Lithium Brine Extraction Water Usage per Tonne, Compared

Two Argentine brine plants that published their numbers used 51.0 and 135.5 cubic metres (m³) of fresh water per tonne of battery-grade lithium carbonate in 2021, plus 537.4 and 319.6 m³ of brine that evaporated in the ponds (Díaz Paz et al., Heliyon, 2025). That is the first thing to know: a “water per tonne” figure can mean fresh water only, brine only, or both added together, and the answer can differ by a factor of 40 across the literature (about 50 m³ of fresh water against about 2,000 m³ of total water). This page puts every published figure next to its basis (per tonne of lithium carbonate equivalent, fresh water or brine), shows where the fresh water is used inside a plant, and gives you a calculator for your own tonnage.

Lithium carbonate equivalent (LCE) is the standard way to count lithium: one tonne of lithium metal equals about 5.323 tonnes of LCE, calculated from atomic masses. Brine is the salty groundwater pumped from a salt flat (a salar); it is sun-evaporated in ponds until the lithium is concentrated enough to process. A cubic metre (m³) is 1,000 litres, so 51 m³ per tonne is 51,000 litres.

How much water does lithium brine extraction use per tonne

There is no single number, because the studies count different things. The table lists the figures we could trace to a document, with the basis stated. All values are per tonne of lithium carbonate (Li₂CO₃) or LCE.

Chart

Published water figures per tonne of lithium carbonate or LCE

m³ per tonne; bases differ (see labels): fresh water, brine, total or lab test

The blue water footprint is the fresh water taken from rivers or aquifers that does not return to the basin; the grey part is the water needed to dilute pollutants, and in this study it covers only the workers’ camp effluent. The Heliyon paper also notes that the industry’s own sustainability reports and the academic estimates sit far apart, and that debating whether a figure is “high or low” needs local context, not only the number.

Albemarle gave the Danish investigators at Danwatch a figure of 400 litres of water per kilogram of lithium but would not explain how it was calculated, and Danwatch found estimates from other sources that differed by orders of magnitude (Danwatch, December 2019). If that figure refers to lithium metal, it would be roughly 75 m³ per tonne of LCE, but because the basis was not disclosed we do not plot it.

Fresh water, brine and total water: what each number counts

Three different quantities get called “water use”, and mixing them is the main reason published numbers disagree.

• Fresh water consumed. Water pumped from rivers or freshwater aquifers around the salar and used for washing, dilution and chemical preparation. This is the number that competes with towns, herders and wetlands. In the Argentine study it is 51.0 m³/t at Olaroz and 135.5 m³/t at Fénix.
• Brine consumed. Salty groundwater pumped into evaporation ponds, where the water leaves as vapour and the lithium stays. Mining companies in the study define “water” as fresh water only and treat brine as a saline liquid with no other use. Local communities see the brine and the fresh water as one connected system, so the authors report brine consumption separately: 537.4 m³/t at Olaroz and 319.6 m³/t at Fénix.
• Total water. Fresh water plus brine. Adding them gives about 588 m³/t at Olaroz and 455 m³/t at Fénix, calculated from the paper’s Table 1. Note that the order reverses: Olaroz uses more brine per tonne but less fresh water.

The paper’s authors argue brine matters because lowering the pressure in a brine aquifer can draw fresh water in from its edges, which would add fresh-water use that the simple footprint does not capture. The hydrology of these basins is not well understood, they say, which is why they treat brine as part of the picture.

Where the fresh water goes inside a brine plant

Evaporation is slow and uses no fresh water. In the two Argentine plants, the evaporative concentration phase typically spans 12 to 18 months, and fresh water is used in the finishing steps: first in “polishing”, where concentrated brine is conditioned for temperature and acidity so that residual calcium, magnesium, boron and arsenic come out of solution, then in precipitation, where fresh water is used to prepare the soda ash solution that turns dissolved lithium into lithium carbonate and to wash the product to higher purity (Díaz Paz et al., 2025). At Fénix, which also makes lithium chloride, fresh water is used to wash lithium off adsorbent columns as well. For the pond stage and the full process in more detail, see our guide to the lithium brine extraction process.

