Nickel, Cobalt, Manganese, Iron and Aluminium
Nickel supplies electrons, cobalt supplies structure, manganese supplies cheapness with a Jahn–Teller liability, iron supplies abundance — and aluminium supplies nothing at all, which is the point.
The Periodic Table of the EV · Part 2 — The Cathode · Chapter 4 · 23 min read
Every cathode metal is chosen for one property and tolerated for its defect. Nickel brings capacity and cation mixing. Cobalt brings structure and a supply chain nobody wants. Manganese brings cheapness and a lattice distortion that cracks particles. Iron brings abundance and a voltage too low to use.
NMC is not a compromise between three good options. It is an attempt to get one property from each element while diluting all three defects.
69 vs 76 pm
Ni²⁺ against Li⁺ — the mixing problem
~6.5 %
Volume change from Jahn–Teller in LMO
2,250×
Iron’s abundance over cobalt
0 mAh/g
Aluminium’s contribution to NCA
4.1 — Nickel — the capacity element
Nickel accesses Ni²⁺, Ni³⁺ and Ni⁴⁺, meaning it can deliver close to two electrons per metal atom. More electrons per unit mass means more capacity, which is why cathodes went NMC111 → 532 → 622 → 811, lifting usable capacity from about 150 to over 200 mAh/g.
4.1.1 — The problem — cation mixing
Ni²⁺ has an ionic radius of 69 pm. Li⁺ is 76 pm. They are nearly identical, and that similarity is expensive.
A layered oxide is a stack of alternating sheets — lithium layer, metal-oxide layer, lithium layer. For that to work, the metal must strongly prefer the metal site. Because Ni²⁺ and Li⁺ are the same size, nickel readily occupies lithium sites instead, forming an antisite defect.
- •Nickel sitting in the lithium layer physically blocks the diffusion pathway — costing both capacity and rate capability.
- •It pins the layers together, reducing interlayer spacing.
- •The effect worsens with nickel content and with synthesis temperature, so the problem grows with exactly the change that was supposed to help.
4.1.2 — The second problem — moisture
Nickel-rich cathodes are hygroscopic. They react with atmospheric moisture and CO₂ to form LiOH and Li₂CO₃ on the surface, which causes gassing in the cell and gelation of the coating slurry in manufacturing. This is why NMC811 production needs dry-room handling that LFP simply does not.
4.2 — Cobalt — the structural element
Low-spin Co³⁺ (t₂g⁶) is exceptionally well suited to octahedral sites in a layered oxide, and it brings four things at once:
- •A strong site preference energy for the octahedral metal site — so almost no cation mixing with lithium.
- •A small, rigid ionic radius of 54.5 pm — so stable layer spacing.
- •Good electronic conductivity through the delocalised t₂g band.
- •Fast Li⁺ diffusion, as a consequence of the well-ordered structure the first three produce.
Important
In NMC, cobalt’s job is not primarily capacity. It is suppressing cation mixing and enabling rate capability. It is the structural glue — which is precisely why removing it is so hard. You are not replacing one property, you are replacing four.
4.2.1 — Why the industry wants less of it
- •Crustal abundance of 25 ppm, and mostly present as a byproduct of nickel and copper mining rather than as a primary ore.
- •Roughly 70 per cent of supply comes from the DRC, with serious artisanal mining and human-rights problems.
- •Highly price-volatile, which makes cell costing difficult years ahead of production.
- •Toxic.
- •At high delithiation, beyond about Li₀.₅CoO₂, the Co⁴⁺ d band overlaps the O 2p band and oxygen is released — the thermal-stability weak point described in chapter 3.
4.3 — Manganese — the cheap stabiliser with a defect
Manganese is abundant at 950 ppm — 38 times cobalt — cheap and non-toxic, and Mn⁴⁺ (t₂g³) is an exceptionally stable, structurally inert configuration. In NMC, manganese is largely electrochemically inactive, sitting as Mn⁴⁺ throughout and acting as a structural pillar while nickel does the redox work.
