Salt Lake Research, a Chinese academic journal devoted to the chemistry of the planet’s saltiest water, put an unusual cover on its latest issue, volume 34, number 3: not a photograph of a playa or an industrial plant, but a micrograph of a polymer film tinted in the pale ochre of aluminum chloride, produced by the laboratory of Meng Xiaorong at Xi’an University of Architecture and Technology, working with a Shaanxi coalfield geology company.
It is a new material for lithium-magnesium separation, the single hardest step in recovering lithium from brine. By the paper’s own introduction, nearly 80% of the world’s lithium resources sit in seawater and salt-lake brine. Magnesium and lithium are chemical cousins that occur together in nearly every salt lake on Earth, and pulling them apart cleanly decides whether a brine deposit becomes a battery supply chain or an expensive pool of undrinkable water.
The paper describes an electrically enhanced membrane that pushes lithium ions across a selective film while magnesium ions stay behind. In the optimized formulation, the film lost 0.4% of its lithium flux over 120 hours of continuous laboratory cycling. It is one piece of a much larger push inside China, where researchers and companies are attacking the magnesium problem from every available angle at once.
Why Lithium and Magnesium Are So Hard to Separate
Lithium sits below sodium in the periodic table and diagonally across from magnesium. That diagonal pairing is not a coincidence of layout. The two ions carry a similar ratio of charge to size, which gives them strikingly similar chemical habits: both combine directly with nitrogen when heated, and both have carbonates and hydroxides that break down rather than melting cleanly.
The two ions are nearly the same size, which is exactly the difficulty. Magnesium’s bare ion is slightly larger, 1.42 angstroms across against lithium’s 1.36, but it carries twice the charge, so it grips water almost four times as tightly, with a hydration energy of about -1,828 kilojoules per mole against lithium’s -475. Breaking that grip is the energy cost every separation has to pay. Membranes sort the ions by the size of their hydrated shells, while extractant molecules sort them by how strongly each ion binds.
Reactivity adds a second complication. Lithium metal is the more reactive of the two, with a standard reduction potential of about -3.04 volts against magnesium’s -2.37 volts, which is why lithium metal is produced by electrolyzing molten salt rather than aqueous solution.
Any chemistry strong enough to grab magnesium aggressively can end up grabbing lithium too, and every lost lithium ion is lost revenue.
The practical measure of the problem is the magnesium-to-lithium ratio in the feed brine. Selective recovery generally becomes difficult above a ratio of about 6 to 1. Many plateau brines in Qinghai run at 15 to 50 to 1, and the brine at Qarhan Salt Lake, the largest in China, has been reported at 500 to 1 before treatment.
Conventional solar evaporation recovers only 30% to 50% of the lithium in such feedstocks and leaves the rest locked in magnesium-rich waste bitters.
Inside the New Electrically Enhanced Membrane
The cover article builds what the authors call a polymer inclusion membrane. They synthesize a polymer-bound lithium extractant, polyvinyl chloride grafted with diethyl phosphonate groups (PVCTE), through the Arbuzov reaction. A second component deserves a note. D2EHPA does not participate directly in carrying lithium. It grips iron(III) so tightly, across both acidic and alkaline conditions, that the iron stays put inside the film instead of leaching out with the feed. Holding the iron in place holds the real carrier in place: the tetrachloroferrate anion, FeCl4–, which coordinates lithium preferentially over magnesium. Blended together and cast into a film, the two components become the selective layer.
The separation does not rest on the iron salt alone. Density functional theory calculations in the paper show the film’s phosphonate groups bind the magnesium complex slightly more strongly than the lithium one, which is the opposite of what selectivity would require. The deciding factor is the water: magnesium holds its hydration shell so tightly that stripping it costs far more energy than lithium pays, and the larger hydrated ion is harder to shove through the polymer matrix anyway.
Coupling the film to an external electric field turns the setup into an electric membrane extraction cell.
The team varied membrane composition, applied voltage, and the pH of both solutions, then optimized the formulation. One result matters more for real-world use than any of the headline numbers: the system worked well with neutral solutions on both sides of the film. Actual salt-lake brine sits close to neutral, so a membrane that only performs in strongly acidic feed has little future outside the laboratory.
At low voltages, lithium was pulled out of the feed but did not make it across the film at all. Real transport only began around 5 volts. From there to 30 volts, lithium ions crossed faster than magnesium ions, and lithium flux climbed from 0.54 to 3.42 x 10-2 moles per square meter per hour while the separation factor reached 9.73 at the top of that range.
