What is a typical CO₂ corrosion rate for carbon steel?

Uninhibited sweet corrosion can exceed 3–10 mm/yr at high CO₂ partial pressure and 60–80 °C, but a protective FeCO₃ scale at higher temperature, higher pH, or oil-wetting commonly drops the in-service rate to well below 0.1 mm/yr. The rate is meaningless without the partial pressure, temperature, pH and water-wetting context.

Is de Waard–Milliams conservative?

The bare 1991 equation is a worst-case (full water-wetting, no protective scale) baseline. The 1995 resistance model and correction factors for scale, pH and oil-wetting bring it closer to field experience. NORSOK M-506 tends to predict lower rates where a dense carbonate scale forms.

How do I convert partial pressure to fugacity?

At low pressure fugacity ≈ partial pressure. Above roughly 30–50 bar a fugacity coefficient a < 1 is applied (f = a·pCO₂); NORSOK M-506 gives a pressure- and temperature-dependent expression for it.

Corrosion · guide

CO₂ corrosion rate of carbon steel

Sweet (CO₂) corrosion sets the corrosion allowance for most carbon-steel oil-and-gas pipework. Two empirical models dominate practice — de Waard–Milliams and NORSOK M-506. Here is what each one actually computes, the corrections that matter, and where the predictions diverge.

The mechanism in one paragraph

Dissolved carbon dioxide hydrates to carbonic acid, which supplies the cathodic reaction that drives iron dissolution. Rate therefore climbs with CO₂ partial pressure and temperature — until a dense iron-carbonate (FeCO₃, siderite) film nucleates on the steel and throttles the diffusion of reactants. Everything in a CO₂ model is some attempt to capture that competition between attack and protective scale.

de Waard–Milliams baseline

The most-cited correlation gives the worst-case rate (fully water-wetted, no protective scale) as a function of temperature T in kelvin and CO₂ partial pressure p_CO₂ in bar:

log₁₀(V_cor) = 5.8 − 1710 / T + 0.67 · log₁₀(p_CO₂) [V_cor in mm/yr]

Read it as an upper bound. The 1995 revision recasts it as a resistance in series — a reaction-controlled term and a mass-transfer-controlled term — so that flow velocity and scale can each become the limiting step:

1 / V_cor = 1 / V_reaction + 1 / V_mass-transfer

Corrections that change the answer by 10×

The baseline is rarely the design rate. The corrections that move the number most:

Scale (temperature) factor — above the scaling temperature (~60–90 °C, depending on pH and Fe²⁺) a protective FeCO₃ layer forms and the rate falls with increasing temperature.
pH factor — higher pH (more bicarbonate, condensed water vs. formation water) suppresses the cathodic reaction.
Oil-wetting — if the steel is wetted by oil rather than water, effective corrosion can be near zero regardless of the chemistry.

NORSOK M-506

The Norwegian standard model is also empirical but tuned to a large flow-loop dataset. It expresses rate as a temperature-tabulated kinetic constant K_t times CO₂ fugacity raised to ~0.62, scaled by a wall-shear-stress term and a pH function. Because its scaling treatment is more aggressive, NORSOK frequently predicts lower rates than bare de Waard–Milliams where a carbonate film is stable.

Use fugacity, not partial pressure, above ~30–50 bar: f_CO₂ = a · p_CO₂ with the fugacity coefficient a < 1.

Open the calculatorCO₂ corrosion calculator →Enter pCO₂, temperature, pH and velocity; compare de Waard–Milliams and NORSOK M-506 side by side with the scaling and pH corrections applied.

How to use the result

Multiply the predicted rate by the design life to size the corrosion allowance, then sanity-check it against inhibitor availability and the likelihood of sustained water-wetting. Treat any single-number prediction as a screening value: real CO₂ corrosion is dominated by whether a protective scale survives the local flow and chemistry, which no point model captures perfectly.

Frequently asked

What is a typical CO₂ corrosion rate for carbon steel?: Uninhibited sweet corrosion can exceed 3–10 mm/yr at high CO₂ partial pressure and 60–80 °C, but a protective FeCO₃ scale at higher temperature, higher pH, or oil-wetting commonly drops the in-service rate to well below 0.1 mm/yr. The rate is meaningless without the partial pressure, temperature, pH and water-wetting context.
Is de Waard–Milliams conservative?: The bare 1991 equation is a worst-case (full water-wetting, no protective scale) baseline. The 1995 resistance model and correction factors for scale, pH and oil-wetting bring it closer to field experience. NORSOK M-506 tends to predict lower rates where a dense carbonate scale forms.
How do I convert partial pressure to fugacity?: At low pressure fugacity ≈ partial pressure. Above roughly 30–50 bar a fugacity coefficient a < 1 is applied (f = a·pCO₂); NORSOK M-506 gives a pressure- and temperature-dependent expression for it.

References

C. de Waard, U. Lotz, D.E. Milliams, "Predictive Model for CO₂ Corrosion Engineering in Wet Natural Gas Pipelines," Corrosion 47(12), 1991.
C. de Waard, U. Lotz, A. Dugstad, "Influence of Liquid Flow Velocity on CO₂ Corrosion: a Semi-Empirical Model," NACE CORROSION/1995, paper 128.
NORSOK Standard M-506, "CO₂ corrosion rate calculation model," Standards Norway (rev. 2017).
S. Nešić, "Key issues related to modelling of internal corrosion of oil and gas pipelines — A review," Corrosion Science 49(12), 2007.

Related guides

Browse all engineering guides →