4.3.1 — Problem one — the Jahn–Teller distortion
The Jahn–Teller theorem states that any non-linear complex with a degenerate electronic ground state will spontaneously distort to remove that degeneracy and lower its energy.
High-spin Mn³⁺ is d⁴: t₂g³ e_g¹. That single electron in the doubly degenerate e_g set is the trigger. The octahedron elongates along one axis, typically by about 10 per cent, splitting the e_g levels and stabilising the occupied one.
Why this matters
In LiMn₂O₄ spinel, deep discharge produces Mn³⁺, and the resulting cooperative distortion changes the crystal symmetry from cubic to tetragonal with roughly a 6.5 per cent volume change. Repeated cycling through that transition cracks particles, and that is the notorious capacity fade of LMO.
4.3.2 — Problem two — dissolution by disproportionation
2 Mn³⁺(solid) → Mn⁴⁺(solid) + Mn²⁺(dissolved)
Mn²⁺ is soluble in the electrolyte. It migrates to the anode, deposits there, and catalytically destroys the SEI — which then reforms, consuming more lithium inventory each time. The reaction is accelerated by trace HF from LiPF₆ hydrolysis and by temperature.
In plain English
This is the mechanism behind a specific, familiar field observation: manganese-based packs age badly in hot climates. A chemistry that is merely mediocre through a European winter degrades quickly through an Indian summer, and the reason is a temperature-accelerated dissolution reaction rather than anything about charging habits.
4.4 — Iron — the abundance element
Iron is the ideal cathode metal on every economic and environmental axis: crustal abundance of 56,300 ppm — 2,250 times cobalt — non-toxic, cheap and thermally stable.
So why is LiFeO₂ not a product? Two reasons. The Fe²⁺/Fe³⁺ redox couple in a simple oxide sits at only about 3.0 V vs Li/Li⁺, too low to be interesting. And layered LiFeO₂ suffers severe cation mixing, converting to a disordered rock-salt structure that does not cycle.
Technical framing
The solution is not to fix iron. It is to change what surrounds it, using an element that contributes no capacity at all. That is chapter 5, and it is the most elegant application of periodic-table reasoning in battery materials science.
4.5 — Aluminium in NCA — the inert dopant
Al³⁺ has a fixed oxidation state: a closed [Ne] shell, no accessible d electrons. It contributes zero capacity.
That is precisely the point. Substituting a few per cent of aluminium for nickel in NCA:
- •Pins the lattice. The strong, short, highly ionic Al–O bond — Al³⁺ is 53.5 pm with high charge density — resists the layer collapse that occurs at high delithiation.
- •Raises the onset temperature of oxygen release and thermal runaway.
- •Suppresses phase transitions during deep charge.
- •Adds essentially no mass, at M = 27.
Titanium, magnesium and zirconium are used in the same way. When you read that a cathode is “cobalt-free,” what has usually replaced cobalt is a dopant package doing cobalt’s structural job by other means.
4.6 — Why NMC is a mixture at all
| Element | What it contributes | What it costs you |
|---|---|---|
| Nickel | Capacity — close to two electrons per atom | Cation mixing with lithium, moisture sensitivity, oxygen release at high charge |
| Cobalt | Structure — site ordering, rate capability, conductivity | Price volatility, 25 ppm abundance, DRC supply concentration, toxicity |
| Manganese | Cost and structural stability as inert Mn⁴⁺ | Jahn–Teller distortion and dissolution wherever Mn³⁺ appears |
| Aluminium (NCA) | Lattice pinning and higher thermal onset | Nothing except the capacity of the site it occupies |
Important
Read that table as a single sentence: no element supplies more than one of the properties a cathode needs, so every commercial cathode is a mixture whose ratios are an argument about which defect you would rather have.
Quick check: test yourself
1.A cell using LMO loses capacity much faster in Chennai than in Manchester. Give the mechanism.
