Hydrogen uptake and hydride formation in Al$_x$CoCrFeNi high-entropy alloys: First-principles, universal-potential, and experimental study

This study combines high-pressure experiments, density-functional theory, and a universal interatomic potential to reveal that aluminum suppresses hydrogen uptake and hydride formation in Al$_x$CoCrFeNi high-entropy alloys primarily through increased solution energies and interstitial site destabilization, with B2 crystal ordering acting as a secondary inhibitory factor.

The Gist

Imagine you have a team of five different friends (Aluminum, Cobalt, Chromium, Iron, and Nickel) who love to hang out together in a chaotic, crowded party. In the world of materials science, this chaotic mix is called a High-Entropy Alloy. It's a super-strong metal that's great for making things that need to last a long time.

But there's a problem: Hydrogen.

Hydrogen is like a tiny, mischievous ghost. When it sneaks into these metal parties, it can cause the metal to crack, crumble, or become brittle. This is called "hydrogen embrittlement." On the flip side, if we can control how the metal absorbs hydrogen, we could use these alloys as super-efficient batteries to store clean energy.

The big question the scientists asked was: "How do we stop the ghost from ruining the party, or how do we invite it in safely?"

To find the answer, they looked at two specific versions of this metal team:

1. The "Low-Aluminum" Team: Mostly Cobalt, Chromium, Iron, and Nickel, with just a tiny bit of Aluminum.
2. The "High-Aluminum" Team: A version where Aluminum is the star of the show.

The Experiment: Squeezing the Metal

The scientists put these metal teams inside a tiny, super-strong diamond box (like a microscopic vice) and squeezed them with the pressure of a mountain, all while pumping in hydrogen gas.

• The Low-Aluminum Team: As soon as the pressure hit a certain point (about 3 GPa, which is like the pressure 30 kilometers underwater), this team swelled up. They happily absorbed the hydrogen ghosts and formed a new, hydrogen-filled structure. They were like sponges soaking up water.
• The High-Aluminum Team: The scientists squeezed this team even harder—up to 50 GPa (pressure deeper than the Mariana Trench!) and even heated them up. Nothing happened. The metal stayed exactly the same size. The hydrogen ghosts tried to get in, but the Aluminum team slammed the door shut. They were completely inert.

The Detective Work: Why did this happen?

To understand why the Aluminum team was so tough, the scientists used two powerful tools:

1. Supercomputers (DFT): These ran complex math to simulate how atoms interact.
2. AI (The "Universal Potential"): Think of this as a highly trained AI assistant. Instead of doing the heavy math for every single atom (which takes forever), the AI learned the rules of physics from millions of examples. It could predict how the metal would behave almost instantly and just as accurately as the supercomputer.

The AI and the Supercomputer agreed on the verdict:

1. The "Aluminum Shield" (Chemistry)

Aluminum is naturally very picky about who it hangs out with. It doesn't like hydrogen. When you add a lot of Aluminum to the mix, it changes the chemical "atmosphere" of the metal. It's like putting a "No Ghosts Allowed" sign on the door. The Aluminum atoms make it energetically expensive (uncomfortable) for hydrogen to squeeze in.

2. The "Crowded Room" (Structure)

The High-Aluminum team arranges themselves in a very specific, orderly pattern (called a B2 structure). It's like a perfectly organized dance floor where everyone knows their spot. There are very few empty spaces (interstitial sites) for the tiny hydrogen ghosts to hide. Even if they try to squeeze in, they get pushed out.

The Low-Aluminum team, however, is more chaotic (a FCC structure). It's like a mosh pit with lots of gaps. Hydrogen can easily slip into the empty spaces between the atoms.

3. The "Room Size" (Volume)

Aluminum atoms are a bit larger than the others. Adding them makes the whole metal structure expand, like inflating a balloon. You might think a bigger balloon would be easier to fill with more stuff, but in this case, the Aluminum atoms are so "hydrogen-phobic" that even the extra space doesn't help. The chemical dislike of hydrogen wins over the extra space.

The Big Takeaway

The most surprising discovery was that the specific shape of the metal (whether it's a cube or a different shape) mattered the least.

If you took the Low-Aluminum team and forced them into the High-Aluminum shape, they would still absorb hydrogen. If you took the High-Aluminum team and forced them into the Low-Aluminum shape, they would still reject hydrogen.

The real boss is the "Recipe" (Composition).

• Too much Aluminum? The metal becomes a fortress against hydrogen.
• Just a little Aluminum? The metal becomes a sponge for hydrogen.

Why does this matter?

This study is like finding the "secret sauce" for designing new materials.

• If you want to build a hydrogen tank for a fuel-cell car, you want a metal that loves hydrogen (Low Aluminum).
• If you want to build a pipeline or a bridge that won't crack from hydrogen exposure, you want a metal that hates hydrogen (High Aluminum).

By using this new AI tool (the Universal Potential), scientists can now quickly test thousands of different "recipes" on a computer to find the perfect metal for the job, without having to build and break them in a lab first. It's a faster, cheaper way to design the materials of the future.

Technical Summary

1. Problem Statement

The study addresses the critical challenge of understanding hydrogen uptake and hydride formation in complex multicomponent alloys, specifically High-Entropy Alloys (HEAs). While hydrogen storage and resistance to hydrogen-induced degradation are vital for structural materials, the mechanisms governing hydrogen solubility in HEAs are not fully understood.

