Nonmagnetic Ground State of Rutile RuO$_2$ from Diffusion Quantum Monte Carlo

Using fixed-node diffusion quantum Monte Carlo, this study resolves the debate over the magnetic ground state of bulk rutile RuO$_2$ by demonstrating that it is nonmagnetic in its pristine structure but can be stabilized into an antiferromagnetic state under 3% compressive strain, thereby reconciling conflicting experimental reports.

The Gist

The Mystery of the "Chameleon" Metal

Imagine you have a piece of metal called Rutile RuO₂. For years, scientists have been arguing over what this metal is actually doing deep inside.

• Team A says: "It's a magnet! Specifically, a weird new kind called an altermagnet that has special properties useful for future computers."
• Team B says: "No, it's not magnetic at all. It's just a regular, non-magnetic metal."

Why the disagreement? Because the tools scientists usually use to predict how atoms behave (called DFT or Density Functional Theory) are like different types of weather forecasters. Some forecasters say it will rain; others say it will be sunny. They all use the same data, but they tweak their formulas slightly, and the results change completely. For this specific metal, the "weather" is so delicate that a tiny change in the math flips the prediction from "Magnetic" to "Not Magnetic."

The "Gold Standard" Detective

To settle the argument, the authors of this paper brought in a super-advanced detective tool called Diffusion Quantum Monte Carlo (DMC).

Think of DFT as a sketch artist who draws a picture based on a quick description. It's fast, but sometimes it gets the details wrong.
DMC is like a high-definition 3D scanner. It doesn't just guess; it simulates the behavior of every single electron in the material with extreme precision. It's computationally expensive (it takes a lot of computer power), but it's the most accurate way to see the truth.

The Verdict: The "Perfect" Metal is Chill

When the authors ran their high-precision DMC simulation on a perfect, untouched piece of RuO₂, the result was clear:

The metal is non-magnetic.

In its natural, pristine state, the atoms are not lining up to create a magnetic field. In fact, the non-magnetic state is energetically "happier" (more stable) than the magnetic state by a small but significant margin.

This explains why some recent experiments (using muons, which are like tiny magnetic spies) couldn't find any magnetism in the bulk material. The "perfect" version of this metal is just a regular, non-magnetic conductor.

The Twist: Strain is the Switch

So, if it's not magnetic, why did other scientists see magnetism?

The paper reveals that RuO₂ is like a tightrope walker. It is standing on the very edge of a cliff between "Non-Magnetic" and "Magnetic." It is so sensitive that a tiny push can knock it over.

That "push" is strain (squeezing or stretching the material).

• The Analogy: Imagine a spring that is barely holding its shape. If you squeeze it just a tiny bit (about 3% compression), it suddenly snaps into a new shape.
• The Result: When the researchers simulated squeezing the metal (compressive strain), the metal did become magnetic. It snapped into an "antiferromagnetic" state (where the magnetic spins cancel each other out but are still ordered).

Why This Matters

This discovery solves a huge puzzle in the world of spintronics (using electron spin for computing):

1. It explains the confusion: The "magnetic" behavior seen in some experiments wasn't a mistake; it was likely caused by the metal being squeezed or stretched by the surface it was grown on (like a thin film on a substrate). The strain acted as the switch.
2. It's a new control knob: We don't need to change the chemical recipe to make this metal magnetic; we just need to squeeze it. This makes RuO₂ a fantastic candidate for tunable devices.
3. It validates the "Altermagnet" idea: The paper shows that while the bulk metal isn't magnetic, the texture of the magnetism (the altermagnetic pattern) is still there in the atomic structure. It just needs that little nudge of strain to wake up.

The Bottom Line

The paper concludes that pure, perfect RuO₂ is non-magnetic. However, it is sitting on a hair-trigger. If you squeeze it even a little bit, it becomes a magnetic material with unique properties.

This means that the "magnetic" RuO₂ we see in labs isn't a different material; it's the same material, just wearing a "strained" outfit. This insight helps scientists stop arguing about what the material is and start focusing on how to tune it for better electronics.

Technical Summary

1. Problem Statement

The magnetic ground state of stoichiometric rutile RuO₂ has been a subject of significant debate.

• The Conflict: RuO₂ was proposed as a candidate for altermagnetism (a symmetry-protected magnetic phase bridging ferromagnetic and antiferromagnetic properties). However, experimental and theoretical results are contradictory:
  • Supporting Magnetism: Early X-ray scattering, neutron diffraction, and Density Functional Theory (DFT) calculations with Hubbard-$U$ corrections ($DFT+U$) suggested antiferromagnetic (AFM) order or spin-polarized states.
  • Supporting Non-magnetism: Recent muon spin rotation ($\mu$SR) measurements failed to detect static magnetic moments in bulk or epitaxial samples. Other bulk-sensitive probes indicate a paramagnetic metallic response.
• Theoretical Limitation: Standard DFT approaches (LDA, GGA, meta-GGA, and hybrid functionals) yield conflicting predictions for RuO₂. The predicted magnetic ordering energy ($E_{AFM} - E_{NM}$) and local Ru moments vary drastically depending on the exchange-correlation functional and the choice of tuning parameters (e.g., the Hubbard $U$ or the exact-exchange fraction $\omega$ in hybrid functionals). This sensitivity makes DFT unreliable for determining the ground state of this itinerant 4d metal, where the relevant energy scales are small.

