CO2 Corrosion Fundamentals

1. CO2 Corrosion Mechanism

Carbon dioxide corrosion occurs when CO2 dissolves in produced water or condensed water to form carbonic acid (H2CO3). Although carbonic acid is weak compared to mineral acids, its ability to provide direct reduction at the cathode makes it significantly more corrosive than expected from pH alone. This is the dominant internal corrosion mechanism in most gas gathering and transmission systems.

Fundamental Chemistry

CO2 Dissolution and Hydration: CO2(g) ↔ CO2(aq) (Henry's law: CO2_aq = KH × pCO2) CO2(aq) + H2O ↔ H2CO3 (slow hydration step, rate limiting) H2CO3 ↔ H+ + HCO3- (first dissociation, pKa ~ 6.4) HCO3- ↔ H+ + CO3(2-) (second dissociation, pKa ~ 10.3) Electrochemical Reactions: Anodic: Fe → Fe(2+) + 2e- (iron dissolution) Cathodic: 2H2CO3 + 2e- → H2 + 2HCO3- (direct reduction) 2H+ + 2e- → H2 (hydrogen evolution) Overall: Fe + CO2 + H2O → FeCO3 + H2 (iron carbonate formation)

Why CO2 Is More Corrosive Than pH Suggests

The cathodic reduction of undissociated H2CO3 provides an additional source of hydrogen ions directly at the metal surface, bypassing the solution pH limitation. This "direct reduction" mechanism was first identified by de Waard and Milliams (1975) and explains why CO2 corrosion rates are 3-10 times higher than what would be expected from hydrochloric acid at the same pH.

Low pCO2

< 0.5 bar (< 7 psia)

Generally non-corrosive. Rates below 1 mpy for carbon steel if water chemistry is favorable.

Moderate pCO2

0.5 – 2 bar (7 – 30 psia)

Corrosive environment. Carbon steel with inhibition is viable. Typical gathering system conditions.

High pCO2

> 2 bar (> 30 psia)

Highly corrosive. Consider CRA materials or very aggressive inhibition programs with continuous monitoring.

Industry impact: CO2 corrosion accounts for roughly 25% of all pipeline integrity failures worldwide. In the US Gulf Coast and Permian Basin, gathering system corrosion repair costs exceed $500 million annually. Accurate prediction using models like de Waard-Milliams is essential for cost-effective design.

2. de Waard-Milliams Model

The de Waard-Milliams model is the most widely used empirical correlation for predicting CO2 corrosion rate of carbon steel. Published in 1975 and updated in 1991 and 1993, it provides the base corrosion rate as a function of temperature and CO2 partial pressure.

Base Equation (1975)

de Waard-Milliams (1975) Base Correlation: log10(Vcor) = 5.8 - 1710 / (T + 273) + 0.67 × log10(pCO2) Where: Vcor = corrosion rate (mm/yr) T = temperature (°C) pCO2 = CO2 partial pressure (bar) Derivation basis: - Activation energy: 1710 × 2.303 × R ~ 32.7 kJ/mol (consistent with dissolution-controlled process) - Pressure exponent 0.67 indicates partial diffusion control - Constant 5.8 calibrated to large laboratory dataset Example: T = 65°C, pCO2 = 2.0 bar: log(Vcor) = 5.8 - 1710/338 + 0.67 × log(2.0) = 5.8 - 5.059 + 0.202 = 0.943 Vcor = 10^0.943 = 8.8 mm/yr = 346 mpy (uninhibited)

Model Evolution

Version	Year	Key Addition	Reference
Original	1975	Base equation: temperature + pCO2	Corrosion 31(5):177
Update 1	1991	pH correction factor, fugacity	Corrosion/91, Paper 577
Update 2	1993	Scale effect, glycol, oil wetting	Corrosion/93, Paper 69
Revision	1995	Top-of-line corrosion, condensation	Corrosion/95, Paper 128

Temperature Dependence

The base rate increases exponentially with temperature due to activation energy of the dissolution process. However, above approximately 60–80°C (140–176°F), FeCO3 scale formation becomes thermodynamically favorable, which can reduce actual corrosion rates dramatically. The base model does not account for this scale effect — that requires a separate correction factor.

Critical limitation: The 1975 base equation predicts ever-increasing corrosion rate with temperature. In practice, the rate peaks around 60–80°C and decreases at higher temperatures due to protective FeCO3 scale. Always apply the scale correction factor when operating above 140°F.

