An in-situ cross-linked polyurethane gel electrolyte reports a 4.97-volt stability window and a lithium transference number of 0.78, holding 81.7 per cent capacity over 500 cycles against a high-nickel cathode.

Lithium-metal batteries promise far more energy per kilogram than the lithium-ion cells in phones and cars, because they replace the graphite anode — which merely hosts lithium — with lithium metal itself. The promise has been understood for decades. The obstacle is that lithium metal and high-voltage cathodes want incompatible things from the electrolyte between them, and conventional liquid electrolytes cannot satisfy both.

A study in Nano-Micro Letters, from a group spanning North China Electric Power University, the Southern University of Science and Technology, the Chinese Academy of Sciences, Sun Yat-Sen University and the battery manufacturer Sunwoda Mobility Energy Technology, reports a gel electrolyte engineered to do both.

The two-sided problem

On the cathode side, the difficulty is voltage. Getting more energy from a cell means charging it higher, and high-nickel cathodes such as NCM811 — roughly 80 per cent nickel, with cobalt and manganese — operate at potentials where ordinary electrolytes begin to oxidise and decompose. On the anode side, the difficulty is shape. Lithium metal does not plate back evenly during charging; it grows needle-like dendrites that can pierce the separator and short the cell.

An electrolyte that resists oxidation at the cathode is typically not the kind that encourages smooth lithium plating at the anode, and vice versa. Most work in the field improves one end and accepts a compromise at the other.

The material reported here, designated G-P3 AR, is a polyurethane gel formed by cross-linking inside the assembled cell rather than being cast as a film and inserted. In-situ formation means the polymer network grows in intimate contact with both electrodes, which avoids the interfacial gaps that plague pre-formed solid electrolytes.

Its design mixes three elements. Polyester segments and polyether segments are combined in the same backbone — polyethers conduct lithium ions well, polyesters resist oxidation — and a boron-based unit is added to regulate anions. Boron is electron-deficient and tends to bind the negative counter-ions, which has a specific and useful consequence described below.

What the numbers say

The electrochemical stability window is reported at 4.97 volts, which is the figure that makes NCM811 usable; that cathode is typically charged to around 4.3 volts, so a window approaching five volts leaves real margin.

The lithium-ion transference number is 0.78. In a conventional liquid electrolyte this figure sits nearer 0.2 to 0.4, meaning most of the current is carried by anions moving the wrong way, which builds up concentration gradients and drives exactly the uneven plating that produces dendrites. Pinning the anions with the boron unit is what raises the number, and a transference number of 0.78 is the single most consequential result in the paper.

Ionic conductivity is 0.78 millisiemens per centimetre at 25 °C — respectable for a gel, though below the several millisiemens typical of liquid electrolytes. The lithium desolvation activation energy is 38.9 kilojoules per mole, a measure of how much energy it takes to strip a lithium ion free of its solvent shell before it enters the electrode.

In lithium-on-lithium symmetric cells, which test plating and stripping in isolation, the electrolyte ran stably for more than 1,000 hours at 0.5 milliamps and 0.5 milliamp-hours per square centimetre. Against NCM811, the cell delivered 157.7 milliamp-hours per gram at a 2C rate and recovered to 205.7 at 0.1C, indicating the capacity loss at speed is not permanent damage.

Cycle life, and the honest comparison

Cycling at a 0.5C charge and 1C discharge, the cell retained 81.7 per cent of its capacity after 500 cycles, against 73.5 per cent for a conventional liquid electrolyte under the same conditions. That is the comparison that matters, and the gel wins it — though both figures are a long way from the thousands of cycles at better than 80 per cent retention that commercial lithium-ion cells achieve.

The more demanding test used a high-loading NCM811 cathode at 2 milliamp-hours per square centimetre, closer to what a practical cell requires, and retention fell to 76.6 per cent after 200 cycles. Thicker electrodes are harder on an electrolyte, and the drop is the expected direction.

A pouch cell — a format resembling an actual product rather than a laboratory coin cell — held 85.5 per cent after 50 cycles, and continued to power an LED while being bent, punctured and sheared. Puncture tolerance is a genuine safety advantage of gel and solid electrolytes over flammable liquid ones, and it is the demonstration most likely to matter commercially.

Fifty cycles, however, is a short test. It establishes that the pouch cell works, not that it lasts.

What is not disclosed

The release contains no direct quotations from any of the twelve listed authors, so none appear in this article. It names no first or corresponding author, so the endnote carries the paper title alone. It lists no funding sources.

It also carries no conflict-of-interest statement, which is a more pointed omission here than usual: one of the listed affiliations, Sunwoda Mobility Energy Technology, is a commercial battery manufacturer with an obvious interest in electrolyte technology. Industry participation in battery research is normal and often valuable, but readers are entitled to see it declared rather than inferred from an affiliation line.

As with the other release from the same journal centre this week, the dates do not line up: the release is dated 7 October 2026 while the paper is given as published on 3 August 2026.

The field is crowded with electrolytes that perform well in coin cells and disappoint at scale. What distinguishes this one is a specific mechanistic claim — that anchoring anions with boron raises the transference number, and that a mixed polyester–polyether backbone handles both electrodes — tied to a measured 0.78 and a 4.97-volt window. Those are numbers other groups can try to reproduce.

---

Notes

1. “Atomic and Molecular Structure Regulated In Situ Cross-Linked Polyurethane Gel Electrolyte for High-Performance Lithium Metal Batteries.” Nano-Micro Letters, 3 August 2026. DOI: 10.1007/s40820-026-02307-4.
2. Shanghai Jiao Tong University Journal Center. “Molecularly engineered gel electrolyte bridges high voltage and lithium-metal stability.” EurekAlert!, 7 October 2026. https://www.eurekalert.org/news-releases/1146949.