How Diamonds Form Deep Inside the Earth

One element, two crystals, one very different fate

Diamond and graphite are both pure carbon — chemically identical, C — yet they are structurally worlds apart. The difference is how each carbon atom hybridizes and bonds:

• Graphite uses sp² hybridization: each carbon bonds to three neighbors in flat hexagonal sheets, with the fourth valence electron delocalized in a π system spread across the sheet. The sheets are held together only by weak van der Waals / dispersion forces (~interlayer spacing 335 pm), which is why graphite cleaves, conducts electricity, and lubricates.
• Diamond uses sp³ hybridization: every carbon bonds to four others at 109.5° in a rigid three-dimensional lattice, with a C–C bond length of 154 pm and a bond enthalpy near 346 kJ/mol. There are no weak planes and no free electrons, so diamond is an electrical insulator, optically transparent, and mechanically the hardest natural solid (Mohs 10).

Because the tetrahedral lattice packs carbon more tightly, diamond is much denser: 3.51 g/cm³ versus graphite's 2.27 g/cm³, corresponding to molar volumes of 3.42 vs 5.30 cm³/mol. That volume difference — diamond is smaller — is the entire reason pressure favors diamond, as we'll see next.

Why pressure, and only pressure, makes diamond stable

At room conditions, the conversion

• C(graphite) → C(diamond)  ΔG° ≈ +2.9 kJ/mol, ΔH° ≈ +1.9 kJ/mol

has a positive Gibbs free energy, so graphite is the stable phase and diamond is only metastable. To flip the sign of ΔG, we exploit the pressure term. At constant temperature, the pressure dependence of the reaction's free energy is governed by the volume change ΔV:

• (∂ΔG/∂P)ₜ = ΔV = V(diamond) − V(graphite) = 3.42 − 5.30 = −1.88 cm³/mol

Because ΔV is negative, squeezing the system lowers the free energy of the reaction. Setting ΔG(P) = ΔG° + ΔV·P = 0 gives an equilibrium pressure of roughly 1.5 GPa at 298 K. This is a beautiful, textbook illustration of Le Chatelier's principle: raise the pressure and the system shifts toward the phase that occupies less volume — the denser diamond.

But there's a catch of temperature. The full graphite–diamond phase boundary (the Berman–Simon line) slopes upward: it takes more pressure to stabilize diamond as things get hotter. At mantle temperatures near 1,100–1,400 °C, the boundary sits at about 4–5 GPa. That combination — high pressure and high temperature — is only found together at depths of roughly 140–200 km in the mantle.

The diamond stability field: cold cratonic roots

Pressure in the Earth increases with depth at about 0.033 GPa per km (~330 bar/km of overburden), so 5 GPa corresponds to roughly 150 km down. But pressure alone isn't enough — the rock must also be cool enough to stay below the phase boundary rather than melting or drifting into graphite's field. That's why virtually all natural diamonds come from beneath cratons, the ancient, thick, tectonically dead cores of continents.

Cratons have deep, cold mantle keels — think of an iceberg's submerged bulk — that reach 200+ km down while staying relatively cold (a geothermal gradient of only ~40 mW/m² surface heat flow). The intersection of that cold geotherm with the diamond stability field creates a wedge of pressure–temperature space, the "diamond window," typically:

• Depth ≈ 150–200 km (some "superdeep" diamonds far deeper)
• Pressure ≈ 4.5–6 GPa
• Temperature ≈ 900–1,400 °C

Under a young, hot orogenic belt the geotherm is too warm and the same depth yields graphite or melt instead. This is why diamond exploration follows craton maps: the chemistry demands old, cold, thick lithosphere.

Where the carbon comes from — a redox story

Getting into the diamond stability field is necessary but not sufficient; you also need a source of carbon and a chemical reason for it to crystallize as a solid. In the mantle, carbon is dissolved in fluids and melts, and whether it precipitates as diamond depends on oxygen fugacity (fO₂) — essentially the redox state of the surroundings. Two main pathways operate:

• Reduction of oxidized carbon. Carbonate or CO₂-bearing fluids can be reduced to elemental carbon by ferrous-iron-bearing mantle minerals. A representative reaction near the enstatite–magnesite–olivine–diamond (EMOD) buffer is:
  MgCO₃ (magnesite) + MgSiO₃ (enstatite) → Mg₂SiO₄ (forsterite) + C (diamond) + O₂
• Oxidation of reduced carbon. In more reducing domains, methane-bearing fluids drop their hydrogen and deposit carbon:
  CH₄ → C (diamond) + 2 H₂

A closely related and widely cited mechanism is redox freezing/melting, where a moving fluid crosses the iron–wüstite (or Fe³⁺/Fe²⁺) redox front and carbon simply falls out of solution. Isotopes back this up: most gem diamonds cluster around a mantle carbon signature of δ¹³C ≈ −5 ‰, while some show light values down to −25 ‰ or below — a fingerprint of ancient subducted organic carbon, meaning your diamond may contain carbon that was once part of Precambrian microbial life, dragged down at a subduction zone and re-crystallized.

