All chapters
NucleosynthesisGoldschmidt classificationCobalt byproduct problemLFP cost advantageLong-run forecasting

Cost looks like an economic variable. Over a long enough horizon it is a periodic one — because in a mature commodity industry, material cost converges toward extraction cost, which converges toward crustal abundance and ore concentration.

Read down the abundance column and the trajectory of the battery industry over the next two decades is legible in advance. Most of it has already started.

12.1Nucleosynthesis sets the baseline

Elemental abundance is not arbitrary. It reflects how the elements were made in stars.

  • Light elements up to iron form by fusion, which releases energy. Iron-56 has the highest binding energy per nucleon, so fusion stops there — which is why iron is abundant and everything heavier is rare.
  • Elements heavier than iron require energy input — supernovae and neutron-star mergers — so they are orders of magnitude scarcer.
  • Lithium, beryllium and boron are anomalously rare for their low atomic number, because they are destroyed in stellar interiors rather than made there. Most lithium comes from Big Bang nucleosynthesis and cosmic-ray spallation.

Important

Lithium’s scarcity is written into the physics of stars. At 20 ppm it is rarer than nickel at 84 and copper at 60, despite being the third element in the periodic table. No mining innovation changes that baseline.

10^-410^-310^-210^-110^010^110^210^310^410^5O461000Si282000Al82300Fe56300Ca41500Na23600Mg23300K20900Ti5650P1050Mn950F585C200Ni84Cu60Nd28Co25Li20Dy5.2Pt0.005Ir0.0004CRUSTAL ABUNDANCE / ppm — LOGARITHMICFe is 2,250× more abundant than Co
Figure 12.1Elemental abundance is written into the physics of stars. Fusion releases energy only up to iron-56, which has the highest binding energy per nucleon — so iron is abundant and everything heavier is rare. Lithium, beryllium and boron are anomalously scarce for such light elements because they are destroyed in stellar interiors rather than made there.

12.2Goldschmidt — where elements ended up on Earth

The Goldschmidt classification sorts elements by the phase they preferentially entered during Earth’s differentiation — and it explains the industry’s supply structure more directly than any market analysis.

ClassMeaningWhere they wentEV-relevant members
Lithophile“Rock-loving”Bond with oxygen; concentrated in the silicate crustLi, Na, K, Mg, Ca, Al, Si, Ti, Mn, the rare earths
Chalcophile“Ore-loving”Prefer sulfur; found in sulfide oresCu, Zn, Pb, and partly Ni and Co
Siderophile“Iron-loving”Dissolved in metallic iron and sank into the coreNi, Co, and the platinum group
AtmophileVolatileIn the atmosphereH, N, noble gases

12.2.1What that explains

  • Lithium is lithophile — concentrated in crustal pegmatites (spodumene) and evaporitic brines. Reachable, but geologically localised, which is the origin of the Australia/Chile/Argentina concentration.
  • Nickel and cobalt are siderophile and chalcophile — most of Earth’s inventory is in the core, unreachable. Crustal deposits are sulfide and laterite ores.
  • Iron, manganese, aluminium and phosphorus are abundant lithophiles — which means LFP’s cost advantage is geochemical destiny rather than a manufacturing achievement.
  • Platinum and iridium are siderophile — 5 ppb and 0.4 ppb crustal abundance. This is why hydrogen fuel cells face a materials ceiling that batteries do not.

Why this matters

The cobalt point deserves emphasis, because it is routinely misunderstood as a political problem. Cobalt is almost always a byproduct of copper or nickel mining rather than the primary target.

That means you cannot scale cobalt production independently of copper demand. It is a structural, geochemical constraint on supply elasticity that no amount of investment fixes quickly — and it is why designing cobalt out proved faster than mining more of it.

CRUST — LITHOPHILEbond with oxygenMANTLECORESIDEROPHILENi · Co · Pt · IrLiNaKCaMgAlSiTiMnPFNdFeDyYCuZnPbchalcophileWHAT THIS EXPLAINS ABOUT SUPPLYLi is lithophile→ concentrated in crustal pegmatites and brines.Reachable, but geologically localised.Ni and Co are siderophile→ most of Earth's inventory sank into the core.What's left is sulfide and laterite ore, and cobaltis almost always a byproduct of Cu or Ni mining.You cannot scale cobalt independently of copper demand.Fe, Mn, Al, P are abundant lithophiles→ LFP's cost advantage is geochemical destiny,not a manufacturing achievement.Pt and Ir are siderophile→ 5 ppb and 0.4 ppb. Hydrogen fuel cells face amaterials ceiling that batteries do not.
Figure 12.2Goldschmidt sorted the elements by the phase they entered during Earth's differentiation — lithophile with the silicate crust, chalcophile with sulfides, siderophile into the metallic core. That 4.5-billion-year-old sorting event is why cobalt is a byproduct and lithium is a primary ore, and therefore why one supply chain is elastic and the other is not.

