Copper-based
chalcogenide semiconductors are particularly attractive
materials for photovoltaic and nonlinear optical applications. Unfortunately,
accurate theoretical understanding of the optical responses of these
materials is hindered by the presence of strongly localized Cu 3d electrons. To assess the validity of calculation schemes
for determining the optical properties of Cu-based chalcogenides,
different density functional theory (DFT) approaches including the
conventional Perdew−Burke−Ernzerhof (PBE) method and
the modified Becke–Johnson (mBJ) potential with and without
Hubbard U correction, and the hybrid HSE06 methods
are employed to evaluate the linear and nonlinear optical performances
of 16 typical Cu-based chalcogenides. We calculate band gaps, dielectric
functions, refractive indices, absorption coefficients, and the second
harmonic generation susceptibilities of these compounds and compare
the results with available experimental measurements. It is clear
that the mBJ + U approach yields band gaps close
to those predicted using the HSE06 method, which is consistent with
the experimental values. More importantly, the fundamental optical
properties in the infrared–visible light region calculated
within the mBJ + U method agree better with experiments
compared with the HSE06 functional. In addition to the linear optical
responses, the mBJ potential with on-site Coulomb U correction to the Cu-3d state is also capable of
describing the optical nonlinearity of copper chalcogenides reasonably.