9.1 High-Temperature Hydrogen Attack (HTHA) Mechanisms & Nelson Curves

High-Temperature Hydrogen Attack Overview — API RP 571 Section 3.36

1. Fundamental Definition and Phenomenological Contrast

High-Temperature Hydrogen Attack (HTHA) is a permanent, irreversible metallurgical degradation mechanism resulting from the exposure of susceptible steels to high partial pressures of hydrogen at elevated operating temperatures. HTHA affects carbon steel, C-0.5Mo, and low-alloy steels (including 1Cr-0.5Mo, 1.25Cr-0.5Mo, and 2.25Cr-1Mo).

Candidates must distinguish HTHA from low-temperature hydrogen mechanisms (such as hydrogen blistering, hydrogen-induced cracking [HIC], stress-oriented hydrogen-induced cracking [SOHIC], and sulfide stress cracking [SSC]):

• Low-Temperature Hydrogen Mechanisms (API RP 571 Section 3.67): Occur where liquid water is present in sour service (API RP 571 notes blistering, HIC, and SOHIC from ambient to about 300 °F / 150 °C, and SSC usually below about 180 °F / 82 °C). The atomic hydrogen ($H^\bullet$) is generated by the electrochemical corrosion reaction ($Fe + H_2S \rightarrow FeS + 2H^\bullet$). Trapped hydrogen atoms recombine into molecular hydrogen gas ($H_2$) within laminations or inclusion interfaces.
• High-Temperature Hydrogen Attack (API RP 571 Section 3.36): Occurs at elevated temperatures (typically above 400 °F / 204 °C) and elevated hydrogen partial pressures ($p_{H_2} > 50\text{ psia}$ / 0.34 MPa) in dry or wet gaseous process environments. The hydrogen originates directly from the gaseous phase through high-temperature catalytic dissociation ($H_2 \rightleftharpoons 2H^\bullet$). Furthermore, the trapped gas within the metal is methane ($CH_4$), not molecular hydrogen, formed via chemical reaction with the steel's internal carbide phase.

---

Chemical Kinetics & The Methane Reaction Mechanism

Gaseous Process Environment (Elevated Temperature > 400 °F, Elevated pH2 > 50 psia)
                  │
                  ▼  Thermal / Catalytic Dissociation on Metal Surface
       H2 (Molecular Gas) ◄───► 2 H• (Nascent Atomic Hydrogen)
                  │
                  ▼  Interstitial Lattice Diffusion (Rapid interstitial migration through BCC ferrite)
        Atomic Hydrogen Permeates Deep into Bulk Steel Matrix
                  │
     ┌────────────┴─────────────────────────────────────────┐
     ▼                                                      ▼
Surface Decarburization                               Internal Decarburization & Methane Reaction
(Near external / internal surface)                    (Deep within bulk steel matrix)
Reaction: C + 2H2 → CH4 ↑                             Reaction: Fe3C + 4H• → 3Fe + CH4
- CH4 vents into process stream                       - CH4 molecule too large to diffuse (0.38 nm)
- Carbon depleted from surface layer                  - Trapped at grain boundaries & inclusions
- Loss of surface hardness & strength                 - Immense pressure builds (>10,000 to >100,000 psi)
- Slight increase in ductility                        - Sub-micron bubbles nucleate and coalesce
- No internal micro-fissuring or cracks               - Intergranular micro-fissuring & macro-rupture

1. Dissociation and Lattice Diffusion

Molecular hydrogen gas ($H_2$) dissociates on the clean, hot catalytic surface of the steel into atomic hydrogen ($H^\bullet$): $H_2 \rightleftharpoons 2H^\bullet$

Because the atomic radius of hydrogen is extremely small (approximately $0.1\text{ nm}$), atomic hydrogen readily dissolves into the body-centered cubic (BCC) ferrite crystal lattice and diffuses interstitially. The rate of hydrogen permeation increases exponentially with increasing metal temperature and hydrogen partial pressure according to Sieverts' Law and Arrhenius diffusion relationships: $C_H = K_S \cdot \sqrt{p_{H_2}} \cdot \exp\left(-\frac{\Delta H_S}{R \cdot T}\right)$

2. Surface Decarburization vs Internal Decarburization

As atomic hydrogen enters the steel, two distinct degradation phenomena can manifest depending on temperature, hydrogen partial pressure, and depth:

1. Surface Decarburization:

   • Occurs when atomic hydrogen reacts with dissolved carbon or metal carbides located directly at or immediately adjacent to the metal surface exposed to the hydrogen gas stream: $C(\text{in solution}) + 2H_2 \rightarrow CH_4 \uparrow$ $Fe_3C + 2H_2 \rightarrow 3Fe + CH_4 \uparrow$
   • The synthesized methane gas readily escapes into the passing process stream. As carbon is removed from the surface, carbon atoms from the near-surface bulk diffuse outward toward the concentration gradient.
   • Consequences: The surface layer transitions from a pearlitic/ferritic structure to pure, carbon-depleted ferrite. This results in a localized reduction in yield strength, tensile strength, and hardness, accompanied by an increase in ductility and impact toughness. Surface decarburization does not produce internal micro-fissuring or cracking.

