Assessing the suitability of the Thomas-Fermi-von Weizsäcker density functional for itinerant magnetism

This study demonstrates that the Thomas-Fermi-von Weizsäcker orbital-free density functional fails to qualitatively describe itinerant magnetism in metals like Al, Pd, Fe, Co, and Ni, revealing fundamental limitations when benchmarked against Kohn-Sham DFT.

The Gist

Imagine you are trying to predict how a crowd of people (electrons) will behave in a giant stadium (a metal). Specifically, you want to know if they will stay calm and mixed together (paramagnetic) or if they will suddenly all start shouting in the same direction, creating a unified roar (ferromagnetism/magnetism).

This paper is a report card on a specific tool scientists use to make these predictions. The tool is called Orbital-Free Density Functional Theory (OF-DFT), and the specific version they tested is the Thomas-Fermi-von Weizsäcker (TFW) functional.

Here is the breakdown of the study using simple analogies:

1. The Two Competing Tools

To understand the problem, we need to compare two ways of simulating electrons:

• The "Gold Standard" (Kohn-Sham DFT): Imagine this is a high-end, slow-motion camera. It tracks every single person in the stadium individually. It knows exactly where everyone is, how fast they are moving, and what they are doing. It is incredibly accurate but requires a massive computer and takes a long time to run. It can handle millions of people, but only if you have a supercomputer.
• The "Fast & Cheap" Tool (Orbital-Free DFT): This is like looking at the stadium from a drone and just counting the density of the crowd. It doesn't track individuals; it just sees a "cloud" of people. It is incredibly fast and can simulate stadiums with millions of people in seconds. However, because it ignores the individuals, it has to guess how the crowd moves based on general rules.

2. The Big Question

The researchers asked: "Can the 'Fast & Cheap' tool (TFW) accurately predict when a metal will become magnetic?"

Magnetism in metals (like Iron or Nickel) is tricky. It happens because of very subtle, sharp details in how the electrons are arranged near the "edge" of their energy levels (the Fermi level). It's like a tightrope walker: a tiny shift in balance decides if they fall (become magnetic) or stay standing (stay non-magnetic).

3. The Experiment

The team tested the TFW tool on five metals:

• Aluminum (Al) & Palladium (Pd): These are "calm" metals that don't want to be magnetic, though Palladium is very close to the edge.
• Iron (Fe), Cobalt (Co), & Nickel (Ni): These are the "loud" metals that are naturally magnetic.

They ran the simulation three ways:

1. The Gold Standard: Full, slow-motion tracking (KS-KS).
2. The Fast Tool: Pure density guessing (OF-OF).
3. The Hybrid: Using the Fast Tool's crowd map, but feeding it into the Gold Standard's brain to calculate the energy (OF-KS).

4. The Results: A Tale of Two Outcomes

The Failure of the Pure Fast Tool (OF-OF):
When the researchers used the pure "Fast & Cheap" tool, it failed miserably at predicting magnetism.

• The Analogy: Imagine trying to predict a riot by looking at the crowd from a drone. The drone sees a smooth, calm cloud of people. It concludes, "Everyone is happy; no riot will happen."
• The Reality: In Iron, Cobalt, and Nickel, the electrons are actually on the verge of a riot (magnetism). The TFW tool was too "blurry" to see the sharp, jagged details of the electron crowd. It told the scientists these metals were stable and non-magnetic, which was completely wrong. It missed the qualitative trend entirely.

The Partial Success of the Hybrid (OF-KS):
When they took the "Fast Tool's" crowd map but ran it through the "Gold Standard's" brain, things got better.

• The Analogy: The drone still gives a blurry map of the crowd, but now a smart analyst looks at that map and says, "Wait, the density here is weird; they might riot."
• The Result: For Iron and Nickel, this hybrid approach correctly predicted they would become magnetic. For Palladium, it got closer to the truth, though not perfect. It fixed the "direction" of the prediction but still lacked the precise numbers.

5. The Conclusion

The paper concludes that the Thomas-Fermi-von Weizsäcker (TFW) functional is not suitable for describing itinerant magnetism (magnetism caused by moving electrons in metals).

Why?
Because magnetism in these metals depends on very sharp, detailed "spikes" in the electron arrangement. The TFW tool is like a low-resolution photo; it smooths out those sharp spikes. Without seeing those spikes, it cannot calculate the delicate balance required to predict magnetism.

The Takeaway:
If you want to simulate a system with millions of atoms to see how a material behaves generally, the "Fast & Cheap" tool is great. But if you want to know if a metal will stick to your fridge (magnetism), you cannot rely on this specific fast tool yet. It's too blurry to see the fine print of magnetic behavior.

The researchers suggest that while the "Fast Tool" can provide a good starting point (the crowd map), you still need the "Gold Standard" brain to interpret it correctly if you want to understand magnetism.

Technical Summary

1. Problem Statement

The paper addresses a critical limitation in Orbital-Free Density Functional Theory (OF-DFT): its ability to describe itinerant magnetism (magnetism arising from the delocalized electrons in metals).

• Context: While OF-DFT offers linear computational scaling (enabling simulations of millions of atoms), it approximates the non-interacting kinetic energy ($T_s$) directly as a functional of the electron density, bypassing the explicit calculation of Kohn-Sham orbitals.
• The Challenge: Itinerant magnetism in transition metals is driven by subtle features in the electronic density of states (DOS) near the Fermi level, specifically the competition between the kinetic energy cost of spin polarization and the exchange-correlation energy gain.
• The Gap: Because OF-DFT does not explicitly resolve electronic states or the DOS, it is unclear if standard kinetic energy functionals, such as the Thomas-Fermi-von Weizs¨acker (TFW) functional, can accurately capture the magnetic instabilities (e.g., the transition from paramagnetic to ferromagnetic states) observed in materials like Fe, Co, and Ni.

