A coupled fully kinetic hydrogen transport and ductile phase-field fracture framework for modeling hydrogen embrittlement

This paper presents a novel coupled fully kinetic hydrogen transport and ductile phase-field fracture framework that successfully models hydrogen embrittlement by capturing the critical role of dislocation segregation and the competition between loading rates and diffusion kinetics, thereby reproducing experimental observations of crack initiation shifts, multiple surface cracking, and the transition from ductile tearing to brittle fracture.

The Gist

Imagine a metal bridge or a pipeline as a strong, silent giant. It's built to hold heavy loads, but there's a hidden enemy: Hydrogen.

Hydrogen is like a tiny, invisible ghost that slips into the metal's structure. Once inside, it doesn't just sit there; it causes the metal to become brittle and snap unexpectedly, often under loads that should be perfectly safe. This is called Hydrogen Embrittlement (HE).

For a long time, engineers have struggled to predict exactly how and where this giant will snap. They knew hydrogen was the culprit, but they didn't have a perfect map to see the damage before it happened.

This paper introduces a new, super-smart digital simulator (a computer model) that acts like a "crystal ball" for metal failure. Here is how it works, explained simply:

1. The Two Main Problems They Solved

The researchers realized that previous models were missing two crucial pieces of the puzzle:

• The "Traffic Jam" Problem: Hydrogen atoms are tiny and move fast, but they love to get stuck (segregate) at specific spots inside the metal called dislocations (think of these as tiny traffic jams or potholes in the metal's atomic road). Old models treated these potholes as static parking spots. This new model treats them like moving highways. It understands that hydrogen flows toward these traffic jams, piling up there and weakening the metal specifically at those spots.
• The "Tearing vs. Snapping" Problem: Metals usually fail in two ways:
  • Ductile (Tearing): Like pulling apart a piece of taffy. It stretches, gets thin, and then rips.
  • Brittle (Snapping): Like breaking a dry twig. It just snaps suddenly.
  • Hydrogen makes metal switch from "taffy" to "twig." The new model uses a special mathematical "switch" (a hyperbolic tangent function) to decide: Is the metal being pulled apart (tension) or squeezed? It ensures the model only counts the "tearing" damage when the metal is actually being stretched, mimicking how real voids (tiny bubbles) grow inside the metal.

2. The "Skin Effect" Analogy

One of the coolest things this model discovered is something called the "Skin Effect."

Imagine you have a chocolate bar.

• Without Hydrogen: If you pull the bar, it usually breaks right in the middle, where the stress is highest.
• With Hydrogen: The hydrogen rushes to the surface of the bar first. It turns the outer "skin" of the chocolate into a brittle, dry shell, while the inside remains soft and stretchy.

When you pull this bar, the soft inside wants to stretch, but the brittle skin wants to snap. This creates a conflict. The result? Instead of one clean break in the middle, the bar develops multiple cracks all around the surface, like a ring of dry skin cracking on a fruit.

The model successfully predicted this exact behavior, matching real-world experiments where metal samples developed these strange "ring cracks" around their necks.

3. The Speed Factor (The Race Against Time)

The model also showed that speed matters.

• Fast Pull (High Speed): If you pull the metal very quickly, the hydrogen doesn't have time to move deep inside. It stays stuck on the surface. The metal breaks from the outside in (the "Skin Effect" described above).
• Slow Pull (Low Speed): If you pull the metal very slowly, the hydrogen has plenty of time to diffuse (spread out) evenly throughout the entire bar. The whole bar becomes equally weak. In this case, it breaks from the center out, just like a normal metal bar without hydrogen.

It's like a race: Hydrogen diffusion vs. The speed of your pull. Who wins determines how the metal breaks.

4. Why This Matters

This new framework is like giving engineers a high-definition X-ray that can see the invisible battle between hydrogen and metal atoms.

• It predicts the unexpected: It can tell us why a pipe might develop a ring of cracks instead of a single break.
• It saves money and lives: By understanding exactly how fast hydrogen moves and where it weakens the metal, engineers can design safer pipelines, cars, and airplanes that won't fail suddenly.
• It's efficient: Unlike older, overly complicated models that take forever to run, this one is streamlined and fast, making it practical for real-world engineering.

In a nutshell: The researchers built a digital twin that understands that hydrogen is a "social butterfly" that loves to hang out at metal defects, and that the speed at which you pull the metal changes the entire story of how it breaks. This helps us build stronger, safer structures in a hydrogen-rich world.

Technical Summary

1. Problem Statement

Hydrogen Embrittlement (HE) is a critical failure mechanism in metallic materials where solute hydrogen causes premature fracture at stress levels below design limits. Existing modeling frameworks face significant limitations:

• Kinetic Limitations: Traditional models often assume local equilibrium (e.g., Oriani's theory) between lattice and trapped hydrogen, treating dislocation-trapped hydrogen as immobile. This neglects transient kinetics and phenomena like "pipe diffusion," where hydrogen moves rapidly along dislocation cores.
• Fracture Modeling Gaps:
  • Cohesive Zone Models (CZM) require pre-defined crack paths.
  • Gurson-Tvergaard-Needleman (GTN) models, while effective for ductile damage, require complex non-local formulations to handle mesh sensitivity and often struggle with numerical convergence due to non-isochoric yield surfaces.
  • Standard Phase-Field Models often rely on elastic strain energy for crack driving forces, failing to adequately capture plasticity-driven ductile damage. Furthermore, they typically neglect explicit hydrogen accumulation at dislocations, relying instead on apparent diffusivity, which fails to reproduce the localized hydrogen solubility critical for HE initiation.
• The Gap: There is a lack of a unified framework that couples fully kinetic hydrogen transport (accounting for dislocation segregation and transient exchange) with a ductile phase-field fracture model capable of capturing the competition between loading rates and diffusion kinetics.

