A lithium-ion cell has a failure mode that almost nobody models, and it only shows up when you charge fast. Lithium ions do not travel alone through the electrolyte — each one drags several solvent molecules with it, from the positive electrode to the negative. Charge slowly and the solvent flows back into the porous cathode, refilling it. Charge quickly and it cannot keep up: the electrode dries out, and the current you were pushing in stops going in. Researchers at the University of Maryland have now named the problem — electro-osmotic drag — and attacked it by redesigning the solvent rather than the electrode. In a 4 Ah pouch cell they reached 80% in 12.8 minutes and 90% in 14.7, inside the US Department of Energy’s target of 15 minutes to 80%, and published the result in Science.
What Electro-Osmotic Drag Does to a Cell
The mechanism is not new to science. Fuel-cell engineers have worked around electro-osmotic drag for two decades: in a polymer membrane, water is dragged along with protons, drying out the anode side. It had simply never been applied to lithium-ion, where the electrolyte is a liquid in a porous electrode rather than a hydrated membrane.
What changes at high current is the balance of two flows. Solvent is carried out of the cathode by the lithium flux; solvent diffuses back in by concentration gradient. At low rates the second flow wins and the electrode stays wet. At high rates the first flow dominates, the pores empty, and the cell can no longer deliver the current its voltage suggests.
That matters because the industry has spent a decade making electrodes thinner, more porous and more structured to shorten the ion path. Every one of those changes makes the drying problem worse, not better.
Why Every Previous Fix Cost Energy Density
The standard playbook for fast charging — thinner electrodes, laser-ablated channels, redesigned tabs, more elaborate cooling plates — buys power by giving up something. Thinner electrodes mean more inactive material per unit of energy. That trade is fine for a power tool and awkward for a car, where range is the thing buyers check first.
This is why “fast charging” announcements and “energy density” announcements rarely arrive together, and why a 10-minute charge usually comes with an asterisk: it applies to a specific cell at a specific temperature from a specific state of charge.
The Maryland result is interesting precisely because it does not make that trade. It changes what the solvent is, not how much of it there is.
What They Actually Changed
Led by Chang-Xin Zhao in the lab of Distinguished University Professor Chunsheng Wang, the team added fluorine-containing chemical groups to the solvent molecules in the electrolyte. Fluorination weakens the attraction between the solvent and the lithium ion, so the solvent is less tightly bound to the ion train and flows back into the cathode more readily.
Electrolyte engineering is not a new idea in this industry — it is the same lever behind additive work in Japan and much of the incremental progress in cycle life. What is new here is the target: not the solid-electrolyte interphase, not salt dissociation, but the solvent’s own mobility under an electric field.
The results, from the university’s own account: 80% in 12.8 minutes, 90% in 14.7 minutes, against a Department of Energy goal of 15 minutes to 80% and a current baseline of 20 to 60 minutes on DC fast chargers. The researchers note that reaching a full 100% quickly remains hard without compromising other performance.
What the Number Is Worth, and What It Is Not
Here is where the discipline matters, because a 12.8-minute result sounds faster than it is.
| Measure | This result | Context |
|---|---|---|
| Cell size | 4 Ah pouch, roughly 15 Wh at nominal voltage | A 200 Ah automotive prismatic is about 740 Wh — some 50 times larger |
| 80% charge time | 12.8 min | CATL’s third-generation Shenxing claims 10–80% in 3 min 44 s |
| Cycles disclosed | none | The release gives no cycle-life figure |
| Test conditions | not stated | No C-rate, temperature or starting state of charge published |
Two of those lines are the important ones. Electrode drying gets harder, not easier, as cells get bigger: a larger electrode has longer path lengths for the return flow and a worse thermal and current distribution. A 4 Ah pouch is the right size to demonstrate a mechanism and the wrong size to demonstrate a product.
And a 4 Ah cell stores about 15 Wh. A 60 kWh pack built from cells that size would need roughly 4,000 of them; built from 740 Wh automotive cells it needs about 80. Nobody is going to assemble a car from four thousand pouches, so the number that matters is whether the effect survives at 50 times the electrode area.
On raw speed, this is not a record. The state of the art in production, from CATL and from BYD’s flash-charging network, is faster. What Maryland has is something the faster numbers do not come with: a physical explanation for why fast charging stops working, and a lever that does not cost energy density.
