Alternative treatment of relativistic effects in linear augmented plane wave (LAPW) method: application to Ac, Th, ThO2 and UO2

This paper proposes alternative relativistic corrections to the linear augmented plane wave (LAPW) method—specifically new basis functions, potential matrix corrections, and refined spin-orbit coupling calculations—which significantly improve the accuracy of predicted lattice constants and elastic moduli for actinides and reveal that uranium dioxide (UO2) is a semimetal with a small band gap.

The Gist

Imagine you are trying to build a perfect 3D model of a complex city using a set of Lego bricks. In the world of physics, this "city" is a solid material (like a metal or an oxide), and the "Lego bricks" are mathematical functions used to describe how electrons move around inside it. This method is called the LAPW method (Linear Augmented Plane Wave).

For most everyday materials, standard Lego bricks work fine. But when you try to model heavy elements like Actinium, Thorium, or Uranium (the "Actinides" at the bottom of the periodic table), things get tricky. These atoms are so massive and their electrons move so fast (close to the speed of light) that standard physics rules start to break down. This is where Relativity comes in.

This paper is like a team of master architects saying, "Hey, our current Lego bricks aren't quite right for these heavy cities. We need to redesign the bricks and the blueprints to get an accurate model."

Here is a breakdown of their three main improvements, explained simply:

1. The "Super-Brick" Upgrade (New Radial Functions)

The Problem:
In the old method, the architects used a "standard average" brick to represent the electron clouds around the heavy atoms. It was like trying to describe a complex shape by just taking the average of two different shapes. For the heavy "6p" electron shells (which are like the inner layers of the city's foundation), this average was slightly off. It missed the specific, weird behavior of electrons moving near the speed of light.

The Fix:
The authors created new, custom-made bricks. Instead of averaging two shapes blindly, they looked at the two specific "real-life" shapes (solutions to the Dirac equation) that electrons actually take when they are moving fast. They blended these two real shapes together in a smarter way.

• The Analogy: Imagine trying to paint a portrait of a person who is running. The old method used a blurry, average photo of a runner. The new method uses a high-definition photo that captures the specific blur of the motion.
• The Result: These new bricks are so good at describing the heavy electron shells that they don't need any "extra" helper bricks (local atomic functions) that previous methods required. It's a more efficient, self-contained system.

2. Fixing the Blueprint Math (Correcting Matrix Elements)

The Problem:
Even with better bricks, the architects realized the math used to snap the bricks together was based on old, non-relativistic rules. It was like using a ruler calibrated for inches to measure a building designed for centimeters. The equations assumed the electrons were slow and lazy, but in these heavy atoms, they are fast and energetic.

The Fix:
They went back to the blueprints and corrected the formulas used to calculate how the bricks connect. They removed the hidden assumptions that only work for slow electrons.

• The Analogy: It's like realizing your GPS was programmed for a flat Earth, but you are driving on a mountain. You have to update the algorithm to account for the curves and slopes, or you'll end up in the wrong place.
• The Result: Small changes in the math led to big changes in the final model. The size of the crystal (lattice constant) and how hard it is to squeeze (bulk modulus) changed significantly—sometimes by amounts that matter a lot in engineering.

3. The "Spin-Orbit" Tweak (Handling the 6p States)

The Problem:
Electrons have a property called "spin" (like a tiny top spinning). In heavy atoms, this spin interacts with their movement (orbit) in a strong way, called Spin-Orbit Coupling. This interaction splits energy levels, like a single road splitting into two lanes.
The old method tried to calculate this split by averaging the two lanes together. But for the specific "6p" electrons in heavy atoms, this averaging was too aggressive. It made the split look much bigger than it actually was.

• The Analogy: Imagine you are trying to guess the height difference between two hills. The old method looked at the average height of both hills and guessed the difference was huge. The new method realized that one hill is actually much steeper than the other, so you should measure the difference based on the steeper hill alone to get the right answer.

The Fix:
They decided to stop averaging for these specific electrons. Instead, they used the mathematical description of just the "steeper" electron state (the 6p3/2 component) to calculate the energy split. This gave a much more realistic result.