Infographic

Brine to lithium carbonate in 4 steps

1. 1Pump and evaporateBrine goes to ponds for 12 to 18 months; no fresh water, but the brine water evaporates.
2. 2PolishFresh water conditions the concentrate so calcium, magnesium, boron and arsenic drop out.
3. 3PrecipitateFresh water makes up the soda ash solution that turns lithium into lithium carbonate.
4. 4Wash and dryFresh water washes the product to higher purity before drying and packing.

About 90% of the water in the brine is lost to the air. A Reuters report on a University of Chile study notes that about 90% of the water is lost to the atmosphere in this method (Reuters via Kitco, 22 August 2024). The Argentine study applies the same reasoning to the plants’ waste stream, which is mostly depleted brine and wash water: it reports that roughly 90% of that wastewater evaporates and only about 10% is prone to infiltrate.

Olaroz and Fénix compared: two Argentine brine plants

These are the only two plants in the sources we reviewed where a researcher rebuilt water use per tonne from the operator’s own environmental impact report, so they make a useful side-by-side. Both make battery-grade lithium carbonate at 97.7% LCE purity from brine pumped from a salt flat, and both draw fresh water from rivers and aquifers around the flat.

Chart

Fresh water versus brine consumed per tonne of lithium carbonate

m³ per tonne of battery-grade Li₂CO₃, 2021 production cycle, two Argentine plants

Source for the table: Díaz Paz et al., Heliyon, 2025 (Table 1 and text). Fénix used 2.7 times more fresh water per tonne and 4 times more in absolute terms, even though its output was only 1.5 times higher. The authors link this to brine chemistry (the two brines carry different mixes of sodium, magnesium, potassium and sulphate), to production level and to process design, and note that Fénix has the extra lithium chloride step. Olaroz is run by Sales de Jujuy; Fénix by Minera del Altiplano, the Argentine unit of what became Arcadium Lithium, which the paper reports Rio Tinto has since acquired.

The paper also compares the fresh water with local supply: the fresh water used at Olaroz equals the consumption of 32,238 residents of the nearest town, Susques, which has fewer than 2,100 inhabitants, and Fénix’s use equals 141,047 residents of Antofagasta de la Sierra, also under 2,100. The authors note that Fénix’s higher footprint matters less locally because that basin has more blue water available.

Salar de Atacama: brine pumped, fresh water caps and ground effects

Chile’s Salar de Atacama is the other well-documented case. Danwatch reported in 2019 that the two operators, SQM and Albemarle, pump nearly 2,000 litres of brine per second, about 63 billion litres a year, and that they also use some fresh water (Danwatch, December 2019). In February 2022 SQM filed a project with Chile’s environmental evaluation service to limit fresh water pumped from four wells to 120 litres per second and to cut brine extraction by 822 litres per second by 2028 (Mining.com, 3 February 2022). To give scale, 120 litres per second running all year is about 3.8 million m³.

A life-cycle study of the Salar de Atacama operation for 2022 and 2023 measured scarcity-weighted impact rather than volumes: 442 m³ world-equivalent per tonne of lithium product (87% carbonate, 13% hydroxide) using the AWARE method, of which concentrated brine production accounts for 326. The final carbonate stage contributes 59.9 m³ world-equivalent per tonne but with minimal scarcity impact because desalinated water is used there; all indicators improved by 9% to 42% against 2020 to 2021 (Marinova et al., Journal of Cleaner Production, 2024). These are impact scores, not cubic metres, so they should not be put in the same column as the table above.

