The world's largest lithium reserves aren't in Chile's salt flats. They're underground, dissolved in brines that conventional extraction can't economically touch — too dilute, too contaminated with other minerals, too expensive to pump to the surface and wait years for evaporation to concentrate. A team at Columbia Engineering has built a process that changes that arithmetic.
What Conventional Extraction Actually Costs
Lithium is not scarce. The planet holds enough to power the entire global vehicle fleet for centuries. The problem is that most of it sits in forms that current extraction technology handles badly. The dominant method today is evaporation: pump brine from underground or surface salt flats into enormous ponds, wait 12 to 24 months for the sun to do the work, then process the concentrated residue. That works reasonably well in the Atacama Desert, where conditions are nearly optimal. It works poorly almost everywhere else.
Evaporation ponds consume water in landscapes where water is scarce. They destroy habitat. They take years to cycle, which means that a new lithium deposit can't deliver supply quickly enough to match demand that is doubling every few years. They also fail entirely on low-grade brines, where lithium concentrations are too low to concentrate economically. A significant fraction of known lithium deposits have never been developed for exactly this reason.
Direct lithium extraction has been the field's aspirational answer for a decade. The idea is simple: pull lithium out of the brine immediately, without waiting for evaporation, using chemistry selective enough to ignore the sodium and potassium ions that make up most of what's dissolved. The hard part is the selectivity. Sodium and potassium are chemically similar to lithium. Getting a molecule that prefers lithium by a wide enough margin to be industrially useful has been, until now, harder than it sounds.
How the Switchable Solvent Works
The process developed by Ngai Yin Yip and colleagues at Columbia Engineering is called S3E: Switchable Solvent Selective Extraction. It relies on a class of solvents with an unusual thermal property. At room temperature, these solvents are hydrophilic: they mix freely with water and, by extension, with brine. When heated to 70°C (158°F), they become hydrophobic. They stop mixing with water and instead form a separate phase that water and dissolved salts can't enter.
That phase transition is the engine of the process. At room temperature, the solvent mixes into the brine and selectively binds lithium ions. At 70°C, the solvent and the lithium it's carrying separate from the brine into their own phase. Cool the solvent phase back down, strip out the lithium, and the solvent reverts to its room-temperature form. Repeat.
Lithium, sodium, and potassium ions are all positively charged and all dissolve readily in water. They differ mainly in size: lithium is the smallest, sodium is intermediate, potassium is the largest. A selective extractant has to exploit those size differences with enough precision to extract lithium at a useful rate while leaving sodium and potassium behind. S3E achieves 10-times selectivity over sodium and 12-times over potassium — ratios that are commercially meaningful rather than merely academically interesting.
The S3E flow looks like this:
Figure 1 — S3E (Switchable Solvent Selective Extraction) process cycle
The energy cost of heating the solvent to 70°C is modest by industrial standards. For comparison, the aluminum smelting that produces the metal in a battery's casing requires roughly 800°C. The solvent reuse loop is the other important efficiency feature: unlike reagents that are consumed in a reaction, the switchable solvent recovers its original state after each cycle and goes back in. At industrial scale, this matters a great deal for operating costs.
The work was published in Joule on January 21, 2026.
What the Numbers Say About What Changes
The specific numbers in the Joule paper are laboratory-scale results. A laboratory demonstration is not a commercial plant. The gap between the two is where most promising extraction technologies have stalled out over the past decade, consumed by engineering challenges that don't show up in a beaker.
But there are reasons this particular approach deserves to be taken seriously at the level of system architecture, not just chemistry. First, the 10x and 12x selectivity ratios represent genuine margins. A lithium brine typically contains roughly 100 times more sodium than lithium by mass. To extract economically, you need selectivity that can handle that ratio without requiring multiple purification passes. S3E's published selectivity is in the range where industrial process engineers start to see a realistic flowsheet.
Second, the process works on low-grade brines. This is the number that changes the resource map. Conventional evaporation extraction is economic only above certain lithium concentrations. S3E's ability to function on dilute brines, not just the high-grade deposits currently mined, is what the research team at Columbia points to as the genuine expansion of the resource base. There are enormous brine deposits around the world — in the American West, in South America outside the Atacama, in geothermal brines in Europe — that current technology can't economically access. A process that can reach them is addressing a different problem from a faster version of what already exists.
The United States Geological Survey estimates global lithium reserves at around 28 million tonnes — but identified resources (deposits known to exist, not all economic to extract) are closer to 98 million tonnes. The gap between reserves and resources is largely a function of what current extraction technology can reach. A method that works on low-grade brines moves some of that resource column into the reserve column.
Between a Paper and a Battery
The honest accounting of where S3E sits right now: it's a proof of concept with strong numbers, published in a rigorous journal, with a mechanism that makes physical sense and selectivity ratios that look commercially interesting. It has not been demonstrated at pilot scale. It has not been optimised for the specific brine chemistries of individual deposits, which vary considerably. It has not been costed at industrial throughput. None of that disqualifies it. That's what the development pipeline between academic research and commercial deployment is for.
The lithium industry is paying attention to a field that a decade ago was mostly theoretical. Several companies are now operating pilot-scale direct lithium extraction plants using different approaches, primarily ion-exchange and electrochemical methods. S3E would compete in that space, with the potential advantage of simpler chemistry: no membranes prone to fouling, no electrodes to replace, just a thermally switchable solvent in a cycle.
Battery demand projections for 2030 suggest lithium supply will need to roughly triple from current levels. Evaporation ponds can't get there in time. The ramp requires methods that can be deployed quickly, in diverse geographies, without the years-long lead times that surface evaporation demands. That's the market S3E is aimed at. Whether it gets there depends on engineering work that hasn't been done yet. The chemistry, at least, is now proven.
Buy me a coffee