The elevator: kimberlite and a race against graphitization

Diamonds form and then sit in the deep lithosphere for hundreds of millions to billions of years. Getting one to the surface requires an eruption fast and cold enough not to destroy it — because once diamond leaves its stability field, the reaction diamond → graphite becomes thermodynamically favorable again. The only saving grace is kinetics: breaking every C–C bond to reorganize the lattice has a huge activation energy, so at low temperature the back-reaction is effectively frozen (that's why your diamond isn't quietly turning to pencil lead).

The delivery vehicle is kimberlite (and rarer lamproite) — a volatile-rich, CO₂- and H₂O-charged ultramafic magma that rises from >150 km. Key facts:

• Ascent is astonishingly fast: estimates run from a few to tens of km/h, with the final approach to the surface possibly hours to days, driven by exsolving CO₂ gas.
• Rapid rise means diamonds spend little time at high temperature outside their stability field, so graphitization is quenched.
• The diamonds are xenocrysts — passengers, not products of the kimberlite. Radiometric dating of mineral inclusions gives diamond ages of 1–3.5 Gyr, far older than the kimberlite eruptions (often < 100 Myr), proving they were picked up en route.

Slow-rising or stalled magma cooks its diamond cargo into graphite or dissolves ("resorbs") it. A successful diamond pipe is therefore a record of a rare, near-supersonic burst of deep magma.

Superdeep diamonds and windows into the deep Earth

A small but scientifically priceless population of superdeep diamonds forms far below the lithospheric keel — in the mantle transition zone (410–660 km) and even the lower mantle (below 660 km), where pressures exceed 24 GPa. We know this because diamonds are perfect time capsules: their rigid lattice traps mineral inclusions that survive the trip up. Famous examples include:

• Ringwoodite — a high-pressure polymorph of olivine found in a diamond in 2014 containing ~1.5 wt% H₂O, direct evidence of a wet deep mantle.
• Bridgmanite / ferropericlase — lower-mantle minerals confirming diamonds sourced below 660 km.
• Ice-VII — a high-pressure form of water, and CaSiO₃-perovskite (davemaoite), both first found as natural samples inside diamonds.

These inclusions make diamonds among the most important geological samples we have: they are essentially free, pristine deliveries of matter from depths no drill can reach (the deepest borehole, Kola, reached only ~12 km). The chemistry of diamond formation thus doubles as a sampling technique for a planet we cannot otherwise touch.

From nature to the lab — and why it matters

Humans learned to reproduce the deep-Earth recipe in the 1950s. Two industrial routes now dominate, and together they dwarf mined output for technical uses:

• HPHT (High-Pressure High-Temperature) — literally a bench-top mantle: a metal-solvent catalyst (Fe, Ni, Co) dissolves carbon at ~5–6 GPa and 1,300–1,600 °C, and diamond crystallizes on a seed, exactly mimicking the natural stability field.
• CVD (Chemical Vapor Deposition) — a clever kinetic trick that bypasses high pressure entirely. A hydrogen-methane plasma at low pressure grows diamond atom by atom:
  CH₄ → C(diamond) + 2 H₂
  Atomic hydrogen selectively etches away any sp² graphite that tries to form, so diamond grows metastably even though graphite is the stable bulk phase.

Why does any of this matter beyond jewelry? Diamond's extreme properties — hardest known material, highest thermal conductivity of any bulk solid (~2,200 W/m·K, five times copper), optical transparency from UV to IR, and chemical inertness — make it indispensable in cutting tools, high-power electronics, quantum sensors (nitrogen-vacancy centers), and pressure research (diamond-anvil cells that recreate mantle and core conditions). And scientifically, natural diamonds remain irreplaceable: they carry the isotopic and mineralogical fingerprints of subducted carbon, deep water, and the redox architecture of the mantle — a chemical archive of how carbon cycles through the whole solid Earth over billions of years.

Two allotropes of pure carbon: why graphite rules the surface and diamond rules the deep mantle.

Property	Graphite	Diamond
Bonding	sp² sheets, delocalized π	sp³ tetrahedra, 154 pm bonds
Density	2.27 g/cm³	3.51 g/cm³
Molar volume	5.30 cm³/mol	3.42 cm³/mol
ΔG°f at 298 K, 1 bar	0 (reference)	+2.9 kJ/mol (metastable)
Stable field	Low P (< ~1.5–2 GPa)	High P (> ~4 GPa at mantle T)
Mohs hardness	1–2	10