Environmental, Food and Medical Applications of Proton-Transfer-Reaction Mass Spectrometry (PTR-MS)

Werner Lindinger*; Ray Fall; Thomas G. Karl

I INTRODUCTION

Numerous gas chromatography (GC) methods have been developed over the past decades for trace gas analysis to such an extent that nearly any volatile can be quantitatively analyzed with high precision even at concentrations far below the parts per trillion by volume (pptv) level. For example, Sturges et al.1 were able to measure atmospheric concentrations of trifluoromethyl sulfur pentafluoride (CF3SF5) at levels of 0.1 pptv, and quite recently the group of Stuart Penkett has performed investigations on air samples from ice cores showing that the mixing ratio of halothane (CF3CHClBr) in the atmosphere has risen from 5 × 10− 4 pptv in the 1950s up to a maximum of 8 × 10− 3 pptv in 1985, and has been declining from then on to today’s level of about 5 × 10− 3 pptv.2 GC methods represent ideal tools when static or slowly changing mixtures are to be analyzed, but on-line monitoring of mixtures with fast varying concentrations—on time scales of a few seconds or minutes—has remained problematic. Mass spectrometry has an extremely fast response, but on-line gas analysis based on conventional mass spectrometry, using electron impact ionization, suffers because of considerable fragmentation of molecular ionic species. Especially when a mixture of organic compounds is to be analyzed, the complexity of break-up patterns puts severe constraints on the quantitative analysis of the concentrations of these components. For instance, electron impact on H2O not only yields H2O+ ions, but also OH+, O+, H2+ and H+, and benzene yields at least 18 different ions when ionized in this way. The proportion of these ions also depends on electron energy. The break-up pattern of ethanol (CH3CH2OH) contains all the ions that also appear in electron impact ionization of methanol (CH3OH), therefore it is nearly impossible to quantify trace amounts of methanol in commercial alcoholic products by conventional mass spectrometry.

Munson and Field3 reported in 1966 on a technique of ionizing molecules by gas phase ion-molecule reactions, which they called chemical ionization (CI). In this way, break-up of the molecules can be greatly reduced or even avoided. Thus, measured ion currents can be correlated with the densities of the respective parent neutral compounds, allowing for on-line monitoring of rather complex gas mixtures. The fundamental principles of gas phase ion chemistry on which CI is based, as well as the instrumentation for CI, have been reviewed in great detail by Harrison.4 The wide variety of CI methods that has been developed includes Medium Pressure Mass Spectrometry, Fourier Transform Mass Spectrometry, Quadrupole Ion Trap Mass Spectrometry, Pulsed Positive Ion-Negative Ion Chemical Ionization, and Atmospheric Pressure Ionization Mass Spectrometry (API-MS). Of these, API-MS5 has developed into a very reliable and widely used technique for analysis of VOCs in flavor release studies and human breath.6 A variety of API-MS applications in these fields of research has been described in a recent volume by Roberts and Taylor.7

A general problem in on-line monitoring is the quantification of concentrations of volatiles investigated. Usually, calibration gases are needed so that, by comparison of the respective ion signals, the actual density of the compound of interest can be evaluated. This can be avoided if the concept of swarm type experiments like that of a flow tube or flow-drift tube8 is applied. These techniques were developed by Ferguson and his group9,10 and similar ones by Adams and Smith11 in order to measure rate constants for Ion-Molecule-Reactions (IMR). Ions travelling in a carrier gas containing traces of reactant gas will be depleted depending on the concentrations of these admixed gases, and from the measured ion declines, the rate constants for the specific IMR can be obtained. Turning this principle around, from the decline of a reactant ion and/or from the quantitative appearance of product ions in a flow experiment, the density of the neutral reactant can be calculated, provided that the rate constant (and reaction time) for the respective IMR is known. This procedure has been used in our laboratory to develop an on-line method for trace gas monitoring and, as the reactions on which the method is based are proton-transfer-reactions, it was named Proton-Transfer-Reaction Mass Spectrometry (PTR-MS).12,13 In this review, a short description of the method will be given followed by results from applications of PTR-MS in the fields of environmental, food and medical research.

II PROTON-TRANSFER-REACTION MASS SPECTROMETRY (PTR-MS)

PTR-MS combines the concept of CI with the swarm technique of the flow tube and flow-drift-tube mentioned above. In a PTR-MS instrument, we apply a CI system which is based on proton-transfer reactions, and preferentially use H3O+ as the primary reactant ion. As discussed earlier,12H3O+ is a most suitable primary reactant ion when air samples containing a wide variety of trace gases or VOCs are to be analyzed. H3O+ ions do not react with any of the natural components of air, as these have proton affinities lower than that of H2O molecules; this is illustrated in Table 1. This table also shows that common VOCs containing a polar functional group or unsaturated bonds (e.g. alkenes, arenes) have proton affinities larger than that of H2O and therefore proton transfer occurs between H3O+ and any of these compounds (see Equation 4). The measured thermal rate constants for proton transfer to VOCs are nearly identical to calculated thermal, collisional limiting values (Table 1), illustrating that proton transfer occurs on every collision.