Getting more out of the spectrum: addressing issues in the analysis of mixtures

NMR analysis methods created by a team at the University of Manchester aim to improve spectroscopic analysis of mixtures.

The analysis of mixtures is an essential part of analytical chemistry, playing an important role in applications ranging across pharmaceuticals, environmental science, and biomedical research, for example. The identification of molecules within mixtures, however, still poses a challenge for analytical chemists; while multiple analytical methods have been developed to attempt to achieve full structure determination, there is no single technique that ‘measures everything’.¹

Against this background, nuclear magnetic resonance (NMR) spectroscopy has emerged as a key core technique, offering non-destructive and non-specific analysis with highly reproducible results. Ongoing developments and improvements in instrumentation mean the range of applications for NMR methods continues to expand.  

“Complex methods have become much more accessible to colleagues in other disciplines, meaning that NMR spectroscopy is now very widely used – and there are mixture analysis applications in almost every area of science,” says Prof. Morris.

Like many other techniques, NMR analysis, while straightforward for simple mixtures, can have limitations with more complex samples because of signal overlap,  though multidimensional NMR offers greater analytical capability.

“Multidimensional NMR can determine structures for molecules as large as proteins, while recent advances in technique and instrumentation allow the non-invasive measurement of localized proton spectra of metabolites in human beings,” says Prof. Morris. NMR can, however, be much less successful when dealing with mixtures.

Prof. Morris’ primary interest is in improving NMR techniques to transcend current limitations in areas such as mixture analysis, developing new tools for applications in chemistry, biochemistry, and medicine.

Best methods, best analysis, best results

More specifically, Prof. Morris and his colleagues at the University of Manchester are looking into novel techniques in high resolution NMR spectroscopy to develop improved methodologies, with a particular emphasis on advancing experimental techniques and data processing in parallel.

 “We’re particularly interested in the integration of the two, matching the best experimental methods with the best analysis techniques,” Prof. Morris says. “We have a great interest in experiments that manipulate the information content of NMR spectra to improve specificity and resolution, allowing us to discriminate between closely-related species in mixtures.”

One of the families of techniques of interest to the research team exploits the different diffusion coefficients of molecules. “A small molecule will move rapidly and randomly in a liquid, through Brownian motion,” says Prof. Morris. “In diffusion-ordered spectroscopy (DOSY), for example, you can sort NMR signals in a spectrum according to diffusion coefficient, essentially classifying them by size. This is one of the ways in which we can simplify the task of analyzing mixtures.”

DOSY does not provide a complete picture, however. “With DOSY, and related techniques, the information you get is indirect, meaning we have to use nonlinear methods to interpret it, and the results are ambiguous,” he says. “If we have two spectral lines that come from different species, each of them will decay exponentially in its own way as we apply diffusion weighting. The trouble is, the sum of two similar curves of that mathematical form looks just like a third, compromise, exponential - it's almost impossible to tell them apart unless you have significantly different diffusion coefficients.”

He and his team, therefore, use more sophisticated data processing methods, including multivariate analysis, to interpret such data. “In particular, there is a need to design experimental methods that encode multiple different sorts of behavior – not just the diffusion coefficients and the spectrum, but things like relaxation properties, or even time dependence, for example during a chemical reaction.”

Prof. Morris says that computers have been used to improve analytical techniques for decades and have had a steadily increasing impact both on the design of instruments and the field in general.  

“When minicomputers were first introduced, they were put into NMR machines to do a single calculation – the Fourier-transform, that lets you listen to the radiofrequency “sound” of the spins and convert that into a spectrum,” he says. “Then manufacturers realized that they could save money by also using the computer for real-time control and, suddenly, if you wanted to do something different in your experiment, you could spend an afternoon writing a computer program, rather than a year building a new piece of apparatus. That flexibility has been enormously important.”

Nowadays, data processing plays a fundamental part in the advancement of NMR methodology. “I would say that 98% of innovation in NMR techniques is actually implemented through the control software for the spectrometers,” he says. “All the manufacturers work with technique development and methodology groups to bring new methods quickly and efficiently to the market.”

Summary of research
The group’s current projects in NMR research include the development of new techniques for spectral editing and selective excitation, and multivariate processing methods for improved data processing, along with a variety of applications of NMR spectroscopy.²

“Our team focuses on ways in which individual parts of an NMR spectrum can be selectively excited, in order to be able to obtain detail from highly complex spectra,” says Prof. Morris. “We match the best experimental methods with the right data analysis techniques to deliver the best possible results.”

Collaborations and applications

Collaborations are an important part of the group's work³,  and provide, Prof. Morris says, a strong incentive to develop improved experimental methods.

“What drives our group is the development of improved tools to give the best equipment a competitive edge – though of course we also have an interest in the eventual scientific applications,” he says. “It's not unusual, for example, for pharmaceutical companies to fund our research into developing new techniques.”