The optimized membrane, containing 40% of the iron-bearing carrier and processing a mixed lithium and magnesium feed at 0.01 moles per liter, delivered an initial lithium flux of 2.952 x 10-2 moles per square meter per hour and a lithium-to-magnesium separation factor of 8.35 at 25 volts.
Above 25 volts, chlorine gas begins appearing at the anode. Push past 30 volts, and the current climbs beyond 0.05 amps, the system shows obvious electrolysis at both electrodes, and the soft film becomes prone to electrical breakdown and leaks. The voltage that drives the separation is also, at the limit, decomposing the brine.
The stability result is the part that matters most for anyone who has followed membrane research. At 25 volts, after five consecutive 24-hour mass transfer cycles, the lithium flux had fallen by less than 0.4%. The film lost 3.5% of its mass in the first cycle, as loose D2EHPA dissolved away, then held nearly constant at 3.8% by the fifth cycle. No iron was detected in either solution, and the operating current stayed near 0.04 amps throughout, with no visible gas bubbles at the electrodes.
Polymer inclusion membranes are notorious for swelling and bleeding extractant into the feed, and many promising laboratory films lose their selectivity within days. A film that holds its flux through five days of continuous duty is a genuine materials advance.
Whether it can run in a commercial plant is a separate question. The authors present the result as guidance for making stable lithium-selective membrane materials, which is a materials science claim rather than a guarantee of industrial performance.
A separation factor of 8.35 looks modest beside the triple and four-digit selectivities reported for some experimental nanofiltration membranes, and the test feed was far more dilute than real plateau brine.
The paper’s own comparison table sets the result against five other membrane-extraction systems reported in the literature. Feed conditions differ widely between the studies, from 0.01 to 3.0 moles per liter of magnesium, so the flux values are indicative rather than directly comparable.
| System | Lithium flux (10-2 mol·m-2·h-1) | Li/Mg factor | Stability |
|---|---|---|---|
| TBP/FeCl4-PIM | 0.08 | 176 | Lost 40% in 72 h |
| P507-TBP-PIM | 0.48 | 10.2 | Lost 15.4% in 168 h |
| TBP-ionic liquid PIM | 0.32 | 2.2 | Lost 6.4% in 60 h |
| TBP-NaBPh4-PIM | 1.73 | 70 | Lost 14.2% in 72 h |
| TBP/NaFeCl4/N503-EME at 40 V | 13 | 0.98 | Lost 7.4% in 72 h |
| PVCTE-DFe-EME at 25 V (this study) | 2.95 | 8.35 | Lost 0.4% in 120 h |
The pattern is consistent. The systems that beat it on selectivity lost between 6% and 40% of their performance within two to three days, and the fastest lithium mover of the group did not separate the ions at all. Trading some selectivity for a film that survives may be the better bargain.
Beyond durability, the film sidesteps the reagent losses of solvent extraction. The carriers stay locked in the polymer instead of being dosed fresh with every batch of brine, which is where much of the cost and waste of liquid extraction comes from.
Five Routes for Splitting Brine, and Where Each One Breaks
There is no single way to remove magnesium from brine. The industry now runs a portfolio of approaches, each with a different failure mode, and the newest Chinese work sits at the membrane and electrochemical end of that spectrum.
| Route | How it works | Reported lithium outcome | Main limitation |
|---|---|---|---|
| Solar evaporation ponds | Sun concentrates brine in stages while salts crystallize out | 30% to 50% recovery over many months | Huge land footprint; lithium lost to waste bitters |
| Adsorption | Selective ion-sieve sorbents capture lithium, then release it with acid | About 82% recovery reported at Qarhan plant scale | Sorbent dissolution, long cycles, high water use |
| Solvent extraction | Organic ligands shuttle magnesium out of the brine | High throughput at industrial scale | Large reagent volumes and phase separation problems |
| Membrane separation | Nanofiltration, electrodialysis and inclusion membranes sort hydrated ions | Separation factors from about 148 to over 26,000 reported | Fouling and long-term film stability |
| Electrochemical | Electric fields precipitate magnesium, then crystallize lithium carbonate | About 83% recovery with 99.6% pure product | Electrode cost is the main bottleneck |
The recovery numbers tell the strategic story. Solar ponds, still the workhorse, throw away roughly half the lithium in the ground. The emerging routes consistently report recovery above 80%, and several of them also produce a cleaner final product.
Each route is a different answer to the same question: how to deliver selective lithium magnesium separation from water that is mostly magnesium. In practice, operators tend to match the method to the brine they actually have.
