Chlorine, Caustic Soda and Hydrochloric Acid Come Out of One Cell

On Tuesday 29 September 2026 Westlake Corporation said it will close its polyvinyl chloride plant in Cologne, Germany, in the first quarter of 2027, citing “weak demand, elevated energy costs and increased import pressure from Asia.” Exactly one week earlier, on 22 September, the Vancouver-based miner Century Lithium said it is choosing between sites in Nevada and Utah for a new chlor-alkali plant that would make “chlorine, hydrochloric acid and sodium hydroxide from sodium chloride, water and electricity.” One company is retiring a large user of chlorine; the other wants to start making it.

Neither announcement mentions caustic soda prices or hydrochloric acid supply. Both still matter to anyone who buys those two chemicals, because chlorine, caustic soda and hydrochloric acid are not three separate markets. They come out of one electrolytic cell in fixed proportions, and the plant that wants the chlorine sets how much of the other two gets made. This article sets out the two announcements, the arithmetic that connects them, and the part that outlasts the news: how the cell a product came from shows up on its certificate, and how to compare strengths when you buy.

It is a market-structure piece. It forecasts no price, recommends no purchase timing, and every figure comes from a company announcement, a government profile, a published specification or a molar mass, all listed at the end.

What happened, in order

Notice what is missing from that table. Neither company announced a chlor-alkali outage, a force majeure or a price change. One is closing a downstream plant and the other is planning an upstream one. Neither moves a drum of caustic soda this quarter. What they show is the structure underneath, and the structure is worth understanding because it explains why caustic soda and hydrochloric acid move in ways that confuse buyers.

One cell, three products

About 95% of US chlorine is made by the chlor-alkali process, according to the EPA. A direct current is passed through a sodium chloride brine. Chlorine gas comes off at the anode, and hydrogen and sodium hydroxide form at the cathode. The EPA’s supply-chain profile writes it in one line:

The third product is made from the first two. Burning the cell’s chlorine in its own hydrogen gives hydrogen chloride gas, which is absorbed in water to make hydrochloric acid. The EPA calls that the thermal method and says it accounts for less than 10% of US hydrochloric acid. The other 90%-plus comes from the chlorine’s customers: “Over 90% of hydrochloric acid is produced as a byproduct from the production of chlorinated solvents, fluorocarbons, isocyanates, organics, magnesium, and vinyl chloride monomer.” In both routes the chlorine came from a salt cell first. The EPA’s summary is that both manufacturing methods “depend on the chlor-alkali industry for inputs.”

The separator between the two halves of the cell is the detail that reaches your certificate. The EPA records that in 2021, membrane cells made 46% of US chlorine, asbestos-diaphragm cells 36% and non-asbestos diaphragm cells 1%, with potassium-chloride cells, metal production and brine-to-bleach plants making the rest. The section on grades below explains why that split shows up in the caustic you buy.

The arithmetic, in both directions

Two molar masses do most of the work. Chlorine is 35.45 grams a mole, and PVC’s repeating unit, C₂H₃Cl, is 62.50, so PVC is 56.7% chlorine by mass. One mole of chlorine gas, 70.91 grams, comes with two moles of sodium hydroxide, 80.00 grams, so the cell makes 1.128 tonnes of caustic soda per tonne of chlorine, plus 0.028 tonnes of hydrogen, from 1.648 tonnes of salt.

Read the Cologne column carefully. 105,600 tonnes is the caustic soda that would be made alongside the chlorine in that PVC. It is not caustic soda leaving the market. Westlake says it will keep supplying PVC from its other German plants, and the release does not say whether any chlor-alkali or vinyl chloride capacity changes. If the PVC is still made, just somewhere else, the chlorine is still made, and so is the caustic. The table gives the size of the link, not a forecast.

The link matters because of who decides the run rate. The EPA’s sodium hydroxide profile is direct about it: “Chlor-alkali producers generally set chlor-alkali production around demand for chlorine. In the past this has created uneven supply and demand patterns for the sodium hydroxide market.” When construction slows and PVC plants close, cells run slower and caustic soda tightens, even if nobody wants less caustic. When chlorine demand is strong, caustic soda becomes the coproduct the producer has to sell. That is why caustic soda prices can move against caustic soda demand. Our caustic soda guide covers the pricing side of that, including the electrochemical unit producers price in.

