MIT's bismuth-electrode seawater magnesium looks cheap on paper, and unproven everywhere else

Researchers at MIT have reported an electrochemical cell that extracts magnesium directly from raw seawater using custom bismuth electrodes, in a study published in ACS Energy Letters ^[1][2]. It matters because magnesium is a critical mineral that runs from concrete into lightweight alloys, and the report states that most supply is dominated by China ^[5][7].

The engineering problem is selectivity, not abundance. Magnesium is the eighth most abundant element in the Earth's crust and the third most plentiful element dissolved in seawater ^[6], but seawater is dominated by sodium, and separating the two has typically required complex filter membranes that clog easily and push up operating costs ^[8]. T. Alan Hatton and Kripa Varanasi built a layered cell that, according to the report, avoids both costly polymer membranes and harsh chemical precipitants ^[3][9]. Micro-channels of seawater are lined with thin bismuth electrodes; a precise electric field temporarily shifts local acidity, which forces magnesium ions out as solid magnesium hydroxide, with real seawater in one channel and an electrolyte in the other ^[10]. Reversing the polarity and adjusting fluid flow redissolves that precipitate into liquid magnesium chloride, and each pass adds yield ^[11].

The reported results: iterative cycles produced an eightfold increase in concentration and a 20-to-1 ratio of magnesium to sodium ions, which the researchers say outperforms every rival electrochemical system on record ^[12]. The headline economics are an estimated baseline cost of $107 per ton, described as a fraction of prevailing market prices ^[13].

Three things about that number deserve flagging. It excludes post-extraction drying costs, which the source acknowledges ^[14]. It is quoted per ton of magnesium chloride, the product the cell actually delivers, and no conversion to a cost per ton of magnesium metal is given ^[21]. And the comparison to market prices is asserted without stating the market price being compared against, so the size of the gap cannot be checked ^[20]. Beyond cost, concentration factor, and ion ratio, the account carries no throughput per unit of electrode area, no electrode lifetime across cycles, no current efficiency, and no cell size ^[19]. Those are precisely the numbers that decide whether a lab cell becomes a plant.

The demand case is the least speculative part. Magnesium is the lightest structural metal, and alloyed with aluminium it goes into car parts, aircraft bodies, and high-performance bicycle frames ^[16]. Demand is rising across healthcare and construction at once, with supplements competing against industrial alloys for the same supply ^[17]. The incumbent routes are unattractive: pulverising ore and firing energy-intensive kilns, or dosing ocean brine with large volumes of chemical additives ^[15]. David Kim, the study's first author, framed the shift as treating seawater as a reservoir rather than only something to desalinate and protect ^[4][18].

For anyone buying magnesium or specifying it into hardware, nothing changes this year. What to watch is whether the group publishes a cost that includes drying, a throughput figure that can be multiplied out to tons per day, and bismuth electrode life over enough cycles to matter. Also worth watching is whether the 20-to-1 magnesium-to-sodium ratio holds on unfiltered intake water over long runs rather than in cycled laboratory passes ^[12].