FEC in Silicon Anode Electrolyte: How Much to Use

Start at 5–15 wt% FEC and you are in the right ballpark — but that percentage says almost nothing about whether your cell survives 500 cycles. FEC is a consumable, not a catalyst: the amount available is set by how much electrolyte is in the cell, not by the number on the formulation sheet. This article converts FEC wt% into grams per amp-hour, shows you how to predict the depletion knee before you build cells, and what to do when the arithmetic says 10% is not enough.

Pick a percentage, then immediately convert it to grams per amp-hour

For most silicon-containing anodes, a sensible starting point is 3–5 wt% FEC for graphite blends holding under 10% silicon, 5–10 wt% for Si-C composite blends, and 10–15 wt% for silicon-dominant anodes.

That is where you start, not where you stop. Cycle life tracks grams of FEC per amp-hour, because FEC is irreversibly consumed every time fresh silicon surface is exposed — and two cells at an identical 10 wt% can differ by more than 20× on that measure.

Why FEC works — and why it runs out

FEC reduces before the other carbonates do — roughly 1.2–1.4 V vs Li/Li⁺ against about 0.8 V for ethylene carbonate — laying down a LiF-rich solid electrolyte interphase (SEI) before EC gets the chance.

That matters because silicon expands roughly 280% by volume on full lithiation to Li₁₅Si₄ (3,579 mAh/g), fracturing the SEI every cycle.

A LiF-rich, partly polymeric FEC-derived film tolerates that abuse better than an EC-derived one, but every crack still exposes bare silicon, and every patch of bare silicon consumes more FEC. This is the asymmetry most articles miss: FEC buys cycles at a fixed exchange rate rather than stabilising the interface permanently. When the reservoir empties, the cell falls off a cliff.

The unit problem: wt%, vol%, and "% of solvent" are three different numbers

Before comparing your formulation to a published one, confirm which basis the percentage uses — the same "10% FEC" can mean 8.7%, 10.0%, or 12.3% by weight of the finished electrolyte.

• FEC density is 1.454 g/cm³ against ~1.15–1.25 g/cm³ for a typical carbonate blend, so 10 vol% FEC is about 12.3 wt%.
• A 1 M LiPF₆ electrolyte is 12–13 wt% salt, so "10 wt% of the solvent" is about 8.7 wt% of the finished electrolyte.
• That spread — 8.7% to 12.3% — is a 40% difference in FEC mass from a labelling choice alone.

Fix it once: state FEC as wt% of the finished electrolyte, salt included, on every batch sheet.

Why coin cell FEC results do not survive the move to pouch cells

A flooded coin cell can carry more than 20 times the FEC per amp-hour of a pouch cell at an identical weight percent. This is the single biggest reason coin-cell cycle life evaporates on scale-up.

 

Cell format (all at 10 wt% FEC)	Capacity	Electrolyte	E/C ratio (g/Ah)	FEC mass	FEC per Ah
Flooded coin half-cell, 12 mm, 2 mAh/cm²	2.3 mAh	100 µL (120 mg)	53	12 mg	5.3 g
Lean coin full cell, 14 mm, 3 mAh/cm²	4.6 mAh	60 µL (72 mg)	15.6	7.2 mg	1.6 g
Single-layer pouch	200 mAh	500 mg	2.5	50 mg	0.25 g
Multilayer pouch / EV format	5 Ah	11 g	2.2	1.1 g	0.22 g

 

The flooded coin cell is not testing your formulation. It is testing your formulation plus a 20× FEC overdose you could never afford in a real cell.

How to calculate the FEC dose your cell actually needs

Convert your target cycle life into cumulative irreversible charge, then convert that charge into grams of FEC using Faraday's law.

FEC has a molar mass of 106.05 g/mol. At one electron per molecule, 1 g of FEC absorbs 0.25 Ah of irreversible charge; at two electrons, 0.51 Ah/g. Real systems sit between, so run both as a bracket. [VERIFY: reported electron stoichiometry for FEC reduction spans 1–2 e⁻ per molecule depending on pathway — confirm against your reference before publishing a single figure.]

The five-step method

1. Measure steady-state coulombic efficiency (CE) after formation, on the anode you intend to ship, with excess FEC so the reading is not already FEC-limited.
2. Calculate cumulative irreversible charge: Q_irr = N × (1 − CE) × Q_cell, where N is your target cycle count.
3. FEC required = Q_irr ÷ 0.25 (conservative) through Q_irr ÷ 0.51 (optimistic), in grams.
4. FEC available = E/C ratio (g/Ah) × FEC wt% × Q_cell.
5. If required exceeds available, you have four levers in descending order of value: raise CE, raise FEC%, raise E/C, or shorten the warranty.