2. Internal Decarburization and Methane Reaction:

   • Occurs when atomic hydrogen diffuses deep into the bulk metallic structure and encounters iron carbides (predominantly cementite, $Fe_3C$) at grain boundaries, pearlite colony interfaces, and non-metallic inclusions: $Fe_3C + 4H^\bullet \rightarrow 3Fe + CH_4$
   • The Methane Trapping Phenomenon: While atomic hydrogen ($0.1\text{ nm}$) diffuses freely through the BCC iron lattice, the synthesized methane molecule ($CH_4$) has an effective kinetic diameter of approximately $0.38\text{ nm}$. The methane molecule is physically too large to diffuse through the interstitial or substitutional spaces of the solid iron lattice.
   • Extreme Internal Pressure: Unable to migrate, the trapped methane molecules accumulate at grain boundaries, structural discontinuities, and manganese sulfide ($MnS$) inclusion interfaces. Thermodynamic equilibrium methane fugacity can build localized pressures ranging from $>10,000\text{ psi}$ to exceeding $100,000\text{ psi}$ ($70\text{ to }>700\text{ MPa}$).
   • Cavity Coalescence: This immense localized internal gas pressure exceeds the high-temperature yield strength and creep strength of the surrounding metal matrix. Sub-micron methane bubbles nucleate, grow via vacancy diffusion, and coalesce into intergranular micro-cavities. Continued methane generation forces these micro-cavities to link up along grain boundaries, forming planar micro-fissures, which ultimately coalesce into macroscopic intergranular cracks and catastrophic brittle-like rupture.

---

The Incubation Period and Damage Evolution

A defining characteristic of HTHA is the existence of an extended incubation period. The incubation period represents the operational timeframe during which methane bubbles nucleate on a sub-micron scale at grain boundaries, but have not yet linked into micro-fissures or caused detectable changes in physical properties.

Incubation Kinetics and Nelson Parameter

The length of the incubation period is an inverse function of operational severity. Depending on the operating temperature, hydrogen partial pressure, residual stress, and metallurgy, the incubation period can range from thousands of hours to over 300,000 hours (30+ years). Operating excursions above design temperatures accelerate methane generation exponentially, irreversibly consuming the incubation life of the asset.

API RP 941 and the Evolution of the Nelson Curves

1. Historical Origin and Empirical Foundation

To establish safe operational envelopes for steels exposed to high-pressure hydrogen service, the refining and petrochemical industry relies upon API Recommended Practice 941: Steels for Hydrogen Service at Elevated Temperatures and Pressures in Petroleum Refineries and Petrochemical Plants.

In 1949, George A. Nelson of the Shell Development Company compiled operating histories, plant inspection reports, and laboratory data from refineries across the globe to construct empirical limit curves. These boundaries—universally known as the Nelson Curves—plot operating metal temperature against hydrogen partial pressure, delineating safe operating zones from zones where HTHA has been documented.

2. Metallurgical Alloy Hierarchy

Resistance to HTHA increases systematically as carbide-stabilizing alloying elements (primarily chromium, molybdenum, and vanadium) are added to the steel. These elements form thermodynamically stable alloy carbides ($Cr_7C_3$, $Cr_{23}C_6$, $Mo_2C$, $V_4C_3$) that resist reduction into methane by atomic hydrogen.

---

3. The C-0.5Mo Steel Debacle and Eventual Discreditation

From the 1950s through the 1970s, carbon-molybdenum steel (C-0.5Mo), specified under ASTM A204 (plate) and ASTM A335 Grade P1 (piping), was extensively specified for hydroprocessing units operating above the carbon steel Nelson curve. However, starting in the late 1970s, numerous refineries experienced unexpected, catastrophic HTHA failures in C-0.5Mo vessels and piping operating well below the published C-0.5Mo Nelson curve.

Metallurgical Root Cause:

Molybdenum alone in low concentrations (0.5 wt%) forms metastable carbides ($Mo_2C$). Under prolonged operating exposure at temperatures between 550 °F and 750 °F (288 °C to 399 °C), these carbides undergo thermal aging and coarsening. Over thousands of hours, the fine, dispersed molybdenum carbides coalesce or decompose, depleting the surrounding ferrite matrix of carbon and leaving unbonded iron carbides ($Fe_3C$). These residual iron carbides react with dissolved atomic hydrogen to produce methane, initiating HTHA at operating points previously considered safe.

Code Action in API RP 941:

• After repeated failures below the published C-0.5Mo curve, API removed C-0.5Mo from the Figure 1 Nelson curves and moved its discussion to an annex.
• C-0.5Mo is not recommended for new equipment in HTHA service.
• Existing C-0.5Mo equipment is evaluated conservatively (many owners assess it against the carbon steel curves) and may need advanced ultrasonic inspection.

---

4. The 2010 Tesoro Anacortes Disaster and Nelson Curve Overhaul

On April 2, 2010, a catastrophic explosion and fire occurred at the Tesoro refinery in Anacortes, Washington, killing seven refinery workers. The event was triggered by the rupture of a carbon steel heat exchanger (designated E-6600E) in the catalytic reformer naphtha hydrotreater unit.

