Why the Anode Is Carbon

Q: Why is graphite used as the anode instead of something with higher capacity?

Because sp² hybridisation delivers four useful properties at once. Three σ bonds in a plane form hexagonal sheets while the unhybridised p_z orbitals overlap into a delocalised π system, giving metallic in-plane conductivity of about 10⁴ S/cm so the anode needs no conductive additive. Between sheets there are only van der Waals forces at 3.35 Å spacing, which is the intercalation gallery lithium slides into. It reaches LiC₆ for 372 mAh/g at roughly 0.1 V vs Li/Li⁺, and expands only about 10 per cent doing so — small enough to survive thousands of cycles.

Q: Why is charging a lithium battery in the cold so damaging?

Graphite operates at about 0.1 V vs Li/Li⁺ while metallic lithium plates at 0.0 V. The entire margin is 100 millivolts. Anything that raises the anode’s overpotential by more than that pushes the local potential below zero and plates metallic lithium instead of intercalating it. Overpotential rises with charge current, with low temperature (diffusion in graphite slows exponentially by Arrhenius), with high state of charge as empty sites become scarce, and with an aged, thick SEI. The cold-charging rule is not a rule of thumb — it is a 100 mV margin being consumed by kinetics.

Q: Why is silicon not simply used instead of graphite?

Because silicon alloys rather than intercalates, and the difference is structural. Intercalation puts guests into pre-existing gaps; alloying rebuilds the crystal. Li₁₅Si₄ gives 3,579 mAh/g — nearly ten times graphite — but requires roughly 280 to 300 per cent volume expansion. Particles fracture, fresh silicon surface is exposed, new SEI forms on it and permanently consumes lithium inventory, the electrode pulverises and loses electronic contact, and the cell swells. Current practice blends just 5 to 10 per cent silicon into graphite to lift capacity about 20 per cent while keeping expansion manageable.

Q: Why does LTO never plate lithium?

Because it operates at 1.55 V vs Li/Li⁺ — far above the 0.0 V plating threshold, so there is no margin to consume even at −20 °C and 10C charge. It also sits above the electrolyte reduction potential, so no SEI forms at all and there is no SEI-driven ageing. Its spinel structure is effectively zero-strain, under 0.2 per cent volume change. The price is paid twice: 1.55 V subtracts directly from cell voltage, giving only about 1.9 V against LFP, and capacity is only 175 mAh/g, so cell-level energy lands at 50 to 90 Wh/kg — lead-acid territory at many times the cost.

Q: Why do sodium-ion cells use hard carbon instead of graphite?

Because sodium does not intercalate graphite in any meaningful amount — NaC₆₄ at best in standard electrolytes — since the binding energy of sodium in the graphite gallery is too weak relative to metallic sodium. Hard carbon is non-graphitisable carbon: short, randomly oriented graphene fragments with nanopores between them, made by pyrolysing sucrose, phenolic resin or biomass. Sodium occupies those disordered defect sites and nanopores for about 300 mAh/g, at the cost of lower first-cycle coulombic efficiency and a sloping rather than flat voltage profile.

sp² hybridisation produces a layered, conductive, van-der-Waals-bonded host that lithium enters cheaply and reversibly. Silicon stores ten times as much and expands 280 per cent doing it.

The Periodic Table of the EV · Part 3 — The Anode and the Electrolyte · Chapter 6 · 23 min read

All chapters

sp² hybridisationLiC₆ and stagingLithium plating marginSilicon expansionLTO and hard carbon

Part 3 — The Anode and the Electrolyte

Why carbon, and what it costs to leave it

sp² hybridisation produces a host lithium enters cheaply and reversibly. Every alternative stores more and behaves worse.

The anode is the one part of a lithium cell where the periodic-table argument is almost embarrassingly simple. Carbon is small, it catenates indefinitely, and one particular hybridisation of it produces a layered solid held together by nothing stronger than van der Waals forces.

That last property is the intercalation gallery. Everything else — conductivity, capacity, the 10 per cent expansion, even the cold-charging rule — follows from it.

372

mAh/g at LiC₆

3.35 → 3.70 Å

Interlayer spacing on lithiation

100 mV

The entire plating margin

~280 %

Silicon’s volume expansion

6.1 — Graphite — structure from hybridisation

Carbon has four valence electrons, is small, and is uniquely capable of catenation — bonding to itself indefinitely.

