Periodic coupled-cluster theory for the ground and excited states with atom-centered basis functions

Periodic coupled-cluster (CC) theory promises to be a reliable, highly accurate electronic structure method in materials science [1]. The all-electron code FHI-aims [2], which employs numeric atom-centered orbitals (NAOs), has recently been interfaced to the CC theory for solids (Cc4s) code [3,4], making CC theory for both the ground state and excited states (in the equation-of-motion CC theory (EOM-CC)) accessible to FHI-aims. For molecules, EOM-CC predicts quasi-particle energies more accurately and reliably than the GW approximation[5], and we expect that this will also hold for bandstructures and gaps of solids. Like most correlated wave function methods, CC methods exhibit excessively slow convergence with the size of the super cell. While in the plane wave basis framework and in the case of the CC ground state, an efficient approach has been suggested for solving this problem [6], no analogous technique exists for NAOs, so far. We present the current state of the CC theory framework available in FHI-aims and possible avenues to address the finite-size error.

[1] G. Booth et al., Nature 493, 365–370 (2013)

[2] The FHI-aims web page, https://fhi-aims.org

[3] F. Hummel et al., JCP, 146, 124105 (2017)

[4] E. Moerman et al., JOSS, 7, 4 (2022)

[5] M. Lange et al.,JCTC, 14, 4224-4236 (2018)

[6] K. Liao, A. Grüneis, JCP, 145, 141102 (2016)