Theoretical and statistical solutions to problems in physical mass spectrometry

Anthony Beardsworth Costa, Purdue University

Abstract

The advent of ambient ionization mass spectrometry in the past decade has revolutionized the way direct analysis is performed. DESI mass spectrometry was the first of these techniques to be introduced, in 2004. Since that time an explosion in the number of ambient ionization techniques has occurred, a testament to the utility of performing analysis with mass spectrometry in open and native environments. The first part of this thesis develops a basic hydrodynamic theory of DESI via the methods of diffuse-interface capturing multiphase fluid dynamics. Results from these simulations confirm that a momentum-transfer event on a wetted surface is sufficient to replicate known progeny droplet properties. This is true even without incorporating the influence of electrostatics. The second part of this thesis develops a multivariate statistical method for unsupervised analysis of DESI in the imaging mode. This work is motivated by the need for a simple visualization method for morphological and chemical variation on a sample surface, as well as enabling a non-expert end-user to rapidly identify the state of an interrogated region of sample without a priori knowledge of the sample or complex, systematic analysis of full mass spectra. An approach based on the development of a uniform coordinate system for a given tissue type and disease state is developed via principal component analysis. It is shown that this method gives excellent agreement with false-color ion images of known biomarkers and histological stains. The final section of this thesis concerns the statistical and quantum mechanical treatment of serine clustering in the gas phase. These clusters are produced by a variety of atmospheric ionization methods, including sublimation/APCI, ESI, ESSI and SSI. They have been implicated in one possible mechanism leading to the origin of homochirality, as certain clusters exhibit remarkable chiral selectivity. A “structural landscape” is developed over a range of relevant cluster sizes, enantiomeric compositions, and ionizing charge states. Structures discovered via an approach based on basin-hopping molecular dynamics are used for further DFT-based optimization and analysis. It is shown that the behavior and stability of these systems is due to major structural rearrangements as a function of size and charge. The experimentally observed chiral selectivity may be understood in part by the unique network of hydrogen bonds facilitated by the serine hydroxyl side chain.

Degree

Ph.D.

Advisors

Cooks, Purdue University.

Subject Area

Analytical chemistry|Physical chemistry

COinS