Chlor-Alkali Process

The half-reactions

Pass direct current through saturated brine and three things happen at three places.

Anode (oxidation):     2 Cl⁻ → Cl₂(g) + 2 e⁻        E° = +1.36 V
Cathode (reduction):   2 H₂O + 2 e⁻ → H₂(g) + 2 OH⁻  E° = −0.83 V
Overall (decomposition): 2 NaCl + 2 H₂O → 2 NaOH + Cl₂ + H₂   E°_cell ≈ −2.19 V

The minus sign on E°_cell means the reaction is non-spontaneous — you have to do electrical work on it. That's the entire point: the process trades electricity for chemical bonds.

The sodium ion is a spectator at the moment of reduction — water is reduced first, because reducing Na⁺ to metallic sodium would require −2.71 V instead of −0.83 V. Instead, Na⁺ migrates from the anode compartment (where Cl⁻ is being consumed) toward the cathode compartment (where OH⁻ is accumulating), driven by the electric field, and the two ions pair up to form NaOH solution.

Three cell technologies

What changes between the three industrial designs is how the anolyte and catholyte are kept apart. Every design must prevent chlorine from contacting NaOH (else bleach forms), prevent hydrogen from contacting chlorine (else explosion), and let the right ions through to balance charge.

	Mercury cell (1892, declining)	Diaphragm cell (1900s, declining)	Membrane cell (1970s, dominant)
Separator	Flowing Hg amalgam (cathode itself)	Asbestos / polymer-fiber mat	Perfluorinated cation-exchange membrane
Cathode reaction	Na⁺ + e⁻ + Hg → Na(Hg) (then 2 Na(Hg) + 2 H₂O → 2 NaOH + H₂ + 2 Hg in a separate decomposer)	2 H₂O + 2 e⁻ → H₂ + 2 OH⁻ in catholyte	2 H₂O + 2 e⁻ → H₂ + 2 OH⁻ in pure-water catholyte
Cell voltage	~4.4 V	~3.5 V	~2.2 V (best 2.05 V)
Energy (kWh per t Cl₂)	~3 600	~3 400	~2 700-3 200
NaOH purity / strength	50 % directly, <100 ppm NaCl	~12 %, ~1 % NaCl, must evaporate & purify	32-35 % directly, <50 ppm NaCl
Hg / asbestos / capital	~1-3 g Hg lost per t Cl₂; environmental phaseout	asbestos health risk; banned in EU 2017	no toxic materials; membrane cost ~$500 m⁻²
Status (2026)	banned EU 2017, last US plant closed 2014	about 5 % of world capacity, declining	about 90 % of world capacity

The mercury cell made the purest caustic — pharma-grade NaOH for batteries and rayon — but the price was 1-3 grams of mercury lost to air, water, and graveyard sludge per tonne of chlorine. Across a 500 kt yr⁻¹ plant that is up to 1.5 t mercury into the environment annually. The 2013 Minamata Convention forced the worldwide phase-out by 2025.

Inside a modern membrane cell

A commercial membrane cell is a stack of 50-200 thin bipolar elements clamped under hundreds of tonnes of force. One element looks like:

Anolyte side (saturated brine, ~25 % NaCl)
                                                    Catholyte side (~32 % NaOH)
+---------+ ___________________ +---------+
|  TI/RuO₂|                    |  Ni/NiO |
|  anode  |     MEMBRANE       | cathode |
|         |  (Nafion-type)     |         |
|  Cl⁻ → Cl₂  ←──── Na⁺ ────→  H₂O → H₂   |
|         |  H₂O blocked       |         |
+---------+                    +---------+
   |                                    |
   anolyte out (depleted brine)         catholyte out (concentrated NaOH)
   + Cl₂(g) up                          + H₂(g) up

The anode is a titanium plate coated with ruthenium-titanium dioxide (a "DSA", dimensionally stable anode), invented by De Nora in 1968 and the single biggest reason chlor-alkali energy use halved that decade. The cathode is nickel, sometimes coated with a porous high-surface-area Ni-Mo or Ni-RuO₂ to reduce hydrogen overpotential. The membrane is a sulfonated/carboxylated perfluoropolymer (Nafion 350 or its equivalents from Asahi Kasei), perforated with hydrophilic channels that pass Na⁺ and exclude OH⁻ and Cl⁻.

Worked example: voltage budget at 5 kA m⁻²

A typical membrane cell runs 5 kA m⁻² with these losses:

• Thermodynamic minimum (E°_cell, corrected for activities): ~2.19 V.
• Anode overpotential (Cl⁻ → Cl₂ on RuO₂/TiO₂): ~50 mV.
• Cathode overpotential (H₂O → H₂ on activated Ni): ~150 mV.
• Membrane IR drop: ~300 mV (resistance × current density).
• Anolyte / catholyte solution IR drop: ~150 mV.
• Hardware contact resistance: ~50 mV.
• Total cell voltage: ~2.20 V.