                                THE 2010 TESORO ANACORTES EVENT
  Equipment: Heat Exchanger E-6600E (Naphtha Hydrotreater Unit)
  Material of Construction: Carbon Steel (ASTM A515 Gr 70), Non-PWHT
  Operating Conditions: Below the then-current API RP 941 carbon steel Nelson curve (per CSB)
  Service: About 38 years (installed in the early 1970s)
                                          │
                                          ▼
  Metallurgical Findings: Extensive intergranular HTHA micro-fissuring and macro-cracks in weld HAZ
  Failure Mechanism: High-Temperature Hydrogen Attack in non-PWHT carbon steel weldments
                                          │
                                          ▼
                         API RP 941 RESPONSE (8th Edition, February 2016)
  Added 12 new data points and a separate, lower curve for
  carbon steel welded without PWHT. The CSB judged the response
  incomplete because RP 941 still did not require safer materials.

Failure Investigation Findings (CSB & API):

1. Long-Term Incubation: The exchanger shell had operated for nearly 40 years without outward indications of damage or wall thinning.
2. Un-PWHT Welds and High Residual Stress: The shell contained non-post-weld heat-treated (non-PWHT) welds. High residual welding tensile stresses (exceeding 50 to 70 ksi / 345 to 483 MPa) combined with operational hoop stress provided the mechanical energy required to accelerate methane bubble growth and cavity dilation.
3. Operating Below the Curve: The CSB found that the exchangers had operated below the then-current API RP 941 carbon steel Nelson curve, which industry had treated as a safe limit for decades.

What Changed in API RP 941 (8th Edition, February 2016):

1. New Non-PWHT Carbon Steel Curve: The 8th edition added 12 new data points and a separate curve for carbon steel welded without PWHT. It sits below the curve that now applies to PWHT carbon steel, reflecting the vulnerability of high-residual-stress weldments.
2. What Did Not Change: The CSB noted that RP 941 still did not require inherently safer materials or verification of actual operating conditions, and it closed its recommendation as "unacceptable action." Many owners therefore apply extra margin or replace carbon steel that operates near the curves.
3. Looking Ahead: Industry work on time-dependent Nelson curves for a future RP 941 edition indicates that older equipment operating just below a curve can still carry HTHA risk after decades of service.

---

Critical Factors Governing HTHA Susceptibility

1. Hydrogen Partial Pressure ($p_{H_2}$):

   • Hydrogen partial pressure represents the thermodynamic driving force for atomic hydrogen dissolution into the steel.
   • Calculated as the total absolute system pressure multiplied by the mole fraction of hydrogen in the process gas stream: $p_{H_2} = P_{\text{total}} \times y_{H_2}$.
   • HTHA generally requires a minimum threshold hydrogen partial pressure of 50 psia (0.34 MPa). Below this pressure, internal methane generation is insufficient to overcome the mechanical restraint of the steel lattice.

2. Operating Metal Temperature:

   • Both the rate of hydrogen dissociation and the diffusion coefficient of hydrogen in steel follow Arrhenius kinetics, increasing exponentially with temperature.
   • Above 400 °F (204 °C), carbon steel becomes vulnerable if hydrogen partial pressure is sufficiently high. The methane reaction kinetics ($Fe_3C + 4H^\bullet \rightarrow 3Fe + CH_4$) accelerate dramatically as temperatures rise.

3. Metallurgy & Carbide-Forming Alloying Elements:

   • Unalloyed carbon steel possesses only iron carbides ($Fe_3C$), which have a relatively high free energy of formation and are readily reduced by hydrogen.
   • Adding Chromium (Cr), Molybdenum (Mo), Vanadium (V), Niobium (Nb), and Titanium (Ti) produces complex alloy carbides ($Cr_{23}C_6$, $Mo_2C$, $V_4C_3$) with lower free energy of formation, making them thermodynamically resistant to methane formation.

4. Stress Level (Applied and Residual Tensile Stress):

   • Tensile stress accelerates vacancy diffusion toward grain boundaries, providing the mechanical volume expansion needed for methane bubbles to nucleate and grow.
   • High residual tensile stresses from welding (typically approaching the yield strength of the material, 50 to 70+ ksi) explain why HTHA cracking initiates predominantly in the heat-affected zone (HAZ) of un-heat-treated weldments.

5. Post-Weld Heat Treatment (PWHT):

   • PWHT at standard temperatures (1150 °F to 1250 °F / 621 °C to 677 °C) serves two critical functions: a) Relieves residual welding stresses, reducing tensile driving force from $>50\text{ ksi}$ down to $<15\text{ to }20\text{ ksi}$. b) Promotes the precipitation and tempering of stable alloy carbides, tying up free carbon and reducing susceptibility to methane synthesis.

6. Cold Work and Plastic Deformation:

   • Unannealed cold bends, cold-formed vessel heads, or mechanically deformed zones contain high dislocation densities that serve as preferential nucleation sites for methane cavities, drastically shortening the incubation time.