2. Methodology

To resolve the controversy, the authors employed Fixed-Node Diffusion Quantum Monte Carlo (FN-DMC), a high-accuracy many-body method that goes beyond mean-field approximations.

• Computational Framework: Calculations were performed using the QMCPACK code.
• Trial Wave Functions: Since FN-DMC energy depends on the nodal surface of the trial wave function, the authors systematically tested different trial orbitals generated from:
  • PBE0($\omega$): Hybrid functionals where the exact-exchange fraction $\omega$ was treated as a variational parameter.
  • DFT+$U$: Including LDA+$U$, PBE+$U$, and SCAN+$U$, where $U$ was varied.
• Variational Approach: Instead of fixing parameters based on standard conventions, the authors used the many-body energy minimization principle to find the optimal $\omega$ or $U$ for each functional family. This allowed for a direct, unbiased comparison of Nonmagnetic (NM) vs. Antiferromagnetic (AFM) states.
• Strain Analysis: The study also applied $\pm 3\%$ uniaxial strain along the $z$-direction (simulating epitaxial growth conditions) to investigate the tunability of the magnetic state.

3. Key Results

A. Ground State Determination (Unstrained Bulk)

• Nonmagnetic Ground State: The FN-DMC calculations definitively identify the nonmagnetic (NM) metallic state as the ground state for pristine, stoichiometric bulk RuO₂.
• Energy Difference: The NM state is lower in energy than the lowest AFM state considered by 23(9) meV per formula unit.
• Optimal Parameters: Within the PBE0 family, the variational minimum occurs at an exact-exchange fraction of $\omega = 0.10$. At this point, the NM solution is stable.
• Rejection of AFM: While DFT+$U$ and high-$\omega$ PBE0 calculations can stabilize AFM states, these states correspond to higher many-body energies in DMC. Furthermore, these high-energy AFM states are predicted to be insulating, which contradicts experimental transport data showing RuO₂ is a metal.

B. Strain-Induced Instability

• Compressive Strain: Applying 3% compressive strain along the $z$-axis reverses the energy balance. The AFM state becomes energetically favorable (negative ordering energy), stabilizing an AFM phase with a significant ordered moment of 0.97(2) $\mu_B$.
• Tensile Strain: Under 3% tensile strain, the NM state remains the ground state.
• Mechanism: Compressive strain drives the system toward a more correlated magnetic regime by shifting the optimal $\omega$ to higher values and enhancing Ru-O localization.

C. Electronic Structure and Charge Density

• Spin Texture: When an AFM state is forced (e.g., at $\omega=0.20$), the spin density exhibits an altermagnetic-like texture with sign-alternating spin lobes related by 90° rotation, preserving the symmetry-consistent local texture even with many-body correlations.
• Charge Redistribution: Comparing the AFM state to the optimal NM state reveals a reorganization of Ru-O hybridization. The AFM state shows enhanced charge density in Ru-$d_{z^2}$ orbitals and O-$p_\pi$ lobes, indicating a shift from itinerant in-plane hybridization to a more localized, out-of-plane configuration.

4. Significance and Implications

• Resolution of Controversy: The study resolves the long-standing debate by proving that stoichiometric bulk RuO₂ is intrinsically nonmagnetic. The conflicting experimental reports of altermagnetism are likely driven by extrinsic factors (such as strain, disorder, or doping) rather than an intrinsic bulk property.
• Strain Tunability: The paper establishes RuO₂ as a system sitting on the verge of a magnetic instability. It demonstrates that strain is a critical "knob" for tuning RuO₂ between an itinerant nonmagnetic metal and a correlated antiferromagnetic insulator. This explains why epitaxial thin films (which often experience strain) show altermagnetic signatures while bulk samples do not.
• Methodological Benchmark: The work highlights the failure of standard DFT functionals for itinerant 4d metals near magnetic instabilities. It demonstrates that FN-DMC is essential for providing a decisive, controlled benchmark to resolve disputes where mean-field theories yield parameter-dependent, contradictory results.
• Design of Altermagnetic Insulators: By linking the magnetic transition to charge disproportionation and metal-insulator transitions, the study offers a rational strategy for designing new altermagnetic insulators by manipulating charge and bond disproportionation in RuO₂-based structures.

Conclusion

The authors conclude that while RuO₂ is not an intrinsic altermagnet in its pristine bulk form, it is a highly tunable material where modest compressive strain can induce a robust antiferromagnetic state. This finding reconciles experimental discrepancies and provides a clear framework for understanding and engineering magnetic properties in transition metal oxides.