Embedded cluster density approximation for scalable high-level exchange-correlation calculations in periodic systems.

Area of Science:

• Computational chemistry
• Materials science
• Quantum mechanics

Background:

• High-level electronic-structure methods are crucial for accurate property prediction but computationally expensive for large systems.
• Scaling these methods to larger systems is essential for broader applicability in materials science and chemistry.

Purpose of the Study:

• Extend the embedded cluster density approximation (ECDA) method to periodic systems.
• Enable accurate and efficient calculation of electronic properties for large, periodic materials.

Main Methods:

• Adapted ECDA, originally for finite systems, to handle periodic boundary conditions.
• Utilized density functional embedding theory to seamlessly embed cluster densities within the larger system.
• Calculated high-level exchange-correlation (XC) energies on local clusters and projected them onto central atoms for system-wide energy construction.

Main Results:

• Demonstrated ECDA's applicability as a nearly black-box method across various bond types in periodic systems.
• Achieved good accuracy with modest cluster sizes, showing efficiency gains.
• Successfully calculated energy differences and generated smooth energy surfaces, highlighting the method's robustness.

Conclusions:

• ECDA provides an efficient and accurate approach for electronic property prediction in large periodic systems.
• The density partitioning in ECDA facilitates local active space definition, leading to appealing computational features.
• Potential convergence challenges in covalent systems due to boundary effects require further investigation.