Incipient magnetic instability in RuO$_2$ with random phase approximation

Using a three-orbital Hubbard model within the Hartree-Fock and random phase approximations, this study identifies commensurate altermagnetic order as the leading magnetic instability in stoichiometric RuO$_2$ without spin-orbit coupling, while noting that finite hole doping or higher temperatures induce incommensurate wave vectors.

The Gist

The Big Picture: Is RuO2 a Magnet or Not?

Imagine Ruthenium Dioxide (RuO₂) as a busy, metallic city. For a long time, scientists thought this city was completely neutral—a "Pauli paramagnet." In this state, the tiny magnetic "compasses" (spins) inside the atoms were just pointing in random directions, canceling each other out. The city was calm and non-magnetic.

However, in recent years, some researchers started shouting, "Wait a minute! This city is actually a magnet!" They claimed the compasses were lining up in a specific pattern called altermagnetism. This is a weird new type of magnetism that acts like a magnet but has no net magnetic pull (like a team where half the players push left and half push right, so the net force is zero, but they are still organized).

The problem? The evidence was messy. Some experiments saw the magnetism, others didn't. It depended on whether they looked at a perfect crystal or a thin film, or if the sample had tiny defects. It was like trying to hear a whisper in a noisy room; sometimes you thought you heard it, sometimes you didn't.

This paper is the team of scientists (Csontosová, Ahn, and Kuneš) stepping in to act as the ultimate audio engineers. They built a super-precise computer model to listen to the "whisper" of the atoms and figure out: Is the city naturally magnetic, or does it need a push to become one?

---

The Tools: The "Hartree-Fock" and "RPA" Microscopes

To solve this, the authors used two powerful theoretical tools:

1. The Hartree-Fock Approximation: Think of this as looking at the city through a standard pair of glasses. It gives you a good, clear picture of how the atoms interact with their immediate neighbors. It's a "mean-field" approach, meaning it assumes everyone is behaving according to the average of the group.
2. The Random Phase Approximation (RPA): This is like putting on noise-canceling headphones that also amplify specific frequencies. It allows the scientists to see how the atoms fluctuate and react to each other in real-time. It helps them spot "instabilities"—moments where the calm city is about to snap into a new, organized state.

They used a simplified map of the city called a 3-orbital Hubbard model. Imagine the atoms in RuO₂ as houses with three specific rooms (orbitals) where the electrons (residents) hang out. The model tracks how these residents move between houses and how they interact when they get crowded.

---

The Discovery: The City is "On the Brink"

The scientists ran their simulations and found something fascinating. The city of RuO₂ is not a magnet in its perfect, calm state. However, it is extremely unstable. It is like a pencil balanced perfectly on its tip. It looks stable, but the slightest nudge (like lowering the temperature or changing the number of residents) will make it fall over into a magnetic order.

Here is what they found:

1. The "Hot Spots" (The Triggers)

In the city's energy map (the Fermi surface), they found three specific neighborhoods, or "hot spots," where the residents are most restless.

• Analogy: Imagine a crowded dance floor. Most people are dancing randomly, but in three specific corners, the music is just right, and the dancers are about to start a synchronized line dance.
• These hot spots are located at specific points in the crystal structure. When the temperature drops, the electrons at these spots decide to lock into a pattern, creating the magnetic order.

2. The "Altermagnet" Dance

When the city finally snaps into order, it doesn't become a normal magnet (where everyone points North). It becomes an altermagnet.

• Analogy: Imagine a checkerboard. In a normal magnet, all the black squares point North, and all the white squares point South. In an altermagnet, the pattern is more complex. The "black squares" might point North, but the "white squares" point East.
• Why it matters: This specific dance creates a huge "spin splitting" (the energy levels for different spins separate by a lot), which explains why some experiments saw huge magnetic effects even if the net magnetism was zero.

3. The Role of Doping (Adding or Removing People)

The scientists tested what happens if you add or remove "residents" (electrons) from the city.

• Hole Doping (Removing people): This is like clearing out some of the crowd. Surprisingly, this makes the city more likely to snap into a magnetic dance. The remaining residents get more restless and organize faster.
• Electron Doping (Adding people): This is like overcrowding the dance floor. It suppresses the magnetism. The extra people get in the way, and the synchronized dance breaks down.
• Real-world connection: This explains why some experiments saw magnetism (they had slightly "hole-doped" samples) and others didn't (they had "electron-doped" or perfect samples).

