Use this URL to cite or link to this record in EThOS:
Title: Tested to destruction : advanced spectroscopic, spectrometric, and chemometric analysis of Scotch whisky
Author: Kew, William
ISNI:       0000 0004 7654 5349
Awarding Body: University of Edinburgh
Current Institution: University of Edinburgh
Date of Award: 2018
Availability of Full Text:
Access from EThOS:
Full text unavailable from EThOS. Please try the link below.
Access from Institution:
Scotch Whisky is a complex mixture comprising thousands of chemical species at a diverse range of concentrations. The identities, origins, and significance of many of these compounds is largely unknown. Routine characterisation of Scotch Whisky mostly utilises techniques such as gas or liquid chromatography (GC, LC) coupled to a FID, UV, or low-resolution mass spectrometry (MS) detector. In this thesis, advanced spectroscopic and spectrometric techniques are investigated as potential complementary means to unravel the chemical complexity of Scotch Whisky. Chemometric methods are applied to decipher the significance or potential origin of many of these compounds. Being predominantly (ca. 99 %) protonated 'solvent' - ethanol and water - 1H and 13C Nuclear Magnetic Resonance (NMR) required solvent suppression to be implemented into the acquisition of high resolution spectra. A 1D NOESY-presaturation sequence was modified and implemented in automation for this purpose. Furthermore, this solvent suppression was coupled with several 1D and 2D homo- and heterocorrelated NMR experiments for the analysis of Scotch Whisky. With limited sample preparation - only addition of buffer - an approximate limit of detection of 50 μm was achieved. The developed NMR methodology was subsequently used for structural elucidation of dozens of compounds in Scotch Whisky. Quantification of these compounds was hindered by variable chemical shifts, signal overlap, and for some compounds the existence of equilibria of different forms. Quantification of ethanol concentrations in model solutions and genuine Scotch Whisky samples was successful. A large set of Scotch Whisky samples were analysed by the solvent suppressed 1D 1H NMR methodology and various statistical techniques including statistical total correlation spectroscopy (STOCSY), independent and principal component analysis (ICA, PCA), and orthogonal partial least squares discriminant analysis (OPLS-DA). Various parameters were modelled, and discrimination of 'malt' or 'blend' status was achieved, whilst maturation wood type discrimination was less successful. High resolution negative mode electrospray ionisation (ESI) Fourier transform ion cyclotron resonance (FTICR) MS was then used to examine a large set of Scotch Whisky samples. Thousands of unique molecular formulae were assigned within a 1 ppm threshold, representing an assignment rate of 72-88 % of the detected peaks in the spectra. Assignments were selectively confirmed by isotopic fine structure (IFS) analysis, and structural information obtained by both quadrupolar isolation and fragmentation, and in-cell isolation and fragmentation. Similar chemometric methods as applied to NMR data were used to model sample parameters, and identification of potential maturation wood marker compounds was achieved. Alternative ionisation sources - including atmospheric pressure chemical- and photo-ionisation (APCI, APPI), and laser desorption ionisation (LDI) - were compared for the analysis of Scotch Whisky by FTICR MS. The differing sources provide complementary compositional information on the sample set, with APCI and LDI being most different to ESI in terms of compounds ionised and spectral profiles. Positive mode ionisation was also successful, but molecular formula assignment was hindered by insufficient resolving power. Late experimentation pushed the achievable resolving power to 2.8 million at m/z 400, however the required approach is significantly more time consuming and prone to signal quality degradation. Formula assignment software, both commercial, open source, and in-house developed, were compared. The commercial and published open source software provided essentially identical results for Scotch Whisky, and whilst the in-house tools assigned fewer species (a subset of those assigned by the other tools), they did so with a smaller mean error of assignment. Various analysis and visualisation tools for MS data of complex mixtures were also developed.
Supervisor: Uhrin, Dusan ; Clarke, David Sponsor: Biotechnology and Biological Sciences Research Council (BBSRC)
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available
Keywords: whisky ; NMR ; mass spectrometry ; chemometric methods ; statistical total correlation spectroscopy ; positive mode ionisation