- Solar evaporation: still the cheapest option where land is flat, dry, and sunny, and where the brine already carries a manageable magnesium load.
- Adsorption: best suited to brines too rich in magnesium for ponds, since lithium-selective sorbents ignore magnesium ions almost completely.
- Membranes: shine at moderate magnesium-to-lithium ratios, provided pretreatment keeps fouling under control.
- Solvent extraction: handles high-magnesium feeds at volume but draws scrutiny for reagent loss and the wastewater it creates.
- Electrochemical: the most adaptable to changing brine chemistry, held back mainly by electrode cost and power demand.
China’s Salt Lake Push Moves From Lab to Plant
The Salt Lake Research cover paper is one node in a dense network of Chinese work on the same problem, and the industrial half of that network is now visible in Qinghai province. Operators at Qarhan Salt Lake report that adsorption coupled with membrane separation has cut the magnesium-to-lithium ratio of their feed from 500 to 1 down to 6 to 1.
A 40,000-ton facility there, described as the largest single-unit salt-lake lithium extraction plant of its kind, combines a continuous ion-exchange moving bed with membrane coupling and reports a comprehensive lithium recovery of 82.4% alongside a roughly 51% cut in energy use.
State-backed projects are moving in the same direction.
A raw-brine processing route developed by Minmetals Salt Lake and Xi’an Lanshen passed provincial expert review in May 2026, with the reviewing panel, chaired by Chinese Academy of Engineering academician Zheng Mianping, endorsing a direct path from untreated brine to battery-grade lithium carbonate at more than double the recovery of conventional methods.
The academic pipeline is producing results at comparable speed. Researchers at the University of Science and Technology of China developed what they call ion distillation, stacking ion-selective membranes into stages that behave like the trays of a distillation column.
Applied to brine from Dongtaijinar Salt Lake with a magnesium-to-lithium ratio above 35, the process lifted selectivity from about 30 in a single stage to 1,104 in two stages and 26,177 in four, yielding lithium products at 99.69% and 99.98% purity as described in the university’s announcement.
A team at the Qingdao Institute of Bioenergy and Bioprocess Technology took a materials route, publishing a deep-dehydration strategy for vermiculite membranes. Heating the clay to 600 degrees Celsius with cesium ions inserted between its layers locks the structure against swelling. In tests on real waste brine from a Qinghai salt lake, the membrane cut the magnesium-to-lithium ratio from 15.1 to 0.5, raised the lithium share of the concentrate from 6.1% to 45.4%, and delivered industrial-grade lithium carbonate at 99.2% purity over more than ten days of continuous operation.
Engineering firms are filing at the same pace. A Chinese patent published in May 2026 describes a closed-loop process that circulates ammonia and carbon dioxide to precipitate magnesium selectively as carbonate while lithium stays in solution, reporting lithium recovery above 90% and a residual magnesium-to-lithium mass ratio below 0.01 in the treated brine.
Magnesium Is Not Waste, It Is the Co-Product
For most of the history of brine mining, magnesium was the contaminant to be discarded. That framing is shifting, and it changes the economics of the entire separation step.
An approach developed by researchers in Australia and China, published in Nature Sustainability, uses EDTA to chelate magnesium during nanofiltration, widening the effective size and charge gap between the two ions. The process recovered about 90% of lithium across two filtration stages, recycled close to 100% of the chelating agent, and converted the removed magnesium into nano-scale magnesium hydroxide, a feedstock used in flame retardants and wastewater treatment.
Modeled on brine from Longmuco salt lake, the combined value of lithium carbonate and magnesium hydroxide came to three to four times the profit of the conventional route.
Other teams capture magnesium as a sellable solid rather than a sludge. A reaction-coupled method traps magnesium ions in magnesium-aluminum layered double hydroxide, a high-value industrial material, while keeping lithium losses below 8%.
The same logic appears in the electrochemical route described in a 2026 study in Environmental Science & Technology, which precipitated magnesium hydroxide without adding chemicals and finished with 82.99% of the lithium recovered as lithium carbonate at 99.6% purity.
Selling the magnesium does more than add revenue. It offsets the cost of the separation itself, reduces the volume of waste bitters that must be stored, and shrinks the land and water footprint compared with evaporation ponds that spread brine across the plateau for a year or more.
What This Means Beyond China
China’s plateau brines are unusually difficult, but the underlying problem is global. Salt-lake brine accounts for about 60% of world lithium reserves, and a high magnesium-to-lithium ratio is the norm rather than the exception across the Atacama in Chile and the wider lithium triangle of Chile, Argentina, and Bolivia.