• Specific Focus: The research investigates the Al$_x$CoCrFeNi system, comparing two distinct compositions:
  • Al$_{0.3}$CoCrFeNi: Stabilizes in a single-phase Face-Centered Cubic (FCC) structure.
  • Al$_3$CoCrFeNi: Stabilizes in an ordered Body-Centered Cubic (B2, CsCl-type) structure.
• Key Question: How do chemical composition (specifically Aluminum content), crystal structure (FCC vs. B2), local chemical ordering, and lattice volume interact to control hydrogen affinity? Previous studies struggled to disentangle these factors due to the computational cost of sampling vast configuration spaces using traditional Density Functional Theory (DFT).

2. Methodology

The authors employed a multi-scale, complementary approach combining high-pressure experiments, DFT, and machine-learned interatomic potentials.

• Experimental Techniques:

  • High-Pressure Compression: Conducted using Diamond Anvil Cells (DAC) and Large-Volume Press (LVP) at PETRA-III (DESY).
  • Media: Experiments used Hydrogen (H$_2$) and Neon (Ne) as pressure-transmitting media to distinguish between pure compression and hydrogen uptake.
  • Conditions: Room temperature up to 55 GPa (DAC) and high-temperature (up to 720°C) up to 10 GPa (LVP) using NH$_3$BH$_3$ as a hydrogen source.
  • Analysis: In-situ X-ray diffraction (XRD) monitored volume changes ($V(p)$) to detect hydride formation (indicated by volume discontinuities or expansion).

• Computational Techniques:

  • Density Functional Theory (DFT): Performed using VASP with PBE-GGA. Used Special Quasi-Random Structures (SQS) for disordered phases and ordered models for B2 phases. Calculated solution energies and equation of state (EOS) parameters.
  • Universal Interatomic Potential (GRACE): Utilized the GRACE (Graph Atomic Cluster Expansion) foundational potential, pretrained on large DFT databases (OMat24). This allowed for exhaustive sampling of interstitial sites and configurations that are computationally prohibitive for DFT alone.
  • Validation: The GRACE potential was rigorously validated against DFT for both hydrogen-free alloys and hydrogenated systems (dilute limit and hydrides).

3. Key Contributions

1. Validation of Universal Potentials for Hydrogen: The study demonstrates that the GRACE foundational potential, trained without explicit H$_2$ energetics, can accurately reproduce DFT-level trends for hydrogen solution energies and hydride formation in chemically complex, disordered alloys. It successfully captures the transition from dilute interstitials to concentrated hydrides.
2. Disentangling Structural vs. Chemical Effects: By systematically varying composition, volume, and chemical order in simulations, the authors isolated the specific contributions of Aluminum content, lattice volume, and crystal structure to hydrogen affinity.
3. Comprehensive Experimental-Theoretical Benchmark: Provided a direct comparison between high-pressure experimental data and theoretical predictions, confirming that volume expansion is a reliable indicator of hydride formation in these systems.

4. Key Results

A. Experimental Findings

• Al$_{0.3}$CoCrFeNi (FCC): At room temperature, this alloy absorbs hydrogen and forms hydrides above 3 GPa. High-temperature experiments yielded a hydrogen content of approximately MH$_{0.9}$. A distinct volume expansion was observed, signaling hydride formation.
• Al$_3$CoCrFeNi (B2): This alloy remained inert to hydrogen uptake even at pressures up to 50 GPa and temperatures up to 150°C. No volume anomalies indicative of hydride formation were detected.

B. Computational Findings

• Energetics: DFT and GRACE calculations confirmed that hydrogen incorporation is energetically favorable in the FCC phase but unfavorable in the B2 phase. The B2 phase exhibits an energetic penalty of approximately 0.3 eV per H atom compared to the FCC phase.
• Volume Effects: The effective volume of hydrogen in the B2 hydride (if formed) is significantly larger (4 Å$^3$/H) than in the FCC hydride (2 Å$^3$/H), indicating a higher volumetric penalty for the B2 structure.
• Decoupling Factors (The "Why"):
  • Aluminum Content (Primary Driver): High Al concentration chemically suppresses hydrogen uptake by increasing solution energies.
  • Chemical Ordering (Secondary Driver): The ordered B2 structure (where Al atoms occupy specific sublattices) further suppresses uptake by reducing the statistical probability of "Al-poor" local environments where hydrogen might sit. In disordered BCC models, low-energy sites exist but are rare; in ordered B2, they are even less accessible.
  • Crystal Structure (Minor Role): Once composition and volume are fixed, the difference between FCC and BCC/B2 lattices has only a modest effect on hydrogen affinity.
  • Volume Expansion: While the larger lattice volume of Al-rich alloys partially compensates for the chemical penalty of Al, it is insufficient to reverse the trend.
• Magnetism: Spin-polarized vs. non-spin-polarized calculations showed that magnetic contributions have a negligible effect on the relative stability of hydrides in these alloys.

5. Significance

• Material Design: The study establishes that chemical composition (specifically Al content) is the dominant factor controlling hydrogen solubility in Al-containing HEAs, rather than the crystal structure itself. This provides a clear guideline for designing hydrogen-resistant structural materials (high Al, ordered B2) or hydrogen storage materials (low Al, disordered FCC).
• Methodological Advancement: It validates the use of "universal" machine-learned potentials (like GRACE) for studying hydrogen in complex alloys. This approach overcomes the computational bottleneck of DFT, enabling the screening of vast compositional and configurational spaces that were previously inaccessible.
• Fundamental Understanding: The work clarifies that the resistance of the B2 phase to hydrogen is not merely a structural issue but a result of the synergistic effect of high Aluminum concentration and long-range chemical ordering, which destabilize interstitial sites and increase solution energies.

In conclusion, the paper demonstrates that while crystal structure plays a role, the Aluminum-driven chemical effects and ordering are the primary determinants of hydrogen behavior in Al$_x$CoCrFeNi high-entropy alloys.