2. Methodology

The authors developed a comprehensive chemo-mechanical framework implemented in the open-source code PHIMATS. The approach utilizes a staggered operator-splitting scheme to solve three coupled sub-problems:

A. Fully Kinetic Hydrogen Transport

Instead of assuming local equilibrium, the model treats hydrogen accumulation at microstructural defects via a diffusion-based mechanism.

• Flux Equation: The hydrogen flux includes terms for lattice diffusion, stress-driven diffusion (hydrostatic stress), and dislocation-driven diffusion. The latter incorporates the spatial gradient of the normalized dislocation density ($\bar{\rho}$) as a thermodynamic driving force.
• Dislocation Evolution: The dislocation density is calculated using a Kocks-Mecking-Estrin evolution model, linking it directly to equivalent plastic strain.
• Boundary Conditions: The model accounts for hydrogen exchange at damaged regions (crack surfaces) via a source/sink term, allowing for both desorption (in hydrogen-free environments) and absorption (in hydrogen-rich environments).

B. Ductile Phase-Field Fracture

The framework uses a geometric phase-field approach (Miehe et al.) but introduces a novel ductile crack driving force to replace standard elastic-energy-based formulations.

• Novel Driving Force: The authors propose a driving force ($\tilde{D}_d$) that combines the positive elastic strain energy density ($\tilde{\psi}_{e+}$) and plastic work density ($\tilde{w}_p$).
• Stress Triaxiality Filter: To mimic void-driven ductile damage, the plastic work contribution is scaled by a hyperbolic tangent function of stress triaxiality ($\tanh(\kappa \langle T \rangle)$). This ensures that plastic dissipation only contributes to fracture under tensile conditions (positive triaxiality), effectively capturing the physics of void growth without the numerical complexity of GTN models.
• Hydrogen Effect: Hydrogen embrittlement is modeled by reducing the critical work density ($w_c$) exponentially based on local hydrogen concentration.

C. Coupling Strategy

The framework couples the mechanical deformation (plasticity and dislocation evolution), hydrogen transport, and phase-field fracture. The dislocation density serves as the bridge, driving hydrogen segregation, which in turn lowers the material's resistance to fracture.

3. Key Contributions

1. Fully Kinetic Transport: The model moves beyond local equilibrium assumptions, explicitly modeling the transient exchange of hydrogen at dislocations and enabling the simulation of pipe diffusion.
2. Phenomenological Ductile Driving Force: A simplified yet robust driving force is proposed that captures stress-state dependency (triaxiality) for ductile damage without requiring complex non-isochoric yield surfaces.
3. Dislocation-Hydrogen Coupling: The framework explicitly links dislocation evolution to hydrogen segregation, resolving the localized increase in hydrogen solubility at dislocation sites, which is crucial for predicting HE initiation.
4. Unified Framework: It successfully integrates these elements to model the transition from ductile tearing to brittle fracture and the competition between loading rates and diffusion kinetics.

4. Results and Validation

The model was validated against experimental data for API X80 steel and carbon steel compact tension (CT) specimens:

• Ductile Damage Verification: In the absence of hydrogen, the model successfully reproduced the "cup-and-cone" fracture morphology in notched round bars and the crack coalescence in double-notched specimens, confirming the validity of the new ductile driving force.
• Hydrogen Pressure Effects (Smooth Round Bars):
  • Hydrogen-Free: Damage initiated at the specimen core and propagated outward (standard ductile failure).
  • Low/High Hydrogen Pressure: The model predicted a shift in damage initiation from the core to the surface. At high pressures (30 MPa), it successfully reproduced multiple circumferential surface cracks in the necking region. This "skin effect" arises from the mechanical incompatibility between the ductile core and the hydrogen-embrittled surface layer, a phenomenon driven by dislocation segregation.
• Strain Rate Effects:
  • High Strain Rates: Limited diffusion time leads to surface-limited cracking.
  • Low Strain Rates: Sufficient time allows hydrogen to diffuse uniformly, eliminating surface concentration gradients and reverting to a single, center-initiated crack. The model captured this transition from surface cracking to bulk-dominated fracture.
• Fracture Toughness (CT Specimens): The model accurately reproduced experimental J-resistance curves, showing a significant drop in fracture resistance with increasing hydrogen pressure. It also captured the transition from widespread ductile tearing (large plastic zone) to localized, embrittled crack propagation with irregular, "jagged" paths.

5. Significance

This work represents a significant advancement in computational materials science for hydrogen embrittlement:

• Mechanistic Insight: It demonstrates that hydrogen segregation at dislocations is not just a secondary effect but a critical mechanism required to explain complex failure morphologies like multiple surface cracking.
• Predictive Capability: The framework resolves the long-standing challenge of modeling the competition between loading rates and diffusion kinetics, explaining why failure modes shift from surface-initiated to center-initiated based on strain rate.
• Computational Efficiency: By avoiding the numerical instabilities of GTN models while capturing essential ductile physics, the proposed phase-field approach offers a more efficient and robust tool for simulating HE in complex engineering geometries.
• Practical Application: The ability to reproduce experimental J-resistance curves and stress-strain behaviors under various hydrogen environments makes this framework a powerful tool for designing hydrogen-resistant materials and assessing structural integrity in hydrogen infrastructure.