Physical effects have also been measured by satellite. A University of Chile study using radar data from 2020 to 2023 found part of the salt flat sinking at 1 to 2 centimetres a year, an area of roughly 8 by 5 kilometres where pumping is most intense, because brine is pumped faster than the aquifer recharges (Reuters via Kitco, 22 August 2024). A separate study in the Proceedings of the Royal Society B, using satellite imagery of mining ponds, linked a 10 to 12% fall in James and Andean flamingo numbers at the Atacama to the water extracted; SQM disagreed with parts of that study and said its own monitoring showed stable flamingo populations (Reuters via Mining.com, 19 May 2022).

▶ The 7 environmental impacts of mining, explained in clay. · 2:31

Does direct lithium extraction use less water

Direct lithium extraction (DLE) is a group of technologies, such as sorbents, membranes and ion exchange, that pull lithium out of brine without evaporation ponds. Developers say they return the spent brine underground. The water story is not simple. A Nature Reviews Earth & Environment review found that fresh-water consumption of DLE needs to be quantified urgently and that many DLE technologies might need larger freshwater volumes than current evaporative practice, which would limit them in arid areas (Vera et al., Nature Reviews Earth & Environment, 2023). The Argentine study cites later work by Halkes and colleagues as consistent with that view.

Operators report the opposite when they test their own brine. Chile’s state miner Enami said tests on Altoandinos brine used 36 m³ per tonne of LCE, which it described as 55 times less than evaporation ponds, with recovery of 92% against 42% for ponds, and 10 hectares of plant area for a 75,000 t/yr plant versus 1,020 hectares for ponds (Argus, June 2025). These are company laboratory results, not commercial operating data. Chile’s national lithium strategy, launched in April 2023, commits the country to move from traditional extraction to DLE and to protect at least 30% of salt lake ecosystems, per the same report.

A 2026 study listed on The Nature Conservancy’s Conservation Gateway modelled supply to 2040 and concluded that wider DLE adoption could substantially cut land conversion, freshwater consumption and greenhouse gas emissions compared with brine evaporation and hard rock (Kalies et al., Journal of Cleaner Production, 2026). It is a scenario model, so it shows what could happen under assumptions, not what operating plants have measured.

Lithium output of the main producing countries

Water per tonne matters more as volumes grow. The US Geological Survey estimates world lithium mine production at 290,000 tonnes of contained lithium in 2025, up from 222,000 in 2024, with Australia (hard rock), China, Chile, Zimbabwe and Argentina the five largest producers listed (USGS, Mineral Commodity Summaries 2026). USGS counts lithium content, not LCE, so multiply by 5.323 to compare with the LCE figures above. Chile’s output of 56,000 t in 2025 is therefore about 298,000 t LCE. The USGS notes that US production comes from a continental brine operation in Nevada, with the output figure withheld to protect company data. Chile and Argentina, whose salt flats supplied the water studies above, are both on the list; the USGS table does not split any country’s output between brine and rock.

Chart

Lithium mine production by country, 2025 estimate

Thousand tonnes of contained lithium (not LCE)

Estimate the water bill for a brine plant

The calculator below turns a per-tonne figure into annual volumes and flows. Enter your own production and pick fresh-water and brine intensities from the table above (the defaults are round examples, not a benchmark and not any company’s data). Compare the fresh water result with the permit your project holds or expects.

How to read a water-per-tonne claim

When a company, article or study gives you a number, ask six questions before using it.

1. Fresh water, brine or both? Fresh-water-only figures can be 10 times lower than totals that include evaporated brine.
2. Per tonne of what? Lithium carbonate, LCE, hydroxide, lithium metal and “product” are different bases. A per-kilogram-of-lithium figure multiplies by 5.323 to become per tonne LCE.
3. Consumed or withdrawn? Consumption is water that does not return to the basin. The Argentine study treats the plant’s wastewater as lost because about 90% of it evaporates.
4. Which year and which plant? A single plant-year (Olaroz and Fénix are both 2021) shows one operating point, not an industry average.
5. Who measured it? An operator’s report, a regulator’s environmental impact report and a peer-reviewed rebuild can give different numbers for the same plant.
6. What is excluded? The Argentine study left out indirect water in fuels, materials and construction, and could not estimate grey water from industrial effluent.