In graphite, carbon is sp² hybridised: three σ bonds in a plane at 120°, forming hexagonal sheets, plus one unhybridised p_z orbital perpendicular to the plane. Those p_z orbitals overlap sideways across the whole sheet to form a delocalised π system.

Two consequences follow directly, and both matter commercially.

•In-plane electronic conductivity is metallic at roughly 10⁴ S/cm, because the π electrons are delocalised. The anode conducts without any conductive additive — unlike every cathode.
•Between sheets there are only van der Waals forces — weak, long-range dispersion attraction, at an interlayer spacing of 3.35 Å.

In plain English

That weak interlayer bonding is the intercalation gallery. Lithium slides in between the sheets at modest energy cost, because there is almost nothing holding them together to overcome.

Figure 6.1 — Two bond strengths in one material. Within the sheet, carbon is held by some of the strongest covalent bonds in chemistry. Between sheets, only weak van der Waals attraction. That asymmetry is the intercalation gallery: strong enough to survive thousands of cycles, weak enough to open for a lithium ion. Sodium, being larger and binding more weakly to the gallery, gains nothing by entering — which is why sodium-ion cells cannot use graphite at all.

6.2 — The numbers: LiC₆, staging and 10 per cent

LiC₆ — one lithium per six carbons — is the fully lithiated stoichiometry.

Worked example 6.1 — Graphite specific capacity

Q = (1 × 96,485) / (3.6 × 72.06)
Q = 96,485 / 259.4
Q = 372 mAh/g

Interlayer spacing expands from 3.35 Å to about 3.70 Å, roughly 10 per cent, giving roughly 10 per cent volume expansion — small enough that the electrode survives thousands of cycles. Hold that number; silicon’s is 28 times larger.

6.2.1 — Staging

Lithium does not fill the galleries randomly. It fills every nth gallery completely, progressing through stage IV → III → II → I as state of charge rises. Each stage is a distinct phase.

That is why graphite has a stepped voltage profile, with plateaus at roughly 0.21 V, 0.12 V and 0.09 V vs Li/Li⁺. Those steps are visible in a good dQ/dV analysis and are a diagnostic fingerprint of anode health — when they blur or shift, something has changed at the anode.

6.3 — The 0.1 V problem — cold charging from first principles

Graphite operates at about 0.1 V vs Li/Li⁺ — beautifully low, which maximises cell voltage. But metallic lithium plates at 0.0 V.

Important

The margin is 100 millivolts. That is the whole safety budget between normal intercalation and permanent, dangerous lithium plating.

Overpotential is the extra voltage needed to drive a reaction at a given rate. Anything that raises the anode’s overpotential by more than 100 mV pushes the local potential below zero and plates metallic lithium instead of intercalating it. Overpotential rises with:

1High charge current — the simple I × R term.
2Low temperature — diffusion in graphite slows exponentially by the Arrhenius relation from chapter 2.
3High state of charge — fewer empty sites remain, so the driving force falls.
4An aged, thick SEI — higher interfacial resistance.

Why this matters

This is the complete, first-principles explanation of the cold-charging prohibition. It is not a rule of thumb, and it is not manufacturer caution. It is a 100 mV margin being consumed by kinetics — and note that all four contributors stack, which is why a cold, fast, high-SoC charge on an aged cell is the specific combination that plates lithium.

Figure 6.2 — Graphite operates at 0.1 V vs Li/Li⁺; metallic lithium plates at 0.0 V. The margin is one hundred millivolts. Anything that raises anode overpotential by more than that — high charge current, low temperature, high state of charge, a thickened SEI — deposits metallic lithium instead of intercalating it. That deposit is permanently lost capacity and a dendrite risk. The cold-charging prohibition is not a rule of thumb; it is this margin being consumed by kinetics. LTO at 1.55 V sits so far above the threshold that it cannot plate at all — and does not even form an SEI.

6.4 — Silicon — why it is tempting and why it is hard

Silicon sits directly below carbon in group 14. It does not intercalate; it alloys. Li₁₅Si₄, the room-temperature phase, gives:

Worked example 6.2 — Silicon specific capacity

Q = (3.75 × 96,485) / (3.6 × 28.09)
Q = 3,579 mAh/g — nearly ten times graphite

Technical framing

The problem is a direct consequence of alloying rather than intercalating. Intercalation puts guests into pre-existing gaps; alloying rebuilds the crystal. Silicon’s lattice must expand by roughly 280 to 300 per cent in volume.