Energy per tonne of chlorine: 2 mol e⁻ per mol Cl₂ × 96 485 C × 2.20 V × (1 / 70.9 g mol⁻¹) × (1 t / 10⁶ g) × (1 kWh / 3.6 × 10⁶ J) × (1 / 0.96 current efficiency) ≈ 3 100 kWh per t Cl₂. At an industrial power price of $30 MWh⁻¹, that is $93 of electricity per tonne of chlorine — the dominant operating cost, easily 60 % of cash cost of production.

Brine preparation: the unglamorous half

Membrane cells are murderously sensitive to feed impurities. Calcium, magnesium, and strontium precipitate inside the membrane, blocking ion channels and dropping current efficiency from 96 % to under 90 % in days. Brine purification is the longest stage by residence time:

1. Saturation. Solid NaCl (rock salt or solar salt) is dissolved in recycled depleted brine to ~25 % w/w.
2. Primary purification. Add Na₂CO₃ (precipitates CaCO₃) and NaOH (precipitates Mg(OH)₂). Settle, filter.
3. Polishing. Pass through a chelating ion-exchange resin (Amberlite IRC-748 type) that captures ppm-level Ca²⁺, Mg²⁺, Sr²⁺ down to < 20 ppb.
4. Acidification. Lower pH to ~2 with HCl to keep CO₃²⁻ from re-precipitating in the cell.
5. Dechlorination of returning anolyte. Strip dissolved Cl₂ before recycle, then add Na₂SO₃ to reduce trace ClO⁻.

Brine purification accounts for roughly 15 % of plant capital and 5 % of operating cost — invisible in flowsheet diagrams, indispensable in practice.

Downstream: turning the streams into products

• Chlorine leaves the anode at ~85 °C, saturated with water. It's cooled to drop water out, dried with concentrated H₂SO₄ (sulfuric acid), compressed to 6-12 bar, and either liquefied to ship or piped directly to a customer (PVC plant, water utility).
• Sodium hydroxide exits at 32-35 % w/w directly from a membrane cell — pulp, soap, and detergent customers can use it as is. Alumina refining and rayon want 50 %, made by triple-effect evaporation. Pharma-grade 50 % requires further mercury-free polishing through ion exchange.
• Hydrogen leaves the cathode at >99 % purity. About half is burned for plant steam; the rest is sold as feedstock for HCl synthesis, hydrogenation, or fuel cells. As renewable power displaces coal, the chlor-alkali H₂ stream is one of the cheapest sources of low-carbon hydrogen.

Variants

• Bipolar membrane stacks — connect membranes back-to-back in series; one rectifier supplies the whole stack at much higher voltage and lower current, simplifying the rectifier and bus-bar.
• Oxygen-depolarized cathode (ODC) — replace H₂ evolution with O₂ reduction (½ O₂ + H₂O + 2 e⁻ → 2 OH⁻). Saves ~1 V (~30 % electricity), at the price of forfeiting the hydrogen co-product. Commercial since ~2010, ~5 % of capacity.
• HCl electrolysis — same hardware, fed with hydrochloric acid instead of brine. Yields Cl₂ + H₂. Used by isocyanate and PVC plants to recover chlorine from HCl by-product streams.
• Solid-electrolyte / SPE chlor-alkali — research-stage, replaces the liquid catholyte with a humidified gas, aiming for < 2 V cells.

Pitfalls and operational hazards

• Mercury legacy contamination. Decommissioned mercury-cell sites have left soil and groundwater contamination requiring decade-long remediation. Plant operators are liable under Minamata-aligned national laws even after closure.
• Asbestos in diaphragm cells. The classic chrysotile-asbestos diaphragm has been replaced by polymer composites (Polyramix) since the 1990s, but legacy plants required full asbestos abatement during conversion.
• Hydrogen-chlorine explosions. A pinhole in the membrane and a momentary current reversal can mix the two gases. Plants run online H₂-in-Cl₂ analyzers and trip the rectifier above 4 % H₂.
• Chlorate buildup. Chlorine that dissolves in the anolyte slowly disproportionates (3 Cl₂ + 6 OH⁻ → 5 Cl⁻ + ClO₃⁻ + 3 H₂O). Anolyte must be acidified and purged of chlorate before recycle.
• Membrane fouling. Iron from corroded piping plates out on the membrane surface. Stainless components and a periodic acidic membrane wash are mandatory.

Coupled economics

Stoichiometry handcuffs the plant: every tonne of chlorine drags 1.13 t of caustic out the door. PVC demand pulls chlorine; alumina refining and pulp pull caustic; the two markets do not move together. Operators publish an Electrochemical Unit (ECU) price — the merchant value of one tonne of chlorine plus 1.13 t of NaOH minus power cost — that swings from $200 to over $1 000 per ECU as PVC and alumina cycles diverge. When chlorine is short, plants run hard and dump caustic into a depressed market; when caustic is short, plants flare or store chlorine. This is one of the textbook examples of an industrial process whose economics are dominated by the by-product, not the namesake product.