The Kroll Process

What the Kroll process does

Titanium is the ninth most abundant element in Earth's crust, yet a bicycle frame made of it costs more than one made of steel. The reason is entirely locked up in how hard it is to reduce the ore to metal. Titanium bonds ferociously to oxygen, nitrogen, and carbon, so the obvious routes — smelting the oxide with carbon, or electrolyzing molten ore — either give a ceramic (TiC, TiN) or never deposit clean metal. The Kroll process is the workaround the entire titanium industry still runs on.

It is a three-act play, and the middle act is the reaction that carries Kroll's name:

1. Chlorination. Rutile ore (TiO₂) is heated with coke and chlorine gas at ~900 °C to make titanium tetrachloride, a colorless volatile liquid (bp 136 °C). Turning the oxide into a chloride is the crucial move: TiCl₄ can be distilled, so it can be purified to a level no solid ore ever could.
2. Magnesium reduction (the Kroll step). Purified TiCl₄ vapor is dripped onto molten magnesium in a sealed steel retort under an inert argon blanket at ~800-900 °C. The magnesium tears all four chlorides off titanium: TiCl₄ + 2 Mg → Ti + 2 MgCl₂. The metal grows as a porous titanium sponge.
3. Purification & recycling. Trapped MgCl₂ and excess Mg are boiled out of the sponge by vacuum distillation. The MgCl₂ is electrolyzed back into magnesium and chlorine, both of which are fed to the front of the plant. The sponge is crushed and vacuum-arc-remelted into an ingot.

Everything downstream depends on step 2 running in a bone-dry, oxygen-free vessel. A trace of O₂ or N₂ at 900 °C dissolves straight into the growing titanium as interstitial impurity, and even a few hundred parts per million of oxygen turns ductile titanium brittle.

The reduction, electron by electron

At its core the Kroll step is a redox reaction: magnesium is oxidized, titanium is reduced. Follow the electrons.

In TiCl₄, titanium sits in the +4 oxidation state — it has handed four electrons to four chlorine atoms. Magnesium metal is 0. The overall transformation moves four electrons from two magnesium atoms onto one titanium atom:

  oxidation (each Mg loses 2 e⁻):   2 Mg  →  2 Mg²⁺  +  4 e⁻
  reduction (Ti⁴⁺ gains 4 e⁻):      Ti⁴⁺  +  4 e⁻  →  Ti⁰
  ─────────────────────────────────────────────────────────
  net:                              TiCl₄  +  2 Mg  →  Ti  +  2 MgCl₂

But titanium doesn't jump from +4 to 0 in a single leap. The reduction climbs down the oxidation-state ladder through lower chlorides, and each rung is a distinct, observable intermediate that colors the melt:

  TiCl₄  ──Mg──▶  TiCl₃  ──Mg──▶  TiCl₂  ──Mg──▶  Ti⁰
   Ti(+4)          Ti(+3)          Ti(+2)          Ti(0)
  colorless        violet           black         metal sponge

• Ti(+4) → Ti(+3). The first electron transfer gives TiCl₃ (violet). This is fast and exothermic; it is why the reaction, once lit, is self-sustaining and must be temperature-controlled.
• Ti(+3) → Ti(+2). A second electron gives TiCl₂ (black), which dissolves in the molten MgCl₂ salt bath and shuttles titanium to fresh magnesium surfaces.
• Ti(+2) → Ti(0). The last two electrons deposit metallic titanium atoms, which nucleate and grow into interlocking dendritic crystals — the sponge.

Because the whole thing happens ~750 °C below titanium's 1668 °C melting point, the metal never pools. It crystallizes in place around the magnesium, threaded with pores full of molten MgCl₂. The reaction is heterogeneous and diffusion-limited: titanium species must travel through the salt to reach magnesium, which is exactly why a single batch takes days, not minutes.

Reagents, conditions, and the argon blanket

The Kroll reduction looks simple on paper. Everything that makes it hard is in the operating envelope.