 

Worked example 1: a 3 Ah Si-C pouch cell targeting 500 cycles

Cell: 3 Ah, E/C = 2.4 g/Ah, so 7.2 g of electrolyte. Anode: Si-C composite blend with measured steady-state CE of 99.92%. FEC at 10 wt% gives 0.72 g.

• Required: 500 × 0.0008 × 3 Ah = 1.2 Ah of cumulative irreversible charge, which needs 2.4–4.8 g of FEC.
• Available: 0.72 g.
• Shortfall: 3.3× to 6.6×.

Invert it and the picture is starker. That 0.72 g supports 0.18–0.36 Ah of irreversible charge, which at 99.92% CE buys 75–150 cycles. Not 500.

Additive percentage cannot fix this. Reaching 500 cycles on 0.72 g of FEC requires CE near 99.98%, and even then you are marginal under the one-electron assumption — which is why silicon programmes live or die on coulombic efficiency rather than additive tuning.

Worked example 2: diagnosing a coin-to-pouch collapse

A lab reports 500 cycles to 80% retention in a coin half-cell at 10 wt% FEC. The same electrode in a single-layer pouch full cell dies before 150 cycles.

Coin half-cell: a 12 mm disc (1.13 cm²) at 2.0 mAh/cm² = 2.26 mAh, with 100 µL (~120 mg) of electrolyte. 10 wt% gives 12 mg FEC, or 5.3 g per Ah. At the 99.5% CE typical of a silicon half-cell, that reservoir covers roughly 530 cycles — exactly what was reported.

Pouch full cell: 2.5 g/Ah at 10 wt% gives 0.25 g FEC per Ah — a budget of about 0.126 Ah of irreversible charge per Ah of capacity. CE improves to 99.9% in the full cell, and 0.126 ÷ 0.001 = 126 cycles.

Diagnosis: the electrode was never the problem. The coin cell simply held 21× the FEC inventory.

The fix that costs nothing: re-run the coin cells at 30 µL to match the pouch cell's E/C ratio before changing a single thing about the formulation. A lean-electrolyte coin cell predicts pouch behaviour far better than a flooded one, and it is the step almost everyone skips.

Starting FEC doses by anode type

Silicon surface area, not silicon weight fraction, sets the consumption rate — a 10% nano-silicon blend at 60 m²/g burns FEC faster than a 40% pore-confined Si-C composite at 4 m²/g.

 

Anode type	Si content	Typical BET	Starting FEC (wt% of electrolyte)	Watch for
Graphite + SiOₓ	3–8%	2–5 m²/g	2–5%	Little measurable benefit above 5%
Graphite + Si-C composite	10–25%	3–8 m²/g	5–8%	Add 1% VC if impedance budget allows
Si-dominant, pore-confined Si-C	50–90%	5–15 m²/g	8–15%	Gassing and cathode-side oxidation
Nano-Si / silicon nanowire	50–100%	20–100 m²/g	10–20%	Depletion-limited life regardless of dose

How to tell FEC is running out before the cell dies

Coulombic efficiency and impedance move first; capacity is the last thing to go, which gives you roughly 20–50 cycles of warning if you watch the right signal. The sequence:

1. CE stops improving and drifts down — the earliest and cheapest signal.
2. DCIR or EIS charge-transfer resistance climbs.
3. Pouch cells thicken as CO₂ evolves (Archimedes displacement or calipers).
4. Capacity knee — sudden, not gradual. By now you are diagnosing, not preventing.

Measurement methods that actually resolve FEC content

• ¹⁹F NMR. Harvest in a glovebox, extract with DMC, quantify against an internal standard such as hexafluorobenzene. Direct, quantitative, about an hour per sample.
• GC-FID or GC-MS. Mid-polarity column (DB-624 or equivalent) with an internal standard. Better when you also need to track co-additives and decomposition products.
• dQ/dV on formation. The FEC reduction peak sits near 1.2–1.5 V vs Li/Li⁺; its area is a free in-situ proxy for FEC availability at cycle one.
• High-precision coulometry. Novonix UHPC systems, out of Jeff Dahn's group at Dalhousie, resolve CE differences standard Arbin, Maccor, Neware or BioLogic channels cannot.

Practical note: run the ¹⁹F NMR on a cell stopped at half of expected life, not only on dead ones. A dead cell tells you FEC is gone; a mid-life cell gives you the consumption rate — the number you can design against.

When more FEC makes things worse

Above roughly 15 wt%, FEC's costs start outrunning its benefits, and four situations call for backing off regardless of what the calculation says.

1. High-voltage cathodes. FEC oxidises above about 4.2–4.3 V, generating CO₂ and HF. High FEC plus NMC811 at 4.35 V plus high-SOC storage reliably produces pouch swelling and transition-metal dissolution.
2. High-temperature storage. FEC dehydrofluorinates thermally, and the HF attacks the cathode surface and any SiOₓ. Cells that pass 25 °C cycling can fail 60 °C storage on FEC loading alone.
3. Low-temperature operation. FEC melts at roughly 18–22 °C. Above 15 wt% it raises viscosity enough to measurably hurt sub-zero rate performance.
4. Fast charge. The thick LiF-rich SEI that buys cycle life also raises impedance — the documented lifetime-versus-rate trade-off between FEC and VC.

