KAIST engineers found the slow step in making V3.5+ vanadium electrolyte sits near an oxidation state of +4.1, and changed methods there.

Vanadium redox flow batteries are among the better options for storing electricity at grid scale. They are durable, they do not degrade quickly with cycling, and they scale by adding tanks of liquid rather than by stacking cells — which makes them well suited to storing many hours of energy, the problem that lithium-ion handles least gracefully.

Their obstacle is cost, and a large share of that cost is the electrolyte: the vanadium solution itself and the process of making it. A study in Advanced Energy Materials from the Korea Advanced Institute of Science and Technology reports cutting the production time for V3.5+ vanadium electrolyte by 67 per cent, to roughly a third of the conventional process.

The V3.5+ designation refers to a mixed-valence solution in which vanadium sits, on average, halfway between the +3 and +4 oxidation states. That average is what allows a single electrolyte to be loaded into both tanks of a flow battery at assembly, rather than preparing separate positive and negative solutions and balancing them afterwards. It is the commercially preferred form for that reason, which is also why the cost of producing it is a live engineering problem rather than an academic one.

Finding the slow step

The improvement came from locating a bottleneck rather than from a new material. The team identified a kinetic constraint at an average vanadium oxidation state of about +4.1 — a specific point partway through the reduction that brings vanadium to the mixed +3.5 state the electrolyte requires.

Conventional production uses chemical reduction throughout. The KAIST process switches to catalytic reduction at about +4.1, bypassing the slowest stage, and the 67 per cent saving follows from skipping it.

“This study combined reaction engineering principles with thermodynamic predictions to identify the rate-determining step,” said Hee-Tak Kim, Professor in the Department of Chemical and Biomolecular Engineering at KAIST and the paper’s corresponding author.

That is a description of method as much as of result. A rate-determining step is the stage that sets the pace of everything else, and a process that uses one mechanism from start to finish will be governed by whichever stage that mechanism handles worst. Identifying where the handicap sits allows a different mechanism to be applied to that stage alone.

The first author is Kyunghwa Seok, a PhD candidate in the same department, with Minseong Kang as second author.

The impurity and the catalyst

Two further results matter more for manufacturing than the headline number.

The process removes residual oxalic acid, an impurity carried over from conventional chemical reduction that can degrade battery performance. Eliminating a contaminant at the production stage is worth more than it sounds, because the alternative is either a purification step or a slow performance penalty over the battery’s service life.

The Pt/C catalyst — platinum on carbon — was reused more than 2,500 times without a notable drop in performance. This is the number that determines whether the approach is economic at all. Platinum is expensive enough that a catalytic step is only viable if the catalyst survives; at 2,500 reuses, the platinum cost is spread thinly enough to stop being the dominant term.

“By scientifically identifying the conditions under which the catalyst operates stably,” Kim said, “we expect this to significantly accelerate the commercialization of large-capacity energy storage technology.”

What 2,500 cycles does and does not tell you

Catalyst stability figures need reading with care. More than 2,500 reuses without notable degradation is a strong laboratory result, and laboratory reuse conditions are cleaner than an industrial feedstock stream, where trace contaminants in commercial vanadium sources are a standard cause of catalyst poisoning.

The release also gives no production scale. A 67 per cent time saving measured at laboratory volumes is a demonstration of the chemistry; whether it survives scale-up depends on heat and mass transfer in a reactor orders of magnitude larger, and on handling platinum catalyst recovery in a continuous process rather than a batch one.

Nor is time the same as cost. Production time is one input to electrolyte price alongside vanadium feedstock, energy, equipment and the capital tied up in the catalyst, and the release does not report a cost figure. Vanadium itself is usually the largest component, and this work does not reduce how much of it a battery needs.

The framing, and the funder

The release presents the work in terms of AI data centres, which are driving a sharp increase in projected electricity demand and in demand for storage to firm it. That framing is the publisher’s rather than the paper’s: the chemistry is indifferent to what the stored electricity is eventually used for, and grid-scale storage has been under pressure from renewables integration for a decade before data centre demand entered the conversation.

The research was supported by Lotte Chemical, a petrochemical company with interests in battery materials. Industrial funding of process engineering work is unremarkable — a company that might build the plant is the natural sponsor of research into making the plant cheaper — but it is worth knowing that the sponsor has a commercial stake in the result.

The paper was published online on 7 May 2026 and was selected as the cover article for the journal’s Issue 34.

The useful summary is that this is a process-engineering advance rather than a new battery. Flow batteries already work; what has kept them from wider deployment is the price of the liquid in the tanks, and a method that makes that liquid faster and cleaner, with a catalyst that lasts, addresses the part of the problem that is actually binding.

---

Notes

1. Seok, K., Kang, M., Kim, H.-T. “Streamlined V3.5+ Electrolyte Production by Leveraging Chemical and Catalytic Reductions.” Advanced Energy Materials, 2026. DOI: 10.1002/aenm.71029.
2. The Korea Advanced Institute of Science and Technology (KAIST). “KAIST brings ‘giant batteries’ closer to commercialization in the AI data center era.” EurekAlert!, 4 August 2026. https://www.eurekalert.org/news-releases/1138754.

Photo by ThisIsEngineering on Pexels.com