The consequences run in sequence, and each one feeds the next:

1Particles fracture from internal stress.
2Fresh silicon surface is exposed to the electrolyte.
3New SEI forms on that fresh surface, permanently consuming lithium inventory.
4Repeated expansion and contraction pulverises the electrode and breaks electronic contact.
5The electrode as a whole swells, pressurising the cell.

Mitigations are all partial: nano-silicon, silicon–carbon composites, silicon oxide (SiO_x, lower capacity but less expansion), void-engineered yolk-shell particles, and — most commonly today — simply blending 5 to 10 per cent silicon into graphite to lift capacity about 20 per cent while keeping expansion manageable.

Figure 6.3 — Ten times the capacity, at the cost of a different reaction mechanism. Today's compromise is blending 5–10% silicon into graphite: roughly 20% more capacity with expansion still manageable. Yolk-shell particles, SiOx and fluoroethylene carbonate additives all exist to manage the same underlying problem.

6.5 — LTO — trading energy for physics

Lithium titanate, Li₄Ti₅O₁₂, operates at 1.55 V vs Li/Li⁺, and that single number buys two things outright.

•It is far above the plating threshold, so lithium plating cannot occur — ever, even at −20 °C and 10C charge. There is no 100 mV margin to consume.
•It is above the electrolyte reduction potential, so no SEI forms at all. No SEI means no SEI growth and no SEI-driven ageing.

Its spinel structure is famously zero-strain, with volume change under 0.2 per cent on cycling. No cracking, no delamination. The result is 10,000 to 20,000 cycles, extreme fast-charge capability and excellent cold performance.

Important

The price is paid twice. The 1.55 V operating potential subtracts directly from cell voltage — a full cell with LFP gives only about 1.9 V — and titanium is heavy at 175 mAh/g. Cell-level energy density lands at 50 to 90 Wh/kg, comparable to lead-acid, at many times the cost.

It survives where cycle count matters more than energy: ultra-fast-charge buses and grid frequency regulation.

6.6 — Hard carbon — the sodium answer

Hard carbon is non-graphitisable carbon: short, randomly oriented graphene fragments with nanopores between them, produced by pyrolysing precursors such as sucrose, phenolic resin or biomass.

Sodium cannot fit graphite’s ordered galleries, for the thermodynamic reason given in chapter 2. It can occupy the disordered defect sites and nanopores of hard carbon, giving about 300 mAh/g.

The four anodes compared

Anode	Capacity	Potential vs Li/Li⁺	Volume change	Where it wins
Graphite	372 mAh/g	~0.1 V	~10 %	The default — cheap, stable, low voltage
Silicon (blended)	3,579 mAh/g pure	~0.4 V	~280 %	Energy density, at 5–10 % blend
LTO	175 mAh/g	1.55 V	<0.2 %	Cycle life, fast charge, cold
Hard carbon	~300 mAh/g	sloping	moderate	Sodium-ion, where graphite fails

The trade-offs for hard carbon are lower first-cycle coulombic efficiency — large surface area means large irreversible SEI — and a sloping rather than flat voltage profile, which makes state-of-charge estimation easier but pack design less forgiving.

Quick check: test yourself

1.Why does a graphite anode need no conductive additive when every cathode does?

Show answer

Because sp² hybridisation leaves one unhybridised p_z orbital per carbon, and those overlap into a delocalised π system across each sheet. In-plane conductivity is metallic at about 10⁴ S/cm, so the active material is its own current path.

2.Name the four contributors to anode overpotential and explain why they are dangerous together.

Show answer

Charge current (I×R), low temperature (Arrhenius-slowed diffusion), high state of charge (fewer empty sites), and a thick aged SEI (interfacial resistance). They stack, and the total budget is only 100 mV before the anode potential goes below 0 V and plates lithium.

3.LTO can charge at 10C at −20 °C forever. Why is it not the default anode?

Show answer

Because its 1.55 V operating potential subtracts directly from cell voltage — about 1.9 V in a full cell with LFP — and its capacity is only 175 mAh/g. Cell-level energy lands at 50–90 Wh/kg, lead-acid territory, at many times the cost.