12.3The abundance table

Crustal abundance, class and EV role

ElementAbundance (ppm)ClassEV role
Oxygen461,000Cathode framework
Silicon282,000LithophilePower electronics, anode additive
Aluminium82,300LithophileCollector, structure, NCA dopant
Iron56,300Lithophile/siderophileLFP, steel, magnets
Calcium41,500LithophileCandidate chemistry
Sodium23,600LithophileNa-ion
Magnesium23,300LithophileCandidate, alloys
Potassium20,900LithophileCandidate
Titanium5,650LithophileLTO, dopant
Phosphorus1,050LithophileLFP, electrolyte salt
Manganese950LithophileNMC, LMO
Fluorine585LithophileSalt, binder, SEI
Carbon200Anode, additive, binder
Nickel84Siderophile/chalcophileNMC, NCA
Copper60ChalcophileCollector, windings
Neodymium28LithophileMagnets
Cobalt25Siderophile/chalcophileNMC, NCA
Lithium20LithophileThe charge carrier
Dysprosium5.2LithophileMagnet coercivity
Platinum0.005SiderophileFuel cells
Iridium0.0004SiderophileElectrolysers

12.4What this predicts about the future

  • LFP gains share. Iron, phosphorus and manganese are three to four orders of magnitude more abundant than cobalt. LFP was always going to win the cost-sensitive segment; the only question was whether nano-engineering could fix its conductivity and 1-D diffusion problems. Chapter 9 showed that it did.
  • Cobalt is engineered out. The direction NMC111 → 811 → cobalt-free is set by geochemistry and executed by crystal chemistry, using Al, Mg, Ti and Zr dopants to replace cobalt’s structural role.
  • Sodium-ion is inevitable somewhere. A thousandfold abundance advantage will find an application — most likely stationary storage and cold-climate or low-cost mobility, where its energy-density deficit is tolerable.
  • Manganese-rich chemistries are the live compromise. Manganese is 38 times more abundant than cobalt, and LMFP raises LFP’s voltage from 3.45 V toward 3.8–4.1 V using the same inductive-effect trick from chapter 5 with a manganese redox centre.
  • Rare-earth-free motors get sustained investment. Dysprosium at 5.2 ppm cannot support a billion-vehicle fleet, whatever the price.

In plain English

The general principle: the periodic table is a long-run price forecast. Chemistries built on abundant elements win the volume segments. Chemistries built on scarce elements retreat to applications that can pay for performance. Neither outcome requires anyone to make a good decision — it happens through cost.

Technical framing

One caveat worth stating. Abundance sets the floor, not the price today. Extraction cost also depends on ore grade, on whether an element is a primary product or a byproduct, on refining capacity, and on where processing is located — which is why lithium at 20 ppm and cobalt at 25 ppm behave completely differently in a market, and why refining capacity has recently mattered more than reserves.

Quick check: test yourself

1.Why does high cobalt demand not simply produce more cobalt mines?

Show answer
Because cobalt is siderophile and chalcophile — most of Earth’s inventory is in the core, and crustal cobalt is almost always a byproduct of copper and nickel mining rather than a primary target. Output is therefore tied to copper and nickel demand, a geochemical constraint on supply elasticity that capital does not quickly relieve.

2.LFP is cheaper than NMC. Is that a manufacturing achievement or something else?

Show answer
Mostly geochemical destiny. Iron at 56,300 ppm, phosphorus at 1,050 and manganese at 950 versus cobalt at 25 — three to four orders of magnitude. The manufacturing achievement was fixing LFP’s conductivity and 1-D diffusion through carbon coating and nano-sizing, which is what allowed the geochemical advantage to be collected.

3.Lithium is 20 ppm and cobalt is 25 ppm — nearly identical. Why do they behave so differently in the market?

Show answer
Because abundance sets a floor, not a price. Lithium is lithophile and concentrated into workable pegmatite and brine deposits; cobalt is siderophile, thinly spread, and produced as a byproduct. Ore grade, byproduct status, refining capacity and processing location all sit between crustal abundance and the invoice.

Chapter summary

Frequently asked questions

Why is lithium scarce despite being the third element?+

Because it is destroyed in stars rather than made in them. Elements up to iron form by fusion, which releases energy, and iron-56 has the highest binding energy per nucleon so fusion stops there — which is why iron is abundant and everything heavier is rare. But lithium, beryllium and boron are anomalously rare for their low atomic number because stellar interiors consume them; most lithium comes from Big Bang nucleosynthesis and cosmic-ray spallation. At 20 ppm, lithium is rarer than nickel at 84 and copper at 60. Its scarcity is written into the physics of stars.

Why can’t cobalt production simply be scaled up?+

Because of where it sits in the Goldschmidt classification. Cobalt is siderophile and chalcophile, meaning most of Earth’s inventory dissolved into metallic iron and sank into the core during differentiation, leaving only sulfide and laterite crustal deposits. Worse, cobalt is almost always a byproduct of copper or nickel mining rather than the primary target, so you cannot scale cobalt output independently of copper demand. That is a structural geochemical constraint on supply elasticity, not a shortage of investment, and no amount of capital fixes it quickly.

Is LFP’s cost advantage permanent?+

Largely, because it is geochemical rather than industrial. Iron at 56,300 ppm, phosphorus at 1,050 and manganese at 950 are three to four orders of magnitude more abundant than cobalt at 25. In a mature commodity industry, material cost converges toward extraction cost, which converges toward crustal abundance and ore concentration. LFP was always going to win the cost-sensitive segment; the only real question was whether nano-engineering could fix its conductivity and one-dimensional diffusion problems, and it did.

What does the abundance table predict about future chemistries?+

Several things that are already happening. Cobalt gets engineered out, with crystal chemistry using aluminium, magnesium, titanium and zirconium dopants to replace cobalt’s structural role. Manganese-rich chemistries such as LMFP and LMR become the compromise, since manganese is 38 times more abundant than cobalt and LMFP lifts LFP’s 3.45 V toward 3.8 to 4.1 V using the same inductive-effect trick with a manganese redox centre. Sodium-ion finds an application somewhere, because a thousandfold abundance advantage always does. And rare-earth-free motors get sustained investment, because dysprosium at 5.2 ppm cannot support a billion-vehicle fleet.

Reviewed by

SG

Sahil Goyal

Co-founder, Wingzman

LinkedIn
SG

Sourabh Goyal

Co-founder, Wingzman

LinkedIn

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.