CO₂ for welding: pure carbon dioxide or an argon mix

Pure CO₂ is the cheapest shielding gas and the one that spatters most. Where the argon mix pays for itself, how much gas a welding post actually uses, and when a workshop outgrows cylinders.

For semi-automatic welding of carbon steel there are two mainstream choices: pure carbon dioxide or an argon–CO₂ mixture. Pure CO₂ costs noticeably less per kilogram, which is why most workshops start with it. The full cost of a weld, however, includes the wire that turns into spatter and the hours spent chipping it off.

What the gas actually changes

The arithmetic that decides is not the price of the cylinder. Pure CO₂ produces spatter, and spatter is wire you bought, melted and then removed by hand. On visible welds and thin sheet the argon mix usually wins on total cost even though the gas itself costs more.

One physical detail that catches people out: CO₂ expands from liquid at the regulator and cools sharply. On a cold day and at high flow, a plain regulator freezes over. Welding posts on pure CO₂ need a heated regulator — or a proper vaporizer if the gas comes from a tank.

Which CO₂ a semi-automatic welder needs

Technical grade is enough — food grade costs more and gives the weld nothing. What matters is dry gas: moisture is the main cause of porosity. In EN ISO 14175 pure carbon dioxide is designated C1; the cylinder or certificate should say so. And one gas per line: a half-used CO₂ cylinder topped up from another source is how moisture and air end up in the weld.

How much gas one post actually uses

The consumption calculation is simple and almost always surprising. Take flow rate in litres per minute, multiply by arc-on time, and convert to kilograms at 0.54 m³ per kilogram.

Multiply by the number of posts. A workshop with ten welders at 50% arc-on time burns about 1.2 tonnes a month — roughly fifty 40-litre cylinders, or two swaps every working day.

Where the gas quietly leaks away

• Flow set «with a margin». Above 16–18 l/min the extra gas does not improve shielding; it creates turbulence that pulls air into the arc.
• Long torch leads left pressurised between welds.
• Post-flow set longer than needed on every single weld.
• Leaks at quick couplings — the classic overnight loss that nobody sees on the meter.

Cutting flow from 18 to 13 l/min on ten posts saves roughly a quarter of the consumption without touching weld quality. That is usually the cheapest improvement available in a welding shop.

When a workshop outgrows cylinders

From about a tonne a month, cylinders start costing more in handling than in gas. The usual step is a stationary tank of 10–20 m³ with a vaporizer and a distribution ring around the shop, so every post takes gas from a wall socket instead of a cylinder standing beside it.

For a shop of that size a 10 m³ tank covers eight to nine months of supply at one tonne a month, which is too slow a turnover — a 20 m³ vessel shared with other processes, or a smaller microbulk unit, usually fits better. Sizing here follows delivery logistics rather than the tank catalogue.

The vaporizer matters more than the tank in this application. Welding demand is spiky: all posts strike an arc at the start of a shift. One post at 12 l/min draws only about 1.3 kg/h, so a 130 kg/h vaporizer covers several dozen simultaneous posts; what sizes it is the peak at shift start plus a frost reserve, not the number of welders.