1 3
Top Curr Chem (Z) (2018) 376:24
5 Outlook and Perspectives
In this review, we have outlined a route towards simulation of third-order nonlinear electronic spectroscopy with accuracy compatible with current experiments.
The focus on 2D electronic spectroscopy is motivated by the huge impact that
its application in the UV region could have on the study of molecular systems
with absolute biological relevance, such as proteins and nucleic acids, the building blocks of life (Sect. 1). 2DUV experiments feature high spectral and temporal
resolution, allowing us to follow the evolution of individual chromophores though
their spectral fingerprints, providing direct signatures of electronic couplings,
charge and energy transfers, and characteristic line shapes and dynamics, which
provide information on PES topology and solvent reorganization timescales. The
density matrix formalism (Sect. 2) provides an appropriate platform to describe
the state-of-matter evolution in the Liouville space when interacting with weak
electric fields, accessing the nonlinear response recorded in 2DES experiments.
We outlined a route towards simulation and prediction of 2DES spectra from
first principles first by employing “drastic” (static) approximations to the working equations within the global eigenstates basis (SOS approach), which enables
us to ascertain the two fundamental ingredients to be accurately computed, i.e.
the electronic transition energies and dipole moments (TEs and TDMs), and then
recovering the missing contributions that shape 2D electronic spectra, comprising the factors that determine the line shapes and the dynamic evolution of the
nonlinear signals. This route is covered by illustrating some practical examples
(Sect. 3), starting with the benchmark studies of excited-state manifolds of the
target biomolecular systems (Sect. 3.1) by means of accurate electronic structure
computations, providing reliable TEs and TDMs estimates. For accurate calculations of these fundamental ingredients, we rely on the multi-configurational wave
functions combined with multi-reference perturbation theory energy corrections
(the CASSCF//PT2 level of theory), which in our opinion offers the optimal compromise between accuracy/completeness and computational cost. We addressed
its performance with respect to active space size, number and nature of simultaneously computed states, basis set and geometric parameters, and demonstrate
its sensitivity to these parameters. CASSCF/CASPT2 enables adaptive protocols
for obtaining converged energies and TDMs, reaching accuracy of 0.2–0.25 eV,
which is considered state of the art in the field of photochemistry, yet corresponding to ~ 2000 cm
−1
, a significant deviation from a spectroscopic point of view.
Therefore, reaching this upper-bound limit of accuracy is essential for reliable
theoretical spectroscopy. We have shown how such accuracy has been obtained
for the TEs and TDMs of monomeric units of aromatic protein side chains and
nucleobases in the gas phase, demonstrating how the (commonly used) fullvalence active spaces yield large discrepancies relative to available experimental cross-sections, and large, as computationally quite expensive, approaches are
required for quantitative comparisons. These high-level computations yielded
important parameters that could be implemented in Hamiltonian models (within a
quasi-particle representation) in order to extend the simulation to large scale, but
103
Reprinted from the journal
Top Curr Chem (Z) (2018) 376:24
5 Outlook and Perspectives
In this review, we have outlined a route towards simulation of third-order nonlinear electronic spectroscopy with accuracy compatible with current experiments.
The focus on 2D electronic spectroscopy is motivated by the huge impact that
its application in the UV region could have on the study of molecular systems
with absolute biological relevance, such as proteins and nucleic acids, the building blocks of life (Sect. 1). 2DUV experiments feature high spectral and temporal
resolution, allowing us to follow the evolution of individual chromophores though
their spectral fingerprints, providing direct signatures of electronic couplings,
charge and energy transfers, and characteristic line shapes and dynamics, which
provide information on PES topology and solvent reorganization timescales. The
density matrix formalism (Sect. 2) provides an appropriate platform to describe
the state-of-matter evolution in the Liouville space when interacting with weak
electric fields, accessing the nonlinear response recorded in 2DES experiments.
We outlined a route towards simulation and prediction of 2DES spectra from
first principles first by employing “drastic” (static) approximations to the working equations within the global eigenstates basis (SOS approach), which enables
us to ascertain the two fundamental ingredients to be accurately computed, i.e.
the electronic transition energies and dipole moments (TEs and TDMs), and then
recovering the missing contributions that shape 2D electronic spectra, comprising the factors that determine the line shapes and the dynamic evolution of the
nonlinear signals. This route is covered by illustrating some practical examples
(Sect. 3), starting with the benchmark studies of excited-state manifolds of the
target biomolecular systems (Sect. 3.1) by means of accurate electronic structure
computations, providing reliable TEs and TDMs estimates. For accurate calculations of these fundamental ingredients, we rely on the multi-configurational wave
functions combined with multi-reference perturbation theory energy corrections
(the CASSCF//PT2 level of theory), which in our opinion offers the optimal compromise between accuracy/completeness and computational cost. We addressed
its performance with respect to active space size, number and nature of simultaneously computed states, basis set and geometric parameters, and demonstrate
its sensitivity to these parameters. CASSCF/CASPT2 enables adaptive protocols
for obtaining converged energies and TDMs, reaching accuracy of 0.2–0.25 eV,
which is considered state of the art in the field of photochemistry, yet corresponding to ~ 2000 cm
−1
, a significant deviation from a spectroscopic point of view.
Therefore, reaching this upper-bound limit of accuracy is essential for reliable
theoretical spectroscopy. We have shown how such accuracy has been obtained
for the TEs and TDMs of monomeric units of aromatic protein side chains and
nucleobases in the gas phase, demonstrating how the (commonly used) fullvalence active spaces yield large discrepancies relative to available experimental cross-sections, and large, as computationally quite expensive, approaches are
required for quantitative comparisons. These high-level computations yielded
important parameters that could be implemented in Hamiltonian models (within a
quasi-particle representation) in order to extend the simulation to large scale, but
103
Reprinted from the journal