Cost is the fifth constraint. FEC runs several times the price of EC and carries real fluorochemical supply-chain exposure; the 2021–2022 crunch is the reference case.

What to do instead of adding more FEC

If the calculation says you are short, the highest-leverage move is raising coulombic efficiency, not raising the additive load.

1. Fix CE at the anode. Lower-BET silicon, carbon coating, pore-confined Si-C architectures and stiffer binders (PAA, PAA-CMC, polyimide) all cut fresh-surface generation. Every 0.01% of CE is worth more than a percentage point of FEC.
2. Add co-additives that share the load. LiPO₂F₂ at 0.5–1 wt%, LiDFOB or LiBOB at 0.5–2 wt%, and VC at 1–2 wt% form films by different pathways, so they do not deplete on the same clock. Check impedance — VC and FEC together can over-thicken the SEI.
3. Change the salt. LiFSI, alone or as a co-salt with LiPF₆, builds a LiF-rich SEI from the anion rather than a sacrificial solvent. Watch aluminium corrosion above roughly 3.8–4.0 V; LiPF₆ co-salt suppresses it.
4. Go fluorinated across the board. Fluorinated co-solvent blends and localised high-concentration electrolytes shift SEI chemistry to the anion, removing the single-additive bottleneck. Higher cost, longer development.
5. Prelithiate — with a caveat. It replaces lost lithium inventory but does nothing to slow FEC consumption, so it masks fade until the FEC runs out and the knee arrives anyway. Useful, not a substitute.

Where this framework breaks down

This calculation is a bounding exercise, not a simulation, and it will mislead you in four ways.

• It assumes FEC is the only species consumed. EC and LiPF₆ decomposition also absorbs irreversible charge, so real FEC lifetime runs longer than the conservative bound.
• Electron stoichiometry is genuinely uncertain. One versus two electrons changes the answer by 2×. Always bracket it.
• SEI growth is partly self-limiting, so consumption is not perfectly linear with cycle count.
• It ignores the cathode. In high-voltage cells, oxidative FEC loss can rival reductive loss at the anode.

Use it to reject formulations that obviously cannot work and to size experiments — not to predict cycle life to two significant figures.

Frequently asked questions

How much FEC should I use in a silicon anode electrolyte?

Start at 3–5 wt% for graphite blends under 10% silicon, 5–10 wt% for Si-C composites, and 10–15 wt% for silicon-dominant anodes. Then convert that percentage into grams of FEC per amp-hour using your electrolyte-to-capacity ratio, and check it against your target cycle life and measured coulombic efficiency.

Does more FEC always mean longer cycle life?

No. Cycle life improves with FEC up to roughly 15 wt%, then flattens or reverses. Excess FEC raises impedance, hurts fast charge, and oxidises at cathodes above about 4.2 V, producing CO₂ and HF. In high-voltage or hot cells, 20% can shorten life rather than extend it.

Why did my cell hold capacity for 200 cycles and then collapse suddenly?

That signature is FEC depletion. Once the reservoir empties, EC-derived SEI takes over and fails quickly on expanding silicon. Check coulombic efficiency across the 20–50 cycles before the knee — it drifts downward first. Confirm by quantifying residual FEC with ¹⁹F NMR.

Can I use FEC and VC together in a silicon cell?

Yes, and it often helps — typically 5–10 wt% FEC with 1–2 wt% VC. The two form films by different pathways, so they do not deplete on the same clock. Watch impedance: the combined SEI can grow thick enough to hurt rate capability.

Does FEC cause gassing in pouch cells?

Yes. FEC generates CO₂ on reduction at the anode and oxidation at the cathode, and releases HF through thermal dehydrofluorination. Gassing scales with FEC content, cut-off voltage and temperature. If your pouches swell during 45–60 °C storage, high FEC loading is the first thing to check.

Is FEC still needed with modern Si-C composite anodes?

Yes, but usually at lower loadings. Pore-confined Si-C composites expose far less silicon surface than nano-silicon, so they consume FEC more slowly — 5–8 wt% often suffices where nano-Si needs 15%. Size the dose against BET surface area, not silicon weight fraction.

Conclusion

FEC percentage is a formulation input. Grams per amp-hour is the design constraint. Convert your target cycle life into cumulative irreversible charge, divide by 0.25–0.51 Ah/g, and compare against what your electrolyte fill actually delivers.

Your next step: take your current best coin cell result and re-run it at an electrolyte volume that matches your pouch cell's E/C ratio. If cycle life collapses, you were testing an FEC reservoir, not an electrode.