What Did They Find? (The "City" Changes)

By applying these three fixes to materials like Thorium (Th), Uranium Dioxide (UO2), and Actinium (Ac), they found some surprising things:

1. Size Matters: Depending on which "bricks" and "blueprints" they used, the predicted size of the crystal could change by up to 0.15 Angstroms (a tiny fraction of a hair's width, but huge in atomic terms). The "stiffness" of the material could change by 26 GPa (a massive amount of pressure). This shows that if you use the wrong math, your predictions for nuclear fuel or reactor materials could be way off.
2. Uranium Dioxide is a "Semimetal": For a long time, scientists debated whether UO2 (used in nuclear fuel) was a metal, an insulator, or something in between. With their new, highly accurate method, they found a tiny "gap" in the energy levels. This means UO2 acts like a semimetal—a material that is almost an insulator but has a tiny leak of conductivity. This is a crucial detail for understanding how nuclear fuel behaves.
3. Actinium is a Mystery: Their calculations showed that standard methods consistently overestimate the size of Actinium crystals. It seems Actinium has a unique "personality" in its electron structure that makes it expand more than expected, and this effect gets worse the more precise your model is.

The Bottom Line

This paper is a "tune-up" for the most powerful tool physicists use to study heavy materials. They didn't just tweak the settings; they redesigned the core components (the bricks) and the instructions (the math) to account for the fact that in heavy atoms, electrons are running a race against light.

By doing this, they ensure that when we design nuclear reactors or study new materials, our computer models are actually telling the truth about how the universe works at the atomic level.

Technical Summary

1. Problem Statement

Calculating the electronic band structure of heavy elements (specifically actinides like Ac, Th, U, and Np) presents significant challenges due to strong relativistic effects. These elements possess 80+ core electrons, where spin-orbit (SO) coupling and scalar relativistic effects are critical.

• Limitations of Current Methods: The standard Linear Augmented Plane Wave (LAPW) method typically employs a "scalar relativistic" approach (Koelling-Harmon/MacDonald-Pickett-Koelling, or KH/MPK) which averages the $j=l-1/2$ and $j=l+1/2$ Dirac solutions.
• Specific Issues:
  1. The canonical KH radial basis functions often fail to accurately describe semicore states (specifically the 6p shell) without adding extra local atomic orbitals (e.g., $p_{1/2}$ local orbitals).
  2. Standard matrix element expressions in LAPW rely on non-relativistic identities that are violated when using relativistic radial functions.
  3. The standard calculation of Spin-Orbit (SO) coupling constants for 6p states often overestimates the energy splitting between $6p_{1/2}$ and $6p_{3/2}$ states.
  4. Many studies treat SO coupling as a perturbation (second-variation step), which introduces approximations and limits accuracy.

2. Methodology

The authors propose a comprehensive alternative treatment within the Full Potential LAPW (FLAPW) framework to improve accuracy while maintaining the general method structure. They applied three specific corrections:

A. New Radial Basis Functions (Explicitly Averaged Dirac - avD)

Instead of using the canonical KH radial functions derived from coupled differential equations with uncontrolled approximations, the authors propose:

• Solving the radial Dirac equation independently for the large components of $j=l-1/2$ and $j=l+1/2$ states.
• Explicitly averaging these two solutions using the standard weighting formula:
  $$P^{av}_l(r) = \frac{l}{2l+1}P_l(r) + \frac{l+1}{2l+1}P_{-l-1}(r)$$
• Key Advantage: For 6p semicore states, this new basis ($P^{av}_{l=1}$) naturally gains more weight from the $p_{1/2}$ Dirac component. This allows the method to accurately describe fully filled 6p bands without needing to add extra $p_{1/2}$ local atomic functions (unlike the standard FLAPW+$p_{1/2}$ scheme).

B. Correction of LAPW Matrix Elements

The authors identified that standard matrix element expressions for the spherically symmetric potential component ($L=0$) rely on the identity:
$$R_{MT}^2 (\dot{u}_l u'_l - \dot{u}'_l u_l) = 1$$
This identity holds only for non-relativistic Schrödinger solutions. For relativistic Dirac-derived functions, this value deviates from unity.

