CAREER: Theories for Magnetic Properties of Lanthanide and Actinide Complexes

  • Shiozaki, Toru (PD/PI)

Project: Research project

Project Details


Overview: The research objective of the program proposed herein is to understand and predict
magnetic properties of lanthanide and actinide complexes by developing fully relativistic electronic
structure theories for sizable (100–150 atomic systems and a few heavy atoms), open-shell
molecules of quasi-degenerate character. Our proposed research aims to realize rational design of
single molecule magnets and magnetic materials. It is proposed to develop electronic structure
theories for such systems on the basis of the fully relativistic four-component Dirac equation because
the spin–orbit and spin–spin couplings in lanthanide and actinide complexes are beyond the
perturbation regime.
Intellectual Merit: The breakthrough that our research promises is to realize predictive electronicstructure
theories for lanthanide and actinide complexes that are applicable to real systems of
chemical interest, enabling us to compare directly with experimental studies. The research projects
have four components: (1) Scalable relativistic multi-configuration theories will be developed to
determine zero-field splitting of f-element complexes. The density-fitted Dirac–Fock algorithm, which
was applied to systems with 100 atoms and a few heavy elements in our prior studies, will be
extended to open-shell theories. (2) An active-space decomposition strategy, recently developed by
us in the non-relativistic framework, will be extended to model the direct exchange and
superexchange interactions between spins on metal centers and bridges. The tensor decomposition
of a coefficient matrix is proposed, with which the theory can be seen as a low-entanglement wave
function ansatz using a fragment active space as a “site.” (3) Theories for f-element complexes
interacting with a magnetic field will be developed to simulate electron paramagnetic resonance
(EPR) spectra and field-induced single molecule magnets. The so-called London orbitals, which
circumvent the gauge-origin problem, will be employed in the fully relativistic framework. (4)
Relativistic multireference electron correlation methods will be implemented into parallel computer
programs with the aid of an automated code generation approach. The focus will be on the
development of electron correlation theories that can treat multiple states with an equal footing.
Broader Impacts: The proposed research will open up a new field in computational research by
enabling simulations of magnetic properties of large heavy-element complexes with fully relativistic
electronic structure theories. The computer programs that implement the proposed theories and
algorithms will be made freely available to the broad community (from theory and algorithm
developers to computational chemists working on applications, and to the public) under the GNU
Public License. In concert with the program development in the proposed research, the educational
component of this proposal seeks to develop web-based learning modules for computational aspects
of physical chemistry. Provided that the computer resources available to us continue to grow
exponentially, and that computation-aided chemistry research is rapidly expanding, it is of vital
importance to prepare undergraduate chemistry students for computational approaches, not as an
optional subject, but as a part of the core physical chemistry program. The modules will take
advantage of the modern object-oriented design to minimize unnecessary learning overhead. They
will be disseminated on the portal website of NSF-supported “,” whose infrastr
Effective start/end date3/1/142/29/20


  • National Science Foundation (CHE-1351598)


Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.