Harbor University Battery Lab Isolates a Safer Electrolyte Blend

The independent voice most quoted in the field’s response is Dr. Wren Auclair, who runs a storage integration group at a state laboratory inland and agreed to serve as the replication partner. She is publicly supportive of the chemistry and publicly skeptical of the framing.

“Cell-level abuse testing is a legitimate science and a poor predictor of installation fires,” Auclair said. “Look at the incident reports. The failures that burn down a substation enclosure start at a busbar, a coolant leak, a bad weld, or a controller that keeps charging into a fault. A cell that will not ignite is genuinely better. It is not the binding constraint, and every dollar the sector moves from pack engineering into electrolyte chemistry on the strength of a nail test is a dollar mis-spent.”

Voss does not dispute the incident data and says he has stopped arguing the point. His position is narrower: at the pack level, the value of the blend is not that a cell will not burn but that a failed cell will not recruit its neighbors, which changes how tightly modules can be spaced and how much thermal barrier material a cabinet needs. That claim requires module-level testing the lab cannot afford to do, and Auclair’s group can.

Field technicians emphasized that sensors fail in ordinary ways: bird strikes, battery sag, firmware quirks. The science depends as much on maintenance schedules as on elegant models.

## Money, patents, and the next eighteen months

The project runs on a 1.4 million state materials grant now in its third of four years, plus 260,000 in in-kind precursor supply from the materials partner, Kestrel Fine Chemicals, which can produce the phosphonate ester at about 90 kilograms a month. Scaling that to anything a manufacturer would recognize is a different company’s problem and, by Voss’s estimate, a decade out.

Patent filings are in progress, with a commitment to non-exclusive licensing for municipal projects in the state. The commitment was written into the grant terms at the insistence of the state energy office and survived an attempt by the university’s technology transfer office to add a revenue threshold above which exclusivity could be negotiated.

Next steps are longer cycle tests and an independent lab replication. Auclair’s group receives the first shipment of cells in September and expects results at 1,000 cycles by next summer.

Public communication remains a deliberate practice. Figures are released with uncertainty bands, and press officers have stopped promising single-number headlines the data cannot support. Students on the team presented posters at a coastal engineering meetup, translating jargon for municipal staff who buy storage systems more often than they read journals. Students on the project are also writing methods sections that double as training manuals for the next cohort; continuity is a research output when grants turn over every two years.

The abuse tests are run in a steel bunker behind the engineering annex, a converted magazine with a blast door and a floor drain, and the lab keeps every failed cell. They sit in labeled trays on a shelf by the door, 71 of them now, each one a small blackened brick with a curl of foil showing at the seam. Ferreira brings visiting funders in there before she brings them to the data, and hands them one.