The same chemistry reappears in geothermal brines in Europe and the United States, in oilfield wastewater, and in the lithium dissolved in seawater, where concentrations are so low that any separation cost multiplies.
Desalination projects already generate mineral-rich leftovers. In Rizhao, a Chinese coastal city, an ocean refinery concept turns seawater into drinking water, hydrogen, and a brine still loaded with critical minerals, which puts the same separation problem in a different pipeline.
That is why improvements travel. A membrane that tolerates weeks of continuous duty, a chelating agent that recycles, or a process that converts the magnesium stream into a product can be adapted to brine anywhere, as our earlier look at how brine mining can secure lithium, magnesium, and freshwater at once set out. The broader shift toward direct lithium extraction and new global discoveries depends on exactly this kind of chemistry landing in commercial plants.
There is an energy angle too. Electrically driven separation only delivers on its environmental promise when the electricity is clean. The Qinghai plateau has abundant solar and wind resources, which is part of why electric membrane and electrochemical routes are being developed there first.
What is notable about the Chinese effort is its coordination. One bottleneck is being pursued simultaneously through membranes, adsorbents, electrochemistry, and closed-loop precipitation, backed by provincial science reviews, patents, and operating plants. The electrically enhanced membrane on that journal cover is a small part of that effort, and the effort itself is the story.
Frequently Asked Questions
Does lithium show a diagonal relationship with magnesium?
Yes. In the periodic table, lithium sits diagonally below beryllium and across from magnesium, and the lithium-magnesium pairing is the classic example of a diagonal relationship. Because the two ions have similar charge-to-size ratios, they show similar chemical behavior, including the tendency to form nitrides directly and to have carbonates and hydroxides that decompose on heating. That resemblance is precisely what makes separating them from brine so difficult.
Why is the magnesium-to-lithium ratio important in brine?
The ratio is the best single predictor of whether a brine deposit is economical to develop. Selective lithium recovery generally becomes difficult above about 6 to 1, and many operating salt lakes run far higher, up to 500 to 1 before treatment. Every unit of magnesium means more reagent, more energy, and more processing stages for the same kilogram of lithium, so most of the innovation in the field is really an effort to lower that ratio cheaply.
How is lithium extracted from salt lake brine?
The classic method pumps brine into shallow solar ponds and lets evaporation concentrate it over many months, which recovers only 30% to 50% of the lithium. Direct lithium extraction shortcuts that process using adsorbents, membranes, or electrochemical cells that capture lithium selectively, after which the concentrated stream is converted to battery-grade lithium carbonate or lithium hydroxide.
China’s newest plants combine these steps, using ion exchange and membrane stages to take raw brine directly to battery-grade product.
Which is more reactive, lithium or magnesium?
Lithium is the more reactive metal. Its standard reduction potential of about -3.04 volts sits well below magnesium’s -2.37 volts, so lithium reacts vigorously with water and with air, while magnesium reacts slowly with cold water and burns with an intense white flame. That high reactivity is also why lithium metal is produced by electrolyzing molten salts rather than aqueous solutions and why gentle, selective chemistry is preferred inside the brine processing plant.
What happens to the magnesium removed from brine?
Historically it was discarded as magnesium-rich waste bitters or magnesia sludge. Newer processes turn it into products: nano-scale magnesium hydroxide for flame retardants and water treatment, magnesium-aluminum layered double hydroxides for industrial catalysis and materials, and magnesium carbonate. Recovering a saleable magnesium product helps pay for the separation step itself.
The Bottom Line
Lithium-magnesium separation decides how much of the world’s brine lithium can actually reach a battery. The chemistry is stubborn because the two ions are near twins, and the industry has responded with a spread of routes whose recovery rates range from under half in solar ponds to about 90% in the newest membrane and electrochemical systems.
The Chinese work highlighted here shows the direction of travel: membranes built for durability rather than record selectivity, electric fields replacing reagents, waste magnesium reframed as a co-product, and laboratory results moving into provincially reviewed industrial projects.
None of these approaches alone solves the magnesium problem. Together they make brine resources that were once considered unworkable into a practical part of the clean energy supply chain.
The developments worth watching over the next two years are pilot plants running real plateau brine through electrically driven membranes for months rather than days, published life-cycle numbers comparing pond brine with direct extraction, and whether the magnesium co-product markets grow quickly enough to matter. If they do, the salt lakes of Qinghai, the Atacama and every other high-magnesium brine basin become far more valuable than they look today.