Why This Stacks Instead of Competing
That is the part worth keeping. Chinese fast-charging progress over the past two years has been largely engineering — tab geometry, cooling, pack architecture, BMS limits — plus incremental chemistry. Those work on the electrode and the system. Electro-osmotic drag is a solvent-level phenomenon, and a solvent fix can sit underneath all of it.
If the mechanism is real and general, every cell maker gets a new variable to tune, and the improvement compounds with whatever they have already done. That is a different kind of contribution from a charging record: records are demonstrations, and this is a parameter.
Wang’s group is already stacking it. Their next step is to combine the fluorinated solvent with an “adaptive electrolyte” that separates in situ to protect both electrodes from fast-charging stress, with a stated target of around 10 minutes. They are also working on calendar life — how long a battery lasts unused — and on wider operating temperatures.
Author’s Take: The headline number here will disappoint anyone benchmarking against Chinese fast-charging claims, and it should. Twelve point eight minutes is roughly three and a half times slower than what CATL says its current cells do. But the comparison is the wrong one to make, because the two things are not competing. CATL is optimising a product; Maryland has found a mechanism that product engineering has been ignoring. The reason to pay attention is the phrase “without sacrificing anything” — Wang’s stated goal, and the exact opposite of how fast charging has been bought for the last decade. If a fluorinated solvent really restores cathode wetting at high current without thinning the electrode, then the next generation of fast-charging cells gets faster without getting shorter-legged, and that has not been on the menu before. The caveats are real: no cycle life, no test conditions, and a cell one-fiftieth the size of the real thing. But a mechanism published in Science tends to show up in products three to five years later, and this one is cheap to test.
The Bottom Line: University of Maryland researchers identified electro-osmotic drag — lithium ions dragging solvent out of the cathode — as a reason fast charging stalls, and fixed it by fluorinating the electrolyte solvent. A 4 Ah pouch cell reached 80% in 12.8 minutes and 90% in 14.7, inside the DOE’s 15-minute target, with no stated energy-density penalty. That is slower than CATL’s production claims and the cell is about one-fiftieth of an automotive cell, and no cycle life was disclosed. The value is the mechanism, not the record.
Sources & Further Reading
- University of Maryland, A. James Clark School of Engineering — “Redesigning Batteries for Ultrafast Charging EVs” — the 12.8-minute and 14.7-minute results, the Department of Energy’s 15-minute target, the 20–60 minute current baseline, the electro-osmotic drag explanation, the fluorinated-solvent approach, the 10-minute goal, and the quotations from Chunsheng Wang and Chang-Xin Zhao.
- Chang-Xin Zhao et al., Science (2026), DOI 10.1126/science.adv1739 — the peer-reviewed paper behind the result. The article is paywalled and EVsays has not accessed the full text; all figures in this piece come from the university’s public account of it.
- US Department of Transportation, Charger Types and Speeds — the 20–60 minute DC fast-charging baseline cited by the university, and the 4–10 hour figure for Level 2 charging.
- CATL third-generation Shenxing battery, launched 21 April 2026 — the 10–80% in 3 minutes 44 seconds claim used for comparison, as reported in Chinese trade media and repeated in coverage of Dongfeng’s bid for first-launch rights in November.
- EVsays — Geely’s fast-charging claim and battery degradation — what repeated fast charging does to a pack, and why charge curves are negotiated with the BMS rather than fixed by the cell.
- EVsays — Japanese electrolyte additive work — the other half of electrolyte engineering, and the more common route to longer life rather than faster charging.
- EVsays — how long an EV battery actually lasts — the cycle-life and calendar-life questions this paper leaves open.
- EVsays — LFP versus NCM — why chemistry choice, not just charging speed, sets what a pack can accept.
Notes: All performance figures come from the University of Maryland’s published account of a paper in Science; EVsays has not seen the paper, the test conditions or any cycle-life data, and the university’s release does not disclose the C-rate, test temperature, starting state of charge or retained energy density. The comparison with CATL’s Shenxing figure is indicative only, since the two results were measured under conditions that have not been published in comparable form. The 4 Ah figure is converted to about 15 Wh using a 3.7 V nominal cell voltage, and the 50-fold comparison against a 200 Ah automotive cell is our arithmetic, as is the four-thousand-cell count for a 60 kWh pack. We hold no position in any company named. Corrections follow our correction policy; sourcing standards are set out in our editorial policy.
