Phase field model for self-climb of prismatic dislocation loops by vacancy pipe diffusion

Xiaohua Niu, Yang Xiang, Xiaodong Yan

Dislocations—line defects in the crystalline structure of metals and ceramics—are the primary drivers of plastic deformation and material aging. While we often think of them gliding through a crystal, they can also "climb" by moving out of their slip planes. This paper introduces a robust numerical framework to simulate a specific, tricky type of motion: self-climb via vacancy pipe diffusion.

TL;DR

Researchers have developed a new phase field model based on the Cahn-Hilliard equation to simulate how prismatic dislocation loops move and merge through "pipe diffusion" (vacancies traveling along the dislocation line itself). Unlike previous methods, this model handles complex topological changes—like multiple loops merging into one—automatically and maintains the physical requirement that the area enclosed by the loops remains constant.

What Problem Does This Paper Address?

In irradiated materials or at moderate temperatures, prismatic dislocation loops move primarily through self-climb. Unlike standard climb (driven by vacancies moving through the bulk of the crystal), self-climb is driven by vacancies moving along the dislocation line (pipe diffusion). This motion is conservative: the total number of atoms/vacancies within the loop doesn't change, so the loop's enclosed area stays constant even as it changes shape or translates.

Previously, researchers used Discrete Dislocation Dynamics (DDD), a "front-tracking" method where the dislocation line is discretized into segments. While accurate for single loops, DDD struggles with:

  1. Topological changes: Merging two loops requires manual intervention and complex "if-then" logic.
  2. Numerical Instability: As loops get close, time steps must become infinitesimally small.
  3. Re-meshing: As a loop moves, the points defining it bunch up on one side, requiring constant, computationally expensive re-gridding.

Methods

The authors turn to the Phase Field Method (PFM) to solve these issues. Instead of tracking the line itself, they use an order parameter ($\phi$) that varies smoothly across the domain (0 inside the loop, 1 outside).

Key components of their approach include:

  • Cahn-Hilliard Framework: They use the Cahn-Hilliard equation because it inherently describes conserved dynamics (perfect for the area-conserving nature of self-climb).
  • Incorporating Elastic Forces: They integrate the Peach-Koehler force (the elastic force dislocations exert on each other) directly into the chemical potential of the phase field.
  • Degenerate Mobility: They use a specific mobility function that ensures diffusion only happens at the "interface" (the dislocation core), mimicking pipe diffusion.
  • Asymptotic Analysis: The authors perform rigorous mathematics to prove that as the "diffuse" interface of their model becomes infinitely sharp, it perfectly recovers the classical physics equations for dislocation velocity ($v \propto -d^2f/ds^2$).

Datasets

As a theoretical and computational physics paper, this work does not use external datasets. Instead, the model is validated against:

  • Theoretical Predictions: Analytical solutions for loop translation velocities and equilibrium shapes.
  • DDD Benchmarks: Results from previous front-tracking simulations (specifically Ref [13] in the paper).
  • Experimental Observations: Qualitative behaviors of dislocation loops observed in electron microscopy.

Key Insights and Ideas

  • The "Corrector" Term: The asymptotic analysis revealed that the phase field framework naturally generates an extra curvature term. Rather than an error, this term acts as a "corrector" that compensates for the fact that the numerical dislocation core in the model is slightly larger than a real physical core.
  • Handling Opposite Orientations: To simulate both vacancy loops (counter-clockwise) and interstitial loops (clockwise) simultaneously, the authors found that using a single order parameter was numerically unstable. They solved this by using two separate order parameters ($\phi_1, \phi_2$) that interact only through long-range elastic force fields.
  • Automatic Topology: The biggest "aha" moment is the demonstration of seven loops merging into one. In DDD, this would be a coding nightmare; here, the math of the Cahn-Hilliard equation handles the merger seamlessly as the loops' elastic fields pull them together.

Key Results

  • Area Conservation: An initial elliptic loop evolved into a perfect circle. The final radius was 54.9b, nearly identical to the theoretical prediction of 56.6b based on area conservation.
  • Translation Accuracy: A circular loop moving under a stress gradient achieved a velocity of $1.95 \times 10^{-5} c_0D_c/b$, showing excellent agreement with the DDD value ($1.94 \times 10^{-5}$).
  • Multi-Loop Coalescence: The model successfully simulated the merging of seven randomly placed loops into a single stable circular loop, matching theoretical area conservation within a 0.3% margin of error.
  • Repelling Behavior: The model correctly captured the repulsive motion between loops with opposite Burgers vectors, a critical feature for simulating complex microstructures.

Limitations and Future Work

  • Dimensionality: The current study is restricted to a 2D climb plane. While many prismatic loops stay in one plane, real-world materials involve 3D networks. Extending this to a full 3D phase field model is the logical next step.
  • Computational Cost: The Cahn-Hilliard equation is a 4th-order PDE, which typically requires very small time steps for stability. While the authors used a pseudospectral method to speed things up, 3D simulations will likely require more advanced semi-implicit time-stepping schemes.
  • Core Resolution: The model assumes a specific core width ($\epsilon$). Exploring how different core energy formulations (beyond the double-well potential) affect the results could provide deeper physical insights.

Reference: Niu, X., Xiang, Y., & Yan, X. (2020). Phase field model for self-climb of prismatic dislocation loops by vacancy pipe diffusion. arXiv preprint arXiv:2012.01612v1.

Summary updated

Abstract

In this paper, we present a phase field model for the self-climb motion of prismatic dislocation loops via vacancy pipe diffusion driven by elastic interactions. This conserved dynamics model is developed under the framework of the Cahn-Hilliard equation with incorporation of the climb force on dislocations, and is based on the dislocation self-climb velocity formulation established in Ref.[1]. The phase field model has the advantage of being able to handle the topological and geometrical changes automatically during the simulations. Asymptotic analysis shows that the proposed phase field model gives the dislocation self-climb velocity accurately in the sharp interface limit. Numerical simulations of evolution, translation, coalescence and repelling of prismatic loops by self-climb show excellent agreement with discrete dislocation dynamics simulation results and the experimental observation.
BibTeX
@article{niu2021phase,
  author = {Xiaohua Niu and Yang Xiang and Xiaodong Yan},
  title = {{Phase field model for self-climb of prismatic dislocation loops by vacancy pipe diffusion}},
  journal = {International Journal of Plasticity},
  year = {2021},
  doi = {10.1016/j.ijplas.2021.102977},
}

Loading related content...

Figure