Semilocal and Hybrid Density Embedding Calculations of Ground-State Charge-Transfer Complexes

TL;DR

This paper evaluates the frozen density embedding method with various exchange-correlation functionals for charge-transfer complexes, finding hybrid functionals provide more accurate and reliable results than semilocal ones.

Contribution

It demonstrates the effectiveness of hybrid XC functionals in embedding calculations for charge-transfer complexes, highlighting their superiority over semilocal functionals.

Findings

Hybrid XC functionals yield results in close agreement with supermolecular calculations.

Semilocal XC functionals show large deviations but can still sometimes give good results due to error cancellation.

Hybrid functionals are recommended for reliable embedding calculations of charge-transfer systems.

Abstract

We apply the frozen density embedding method, using a full relaxation of embedded densities through a freeze-and-thaw procedure, to study the electronic structure of several benchmark ground-state charge-transfer complexes, in order to assess the merits and limitations of the approach for this class of systems. The calculations are performed using both semilocal and hybrid exchange-correlation (XC) functionals. The results show that embedding calculations using semilocal XC functionals yield rather large deviations with respect to the corresponding supermolecular calculations. Due to a large error cancellation effect, however, they can often provide a relatively good description of the electronic structure of charge-transfer complexes, in contrast to supermolecular calculations performed at the same level of theory. On the contrary, when hybrid XC functionals are employed, both embedding and supermolecular calculations agree very well with each other and with the reference benchmark results. In conclusion, for the study of ground-state charge-transfer complexes via embedding calculations hybrid XC functionals are the method of choice due to their higher reliability and superior performance.