Show answer
2.Why does raising nickel content require a dry room?
Show answer
3.What four things would a cobalt substitute have to replace?
Show answer
Chapter summary
- ✓Ni²⁺ at 69 pm is so close to Li⁺ at 76 pm that it invades the lithium layer, blocking diffusion — and the problem grows with exactly the nickel content that raises capacity.
- ✓Cobalt’s value is structural, not capacitive: site ordering, rigidity, conductivity and rate. Replacing it means replacing four properties at once.
- ✓Manganese is safe and cheap as inert Mn⁴⁺ and ruinous as Mn³⁺ — Jahn–Teller distortion cracks particles, and disproportionation dissolves manganese into the electrolyte.
- ✓Iron is ideal on cost and abundance but its couple sits at only ~3.0 V in a simple oxide, and layered LiFeO₂ disorders into rock-salt.
- ✓Aluminium in NCA contributes zero capacity by design — it pins the lattice, raises the thermal onset and adds almost no mass.
Frequently asked questions
What is cation mixing and why does it get worse with more nickel?+
A layered oxide is alternating sheets of lithium and metal oxide, which only works if the metal strongly prefers the metal site. Ni²⁺ has an ionic radius of 69 pm and Li⁺ is 76 pm — nearly identical — so nickel readily occupies lithium sites as an antisite defect. Nickel sitting in the lithium layer physically blocks the diffusion pathway, costing capacity and rate, and pins the layers together, reducing interlayer spacing. The effect worsens with nickel content and with synthesis temperature, which is the structural price of going from NMC111 to 811.
What is the Jahn–Teller distortion and why does it damage LMO?+
The Jahn–Teller theorem says any non-linear complex with a degenerate electronic ground state will spontaneously distort to remove that degeneracy. High-spin Mn³⁺ is d⁴ — t₂g³ e_g¹ — and that single electron in the doubly degenerate e_g set is the trigger, so the octahedron elongates about 10 per cent along one axis. In LiMn₂O₄ spinel, deep discharge produces Mn³⁺ and the cooperative distortion changes the crystal from cubic to tetragonal with roughly a 6.5 per cent volume change. Repeated cycling through that transition cracks particles, which is the notorious capacity fade of LMO.
Why do manganese-based packs age badly in hot climates?+
Because of disproportionation: 2 Mn³⁺(solid) → Mn⁴⁺(solid) + Mn²⁺(dissolved). Mn²⁺ is soluble in the electrolyte, migrates to the anode, deposits there and catalytically destroys the SEI — which then reforms, consuming more lithium inventory. The reaction is accelerated by temperature and by trace HF from LiPF₆ hydrolysis, so a chemistry that is merely mediocre in a temperate climate degrades quickly in Indian summer conditions.
Why is cobalt so hard to remove from cathodes?+
Because its job is structural, not capacitive. Low-spin Co³⁺ has a strong site preference energy for the octahedral metal site, so it barely mixes with lithium; it is small and rigid at 54.5 pm, keeping layer spacing stable; it conducts electrons well through the delocalised t₂g band; and the well-ordered structure it produces gives fast lithium diffusion. In NMC, cobalt suppresses cation mixing and enables rate capability. Replacing it means replacing four properties at once, which is why the substitutes are dopant packages of aluminium, magnesium, titanium and zirconium rather than a single element.
Why does NCA contain aluminium if aluminium stores no charge?+
That is precisely why it works. Al³⁺ has a fixed oxidation state — a closed [Ne] shell with no accessible d electrons — so it contributes zero capacity. What it contributes is structure: the strong, short, highly ionic Al–O bond pins the lattice against the layer collapse that occurs at high delithiation, raises the onset temperature of oxygen release and thermal runaway, suppresses phase transitions during deep charge, and adds almost no mass at M = 27. Titanium, magnesium and zirconium are used the same way.
Reviewed by
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.