Two-Dimensional Separation via Hybrid Liquid Chromatography and Differential Ion Mobility Spectrometry for PFAS Characterization

Abstract

This thesis details the development and implementation of differential mobility spectrometry (DMS) methods for the separation of per- and polyfluoroalkyl substances (PFAS). PFAS have become ubiquitous environmental pollutants, posing significant risks to ecosystems and human health. The complexity of PFAS matrices in environmental samples necessitates separation prior to mass spectrometric analysis because co-elution of compounds can cause ion suppression and compromise analyte identification and quantification accuracy. Although liquid chromatography (LC) is commonly used in PFAS analyses, some PFAS species co-elute and could benefit from an additional orthogonal dimension of separation.

In Chapter 3 I explore the effects of solvent modifier on DMS behaviour for 224 compounds in negative mode electrospray ionization (ESI) mass spectrometry (MS). The data procured from these measurements will be used for machine learning (ML) purposes to predict the DMS behaviour of emerging environmental pollutants. Prior to this study, our library of DMS data was composed entirely of compounds that were measured in positive mode ESI MS and the distribution of observed dispersion behaviour was heavily skewed towards one behaviour type. Incorporation of the negative mode ESI data not only provided a better overall distribution of dispersion behaviour, but also allows for future ML models to be applicable for anions and cations alike. The results of this chapter also provide insight into the ion-neutral interactions that occur as analytes transit the DMS cell. From this it can be determined how different classes of compounds interact with various solvent modifiers, and how their analytical separation is influenced by the choice of modifier. This allowed us to determine the instrument conditions that lead to the optimal separation of the studied PFAS.

In Chapter 4, I utilize the optimal separation conditions determined in Chapter 3 in a hybrid LC×DMS-MS2 method. Here, I employ DMS following LC separation to analyse 34 PFAS species. Upon incorporating DMS in a 2D separation scheme, I observed baseline resolution of 29 compounds in the 2D space, with only two and three compounds co-eluting, respectively. In comparison, only 5 compounds were baseline resolved in 1-dimensional LC experiments. Because DMS measurements are acquired within seconds, targeted 2D LC×DMS-MS2 analyses operate on the same timescale as 1D LC-MS2 analysis. Additionally, limits of quantitation approach those observed in state-of-the-art LC-MS2 methods. Moreover, distinct trends observed in the 2D separation space for the various PFAS subclasses could enable analyte identification in future non-targeted analyses.