• Feedstock: TiCl₄, distilled to remove FeCl₃, SiCl₄, VOCl₃ and other volatile chloride impurities. Vanadium is the tricky one — VCl₄ boils close to TiCl₄, so it is reduced out with H₂S or copper wire before use.
• Reductant: molten magnesium, held above its 650 °C melting point. Magnesium is charged in slight excess so no unreacted TiCl₄ escapes.
• Temperature: ~800-900 °C. Hot enough to keep Mg and MgCl₂ molten and the kinetics brisk; cool enough that titanium stays solid and doesn't weld to the steel retort.
• Atmosphere: high-purity argon (not nitrogen — titanium reacts with N₂ to form TiN). The retort is a sealed, leak-tight steel vessel. Oxygen ingress is the cardinal sin.
• Vessel: a mild-steel or stainless retort. Titanium slowly attacks steel, contaminating the sponge with iron near the wall, so the reactor is a consumable of limited lifetime.
• Duration: a full cycle — charge, react, cool, distill, break out sponge — runs several days and yields on the order of a few to ~10 tonnes per reactor.

Note that the magnesium is not a catalyst — it is a stoichiometric reagent, fully consumed to MgCl₂. Calling it "the reductant" is exact. What makes the economics survivable is that the MgCl₂ is not waste; it is the feedstock for the next batch's magnesium.

Worked example: the mass and energy of one tonne

How much magnesium does it take to make a tonne of titanium sponge? Run the stoichiometry off the balanced equation.

    TiCl₄  +  2 Mg  →  Ti  +  2 MgCl₂
   M(Ti) = 47.87 g/mol   M(Mg) = 24.31 g/mol

• Moles of Ti in 1 tonne: 1,000,000 g ÷ 47.87 g/mol ≈ 20,890 mol Ti.
• Magnesium required: 2 mol Mg per mol Ti → 41,780 mol × 24.31 g/mol ≈ 1,016 kg Mg (about 1 tonne of magnesium per tonne of titanium).
• MgCl₂ produced: 41,780 mol × 95.21 g/mol ≈ 3,978 kg of magnesium chloride to recycle.
• TiCl₄ consumed: 20,890 mol × 189.68 g/mol ≈ 3,963 kg of titanium tetrachloride.

So every tonne of titanium metal is quite literally built from ~4 tonnes of TiCl₄ and a tonne of magnesium, generating ~4 tonnes of MgCl₂ that must be electrolyzed. The electrolysis to regenerate that magnesium consumes on the order of 12-18 kWh per kg of Mg — a huge slab of the total energy bill, and the reason titanium's embodied energy (~360-750 MJ/kg) dwarfs steel's (~25 MJ/kg).

Kroll vs. the alternatives

	Kroll process	Hunter process	Hall-Héroult (aluminium)
Metal produced	Titanium	Titanium	Aluminium
Reductant	Mg (magnesium)	Na (sodium)	Electrons (electrolysis)
Key reaction	TiCl₄ + 2 Mg → Ti + 2 MgCl₂	TiCl₄ + 4 Na → Ti + 4 NaCl	2 Al₂O₃ → 4 Al + 3 O₂
Feedstock	Distilled TiCl₄	Distilled TiCl₄	Al₂O₃ in molten cryolite
Mode	Batch	Batch	Continuous
Temperature	~800-900 °C	~700-800 °C	~960 °C
Product form	Porous sponge	Porous sponge (finer)	Molten metal (tapped)
Byproduct recycling	MgCl₂ → Mg + Cl₂	NaCl → Na + Cl₂	Carbon anode → CO₂
Status	Dominant since ~1950	Niche high-purity only	Universal for Al
Relative metal cost	Very high	Very high	Moderate

The instructive contrast is the bottom-right column. Aluminium can be made continuously by electrolysis because its oxide dissolves cleanly in molten cryolite and the metal tapped as a liquid. Titanium oxide has no such well-behaved electrolysis, so titanium is stuck with a batch, chloride-detour, thermochemical reduction. That single difference is most of why aluminium is a commodity and titanium is a premium metal.

