61st Annual Report on Research 2016

Our project aims to establish a complete simulation tool chain, which calculates spin coupling and spin relaxation observables in nuclear magnetic resonance (NMR) from first principles, relevant to key problems in oil borehole NMR and in hyperpolarized NMR.

In the first reporting period, we focused on three main objectives:

Objective 1: Coupling FHI-aims molecular simulations and SPINACH

Nuclear spin relaxation, J-couplings and chemical shifts, of statistical ensembles of spins in disordered and/or liquid environments, are key output observables accessible by NMR. These and other observable properties can be predicted by spin dynamics simulations, which rely on molecular-scale NMR shielding tensors and interaction parameters as input. The SPINACH program is a leading package for such simulations. Assisted by its main author (Ilya Kuprov, Southampton), we completed an interface between the FHI-aims all-electron electronic structure code and SPINACH. FHI-aims provides details of the spin system (atom types, positions, NMR shielding and J-tensors) in the standardized SpinXML format, introduced by Kuprov. This work established the technical base for first-principles parameterized spin dynamics simulations based on FHI-aims and SPINACH, which we already use for production simulations of advanced molecular architectures for hyperpolarization (see below).

Objective 2: NMR observables based on FHI-aims' numeric atom-centered basis sets

We created a first working implementation of NMR chemical shieldings and the indirect spin-spin J-coupling tensors for molecular systems in FHI-aims. FHI-aims is an all-electron code for electronic structure simulations of molecules and solids across the periodic table. FHI-aims uses numerically tabulated atom-centered orbital (NAO) basis functions for the numerical discretization, matching the numerical accuracy of the best available benchmark approaches, while enabling calculations of very large systems (currently, thousands of atoms). In this reporting period, the NMR Hamiltonian matrix elements have been implemented and validated in their non-relativistic form, i.e., suitable for light elements, and using density-functional perturbation theory (DFPT) based on a local density approximation kernel. We find (e.g., in the table of content figure of this report) that the high accuracy of the NAO basis sets near the nucleus translates to faster convergence for NMR parameters, as opposed to traditional basis sets, i.e., a performance advantage for applications to large systems.

Objective 3: Long-lived spin states

In the first period, the team focused on target molecules with long-lived nuclear spin states based on 13C or 15N. A technique known as SABRE (Signal Enhancement By Reversible Exchange) enables the transfer of the nuclear spin polarization of parahydrogen (p-H2) directly and highly efficiently to the target molecules. SABRE was recently extended by the Warren group and collaborators by utilizing a shield to reduce Earth's magnetic field by about 99% (SABRE-SHEATH, "SABRE in SHield Enables Alignment Transfer to Heteronuclei"). In the current project, we employed SABRE-SHEATH to demonstrate the hyperpolarization of 15N2-diazirine containing molecular tags in a first joint publication in the journal Science Advances. We are now extending these efforts towards diphenylacetylene (DPA) derived molecules, which could be compatible with future medical MRI applications and beyond.

Project impact on the PI and their teams

The project is already successfully facilitating the entry of Blum and his group into the field of NMR based science. Thanks to the collaboration with co-PI Warren and Duke Assistant Research Professor Thomas Theis, who leads the NMR work in the Warren group, a first joint publication appeared and a second publication is in preparation. The theoretical work on NMR within the numeric atom-centered basis set strategy of FHI-aims will lead to a third publication. A pre-proposal for a follow-up funding opportunity is currently under review. Thus, the project is well on the way to accomplish its longer-term goals for the PIs and their team members.

In the Blum group, in line with the original work plan of the proposal, two post-doctoral researchers led the work. Dr. William Huhn initiated the project, reflected in the team's first publication. Dr. Raul Laasner, who joined Blum's group in December 2009, continues the implementation work. Huhn and Laasner joined Blum to present this work at the international FHI-aims "Users' and Developers' Workshop," held in Munich, Germany, in July 2016, to an audience including other junior theorists and some of the most senior scientist in electronic structure theory today. Additionally, the stay in Munich facilitated a collaboration with the group of Prof. Karsten Reuter (TU Munich), aiming to export NMR matrix elements from FHI-aims into the magnetic resonance module (MAG) of the computational chemistry program ReSpect 1.2, a quantum-chemical tool for nuclear spectroscopy with functionality that reaches well beyond this project. Simone Koehler, a Ph.D. student in Munich and at Forschungszentrum Juelich, will carry this work forward as a visiting student at Duke University in the second project period. Finally, Duke undergraduate senior Wolfgang Seiya helped with developing a molecular structure search strategy for this project in the context of an independent research project in Fall 2015. In Warren's group, part of the effort of postdoctoral researcher Johannes Colell is supported by this project. Profs. Warren and Theis have presented the work in many talks and posters at various conferences including ENC (Experimental NMR Conference), EUROMAR (European magnetic resonance meeting), WMIC (World Molecular Imaging Conferences), ISMRM (International Society for Magnetic Resonance in Medicine). The reactions to our work are very strong and positive, leading to many new collaborations with biochemists and doctors, as well as investors interested in commercialization. Within Duke University, the work has lead to very close collaboration with synthetic chemists making molecular markers using the developed tagging strategies.

Future work

In the second project period, we will complete the implementation of NMR parameters in FHI-aims as planned, including an appropriate scalar relativistic treatment, DFPT for gradient and hybrid density functionals, and periodic boundary conditions. We will broaden the spectrum of applications as planned in our original proposal. Finally, we noticed that additional parameters such as molecular magnetizabilities - potentially of interest for spintronics and quantum computing - are easily accessible as byproducts of our NMR implementation work; this may open another, initially unanticipated avenue for new research directions based on this project.