3. pH & Fugacity Corrections

The 1991 and 1993 updates to the de Waard model introduced two critical corrections that significantly improve prediction accuracy at higher pressures and non-equilibrium pH conditions.

pH Correction

pH Correction Factor (de Waard-Lotz, 1991): f_pH = 10^(pH_sat - pH_actual) Where: pH_sat = saturated pH from CO2 equilibrium pH_actual = measured or calculated pH Effect: - If pH_actual = pH_sat: f_pH = 1.0 (base rate unchanged) - If pH_actual < pH_sat: f_pH > 1.0 (rate increases) - If pH_actual > pH_sat: f_pH < 1.0 (rate decreases) When pH exceeds saturation: Higher pH (from bicarbonate, amines, etc.) reduces the driving force for cathodic reaction and promotes FeCO3 scale stability. pH increase of 1 unit reduces rate by 10x.

Saturated pH Calculation

CO2-Saturated Water pH: The saturated pH is calculated from CO2 solubility and dissociation equilibria: [CO2]_aq = K_H × pCO2 (Henry's law) [H+] = sqrt(K1 × [CO2]_aq) (first dissociation) pH_sat = -log10([H+]) K_H = Henry's constant (temperature dependent) K1 = First dissociation constant of H2CO3 Typical saturated pH values: pCO2 = 0.1 bar, 25°C: pH_sat ~ 4.9 pCO2 = 1.0 bar, 25°C: pH_sat ~ 3.9 pCO2 = 1.0 bar, 80°C: pH_sat ~ 4.2 pCO2 = 10 bar, 80°C: pH_sat ~ 3.5

Fugacity Correction

CO2 Fugacity vs. Partial Pressure: fCO2 = phi × pCO2 Where: fCO2 = CO2 fugacity (bar) - "effective" partial pressure phi = fugacity coefficient (dimensionless) pCO2 = CO2 partial pressure (bar) Fugacity coefficient from Pitzer correlation: ln(phi) = (B0 + omega × B1) × Pr / Tr Tc_CO2 = 304.2 K, Pc_CO2 = 73.8 bar, omega = 0.225 Significance: At low pressure (P < 20 bar): phi ~ 0.95-1.0 (negligible) At moderate pressure (20-70 bar): phi ~ 0.85-0.95 At high pressure (> 70 bar): phi ~ 0.70-0.85 Using fugacity instead of partial pressure improves accuracy at elevated pressures by accounting for non-ideal gas behavior of CO2.

Engineering practice: At pressures below 500 psig, the fugacity correction is minor (less than 5%). Above 1000 psig, always use fugacity rather than partial pressure. For custody-grade predictions, use a proper equation of state (SRK or PR) to determine CO2 fugacity.

4. FeCO3 Protective Scale Formation

Iron carbonate (siderite, FeCO3) scale is the single most important factor that determines whether CO2 corrosion is manageable or catastrophic. When conditions favor dense, adherent scale formation, corrosion rates can decrease by 90% or more compared to bare steel predictions.

Scale Formation Thermodynamics

FeCO3 Precipitation: Fe(2+) + CO3(2-) → FeCO3(s) Supersaturation ratio: S = [Fe(2+)] × [CO3(2-)] / Ksp Where Ksp = solubility product of FeCO3 Ksp temperature dependence: log(Ksp) = -59.3498 - 0.041377×T - 2.1963/T + 24.5724×log(T) (T in Kelvin) Scale forms when S > 1 (supersaturated) Favorable conditions for protective scale: - Temperature > 60°C (140°F) - pH > 5.5 - Low flow velocity (no shear stripping) - Fe(2+) concentration > 5-10 ppm - Low chloride (Cl- disrupts scale)

Scale Protection Factors

Scale Condition	Factor	Description	Typical Temperature
No scale	1.0	Bare metal, full corrosion rate	< 40°C (100°F)
Thin scale	0.5	Partial coverage, porous	40–70°C (100–160°F)
Protective scale	0.1	Dense, adherent, continuous FeCO3	> 70°C (160°F)

Scale Stability Risks

High velocity: Wall shear stress above 10–20 Pa can strip protective scale, especially in slug flow. Localized corrosion at bare patches can be 5–10x the average rate.
Temperature cycling: Thermal expansion mismatch between scale and steel can cause spalling during shutdowns and restarts.
Chloride ions: Cl- concentrations above 5,000–10,000 ppm can destabilize FeCO3 scale and promote localized pitting.
Acetic acid (HAc): Organic acids from formation water dissolve FeCO3 scale and increase corrosion rate even at higher temperatures.

Design implication: Do not rely on FeCO3 scale protection for corrosion allowance calculations unless you have field evidence (coupon data, ILI) confirming scale integrity at the specific operating conditions. The de Waard model with scale factor = 1.0 (no scale) gives the conservative design case.