The Century plant shows the same link from the other side. A lithium project needs acid and base reagents, and its developer found that the cheapest way to get them is to build one cell that makes all three. The release describes chlor-alkali products as “essential inputs across the critical minerals supply chain, supporting lithium and battery-materials processing, rare earth element separation, and precious and specialty metals production.” It is a plan, not a plant. Sites are under due diligence and financing is the next step.

Why hydrochloric acid follows chlorine even more tightly

Caustic soda is made on purpose in a fixed ratio. Hydrochloric acid mostly is not made on purpose. When more than 90% of a chemical is the byproduct of making other chemicals, its supply depends on the output of vinyl chloride, isocyanates for polyurethane foam, chlorinated solvents and fluorocarbons. It does not depend on how much hydrochloric acid people want. The EPA lists the main US uses as oil and gas well acidification (30% of domestic consumption in 2011), steel pickling and food processing (19% each).

That is also why a disruption upstream shows up in hydrochloric acid, and late. The EPA records that Winter Storm Uri in February 2021 caused “a temporary loss of approximately 28% of domestic chlor-alkali production capacity.” Losing that much chlor-alkali capacity cuts caustic soda directly, and can cut hydrochloric acid again as the chlorinated-product plants that make it as a byproduct slow down.

For Cologne specifically, nothing above implies a change in hydrochloric acid. The release gives PVC capacity only and says nothing about vinyl chloride monomer or chlorine at the site. The point is general: if you buy hydrochloric acid, the plants that matter to you are the ones that make chlorinated products, and their announcements are worth reading.

The cell decides your grade

This is the part of the story to keep, because it does not depend on any one announcement. Membrane and diaphragm cells both make 50% caustic soda, but not the same 50% caustic soda.

In a membrane cell, an ion-exchange membrane lets sodium ions through and blocks chloride, so the caustic leaves the cell at about 33% with little salt in it and is evaporated to 50%. In a diaphragm cell, brine soaks through a porous diaphragm, so the caustic leaves weak and salty. Evaporating it to 50% drops most of the salt out as crystals, but not all of it. One US supplier publishes both grades on a single sheet, revised in June 2026:

That is a 110-fold difference in the salt limit at the same assay. If you dose caustic to adjust pH in a cooling loop, the salt makes no difference. If the chloride ends up in a product, a bath or a stainless-steel system that is sensitive to chloride, it can decide whether the job works. The assay will not show it; only the sodium chloride line on the certificate will.

The split is also changing. The EPA’s 2024 asbestos rule set a phase-out of chrysotile asbestos diaphragms in chlor-alkali production, with conversion deadlines five, eight and twelve years after the rule took effect in May 2024. In June 2025 the agency told the Fifth Circuit it would reconsider parts of that rule, including the chlor-alkali provisions. Whatever happens to the dates, the direction is toward membrane cells, which would make membrane-cell caustic a larger share of what is sold.

Hydrochloric acid has a version of the same question. Byproduct acid and thermal acid can both meet a 37% assay. What differs is the trace profile, and the route is rarely printed on the label. The practical line between them is the specification. A reagent-grade acid is made and tested to the American Chemical Society monograph. A technical-grade acid is sold on assay and appearance. Our ACS reagent versus technical grade guide walks through which applications need the reagent specification. The quickest incoming check is still color: good acid is water-white, and a yellow tint usually means iron.

Buy by the kilogram, not by the gallon

Concentrated caustic soda and hydrochloric acid are denser than water, and different strengths carry very different amounts of the active chemical per gallon. Comparing two listings by the gallon compares water as much as chemistry. Using the densities published on our product pages, at 20 °C:

A gallon of 37% acid carries about 23% more hydrogen chloride than a gallon of 31%, not 6% more, because the stronger acid is also denser. To compare two quotes, divide each price by the kilograms of active chemical in the container, not by the gallons or the percentage. Do the same for flake against liquid: caustic soda flake is nearly all NaOH, so a 55-pound bag holds about as much caustic as eight and a half gallons of 50% solution. The flake costs you a dissolving step, and dissolving it releases a lot of heat.