Where Kroll titanium ends up

• Jet engines & airframes. Roughly half of all titanium goes into aerospace. Fan blades, discs, and compressor sections of turbofan engines exploit titanium's strength-to-weight ratio and its resistance to fatigue at temperatures where aluminium would soften. A Boeing 787 airframe is ~15% titanium by weight.
• Medical implants. Titanium's biocompatibility — it forms a passive TiO₂ film that the body tolerates — makes it the standard for hip stems, dental implants, and bone plates. Every one of those started as Kroll sponge.
• Chemical plant & desalination. Titanium heat exchangers and reactor liners shrug off hot seawater and chlorine that would eat steel, so the metal that needs chlorine to be born is prized for surviving it.
• Consumer & sporting goods. Bike frames, eyeglass frames, watch cases, and premium laptop shells all trade on the strength-to-weight and corrosion resistance — the visible tip of a mostly-aerospace industry.
• The pigment paradox. The vast majority of mined titanium ore never becomes metal at all — it is oxidized to TiO₂ white pigment for paint, paper, and sunscreen. Only a few percent takes the Kroll detour to become the metal.

Limitations and side reactions

The Kroll process is famous for being both indispensable and deeply inefficient. Its weak points are well understood:

• It's a batch process. No continuous flow — every reactor must be charged, run, cooled, opened, and cleaned out. This alone caps throughput and inflates cost.
• Oxygen and nitrogen pickup. The single hardest requirement is keeping O₂ and N₂ out. Interstitial oxygen above ~0.2-0.3 wt% embrittles titanium; nitrogen forms hard, brittle TiN inclusions. This is why aerospace-grade sponge demands ultra-tight atmosphere control.
• Iron contamination. Titanium slowly attacks the steel retort, so the sponge nearest the wall is iron-rich and downgraded. The reactor is effectively a consumable.
• Incomplete reduction. If temperature or Mg availability drops, lower chlorides (TiCl₂, TiCl₃) survive in the sponge and must be distilled out; leftover TiCl₂ is pyrophoric on air exposure.
• Energy and CO₂. Between the 900 °C reduction, the vacuum distillation, the MgCl₂ electrolysis, and two vacuum-arc remelts, the process is enormously energy-intensive — the reason decades of research (the FFC-Cambridge molten-salt electrolysis, the Armstrong process, TiRO™, and others) have chased a continuous replacement. None has yet displaced Kroll at scale.

Who invented it, and when

William Justin Kroll (1889-1973), a metallurgist from Luxembourg, developed the magnesium-reduction route in the 1930s and patented it in 1940. He had earlier worked on reducing zirconium and titanium halides and recognized that magnesium gave a cleaner, more recyclable byproduct than the sodium used in Matthew Hunter's 1910 process. Kroll first produced ductile titanium in his own lab, then licensed and demonstrated the method to the U.S. Bureau of Mines.

The timing was decisive. Titanium's exceptional strength-to-weight ratio made it strategically irresistible for aircraft and, soon, jet engines, and the Cold War arms race funded a rapid scale-up through the late 1940s and 1950s. By the mid-1950s the Kroll process had all but retired the Hunter process for commercial titanium. Kroll himself grew critical of how little the process improved after adoption, predicting it would eventually be replaced by direct electrolysis — a prediction that, more than 80 years on, still hasn't come true. The process that bears his name remains, in its essentials, exactly the one he described in 1940.

Safety and industrial notes

• TiCl₄ fumes violently in moist air. It hydrolyzes to TiO₂ and HCl, producing dense white smoke (historically used for skywriting and naval smokescreens). It must be handled anhydrous, in sealed lines, with full respiratory protection.
• Molten magnesium burns fiercely. A magnesium fire cannot be extinguished with water or CO₂; it demands dry Class-D media. The reduction retort is engineered to keep magnesium sealed under argon at all times.
• Titanium sponge and fines are flammable. Finely divided titanium ignites in air and burns in nitrogen and CO₂. Crushing and blending sponge is done with inert-gas precautions.
• The argon must be dry and oxygen-free. The whole plant's product quality rides on atmosphere integrity — leak detection and gas purity monitoring are continuous.