Natural aging and vacancy trapping in Al-6xxx

Abhinav Jain, Michele Ceriotti, W.a. CurtinJournal of Materials research/Pratt's Guide to Venture Capital Sources 2023

TL;DR

This paper uses advanced computational methods to unravel the atomic-level mechanisms behind "natural aging" in Al-6xxx alloys, a process that hinders subsequent strengthening. They discover that clusters formed during natural aging trap vacancies, acting as "prisons" that significantly slow down further growth and explain why these clusters must be dissolved before artificial aging can effectively strengthen the material.

What Problem Does This Paper Address?

Al-6xxx alloys are crucial for their high strength-to-weight ratio, especially in automotive applications. Their strength is typically achieved through a controlled heat treatment process called artificial aging (AA), which precipitates specific Mg-Si structures (β"). However, a phenomenon called "natural aging" (NA) occurs at room temperature between manufacturing steps and AA. NA leads to the formation of undesirable solute clusters that reduce the final achievable strength and delay the AA process. While experimental observations have shown these effects, the underlying atomic-scale mechanisms of NA cluster formation, their evolution, and their interaction with vacancies have remained unclear. This paper aims to provide atomistic insights into these processes to understand why NA is detrimental and how it might be mitigated.

Methods

The researchers employed a combination of advanced computational techniques:

  • Chemically-accurate Neural Network Potential (NNP): They utilized a recently developed NNP specifically for Al-Mg-Si alloys. This potential allows for highly accurate calculations of energies and forces, approaching the accuracy of more computationally expensive Density Functional Theory (DFT) but at a significantly lower cost. This is crucial for simulating larger systems and longer timescales.
  • Kinetic Monte Carlo (kMC) Simulations: Quasi-on-lattice kMC simulations were performed at room temperature to model the time evolution of clustering and vacancy dynamics. This method allows for the exploration of kinetic pathways and the emergence of phenomena over time.
  • DFT Calculations: DFT was used to validate the NNP and to calculate formation energies for various cluster configurations, providing a high-fidelity benchmark.

Datasets

Not explicitly discussed as separate datasets for training or evaluation. The study relies on the NNP developed in previous work [21] and uses DFT calculations for validation.

Key Insights and Ideas

  • "Vacancy Prisons": The most significant finding is the confirmation and detailed characterization of "vacancy prisons." Clusters formed during NA trap vacancies, hindering their movement and thus significantly slowing down the diffusion of solute atoms (Mg and Si) needed for further cluster growth.
  • Non-β" Precursors: The study reveals that the dominant clusters formed during NA are not direct precursors to the desired β" precipitates. Instead, they are distinct, lower-energy structures that require chemical rearrangement (dissolution) before β" can form.
  • Spontaneous Vacancy-Interstitial Formation: For certain cluster sizes (14-18 atoms), the formation of vacancy-Mg interstitial pairs is observed, which are precursors to the β" structure. This suggests a potential pathway for β" formation, but it's complicated by the presence of Si atoms.
  • Si's Stabilizing Role: The presence of Si atoms in specific positions within these non-β" clusters significantly stabilizes them, making them energetically favorable and harder to break down. This explains why these clusters must be dissolved during AA.
  • Kinetics vs. Thermodynamics: The research highlights that while thermodynamically favorable clusters form, their growth is kinetically limited by vacancy trapping, explaining the observed saturation of cluster size and hardness during NA.

Key Results

  • Cluster Formation: 8-14 solute atom clusters form within approximately 1000 seconds of natural aging.
  • Vacancy Trapping Energies: Vacancy binding energies to clusters are found to be around 0.3 eV for surface sites and up to 0.5 eV for interior sites, with larger clusters exhibiting deeper traps.
  • Cluster Size Saturation: The growth of clusters slows considerably due to vacancy trapping, leading to a saturation in cluster size (typically 8-15 atoms) and alloy hardness during NA.
  • NNP Accuracy: The NNP demonstrates excellent agreement with DFT for cluster energetics, validating its use for these simulations.
  • Dissolution Requirement: The findings strongly support the experimental observation that NA clusters must be dissolved during artificial aging before the formation of beneficial β" precipitates can occur.