5. Correction Factors

The 1993 model update introduced multiplicative correction factors to account for real-world conditions that deviate from the laboratory baseline. The final corrected rate is:

Corrected Corrosion Rate: Vcor_final = Vcor_base × f_pH × f_scale × f_Cr × f_oil × f_inh × f_wc × f_glycol Where each factor is less than or equal to 1.0 (all reduce rate): f_pH = pH correction (can increase if pH < pH_sat) f_scale = FeCO3 scale protection (0.1 to 1.0) f_Cr = Chrome alloy reduction (0.01 to 1.0) f_oil = Oil wetting reduction (0.0 to 1.0) f_inh = Inhibitor efficiency (0.01 to 1.0) f_wc = Water cut fraction (0.0 to 1.0) f_glycol = Glycol dilution (0.2 to 1.0)

Chrome Alloy Factors

Steel Type	Factor (f_Cr)	Relative Cost	Application
Carbon Steel (API 5L)	1.00	1.0x	Standard, with inhibition
1% Cr (ASTM A335 P1)	0.50	1.3x	Moderate CO2, reduces rate 50%
3% Cr	0.30	1.6x	Higher CO2, lower chrome option
5% Cr (ASTM A335 P5)	0.10	2.0x	High CO2, reduces rate 90%
13% Cr (Type 410/420)	0.01	3.0x	Nearly immune to CO2 corrosion

Oil Wetting

Corrosion only occurs where water contacts the steel surface. In multiphase flow, the wetting regime determines the effective corrosion area:

Water-wet (f_oil = 1.0): Continuous water phase contacts the pipe wall. Worst case. Occurs at water cut above 30–40% or low velocity in stratified flow.
Mixed wetting (f_oil = 0.3): Intermittent water contact. Typical of slug flow with moderate water cut.
Oil-wet (f_oil = 0.0): Continuous oil film prevents water contact. Occurs at very low water cut (<5–10%) with velocity above 1 m/s. No corrosion.

Glycol Correction

Glycol Correction: f_glycol = 1 - (glycol_wt% / 100) Basis: Glycol (MEG or DEG) reduces water activity and dilutes the corrosive aqueous phase. At 50 wt% glycol, the correction factor is 0.5 (halves the rate). Typical glycol concentrations: Hydrate inhibition: 30-50 wt% MEG pH stabilization + MEG: 80+ wt% MEG (very low corrosion) Note: At glycol concentrations above 80%, corrosion becomes negligible. This is the basis of the "pH stabilization + MEG" strategy used in subsea pipelines.

6. NORSOK M-506 Model

NORSOK M-506 is the Norwegian petroleum industry standard for CO2 corrosion rate prediction. It is an empirical model that uses CO2 fugacity, temperature, pH, and wall shear stress as primary inputs. It is widely used for North Sea applications and provides an independent comparison to de Waard-Milliams.

Model Structure

NORSOK M-506 (Simplified): CR = Kt × fCO2^0.62 × f_pH × f_shear Where: Kt = temperature function (three ranges) fCO2 = CO2 fugacity (bar) f_pH = pH correction = 10^(0.32 × (3.71 - pH)) f_shear = wall shear stress correction Temperature function Kt: T <= 20°C: log(Kt) = 4.93 - 1119/T_K T 20-150°C: log(Kt) = 5.0 - 1290/T_K T > 150°C: log(Kt) = 3.7 - 830/T_K Key difference from de Waard: - Uses fugacity natively (not partial pressure) - Includes explicit wall shear stress term - Three-segment temperature function (captures scale effect)

de Waard vs. NORSOK Comparison

Feature	de Waard-Milliams	NORSOK M-506
CO2 input	Partial pressure + fugacity correction	Fugacity directly
Temperature	Single Arrhenius expression	Three-segment function
pH	Correction factor (1991 update)	Built-in pH term
Flow effect	Not explicit in base model	Wall shear stress term
Scale effect	External correction factor	Implicitly in temperature function
Typical use	North America, global	North Sea, European operations

Best practice: Run both models and use the more conservative result for design. When models disagree by more than 50%, investigate which assumptions better match your specific conditions (temperature range, pressure, scale likelihood).

7. Material Selection for CO2 Service

Material selection is the most fundamental corrosion mitigation decision, made at the design stage. The choice between carbon steel (with inhibition) and corrosion-resistant alloys (CRAs) depends on severity, reliability requirements, and life-cycle economics.