What this means on a specification

Three practical points follow from the structure rather than the news.

First, ask which cell. “Technical grade” on a caustic soda label does not say whether the material came from a membrane or a diaphragm cell. If chloride matters to you, ask for the sodium chloride value on the certificate for the lot you will receive, not the typical value on the data sheet.

Second, read chlorine news as caustic and acid news. A PVC closure, a polyurethane expansion or a new chlor-alkali plant is not about your chemicals on its face. It can still change the run rate of the cells that make them. That reaches your price with a lag, and the direction is not always the obvious one.

Third, do not buy ahead of need on a headline. Neither of this week’s announcements changes a single tonne of US supply this quarter. One closure is in Germany in 2027; the other plant is not yet financed.

A word on what this article does not say. It describes how chlorine, caustic soda and hydrochloric acid are made, what two companies announced, and what government profiles and one producer’s published specification say. It makes no claim about ours: which plant or cell type made the caustic soda and hydrochloric acid we list, or whether our inventory is affected by any of this. Nobody writing this has checked either, and a market-structure article is the wrong place to guess.

Common questions

Why are chlorine, caustic soda and hydrochloric acid made together?

Because one electrolytic cell makes them. Passing a current through salt brine gives chlorine at one electrode and sodium hydroxide plus hydrogen at the other, 1.128 tonnes of sodium hydroxide per tonne of chlorine. Hydrochloric acid is then made from the chlorine, either by burning it in the hydrogen or, for more than 90% of US supply, as a byproduct of plants that use chlorine.

Does the Westlake Cologne closure mean less caustic soda?

Not necessarily. Cologne’s 165,000 tonnes of PVC contain about 93,600 tonnes of chlorine, and making that much chlorine also makes about 105,600 tonnes of caustic soda. Westlake says it will keep supplying PVC from its other German plants, and its release does not mention any chlor-alkali change. That figure is the size of the link, not a forecast of lost supply.

What is the Century Lithium chlor-alkali plant?

A planned plant in Nevada or Utah, announced 22 September 2026. It would make chlorine, hydrochloric acid and sodium hydroxide from salt, water and electricity, starting at 300 short tons of chlorine a day with room to double to 600. The company reports eight non-binding offtake memoranda and says project financing is the next step. No site has been announced yet.

What is the difference between membrane-grade and diaphragm-grade caustic soda?

The cell that made it, which shows up as salt. One US supplier’s June 2026 specification lists 50% membrane-grade caustic at 0.01% sodium chloride and 4 ppm iron maximum, and 50% diaphragm grade at 1.10% sodium chloride and 9 ppm iron, at the same assay. For pH adjustment the difference rarely matters. For chloride-sensitive work, read the sodium chloride line on the certificate.

Is hydrochloric acid 37% better value than 31%?

Compare by kilograms of hydrogen chloride, not by gallons. At the densities on our product pages, a US gallon of 37% acid holds about 1.67 kg of HCl and a gallon of 31% about 1.36 kg, roughly 23% more, not 6%. Divide each price by those figures. Which wins depends on the two prices you are comparing and whether you need the stronger acid at all.

Why does caustic soda price sometimes move when demand has not changed?

Because producers run their cells for chlorine. The US EPA says “chlor-alkali producers generally set chlor-alkali production around demand for chlorine,” which has “created uneven supply and demand patterns for the sodium hydroxide market.” When PVC and other chlorine outlets slow down, the cells slow with them, and caustic soda can tighten even though caustic demand has not changed.

The caustic soda and hydrochloric acid we stock

Sodium Hydroxide 50% Membrane Grade

The low-salt 50% from the membrane cell, for chloride-sensitive work. Quart to 15-gallon drum.

Sodium Hydroxide Flakes

The most caustic per pound, if you can dissolve it on site. 2 lb to 55 lb bags; also ACS flake and 50% ACS solution.

Hydrochloric Acid 37% Technical Grade

Full-strength acid, 1.67 kg HCl per gallon. Also as 37% ACS reagent grade.

Hydrochloric Acid 31% Technical Grade

The working-strength muriatic acid for pickling, etching and pH control, 1.36 kg HCl per gallon.