The simulation of the color optical properties of molecular dyes in liquid solution requires the calculation of time evolution of the solute absorption spectra fluctuating in the solvent at finite temperature. Time-averaged spectra can be directly evaluated by combining ab initio Car-Parrinello molecular dynamics and time-dependent density functional theory calculations. The inclusion of hybrid exchange-correlation functionals, necessary for the prediction of the correct transition frequencies, prevents one from using these techniques for the simulation of the optical properties of large realistic systems. Here we present an alternative approach for the prediction of the color of natural dyes in solution with a low computational cost. We applied this approach to representative anthocyanin dyes: the excellent agreement between the simulated and the experimental colors makes this method a straightforward and inexpensive tool for the high-throughput prediction of colors of molecules in liquid solvents.

We present a new release of the turboTDDFT code featuring an implementation of hybrid functionals, a recently introduced pseudo-Hermitian variant of the Liouville-Lanczos approach to time-dependent density-functional perturbation theory, and a newly developed Davidson-like algorithm to compute selected interior eigen-values/vectors of the Liouvillian super-operator. Our implementation is thoroughly validated against benchmark calculations performed on the cyanin (C 21 O 11 H 21 ) molecule using the Gaussian 09 and turboTDDFT 1.0 codes. External routines/libraries: turboTDDFT 2.0 is a tightly integrated component of the Quantum ESPRESSO distribution and requires the standard libraries linked by it: BLAS, LAPACK, FFTW, MPI. Nature of problem: Calculation of the optical absorption spectra of molecular systems. Solution method: Electronic excited states are addressed by linearized time-dependent density-functional theory within the plane-wave pseudo-potential method. The dynamical polarizability can be computed in terms of the resolvent of the Liouvillian super-operator, using a pseudo-Hermitian variant of the Lanczos recursion scheme. As an alternative, individual eigenvalues of the Liouvillian can be computed via a newly introduced variant of the Davidson method. In both cases, hybrid functionals can now be used. Restrictions: Spin-restricted formalism. Linear-response regime. Adiabatic XC kernels only. Hybrid functional are only accessible using norm-conserving pseudo-potentials. Unusual features: No virtual orbitals are used, nor even calculated. Within the Lanczos method a single recursion gives access to the whole optical spectrum; when computing individual excitations usinf the Davidson method, interior eigenvalues can be easily targeted.

We show how accurate benchmark values of the surface formation energy of crystalline lithium hydride can be computed by the complementary techniques of quantum Monte Carlo (QMC) and wavefunctionbased molecular quantum chemistry. To demonstrate the high accuracy of the QMC techniques, we present a detailed study of the energetics of the bulk LiH crystal, using both pseudopotential and all-electron approaches. We show that the equilibrium lattice parameter agrees with experiment to within 0.03 %, which is around the experimental uncertainty, and the cohesive energy agrees to within around 10 meV per formula unit. QMC in periodic slab geometry is used to compute the formation energy of the LiH (001) surface, and we show that the value can be accurately converged with respect to slab thickness and other technical parameters. The quantum chemistry calculations build on the recently developed hierarchical scheme for computing the correlation energy of a crystal to high precision. We show that the hierarchical scheme allows the accurate calculation of the surface formation energy, and we present results that are well converged with respect to basis set and with respect to the level of correlation treatment. The QMC and hierarchical results for the surface formation energy agree to within about 1 %.

No abstract