Research in the Neugebauer Group
We develop focused quantum-chemical methods and apply quantum chemistry to a wide range of topics:
Our Research Areas:
Quantum chemical methods suffer from an unfavorable scaling of computer time with the system size, e.g., the number of atoms in a molecular system. The idea of subsystem DFT is to circumvent this problem by decomposing the electron density into subsystem contributions. This has a number of advantages, both from a conceptual point of view and for the computational effort in applications to large molecular systems. Frozen-Density Embedding (FDE) can be regarded as a special variant of subsystem DFT in which the properties of a certain active (embedded) subsystem in a larger environment can be determined.
Our group develops and FDE and subsystem DFT methods for the investigation of the electronic structure and properties of molecular aggregates. Recent aspects of this work include:
- Potential-energy surfaces from sDFT: We benchmark existing approximations within sDFT and try to improve PESs from sDFT. We have developed analytical gradient implementations for sDFT in the ADF package for ground and excited states.
- "Exact" embedding: We develop, implement, and apply "exact" embedding strategies based on potential reconstruction and projection. We have recently shown that external orthogonality is not a formal requirement for sDFT to be exact.
- Embedding correlated wavefunctions: We work on WF-in-DFT embedding approaches for higher-accuracy descriptions of molecules in complex environments. Together with our collaboration partners, we have implemented WF-in-DFT approaches for wavefunction methods such as DMRG, CASSCF, CASPT2, CC2/CCSD/CC3, and QMC. We have re-derived expressions for excitation energies including state-specific embedding potentials for WF-in-DFT embedding, and addressed question related to the non-orthogonality of embedded wavefunctions with such state-specific potentials.
- Molecular properties from sDFT: We have derived and implemented a linear-response subsystem TDDFT for the calculation of excited states (see also below) and other response properties (UV/Vis spectra, circular dichroism, frequency-dependent polarizabilities, optical rotatory dispersion)
- sDFT as a constrained-DFT variant: We have shown that sDFT with approximate non-additive kinetic-energy functionals can pragmatically be used to converge to quasi-diabatic electronic states with localized charge or spin. This allows, for example, to study charge transfer based on sDFT.
Related to sDFT (see above), we employ the so-called 3-partitioning FDE (3-FDE) approach for electronic-structure calculations on proteins. 3-FDE can be thought of as a sDFT with buffer regions, which allow to cut covalent bonds. The "dangling bonds" are satured with a buffer fragment (capping group). The electron densities of the capping groups (combined to cap molecules) are subsequently subtracted from the sum-of-capped-fragment density to arrive at a correct total density. Our contributions to this field include:
- Calculations of pigment site energies in a 3-FDE environment (including relaxation effects)
- A benchmark of electron densities from 3-FDE (avoiding failures of KS-DFT)
- A local variant of the COSMO model (needed for modeling protein solvation)
Part of our research addresses the magnetic properties of organic radicals. In particular, we are working on the following topics:
- Black-box prediction of macroscopic magnetism: We have recently implemented a black-box version of the so-called first-principles bottom-up approach, which allows to predict magnetic susceptibilities and heat capacities as a function of temperature based on a crystal structure for crystalline organic radicals.
- Modeling ESR spectra: We simulate ESR spectra of organic radicals based on electronic-structure theory.
- Spin-density distributions: We are working on accurate yet efficient methods for the prediction of spin densities of organic systems.
The aim of this research project is the development of quantum chemical methods that can be applied in studies of natural and artificial photosynthetic systems. These methods are used to study the mechanisms of initial steps in photosynthesis, namely absorption of light by pigment molecules, energy transfer to a reaction center, and initial charge-separation steps.
Recent efforts in this project concern:
Protein-pigment interactions and exciton couplings: We have studied excited states of pigments and pigment networks in (essentially) complete natural light-harvesting antennae such as the Fenna-Matthews-Olson complex
Screening effects on energy transfer: We have developed explicit quantum chemical methods to include environmental screening effects on excited states and excitonic couplings as needed in energy-transfer model theories
Charge- and spin-density distributions in photosynthetic reaction centers: We have conducted a benchmark study on spin densities for radical cations of photosynthetic pigment models. In addition, we have studied environmental effects on the spin-density asymmetry in the special pair of purple bacteria, and provided support for experimental studies on the spin density in photosynthetic reaction centers.
Resonance Raman studies on photosynthetic systems: We have used (selective) theoretical resonance Raman approaches to unravel structural details of spheroidene in the reaction center of Rhodobacter spheroides and to investigate the reaction coordinate for the initial charge separation in the special pair of bacterial reaction centers
Excited-state calculations for molecules in solvents or adsorbates, but also for core excitations, often suffer from the fact that the computational effort is large, not only because of the system size, but also because many excited states need to be calculated, many more than actually of interest. We are therefore interested in selective excited-state methods applicable to large systems.
Recent developments and studies include:
Wavefunction/DFT embedding schemes for excitation energies with state-specfic embedding potentials
Selective excited-state algorithms including applications to adsorbates, solvated systems, and core excitations
Vibronic-structure tracking (VST) for selective determination of vibrational normal modes with high impact on the vibronic structure
Subsystem TDDFT for electronic excitations including work on the corresponding Tamm-Dancoff approximation and applications to networks of coupled chromophores
3-FDE-TDDFT for excitations of pigments in proteins (see also above)
We have recently started looking into on-surface chemistry phenomena in close connection to experiment. This mainly includes
- Structural characterization of molecules on surfaces
- Energy profiles and transition-state location for on-surface reactions
- Participation of surface metal atoms in those reactions
Our group is involved in the development of the massively parallel theoretical vibrational spectroscopy package MoViPac, which comprises the two programs SNF (for standard frequency analyses) and Akira (for mode-tracking calculations). Previous work concerned special types of vibrational spectroscopy such as vibrational circular dichroism (VCD), resonance Raman (RR), and vibrational resonance Raman optical activity (VRROA). Methodologically, we have contributed selective algorithms for tracking characteristic motions, high-intensity vibrations, or vibrations with large impact on the vibronic structure of UV/Vis spectra.
Program references, descriptions and illustrative applications can be found in:
SNF: efficient frequency analysis for massively parallel computations [J. Neugebauer, M. Reiher, C. Kind, B.A. Hess, J. Comput. Chem. 23 (2002), 895]
Akira: Mode-tracking of pre-selected molecular vibrations in large systems [M. Reiher, J. Neugebauer, J. Chem. Phys. 118 (2003), 1634]
MoViPac: Vibrational Spectroscopy with a Robust Meta-Program for Massively Parallel Standard and Inverse Calculations [T. Weymuth, M. Haag, K. Kiewisch, S. Luber, S. Schenk, C. Jacob, C. Herrmann, J. Neugebauer, M. Reiher, J. Comput. Chem. 33 (2012), 2186]