Multiple independent NMR dimensions: smart experiments for complicated problems
Lead Research Organisation:
University of Manchester
Department Name: Chemistry
Abstract
If successful, this work will more than double the chemical resolution of a wide range of NMR methods. Why does this matter?
Nuclear Magnetic Resonance (NMR) spectroscopy is by far the most important and widely-used tool for determining the chemical structures of species in solution, but it relies absolutely on the ability to distinguish between the signals of nuclear spins in different chemical environments. The more complex a chemical material or mixture, the more different environments there are, and the more NMR struggles to resolve the differences between them. This limits both the size of a molecule, and the degree of complexity of a mixture, that can usefully be studied by NMR.
The classic way to improve the resolution of NMR is to increase the strength of the magnetic field used, but this is limited by magnet technology. It has taken 30 years for improvements in magnet strength to double the basic resolution of NMR spectroscopy, and the strongest magnets now cost many millions of pounds. A more efficient - and much cheaper - way to improve resolution is to use more sophisticated experiments, exciting the nuclear spins with multiple pulses of radio waves and then disentangling their responses using the mathematics of the Fourier transform. This has proved very effective, and now enables chemists and bioscientists to solve problems that are far out of the reach of basic NMR methods. Frustratingly, though, we still continually come up against the limits of resolution in NMR, whether in large molecules or in complex mixtures of small ones.
What can we do to give us chemical resolution when existing methods reach their limits, and the signals from different chemical sites are so intermingled that we cannot tell them apart? This project will go a significant step further, encoding extra information into these overlapping signals and then using advanced statistical methods, so-called matrix tensor decomposition, to disentangle them. The most powerful algorithms, such as parallel factor analysis (PARAFAC), need the experimental data to vary independently in three or more different ways. We will design new experiments that are tailored to produce data for tensor analysis, using multiple independent NMR dimensions ("MIND"). Using modular pulse sequence elements will allow this MIND approach to be incorporated into a range of existing experiments, multiplying their resolving power. We will produce all that is needed to allow end users to implement these new methods easily: computer code to control the spectrometer, processing software to analyse the data, and illustrative examples. These new MIND NMR experiments will have wide application across a range of academic and industrial research areas, including chemistry, biochemistry, biology, pharmacy, petrochemistry, agrochemistry, healthcare, and flavours and fragrances.
Nuclear Magnetic Resonance (NMR) spectroscopy is by far the most important and widely-used tool for determining the chemical structures of species in solution, but it relies absolutely on the ability to distinguish between the signals of nuclear spins in different chemical environments. The more complex a chemical material or mixture, the more different environments there are, and the more NMR struggles to resolve the differences between them. This limits both the size of a molecule, and the degree of complexity of a mixture, that can usefully be studied by NMR.
The classic way to improve the resolution of NMR is to increase the strength of the magnetic field used, but this is limited by magnet technology. It has taken 30 years for improvements in magnet strength to double the basic resolution of NMR spectroscopy, and the strongest magnets now cost many millions of pounds. A more efficient - and much cheaper - way to improve resolution is to use more sophisticated experiments, exciting the nuclear spins with multiple pulses of radio waves and then disentangling their responses using the mathematics of the Fourier transform. This has proved very effective, and now enables chemists and bioscientists to solve problems that are far out of the reach of basic NMR methods. Frustratingly, though, we still continually come up against the limits of resolution in NMR, whether in large molecules or in complex mixtures of small ones.
What can we do to give us chemical resolution when existing methods reach their limits, and the signals from different chemical sites are so intermingled that we cannot tell them apart? This project will go a significant step further, encoding extra information into these overlapping signals and then using advanced statistical methods, so-called matrix tensor decomposition, to disentangle them. The most powerful algorithms, such as parallel factor analysis (PARAFAC), need the experimental data to vary independently in three or more different ways. We will design new experiments that are tailored to produce data for tensor analysis, using multiple independent NMR dimensions ("MIND"). Using modular pulse sequence elements will allow this MIND approach to be incorporated into a range of existing experiments, multiplying their resolving power. We will produce all that is needed to allow end users to implement these new methods easily: computer code to control the spectrometer, processing software to analyse the data, and illustrative examples. These new MIND NMR experiments will have wide application across a range of academic and industrial research areas, including chemistry, biochemistry, biology, pharmacy, petrochemistry, agrochemistry, healthcare, and flavours and fragrances.
