain polyunsaturated fatty acids: EPA content climbed from 10 to 47.5 milligrams per gram over ten days, and DHA reached 23.5 milligrams per gram, the highest values among all conditions tested. The authors frame the process as two-stage: a rapid 48-hour brine pre-treatment for silicon removal and lithium retention, followed by a longer biomass valorisation phase for omega-3 production.

A preliminary mass balance illustrates the scale of the opportunity and its limits. Treating 1,000 litres of brine would require 50 grams of dry biomass as inoculum and about 200 grams of urea, costing roughly £3 at agricultural grades, while producing 0.5 kilograms of dry biomass containing about 25 grams of EPA and 12 grams of DHA, with more than 99 percent of dissolved lithium retained in the liquid phase. A crude valuation of the raw lipid co-product yields about £1.10 per 1,000 litres, before any downstream processing. The authors stress that this figure excludes capital and operating costs such as brine cooling, mixing, illumination, harvesting and lipid extraction, and that trace metals largely remained in solution, raising regulatory and purity questions for biomass grown in mineral-rich geothermal fluids. The study also tested only a single brine under laboratory conditions, and lithium recovery was not directly evaluated through an actual direct lithium extraction process after pre-treatment. Scaling up, screening brines of varying chemistry, coupling the treatment to real extraction systems, and running techno-economic and life cycle assessments are the stated next steps. Even so, the concept of a biological pre-treatment that tackles silicon scaling, spares lithium and yields nutraceutical-grade omega-3s offers an unusually circular vision for geothermal resource processing.

Subject of Research: Diatom-based biological pre-treatment of geothermal brine for silicon removal, lithium retention and omega-3 fatty acid co-production

Article Title: Diatom-based treatment of geothermal brines to selectively remove silicon, retain lithium for downstream recovery, and produce omega-3 fatty acids

Article References: Rehmanji, M., Skeffington, A., Hudson-Edwards, K. A., & Newsome, L. (2026). Diatom-based treatment of geothermal brines to selectively remove silicon, retain lithium for downstream recovery, and produce omega-3 fatty acids. iScience, 29(11), Article 117965. https://doi.org/10.1016/j.isci.2026.117965

Image Credits: AI Generated

DOI: 10.1016/j.isci.2026.117965

Keywords: geothermal brine, diatoms, Halamphora sp, silicon removal, lithium extraction, omega-3 fatty acids, EPA, DHA, urea supplementation, biomineralisation, direct lithium extraction, circular bioeconomy