Chapter summary

✓Graphite works because sp² hybridisation gives metallic in-plane conductivity and van-der-Waals-bonded sheets 3.35 Å apart — a gallery lithium enters cheaply.
✓LiC₆ gives 372 mAh/g at ~0.1 V with only ~10 per cent volume expansion, and staging produces the stepped voltage profile used for anode diagnostics.
✓The cold-charging rule is a 100 mV margin: graphite sits at 0.1 V, lithium plates at 0.0 V, and current, cold, high SoC and an aged SEI all consume that budget together.
✓Silicon stores ten times as much because it alloys rather than intercalates — and expands ~280 per cent for exactly the same reason. Today it is blended at 5–10 per cent.
✓LTO buys immunity to plating and to SEI ageing by sitting at 1.55 V, and pays for it twice in cell voltage and in capacity.

Frequently asked questions

Why is graphite used as the anode instead of something with higher capacity?+

Because sp² hybridisation delivers four useful properties at once. Three σ bonds in a plane form hexagonal sheets while the unhybridised p_z orbitals overlap into a delocalised π system, giving metallic in-plane conductivity of about 10⁴ S/cm so the anode needs no conductive additive. Between sheets there are only van der Waals forces at 3.35 Å spacing, which is the intercalation gallery lithium slides into. It reaches LiC₆ for 372 mAh/g at roughly 0.1 V vs Li/Li⁺, and expands only about 10 per cent doing so — small enough to survive thousands of cycles.

Why is charging a lithium battery in the cold so damaging?+

Graphite operates at about 0.1 V vs Li/Li⁺ while metallic lithium plates at 0.0 V. The entire margin is 100 millivolts. Anything that raises the anode’s overpotential by more than that pushes the local potential below zero and plates metallic lithium instead of intercalating it. Overpotential rises with charge current, with low temperature (diffusion in graphite slows exponentially by Arrhenius), with high state of charge as empty sites become scarce, and with an aged, thick SEI. The cold-charging rule is not a rule of thumb — it is a 100 mV margin being consumed by kinetics.

Why is silicon not simply used instead of graphite?+

Because silicon alloys rather than intercalates, and the difference is structural. Intercalation puts guests into pre-existing gaps; alloying rebuilds the crystal. Li₁₅Si₄ gives 3,579 mAh/g — nearly ten times graphite — but requires roughly 280 to 300 per cent volume expansion. Particles fracture, fresh silicon surface is exposed, new SEI forms on it and permanently consumes lithium inventory, the electrode pulverises and loses electronic contact, and the cell swells. Current practice blends just 5 to 10 per cent silicon into graphite to lift capacity about 20 per cent while keeping expansion manageable.

Why does LTO never plate lithium?+

Because it operates at 1.55 V vs Li/Li⁺ — far above the 0.0 V plating threshold, so there is no margin to consume even at −20 °C and 10C charge. It also sits above the electrolyte reduction potential, so no SEI forms at all and there is no SEI-driven ageing. Its spinel structure is effectively zero-strain, under 0.2 per cent volume change. The price is paid twice: 1.55 V subtracts directly from cell voltage, giving only about 1.9 V against LFP, and capacity is only 175 mAh/g, so cell-level energy lands at 50 to 90 Wh/kg — lead-acid territory at many times the cost.

Why do sodium-ion cells use hard carbon instead of graphite?+

Because sodium does not intercalate graphite in any meaningful amount — NaC₆₄ at best in standard electrolytes — since the binding energy of sodium in the graphite gallery is too weak relative to metallic sodium. Hard carbon is non-graphitisable carbon: short, randomly oriented graphene fragments with nanopores between them, made by pyrolysing sucrose, phenolic resin or biomass. Sodium occupies those disordered defect sites and nanopores for about 300 mAh/g, at the cost of lower first-cycle coulombic efficiency and a sloping rather than flat voltage profile.

Reviewed by

Sahil Goyal

Co-founder, Wingzman

Sourabh Goyal

Co-founder, Wingzman

The Periodic Table of the EV is an original educational series on the materials science of electric vehicles. All values are standard-condition literature figures for representative materials, not measured data from a specific product, and sources differ on several of them. Always verify against the specific material datasheet in use before making design, purchasing or certification decisions.