Material Selection Matrix

pCO2 Range	Temperature	Recommended Material	Expected Rate
< 0.5 bar	Any	Carbon steel, no inhibitor	< 1 mpy
0.5 – 2 bar	< 60°C	Carbon steel + inhibitor	2 – 10 mpy (inhibited)
0.5 – 2 bar	> 60°C	CS + inhibitor (scale helps)	1 – 5 mpy (with scale)
2 – 10 bar	< 60°C	1-3% Cr or CS + aggressive inhibition	5 – 20 mpy (CS uninhibited)
2 – 10 bar	> 60°C	5% Cr or 13% Cr	< 1 mpy (13% Cr)
> 10 bar	Any	13% Cr or Duplex 2205	< 0.1 mpy

Economic crossover: For a 20-year design life, the breakeven between carbon steel + inhibitor program and 13% Cr typically occurs at a predicted uninhibited rate of 30–50 mpy. Above this threshold, CRA often wins on life-cycle cost despite 3x higher initial material cost, because inhibitor and monitoring costs accumulate over the project life.

8. Mitigation Strategies

Corrosion Inhibitor Programs

Film-forming corrosion inhibitors are the primary internal corrosion control for carbon steel pipelines in CO2 service. Well-managed programs achieve 80–95% efficiency.

Inhibitor Type	Mechanism	Typical Efficiency	Application
Imidazolines	Hydrophobic film on steel	85 – 95%	General CO2 service, most common
Amido-amines	Chemisorbed barrier	80 – 90%	High-temperature CO2 service
Quaternary ammonium	Cationic surfactant film	75 – 90%	Water-dominated systems
Film-forming amines	Persistent barrier	80 – 95%	Dry gas, intermittent water

pH Stabilization

pH Stabilization Strategy: By raising pH above the saturation pH for FeCO3, the supersaturation condition promotes dense, adherent scale formation that provides long-term protection. Methods: 1. Sodium bicarbonate injection: raises pH directly 2. MEG with pH stabilization: glycol + NaOH/NaHCO3 3. MDEA (methyldiethanolamine): amine-based pH control Target pH: 6.0 – 6.5 (sufficient for FeCO3 stability) Benefit: pH increase from 4.0 to 6.0 reduces base corrosion rate by factor of 100 (two pH units = 100x). Combined with FeCO3 scale, total reduction can be 1000x.

Velocity Control

Maximum velocity: Keep below API RP 14E erosional velocity limit to prevent erosion-corrosion and scale stripping.
Minimum velocity: Maintain above 1 m/s (3 ft/s) to prevent water accumulation at low points (water dropout velocity).
Slug flow: Slug impact forces can strip protective FeCO3 scale. Slug mitigation (slug catchers, pipeline profiling) reduces localized attack.

9. Design Practice

Corrosion Allowance

Corrosion Allowance Calculation: CA = Vcor_corrected × Design_Life × Safety_Factor Typical safety factor: 1.2 (20% margin) Round up to nearest 1/16 inch per pipe specification. Example: Corrected rate: 3.5 mpy = 0.089 mm/yr Design life: 25 years Safety factor: 1.2 CA = 3.5 × 25 × 1.2 / 1000 = 0.105 in Rounded up: 0.125 in (1/8") = 2/16" Total wall thickness: t_total = t_pressure + CA + t_manufacturing_tolerance

Monitoring Program Design

Predicted Rate	Inspection Interval	Monitoring Method
< 1 mpy	5 – 10 years	UT thickness, coupons (annual)
1 – 5 mpy	3 – 5 years	UT thickness, coupons, ER probes
5 – 10 mpy	1 – 3 years	ILI, coupons, online probes, iron counts
> 10 mpy	6 – 12 months	ILI, continuous LPR/ER, aggressive program

10. Industry Standards

Standard	Title	Relevance
NACE SP0106	Internal Corrosion Direct Assessment for Pipelines	ICDA methodology using CO2 models
NORSOK M-506	CO2 Corrosion Rate Calculation Model	Alternative prediction model
NACE MR0175	Materials for H2S (= ISO 15156)	Material requirements if H2S present
API 5L	Line Pipe Specification	Carbon steel and CRA pipe grades
ASME B31.8	Gas Transmission Pipelines	Wall thickness + corrosion allowance
API 570	Piping Inspection Code	Inspection intervals from corrosion rate
API RP 14E	Erosional Velocity	Maximum velocity limits
NACE SP0775	Corrosion Coupon Programs	Field coupon installation and analysis
DNV-RP-F101	Corroded Pipelines	Fitness-for-service assessment

log(CR) = 5.8 − 1710/T + 0.67·log(pCO2)

pCO2 > 2 bar

NACE SP0106 · NORSOK M-506

Contents

Use this guide when you need to: