Battery chemistry

A cell is two different conductors, usually metals, dipping into an electrolyte (a solution or paste containing ions). Strictly, a battery is two or more cells joined together.

The more reactive metal loses electrons and becomes the negative terminal. Electrons flow from it through the external wire to the less reactive side. Put zinc and copper into salt water and you get a cell of about 1 V. Swap the copper for silver and the voltage goes up, because the two metals are further apart in reactivity. The voltage depends on which materials are used and on the electrolyte.

In a non-rechargeable cell, such as an alkaline AA battery, the chemicals are gradually used up and the voltage eventually falls. In a rechargeable cell, connecting a charger drives the reaction in reverse and rebuilds the original chemicals. Lithium-ion cells in phones work this way.

A fuel cell is different again: it never runs flat as long as fuel is supplied. A hydrogen fuel cell combines hydrogen and oxygen to make water, producing electricity with no carbon dioxide at the point of use, but the hydrogen has to be made and stored, which costs energy.

Every cell pairs an oxidation with a reduction. Electrons leave the electrode where oxidation happens (the negative electrode on discharge) and are taken in where reduction happens. The standard cell potential comes from the electrode potentials:

For a zinc–copper cell this is +0.34 - (-0.76) = 1.10 V.

Commercial cells are judged on voltage, how much charge they store per kilogram, cost and whether they can be recharged safely. Lithium is attractive because it has the most negative standard electrode potential, -3.04 V, and is the lightest metal, so a lot of energy fits in little mass.

In a lithium-ion cell the lithium is never present as the bare metal. On discharge, lithium atoms stored in a graphite negative electrode give up electrons and travel as Li^+ ions through a non-aqueous electrolyte to the positive electrode, typically a cobalt oxide layer, where cobalt is reduced:

Charging reverses both processes. Water cannot be the solvent because a voltage of around 4 V would electrolyse it; organic carbonates are used instead, with a lithium salt such as LiPF_6 dissolved in them.

The lead–acid battery in a petrol car works on the same principles: lead and lead(IV) oxide plates both turn into lead(II) sulfate on discharge in sulfuric acid, and the reverse on charging. It is heavy, but cheap and able to deliver very high currents to a starter motor.

Intercalation

The key idea in lithium-ion cells is that both electrodes are hosts. Lithium ions slide in and out of their structures (intercalation) while the host's framework stays largely intact. Graphite takes up lithium between its graphene sheets to a limiting composition LiC_6. Layered LiCoO_2 has lithium ions between CoO_2 slabs; removing lithium oxidises Co(III) to Co(IV). Other cathodes use the same logic with different frameworks: spinel LiMn_2O_4 with three-dimensional channels and olivine LiFePO_4 with one-dimensional ones.

Capacity and energy

The theoretical specific capacity follows from Faraday's law. For n electrons per formula unit of molar mass M (g mol^{-1}),

the 3.6 converting coulombs to mAh. Worked examples:

• LiFePO_4, M = 157.8: Q = 96\,485/(3.6 \times 157.8) \approx 170 mAh g^{-1}, close to what is achieved in practice.
• LiCoO_2, M = 97.9: Q \approx 274 mAh g^{-1}, but only about half the lithium can be removed before the layered structure becomes unstable, so practical capacity is near 140 mAh g^{-1}.
• Graphite, per C_6 (M = 72.1): Q \approx 372 mAh g^{-1}.

Energy is charge times voltage, E = nFE_{\text{cell}} per mole of reaction. The cell voltage reflects the difference in lithium chemical potential between the electrodes, E_{\text{cell}} = -(\mu_{\text{Li}}^{+} - \mu_{\text{Li}}^{-})/F. LiFePO_4 runs at about 3.4 V against lithium, layered oxides near 3.9 V, so a layered oxide stores more energy per gram but LiFePO_4 is cheaper, cobalt-free and thermally more stable.

The SEI and degradation

Graphite operates at about 0.1 V against Li^+/Li, below the reduction potential of the carbonate solvents. On first charge the electrolyte is reduced, forming a thin solid electrolyte interphase of lithium carbonate, alkyl carbonates and LiF. A good SEI conducts Li^+ but not electrons, so it stops further reduction; ethylene carbonate is included because it forms a particularly good one. Its formation consumes lithium irreversibly, which is why first-cycle efficiency is below 100%.

Ageing comes from slow SEI growth, lithium plating on graphite during fast or cold charging (which can grow dendrites and short the cell), transition-metal dissolution from the cathode, and cracking of cathode particles as they expand and contract. At high temperature, exothermic reactions between the charged electrodes and the electrolyte can feed on themselves: thermal runaway.

Current research pushes on energy density, cost, safety and supply of critical elements together.

On the cathode side, layered oxides have moved from LiCoO_2 towards nickel-rich LiNi_xMn_yCo_zO_2 (NMC811 and beyond) to cut cobalt. Higher nickel brings more capacity but also oxygen release, surface reconstruction to rock-salt phases and intergranular cracking; single-crystal particles and coatings are the main responses. Lithium- and manganese-rich oxides and disordered rock-salts reach higher capacities partly through oxygen (anionic) redox, but voltage fade and hysteresis remain poorly understood. LiFePO_4 has returned strongly for vehicles and grid storage because of cost and safety.

On the anode side, silicon alloys with lithium to about Li_{15}Si_4, giving roughly ten times the specific capacity of graphite, but swells by around 300%, cracking particles and endlessly re-forming SEI. Most commercial cells add only a few percent of silicon. Lithium metal would be better still, which is the main motive for solid-state batteries: garnet Li_7La_3Zr_2O_{12} and sulfide conductors such as argyrodite Li_6PS_5Cl reach room-temperature ionic conductivities comparable to liquids. Dendrite growth along grain boundaries, interfacial contact loss and the stack pressure needed are unsolved engineering problems.

Sodium-ion cells, using hard carbon anodes and layered oxide or Prussian-blue-analogue cathodes, entered commercial production in the 2020s for low-cost applications.

Methods matter as much as materials. Operando X-ray diffraction, solid-state NMR of ^7Li, cryo-electron microscopy of the SEI and differential electrochemical mass spectrometry follow processes as they happen. Data-driven models predict cycle life from the first hundred cycles, and DFT screening of voltages and diffusion barriers guides which compositions are worth making.