Research

Matrix-assisted laser desorption/ionization (MALDI) and matrix-assisted ionization (MAI) both generate gas-phase ions from a solid matrix, the former driven by a laser pulse and the latter by sublimation alone. Both are widely applied to biological molecules, yet neither mechanism is settled. For MALDI, multiphoton ionization, energy pooling, preformed ions and thermal ionization coexist in the literature, and their predictions conflict under certain conditions. MAI presents a further difficulty: the phase transition itself separates charge, and the resulting charge-state distribution resembles that of electrospray, which remains inadequately explained. In consequence, matrix selection relies largely on experience, sensitivity cannot be estimated in advance, and the design of new ionization methods lacks a firm basis.

The primary observable is the absolute ionization efficiency, that is, the ratio of ions to neutrals, rather than signal intensity alone. Laser pulse width and fluence, matrix identity, and the temperature and pressure of the inlet are varied systematically so that each mechanistic model can be tested against its own predictions. For MALDI this yields a quantitative description based on thermally induced proton transfer (TIPT), in which the ion yield follows from the thermodynamic properties of the matrix and the laser-induced temperature. For MAI the focus is the spontaneous charge separation that accompanies sublimation, and how pressure, temperature and matrix properties determine ion formation and charge-state distribution.

TIPT reproduces measured ion yields across independent experiments. For MAI, sublimation-driven charge separation is found to be general rather than specific to one matrix, and operates at atmospheric, sub-atmospheric and vacuum conditions alike. These findings have been applied to ion source design, giving ionization schemes that require neither high voltage nor a laser and are compatible with portable instruments. Taken together, the results indicate that ion formation from solid samples is governed primarily by thermodynamics and phase change.

Carbohydrate analysis by mass spectrometry is limited by several factors: low ionization efficiency, signal division among alkali-metal adducts, and the ring-shaped deposit left as a droplet dries, which distributes analyte unevenly across the sample spot. Intensity therefore varies substantially between positions on the same sample, and quantification is not reproducible. Isomers of similar mass further complicate identification. Together these limitations are the principal obstacle to rapid screening from clinical specimens such as serum and urine.

The rapidly freeze-drying droplet (RFDD) method was developed at the sample-preparation stage. Matrix and analyte solidify together at low temperature, preserving a homogeneous distribution of preformed sodium adducts and improving the spatial consistency of signal within a spot. With the measurement made reliable, the full spectrum is treated as a high-dimensional feature set and interpreted by machine-learning classifiers, reducing dependence on the intensity of any single ion.

RFDD raises ion intensity by more than two orders of magnitude and improves the uniformity of signal within a sample spot; calibration in a serum matrix reaches a linearity of R² = 0.99. The workflow is currently being applied to urinary liquid biopsy, where its feasibility for early cancer detection and metastasis monitoring is under evaluation.

The efficiency of plastic recycling is limited chiefly at the sorting stage. Unless a mixed polymer stream can be identified quickly and at low cost, the quality of the recyclate cannot be improved. Near-infrared sorting, the current standard, has limited capability for dark-coloured plastics and multilayer composites, which constitute a substantial fraction of real waste streams.

Mass spectrometry serves as a fingerprinting tool for polymers: pyrolysis and ionization yield a characteristic fragment pattern for each material, from which a classification model assigns the polymer type. The method is designed for use on site, with attention to instrument size, cost and analysis speed rather than laboratory-grade precision. The portable instruments and ion sources developed earlier in the group provide the technical basis.

The work is at the method-development stage, supported by an NSTC grant. Reference spectra for polymer standards are being compiled and the classification models validated.

Key intermediates in catalytic cycles are typically short-lived and dilute, hence difficult to observe directly, and their existence is inferred from kinetics and computation. Where several competing mechanistic proposals account equally well for the available data, indirect evidence cannot discriminate between them, and catalyst design is constrained accordingly.

Customized inlet systems introduce the reacting solution directly into the mass spectrometer. Cold-spray ionization and temperature-controlled inlets reduce decomposition of transient species during transfer, allowing them to be detected while the reaction proceeds. The work is carried out with synthetic, electrochemical and computational groups, with the aim of supplying evidence that discriminates between competing mechanistic proposals.

In dinuclear cobalt water oxidation, peroxo-bridged species were observed and the change in oxidation state of a non-innocent ligand established, accounting for the trade-off between reaction rate and overpotential. In high-valent iron oxo chemistry, the conditions separating homolytic from heterolytic O–O cleavage were resolved. In palladium precatalyst systems, the formation of the active species was followed directly.