New Quantum Chemistry Methods

How to model your NMR spectra: What basis set and level of theory should you use?

Keywords: Computational Chemistry, Link with Experiment, Basis Set Design, NMR spectroscopy

NMR is a ubiquitous technique extensively utilised throughout the sciences, but most especially by the synthetic chemistry community. There are two issues associated with NMR spectroscopy: assigning the spectra accurately, and understanding what structural information the spectra provides. Accurate computational chemistry calculations can assist with both these tasks for difficult cases like fluorinecontaining compounds. However, existing calculations often use sub-optimal basis sets that do not adequately describe the electron core region, with scaling factors used to fix these inadequacies. In this project, you will explore the quality of NMR spectral predictions using existing and novel basis sets used alongside popular density functional approximations.

Utilising & Developing Mixed Ramp-Gaussian Basis Sets for NMR Calculations & Larger elements

A project probably most suitable for a PhD student interested in method development in quantum chemistry.

In my PhD, I developed methodologies for designing & efficiently integrating a mixed Ramp-Gaussian basis set for molecular electronic structure theory involving first-row atoms (s and p orbitals only).

Unlike Gaussian functions, Ramp basis functions have a non-zero nuclear-electron cusp and so are much more efficient at representing the inner core region. Mixed Ramp-Gaussian basis sets have two main advantages:

  • potential for faster electronic structure quantum chemistry calculations in cases where two-electron integral evaluation is the limiting factor and basis set size is small

  • improve quality of calculations involving core electrons; e.g. NMR and EPR parameters, and excitation of core electrons.

Technically, this project would involve:

  • Expanding existing integral evaluation code to d orbitals and beyond

  • Integrating existing integral evaluation code into quantum chemistry code, e.g. QChem

  • Improving integral evaluation routines, particularly for short-range integrals

  • Developing & testing mixed Ramp-Gaussian basis set for second row atoms

  • Applying mixed Ramp-Gaussian basis sets to calculate NMR and EPR parameters, and compare against experiment

In this project, I will be collaborating with Luke Hunter on fluoro-organic compounds, Shelli McAlpine on macrocyclic peptides and the UNSW Mark Wainwright Analytic Centre on their NMR spectroscopy analysis.

Elucidating the structural-spectral relation for complex NMR spectra of peptides using computational chemistry with specialised and novel basis sets

(collaborations with Shelli McAlpine and Luke Hunter).

Keywords: Computational Chemistry, Link with Experiment, Basis Set Design, NMR spectroscopy

NMR is a ubiquitous technique extensively utilised throughout the sciences, but most especially by the synthetic chemistry community. There are two issues associated with NMR spectroscopy: assigning the spectra accurately, and understanding what the assigned spectra tells us about the structure of the molecule. Accurate computational chemistry calculations can assist with both these tasks for difficult cases like fluorine-containing compounds and macrocyclic peptides. However, existing calculations often use sub-optimal basis sets that do not adequately describe the electron core region (See figure), with scaling factors used to fix these inadequacies.

This project has two main aims: (1) assessing the effectiveness of different basis set choices in understanding NMR spectra of fluoro-containing compounds and macrocyclic peptides, including potentially the use of novel ramp-Gaussian basis sets, and (2) use calculations to interpret experimental NMR spectra to determine molecular conformations.

Investigations into Multi-Reference Electronic Structure Quantum Chemistry for Transition Metal Diatomics in Cool Stars & Hot Exoplanets

(Can be co-supervised with UNSW astronomers)

This project can be expanded or contracted in scope for projects appropriate for any level from undergraduate to post-doctoral.

Weird transition metal diatomic species such as TiO, VO, CrH and FeH play a critical role in determining the spectral and radiative characteristics of atmospheres with temperatures from around 1000 - 4000 K. Astronomers depend on accurate data on these molecules; this is particularly relevant in the current scientific context with the high profile search for habitable planets around red dwarfs and new telescopes offering unparalleled opportunities to truly characterise and understand hot Jupiter exoplanets.

Molecular data for these molecules requires a few components

  • Accurate electronic structure calculations of particularly dipole & transition moment curves, but also potential energy and coupling curves.

  • Accurate nuclear motion calculations of rovibronic structure of diatomic molecules with many coupled electronic states

  • Experimental data

The second goal is achieved well by the DUO program, developed by Dr Sergey Yurchenko & co-workers at the UCL ExoMol group.

The third goal is being actively pursued by experimentalists worldwide; they have some fantastic data (which I am helping to collate with ORBYTS, MARVEL and DC projects).

The first goal is the biggest problem. The electronic states of these molecules are extremely complex - technically, they have significant multi-reference character, i.e. a single set of orbitals is insufficient to describe the electronic state. This multi-reference nature necessitates a different set of electronic structure methods, many under active investigation and development; Multi-reference Configuration Interaction (MRCI), Multi-reference Coupled Cluster (MRCC), Density Matrix Renormalisation Group (DMRG), Stochastic Multi-Configurational Self-Consistent Field (Stochastic MC-SCF), Non-orthogonal Configuration Interaction (NOCI) and more.

This project will aim to compare and contrast a variety of these modern multi-reference calculation methods for transition metal diatomics, using VO, TiO, CrH and FeH as test systems. This will require installation (and often compilation) and running of a variety of computational chemistry codes; it will thus suit a student interested in high performance computing and software development.

Once the advantages and disadvantages of existing methods are well understood in the context of small transition metal molecules, we can start to propose and develop new methods to strive towards spectroscopic accuracy for these (and related) fascinating and very important molecules.