4. The "Staggered Potential" (Breaking the Symmetry)

Finally, they tested what happens if the city isn't perfectly symmetrical (like if the buildings on one side were slightly different from the other).

• Analogy: Imagine the dance floor has a slight tilt.
• Result: Instead of stopping the dance, the tilt actually helped the dancers get organized! This suggests that in real-world samples (which often have defects or strain), the magnetic order might be easier to trigger than in a perfect crystal. This explains why thin films (which are often strained) show magnetism while bulk crystals might not.

---

The Conclusion: Why This Matters

The paper resolves the controversy by saying: "RuO₂ is not naturally a strong magnet, but it is a 'magnetic sleeper.'"

• In a perfect, pure crystal at room temperature, it stays calm (non-magnetic).
• But it is so close to the edge that tiny changes—like cooling it down, adding a tiny bit of impurities (doping), or straining the material—will wake it up and make it dance in that unique altermagnetic style.

The Takeaway for Everyone:
Think of RuO₂ like a sleeping dragon. It looks like a rock (non-magnetic), but it has a heartbeat. If you poke it just right (with the right temperature or doping), it wakes up and breathes fire (magnetic instability). The scientists didn't just find the dragon; they figured out exactly what kind of poke it takes to wake it up, and why sometimes it wakes up and sometimes it doesn't, depending on the environment.

This helps engineers design better spintronic devices (the next generation of computers that use spin instead of charge) because they now know exactly how to control this "sleeping dragon" to make it do what they want.

Technical Summary

1. Problem Statement

Ruthenium dioxide (RuO$_2$) has been a subject of intense debate regarding its ground state. While early experiments suggested it was a Pauli paramagnet, subsequent studies reported antiferromagnetic (AFM) order with small Ru moments ($0.05 \mu_B$) and identified it as a prototype altermagnet (a phase breaking time-reversal symmetry but preserving crystal rotation symmetry, leading to large spin splitting). However, recent high-precision measurements (neutron diffraction, $\mu$SR, transport) on bulk single crystals suggest a non-magnetic (NM) ground state, attributing previous magnetic signals to surface effects, strain, or doping.

The central problem addressed in this work is to theoretically determine the incipient magnetic instability of stoichiometric RuO$_2$. Specifically, the authors aim to:

• Identify the dominant response channel (spin vs. orbital) and the nature of the ordering vector.
• Understand the microscopic origin of the instability in terms of band structure and Fermi surface (FS) topology.
• Clarify the distinction between altermagnetic and conventional antiferromagnetic mechanisms (Slater mechanism).
• Investigate the sensitivity of the magnetic order to hole/electron doping and symmetry-breaking potentials (simulating strain or surfaces).

2. Methodology

The authors employ a three-orbital Hubbard model focusing on the Ru $t_{2g}$ orbitals ($d_{xy}, d_{yz}, d_{zx}$) within a two-atom unit cell.

• Approximations:
  • Hartree-Fock (HF): Used to treat electronic correlations statically. The self-energy $\Sigma_{HF}$ is treated as static, and spin-orbit coupling (SOC) is neglected to isolate the effects of electron-electron interactions and lattice symmetry.
  • Random Phase Approximation (RPA): Used to calculate the static susceptibility $\hat{\chi}(q, \omega=0)$ and magnon spectra. This allows the authors to probe the instability of the non-magnetic phase without assuming a pre-existing magnetic order.
• Model Hamiltonian:
  • Includes a kinetic term ($H_0$) derived from DFT-based Wannier functions.
  • Includes local interactions ($H_{int}$) described by the Slater-Kanamori Hamiltonian with on-site Coulomb repulsion $U$ and Hund's coupling $J_H$.
• Analysis Techniques:
  • Eigenvalue Analysis: The leading eigenvalue $\Lambda(q)$ of the susceptibility matrix is analyzed to identify the wave vector $q$ of the instability.
  • Eigenvector Overlap: A scalar product is defined to quantify how much the leading eigenvector resembles a Zeeman splitting (spin response) versus other channels.
  • Condensation Energy ($\eta(k)$): A $k$-resolved analysis of the energy gain upon ordering is performed to identify "hot spots" in the Brillouin zone driving the instability.
  • Unfolding: The susceptibility is unfolded to a one-atom unit cell to facilitate comparison with experimental FS data.