Prof. Morris collaborated with colleagues from the University of Florence to study artemisin, a vital antimalarial drug, and the micronutrient curcumin. Both of these show poor solubility, which impacts absorption, leading to issues with drug release times and interactions with other pharmaceuticals. DOSY allowed the researchers to analyze the solubilization capacity of micelles of surfactant for the two substances. They reported an enhancement of aqueous solubility by between 25- and 50-fold, and also showed that using a surfactant based on ascorbic acid (vitamin C) protected artemisinin from oxidization.⁴

The team’s work with Stockholm University on breast milk underlines the importance of combining new experimental methods with new computational approaches. This study focused on human milk oligosaccharides, which are important for infant development but are structurally complex, causing signals to overlap when using proton NMR spectroscopy (1H NMR). The researchers demonstrated that when combining pure shift NMR – which can disentangle overlapped spectra – with the computational approach Computer Assisted SPectrum Evaluation of Regular polysaccharides (CASPER), it was possible to obtain increased resolution, and faster and more reliable analysis than with conventional methods.⁵

In addition, the group has developed the general NMR analysis toolbox (GNAT), an ongoing open software project that imports data from NMR manufacturers and contains both basic and advanced processing tools.⁶  “GNAT is intended primarily for the processing of multicomponent high resolution NMR datasets such as those acquired in diffusion, relaxation and kinetic experiments, but it is continually being extended, for example to 3D NMR methods,” Prof. Morris says.

This commitment to open science is also reflected in the team’s publications, which are accompanied by complete libraries of experimental data and software. “If we invent a new experiment, we publish it with code that allows people to implement the new method and analyze its results on commercial instruments, meaning they can be applying new techniques within days of publication.”

Prof. Morris says he enjoys the opportunities and advantages that arise from collaboration, particularly within the NMR field. “One of the limiting factors in research is often the availability of the experimental tools that are needed. In NMR, the community is remarkably open and constructive, because everybody recognizes that they benefit from sharing techniques,” he says.

NMR: a Goldilocks method?

Prof. Morris says that the team at the University of Manchester will continue to develop more sophisticated spectroscopic methods to address issues in mixture analysis, as well as developing computational methods that allow better interpretation of data.

“We use commercial instruments, but historically we have frequently modified them, and some of our modifications have in turn helped to further develop instrumentation,” he says. “We are constantly trying to go beyond the design parameters of the instrument to get the most out of it.”

Prof. Morris also believes there is the potential for NMR to drive further innovation. “There are very few areas that have been so long-lived or have had such a capacity to reinvent themselves every few years,” he says. “NMR is incredibly powerful and flexible, and the fundamental physics is remarkably helpful to us in terms of orders and magnitude – it’s really a Goldilocks phenomenon, where the strengths of interactions and the timescales on which things happen are ‘just right’ for the chemist.”

He says he is “constantly surprised” at the range of science where people find the team’s NMR techniques useful. “A search in the Science Citation Index shows that there are enzymologists, biologists and radiologists all using methods that we've developed,” he says. “That’s one of our aims – to make complex analyses available to the wider community.”

About Professor Gareth Morris

Professor Gareth Morris is Professor of Physical Chemistry at The University of Manchester, UK. He was successively an undergraduate, postgraduate and research fellow at Magdalen College Oxford from 1972 to 1981. In 1978-9, he spent a year at the University of British Columbia as an Izaak Walton Killam postdoctoral fellow. In 1982, he moved to Manchester, where he is now Professor of Physical Chemistry. He was awarded the RSC Corday-Morgan prize and medal in 1988, a Leverhulme Fellowship in 1996, and the RSC Industrially-Sponsored Award in Magnetic Resonance Spectroscopy in 2001, and in 2010 was an invited professor at the Université Pierre et Marie Curie. He was awarded the Russell Varian Prize for NMR in 2011, the James N. Shoolery Award in 2015, and the Günther Laukien Prize in 2021. He was elected a Fellow of the Royal Society in 2014.

¹Gates EL et al. Ultra-selective, ultra-clean 1D rotating-frame Overhauser effect spectroscopy. Chem. Commun. 2023; 59, 5854. DOI: https://doi.org/10.1039/d3cc00550j

²Manchester NMR Methodology Group. Current research. https://www.nmr.chemistry.manchester.ac.uk/?q=node/2. Last accessed Sept 5th, 2023.

³Manchester NMR Methodology Group. Collaborations on NMR applications. https://nmr.chemistry.manchester.ac.uk/?q=node/228. Last accessed Sept 5th, 2023.

⁴Lapenna S et al. Novel artemisinin and curcumin micellar formulations: Drug solubility studies by NMR spectroscopy. J. Pharm. Sci. 2009; 98 (10), 3666-3675. DOI: http://dx.doi.org/10.1002/jps.21685 

⁵Smith, MJ et al. Resolving the complexity in human milk oligosaccharides using pure shift NMR methods and CASPER. Org. Biomol. Chem. 2023; 21, 3984-3990. DOI: https://doi.org/10.1039/D3OB00421J

⁶Manchester NMR Methodology Group. GNAT (General NMR Analysis Toolbox). https://nmr.chemistry.manchester.ac.uk/?q=node/430. Last accessed Sept 5th, 2023.a