2. Methodology

The authors utilized the M-SPARC code (a real-space finite-difference implementation of DFT) to perform a systematic benchmark study.

• Systems Studied:
  • Paramagnets: Aluminum (Al, weak paramagnet) and Palladium (Pd, strongly enhanced paramagnet near ferromagnetic instability).
  • Ferromagnets: Iron (Fe, bcc), Cobalt (Co, hcp), and Nickel (Ni, fcc).
• Calculation Schemes: Three distinct approaches were compared:
  1. KS-KS: Self-consistent Kohn-Sham DFT (serving as the ground-truth benchmark).
  2. OF-OF: Self-consistent Orbital-Free DFT using the TFW kinetic energy functional.
  3. OF-KS: A non-self-consistent approach where the Kohn-Sham functional (including exchange-correlation) is evaluated on the ground-state electron densities obtained from the OF-OF calculation.
• Metric for Stability: Magnetic stability was assessed by calculating the inverse uniform static spin susceptibility ($\chi^{-1}$), derived from the second derivative of the total energy with respect to net magnetization ($M$) at $M=0$:
  $$\chi^{-1} = \left. \frac{d^2E}{dM^2} \right|_{M=0}$$
  • $\chi^{-1} > 0$: Stable paramagnetic state.
  • $\chi^{-1} < 0$: Unstable paramagnetic state (tendency toward ferromagnetism).
• Parameters: Calculations used the Local Spin-Density Approximation (LSDA) for exchange-correlation. Lattice constants were set to experimental values.

3. Key Contributions

• Systematic Benchmarking: The study provides the first rigorous assessment of the TFW functional's performance specifically for itinerant magnetism across a diverse set of $3d$ transition metals and simple metals.
• Decoupling Density vs. Functional: By introducing the OF-KS scheme, the authors successfully isolated the error source. They demonstrated that the failure of OF-DFT in magnetism stems primarily from the kinetic energy functional approximation (TFW) rather than the quality of the ground-state electron density itself.
• Quantitative Analysis of Failure: The paper quantifies the magnitude of error in magnetic susceptibility, showing that self-consistent OF-DFT fails not just quantitatively but often qualitatively (predicting stability where instability exists).

4. Results

A. Paramagnetic Metals (Al and Pd)

• Aluminum (Al):
  • KS-KS: Correctly predicts a stable paramagnetic state ($\chi^{-1} > 0$).
  • OF-OF: Predicts stability but significantly overestimates the magnetic curvature (underestimates susceptibility).
  • OF-KS: Almost perfectly restores the KS benchmark results, indicating the OF density is accurate for Al, but the TFW functional introduces the error.
• Palladium (Pd):
  • KS-KS: Predicts a very small $\chi^{-1}$, indicating the system is near the threshold of ferromagnetic instability (shallow energy curvature).
  • OF-OF: Fails dramatically, predicting a large positive $\chi^{-1}$ (steep curvature), incorrectly suggesting Pd is far from instability.
  • OF-KS: Improves the result significantly (reducing the error by ~70%) but does not fully recover the delicate near-critical behavior of Pd.

B. Ferromagnetic Metals (Fe, Co, Ni)

• KS-KS: Correctly predicts instability ($\chi^{-1} < 0$) for all three, consistent with their ferromagnetic ground states.
• OF-OF: Complete Qualitative Failure. It predicts $\chi^{-1} > 0$ for all three metals, incorrectly concluding they are stable paramagnets. The TFW functional overestimates the kinetic energy cost of spin polarization, preventing the system from becoming ferromagnetic.
• OF-KS:
  • Fe & Ni: Successfully recovers the correct sign ($\chi^{-1} < 0$), restoring the qualitative prediction of ferromagnetism.
  • Co: Reduces $\chi^{-1}$ significantly, placing the system very close to the instability threshold, though it remains slightly positive.
  • Conclusion: The OF densities contain sufficient information to support magnetic ordering if evaluated with a more accurate energy functional (KS), but the TFW functional itself cannot capture the necessary curvature.

5. Significance and Conclusion

• Fundamental Limitation: The TFW kinetic energy functional is fundamentally unsuitable for describing itinerant magnetism in transition metals. It fails to capture the sharp variations in the density of states near the Fermi level (narrow $d$-bands) that drive magnetic instabilities.
• Mechanism of Failure: The error arises because the TFW functional approximates the kinetic energy based on local density and gradients, which smooths out the electronic features necessary to lower the energy cost of spin polarization.
• Pathway for Improvement: The success of the OF-KS (hybrid) approach suggests that future developments in OF-DFT for magnetic materials should focus on:
  1. Developing more advanced kinetic energy functionals capable of capturing non-local effects and sharp DOS features.
  2. Utilizing hybrid schemes where high-fidelity OF densities are used as a starting point for more accurate energy evaluations, potentially enabling efficient simulations of large magnetic systems.

In summary, while OF-DFT remains a powerful tool for large-scale structural simulations, the TFW functional cannot be relied upon for predicting magnetic properties in itinerant systems without significant functional improvements or hybrid correction schemes.