Limitations and Future Work

  • kMC Time Scales: The authors acknowledge that the kMC time scales are estimated and rely on assumptions about attempt frequency and vacancy concentration. Quantitative agreement on absolute time scales might require further refinement.
  • NNP Limitations: While generally accurate, the NNP showed some deviations for specific initial clusters, which might influence interpretations for larger clusters.
  • Future Directions:
    • Extend kMC simulations to longer times and larger system sizes.
  • Investigate artificial aging at higher temperatures (443 K) to study cluster dissolution and β" nucleation/growth.
  • Develop improved kMC algorithms by incorporating more accurate local composition-dependent vacancy migration barriers.
  • Explore promising heat-treatment schedules to minimize NA and optimize AA.
  • Develop new NNPs for other Al alloys and for studying different aspects of alloy processing.

Summary updated

Abstract

Abstract Undesirable natural aging (NA) in Al-6xxx delays subsequent artificial aging (AA) but the size, composition, and evolution of clustering are challenging to measure. Here, atomistic details of early-stage clustering in Al–1%Mg–0.6%Si during NA are studied computationally using a chemically-accurate neural-network potential. Feasible growth paths for the preferred $$\beta ''$$ <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"> <mml:msup> <mml:mi>β</mml:mi> <mml:mrow> <mml:mo>′</mml:mo> <mml:mo>′</mml:mo> </mml:mrow> </mml:msup> </mml:math> precipitates identify: dominant clusters differing from $$\beta ''$$ <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"> <mml:msup> <mml:mi>β</mml:mi> <mml:mrow> <mml:mo>′</mml:mo> <mml:mo>′</mml:mo> </mml:mrow> </mml:msup> </mml:math> motifs; spontaneous vacancy-interstitial formation creating 14–18 solute atom $$\beta ''$$ <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"> <mml:msup> <mml:mi>β</mml:mi> <mml:mrow> <mml:mo>′</mml:mo> <mml:mo>′</mml:mo> </mml:mrow> </mml:msup> </mml:math> -like motifs; and lower-energy clusters requiring chemical re-arrangement to form $$\beta ''$$ <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"> <mml:msup> <mml:mi>β</mml:mi> <mml:mrow> <mml:mo>′</mml:mo> <mml:mo>′</mml:mo> </mml:mrow> </mml:msup> </mml:math> nuclei. Quasi-on-lattice kinetic Monte Carlo simulations reveal that 8–14 solute atom clusters form within 1000 s but that growth slows considerably due to vacancy trapping inside clusters, with trapping energies of 0.3 $$-$$ <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"> <mml:mo>-</mml:mo> </mml:math> 0.5 eV. These findings rationalize why cluster growth and alloy hardness saturate during NA, confirm the concept of “vacancy prisons”, and suggest why clusters must be dissolved during AA before formation of $$\beta ''$$ <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"> <mml:msup> <mml:mi>β</mml:mi> <mml:mrow> <mml:mo>′</mml:mo> <mml:mo>′</mml:mo> </mml:mrow> </mml:msup> </mml:math> . This atomistic understanding of NA may enable design of strategies to mitigate negative effects of NA. Graphical abstract Energy and RMS vacancy displacement vs simulation time for three different kMC simulations of Al matrix containing Mg (green) and Si atoms(orange), and 1 vacancy (pink). Geometries extracted from the third trajectory at four different times show solute clustering and vacancy trapping.
BibTeX
@article{jain2023natural,
  author = {Abhinav Jain and Michele Ceriotti and W.A. Curtin},
  title = {{Natural aging and vacancy trapping in Al-6xxx}},
  journal = {Journal of Materials research/Pratt's Guide to Venture Capital Sources},
  year = {2023},
}
Figure