Solid-Electrolyte Interphase SEI

What it is (precise)

When a fresh Cell is charged for the first time, the graphite Anode is driven to a potential around 0.1 V versus lithium, well below the roughly 1 V reduction limit of the carbonate Electrolyte (ethylene carbonate and linear carbonates with a LiPF6 salt).

The electrolyte reduces at the anode surface and deposits a solid film built from lithium carbonate, lithium fluoride, lithium alkyl carbonates and polymeric species. This film is the SEI. It forms on the anode, not the Cathode, and it sits between the active particle surface and the liquid electrolyte, distinct from the Separator, which is a bulk mechanical membrane between electrodes.

The layer is self-limiting by design: once it covers the anode it blocks electrons from reaching the electrolyte, which stops further reduction, while still allowing lithium ions to shuttle through. It is never perfectly stable. Each Cycle expands and contracts the graphite by roughly 10% in volume, so the film cracks, dissolves and re-forms locally, consuming a small amount of cyclable lithium each time.

In stationary Lithium Iron Phosphate cells the picture is essentially the same as in NMC cells, because SEI behaviour is governed by the graphite anode, not the cathode; this is one reason LFP and NMC calendar-fade models share the same mathematical form.

Why it matters in a real grid-scale project

SEI growth is the dominant mechanism behind calendar fade, the capacity a battery loses just sitting at a given State of Charge and temperature, independent of cycling. For a utility BESS that spends long stretches at rest between dispatch events, calendar aging can rival or exceed cycle aging over a 15-20 year project life; a system cycling once per day at 0.25C often loses more capacity to time than to throughput. That balance directly sets the augmentation schedule, the overbuild in the initial energy sizing, and the capacity-maintenance reserve priced into the long-term service agreement.

Because SEI reactions are thermally activated, growth follows an Arrhenius-like law: as a rule of thumb the rate roughly doubles for every 10-15 C rise in cell temperature, and it accelerates further at high State of Charge because the anode sits at a lower, more reactive potential.

That is the engineering justification for HVAC sizing inside the container, for energy-management strategies that avoid parking fleets at 100% SOC, and for the temperature and SOC envelope written into the capacity warranty. SEI thickening also raises internal resistance, which degrades round-trip efficiency at the POI and shows up as a slowly widening gap between DC energy and delivered AC energy.

SEI is also safety-relevant. Exothermic SEI decomposition is typically the first reaction in the thermal-runaway cascade, with onset on the order of 80-120 C, well before separator collapse or cathode oxygen release.

This is why cell and module propagation behaviour is characterised under UL 9540A, the fire-propagation test method whose data feeds NFPA 855 installation requirements, with NFPA 68 and NFPA 69 covering deflagration venting and prevention for the enclosure. The SEI itself is never named in those documents, but the abuse behaviour it initiates is exactly what they are built to bound.

Typical values and standards

Treat published figures as ranges. SEI formation consumes irreversible capacity in the first cycles: graphite cells typically show a first-cycle coulombic efficiency of roughly 90-95%, meaning 5-10% of the lithium inventory is locked into the initial film.

That is why every manufacturer runs a formation-and-aging step, days to weeks of controlled low-rate cycling and rest that can account for a meaningful share of cell production cost and factory footprint. Silicon-blended anodes, which swell far more than graphite, lose more lithium to SEI re-formation and are still rare in stationary products for exactly that reason.

In service, calendar fade at a 25 C reference and moderate resting SOC is commonly in the 1-2% per year range and, to first order, grows with the square root of time, fast at first, then flattening as the film thickens and slows its own growth. Warranty models usually encode separate calendar and cycle fade terms that are summed, and end of life is conventionally taken at 70-80% of beginning-of-life energy.

On the standards side, IEC 62619 and UL 1973 cover cell and pack safety for stationary use, UL 9540 certifies the complete ESS product, and UL 9540A is the propagation test method; none measures the SEI directly, but the aging and abuse behaviour it governs is what their data characterise.

How it shows up in specs, studies and contracts

You will almost never see the letters SEI in project documents, but its fingerprints are everywhere. Cell datasheets carry it as the capacity-retention curve and the calendar-life table at stated temperature and SOC.

Capacity warranties carry it as the guaranteed State of Health trajectory, with operating conditions, average cell temperature, resting SOC limits, and annual throughput caps, that exist purely to keep SEI growth inside the vendor's degradation model. Degradation reports from independent engineers carry it as the calendar-fade term, usually a t-to-the-power-0.5 expression fitted to accelerated aging data.

Practical checks for a working engineer: confirm the reference conditions behind any retention curve, since a curve at 25 C and 50% resting SOC will look far better than real fleet behaviour at 35 C and 90% SOC; ask whether calendar and cycle fade are treated as additive and at what temperature the calendar term was fitted; check whether the warranty tracks cell temperature or ambient, because the difference flows straight through the HVAC design; and verify that the dispatch profile assumed in the degradation model matches the actual market application, since a frequency-response asset resting at mid-SOC ages very differently from an arbitrage asset parked full overnight.

Common pitfalls

Do not assume SEI lessons transfer unchanged across chemistries. Sodium-ion cells form an SEI too, but on hard-carbon anodes with different, generally more soluble film chemistry, which is one of the technology's open durability questions, and sodium-ion cells tolerate storage at 0 V, something the lithium-ion SEI does not permit.

Also avoid reading calendar fade as harmless because the system is idle: an oversized BESS held at high SOC in a hot climate can lose capacity faster while resting than a hard-cycled system in a mild one. Finally, remember the SEI consumes lithium inventory, not active material; the electrodes may be intact while usable energy and Depth of Discharge headroom quietly shrink.