• Correction: They derived new expressions for the coefficients $a_l$ and $b_l$ and the matrix element $\gamma_l$ that explicitly account for this deviation factor ($F(l) \neq 0$), ensuring mathematical consistency with relativistic basis functions.

C. Refined Spin-Orbit (SO) Coupling Treatment

• Full Treatment: The authors perform SO coupling calculations in the full basis set rather than using the perturbative second-variation method.
• 6p State Specifics: They found that the standard averaged approach overestimates the 6p SO splitting (by ~17–27%).
• Proposed Fix: For 6p semicore states, they calculate the SO coupling constant $\zeta(p)$ using only the $6p_{3/2}$ radial component (and its energy derivative). This yields a much more realistic energy splitting compared to the actual Dirac eigenvalues.
• Other States: For $d$, $f$, and higher $\ell$ states, the new explicitly averaged ($avD$) functions are used, as they provide superior accuracy for those splittings compared to KH functions.

3. Key Contributions

1. Alternative Basis Set: Introduction of the "avD" basis set (explicitly averaged Dirac solutions) which improves the description of semicore 6p bands without auxiliary local orbitals.
2. Mathematical Rigor: Correction of LAPW matrix elements to remove hidden non-relativistic assumptions, ensuring consistency when using relativistic basis functions.
3. SO Coupling Optimization: A specific protocol for calculating 6p SO constants using only the $p_{3/2}$ component to correct systematic overestimation of splitting energies.
4. Full SO Implementation: Moving away from the second-variation approximation to a full diagonalization of the SO Hamiltonian in the complete basis set.

4. Results

The authors applied these methods to Actinium (Ac), Thorium (Th), Thorium Dioxide (ThO$_2$), and Uranium Dioxide (UO$_2$) using various DFT functionals (LDA, PBE, PBEsol).

• Sensitivity to Relativistic Treatment: The choice of relativistic treatment significantly impacts physical properties:

  • Lattice Constants ($a$): Variations up to 0.15 Å were observed depending on the basis set and SO treatment.
  • Bulk Modulus ($B$): Variations up to 26 GPa were found.
  • Trends: The direction of change (increase or decrease) is material-dependent; for example, SO coupling reduced lattice constants for Ac/Th but increased them for UO$_2$.

• Specific Material Findings:

  • Actinium (Ac): DFT calculations consistently overestimate the lattice constant of fcc Ac compared to experiment. This error persists even with improved basis sets, suggesting intrinsic limitations in describing Ac's band structure or phonon properties within standard DFT.
  • Thorium (Th): The new methods showed significant deviations in bulk modulus compared to canonical KH results.
  • Uranium Dioxide (UO$_2$):
    • With full SO coupling, UO$_2$ exhibits a small energy gap (0.2 – 0.4 eV) at the Fermi level between the highest occupied and lowest unoccupied 5f bands.
    • This gap persists for all $\vec{k}$-vectors.
    • Classification: Based on these calculations, UO$_2$ should be classified as a semimetal, not a metal or a standard insulator. (The authors note that experimental insulating behavior might arise from correlation effects or phonons not included in this DFT study).

5. Significance

• Accuracy: The study demonstrates that "small" relativistic details (like the specific averaging of radial functions or matrix element corrections) have macroscopic consequences on predicted material properties (lattice constants and elasticity).
• Methodological Alternative: The proposed "avD" approach offers a viable alternative to the standard "FLAPW + local orbitals" strategy, potentially simplifying calculations for heavy elements by removing the need for ad-hoc local orbital additions for semicore states.
• UO$_2$ Physics: The finding that UO$_2$ is a semimetal in full-relativistic DFT challenges the common view of it as a simple insulator or metal, highlighting the critical role of full SO coupling in actinide oxides.
• Benchmarking: The work provides a rigorous benchmark for future high-precision calculations of actinide materials, emphasizing that standard scalar relativistic approximations may be insufficient for quantitative predictions.