Screening a salar from space before fieldwork

Satellite data does not measure a plant’s water use directly, but it can show pond area, changes in surface water and vegetation around a salar, and ground movement; the studies above used satellite radar and optical imagery to detect sinking and pond water volume. SQM pointed out that satellite analysis can over- or under-estimate water use, which is why on-the-ground monitoring is still needed. Farmonaut offers satellite-based mineral detection that screens a licence area, including lithium targets, before fieldwork; results are exploration targets and not a resource estimate or an environmental clearance. Consultants can send a boundary through the mining query form or contact us. For the geology behind brine, rock and clay lithium, see lithium deposits explained, and for a salar outside the Argentina-Chile pair, Bolivia’s lithium reserves.

▶ How satellites find lithium, explained simply. · 10:49

Frequently asked questions

How much water does it take to produce one tonne of lithium from brine?

Measured fresh-water use at two Argentine plants was 51.0 and 135.5 m³ per tonne of lithium carbonate in 2021, plus 537.4 and 319.6 m³ of brine (Díaz Paz et al., Heliyon, 2025). Earlier studies cited there give about 5 to 50 m³ of fresh water and about 500 m³ during evaporation; one estimate puts the total near 2,000 m³.

Is brine counted as water in lithium statistics?

Usually not. Operators in the Argentine study define water as fresh water suitable for industrial use and brine as a saline liquid. The authors argue both should be reported because the two are hydrologically linked, so check which one a figure covers.

Why do estimates range from 50 to 2,000 m³ per tonne?

Mostly because of what is counted. The low figures are fresh water only; the high ones include evaporated brine. Brine chemistry, lithium grade, process design and the year of production also differ between plants. At the two Argentine plants, with the same product purity, fresh-water use differed by 2.7 times.

How much water does one brine plant use per day?

Olaroz used 643,439 m³ of fresh water in 2021, about 1,760 m³ a day on average, and Fénix 2,574,107 m³, about 7,050 m³ a day (calculated from Díaz Paz et al., 2025). Their brine consumption was 6.8 and 6.1 million m³ a year.

What does LCE mean?

LCE is lithium carbonate equivalent, the tonnage of lithium carbonate that would contain the same lithium. One tonne of lithium metal is about 5.323 tonnes of LCE. Always check whether a figure is per tonne of lithium or per tonne of LCE.

Does direct lithium extraction use less water than evaporation ponds?

Reports conflict. Enami’s laboratory tests gave 36 m³ per tonne of LCE, and a 2026 scenario model found DLE could reduce fresh-water use. A 2023 peer-reviewed review warned that many DLE technologies might need more fresh water than ponds, and that their consumption needs to be quantified.

Has brine pumping affected the Salar de Atacama?

A University of Chile study found a zone of about 8 by 5 km sinking at 1 to 2 cm a year where pumping is most intense, and a Royal Society B study linked lower pond water to a 10 to 12% fall in two flamingo species. SQM disputed the flamingo findings.

Why does Chile want to move to direct lithium extraction?

Its national lithium strategy, launched in April 2023, aims to raise output while reducing impact on salt lake ecosystems, and commits to a shift from traditional methods to DLE and to protecting at least 30% of salt lake ecosystems (Argus, June 2025).

Can satellites measure how much water a lithium plant uses?

Not directly. Researchers have used satellite imagery to estimate pond water and radar to measure ground sinking, but SQM has said satellite analysis can over- or under-estimate water use. Metered flows and well data remain the reference.

What unit should I use to compare projects?

Use cubic metres of fresh water per tonne of LCE, and give brine separately. State the year and whether the figure is withdrawal or consumption.