3. Key Contributions and Results

A. Dominance of Spin Susceptibility and Instability Vector

• Spin Channel: The analysis confirms that the spin susceptibility is the dominant response channel. The leading eigenvectors of the susceptibility matrix have a $\approx 99\%$ overlap with the Zeeman splitting field, confirming that the instability is magnetic in nature.
• Ordering Vector:
  • Stoichiometric (Undoped): At sufficiently low temperatures, the system exhibits a commensurate altermagnetic order with a wave vector $Q = (0, 0, 0)$ (or equivalent high-symmetry points depending on the specific parameter set).
  • Doped/High-T: With hole doping ($n < 0$) or at higher temperatures, the instability shifts to incommensurate wave vectors along the $c$-axis ($Q_z \approx 0.2 \times 2\pi/c$).
• RPA vs. Non-Interacting: The authors demonstrate that the non-interacting spin bubble ($\chi_0$) is a poor proxy for the full RPA susceptibility. The multi-band nature of RuO$_2$ is crucial; the position of the susceptibility maximum in the interacting system cannot be simply deduced from the non-interacting FS nesting.

B. Microscopic Origin: "Hot Spots" and Band Splitting

The study identifies three primary "hot spots" in the Brillouin zone that drive the magnetic ordering:

1. $K_2$ point: Associated with flat bands near the Fermi level.
2. $K_4$ point (on the $X$-$R$ line): The most prominent hot spot. It arises from $d_{x^2-y^2}$ bands that are quasi-degenerate and touch the Fermi level. The Fermi velocity here is significantly low, enhancing the density of states.
3. $K_5$ point: Associated with a quadratic nodal point of $d_{xz}$ bands.

Mechanism of Splitting:
The paper highlights a fundamental difference between Altermagnets and Slater Antiferromagnets:

• Slater AFM: Band splitting occurs primarily via avoided crossings of back-folded bands near the Fermi level. Isolated bands are insensitive to the staggered field.
• Altermagnet (RuO$_2$): Due to strong site polarization (where Bloch states have unequal weight on the two Ru sublattices), isolated bands can undergo significant spin splitting even without band crossings. The system behaves effectively as two coupled ferromagnets polarized in opposite directions. This mechanism stabilizes the order even when the "Slater" mechanism (gapping of crossings) is suppressed.

C. Effects of Doping and Staggered Potential

• Doping:
  • Hole Doping ($n < 0$): Enhances the magnetic moment and the tendency to order. It shifts the Fermi level to regions of higher density of states (specifically affecting the $d_{x^2-y^2}$ tubular network), reinforcing the $K_4$ hot spot.
  • Electron Doping ($n > 0$): Suppresses the magnetic moment.
• Staggered Potential ($\Delta$):
  • The authors introduce a site-dependent potential to simulate symmetry breaking (e.g., surface effects or strain).
  • Counter-intuitive Result: Contrary to the expectation that a staggered potential would inhibit Slater-type AFM by opening gaps that prevent the energy gain from ordering, the potential enhances the magnetic order in RuO$_2$.
  • Reason: The enhancement is driven by the ferromagnetic-like splitting mechanism enabled by site polarization. The positive effect of hole doping on one sublattice outweighs the negative effect on the other, leading to a net increase in the ordering tendency. This explains why magnetic order is often observed in strained thin films or surfaces where symmetry is broken.

4. Significance and Conclusion

• Resolution of Controversy: The study provides a theoretical framework explaining why RuO$_2$ is often non-magnetic in bulk (where the instability is weak and requires low temperatures or specific doping) but becomes magnetic in thin films or strained samples (where symmetry breaking and doping effects enhance the instability).
• Altermagnetism Mechanism: It clarifies that altermagnetism in RuO$_2$ is not solely driven by the gapping of nodal lines (as in Slater AFM) but significantly by site-polarization-induced spin splitting. This distinguishes altermagnets from conventional antiferromagnets.
• Experimental Guidance: The results suggest that the magnetic state of RuO$_2$ is highly sensitive to stoichiometry and strain. The prediction of incommensurate order under doping provides a target for future experimental verification in doped samples.
• Methodological Insight: The work underscores the necessity of using RPA over simple Stoner or non-interacting models for multi-band systems, as the interplay between bands drastically alters the susceptibility landscape.

In summary, the paper establishes that RuO$_2$ possesses a strong incipient magnetic instability driven by specific Fermi surface features and site-polarized band splitting. While the stoichiometric bulk may remain non-magnetic at accessible temperatures, the system is on the verge of altermagnetic ordering, a state that is readily stabilized by the symmetry-breaking conditions common in epitaxial thin films.