Mass spectrometric sampling of flames: how ionic equilibria in flames produce sampling falsifications and “fake” ions, but provide kinetic and thermodynamic data on the reaction occurring

Abstract

Continuously sampling a flame, burning at 1 atm., for mass spectrometry at ≈ 10–8 atm. seriously disturbs the flame. Not only are a flame's temperature and velocity altered, often the composition of a sample is falsified. Thus, “fake” ions appear, even when sampling as quickly as possible, i.e. supersonically, to quench chemical reactions. However, studying these spurious ions is fruitful. They arise, because a sample is unavoidably cooled; the drop in temperature causes a rapid chemical equilibrium to shift position and change the sample's composition. That ions react faster than neutrals (to perturb a sample) magnifies the problem for ions. When continuously sampling a flame, burning at 1 atm., through an inlet at the tip of a hollow, metallic nozzle, cooling can occur in three ways during the formation of a beam for mass spectrometry. Firstly, before a sample passes through the inlet hole to enter the supersonic expansion into the first vacuum chamber of the mass spectrometer, it loses heat to the cooler, sampling nozzle, usually conical in shape. By detecting spurious ions from a flame, this drop in temperature has been measured to be greatest (≈ 400 K) for the smallest orifices. This cooling becomes smaller for larger holes and is trivial for diameters above 150 µm. Secondly, a sample cools (by maybe ≈ 300 K), whatever the orifice's size, on being accelerated to the local speed of sound in the narrowest part, i.e. the throat of the inlet orifice. Thirdly, the drop in temperature in the subsequent, near-adiabatic expansion inside the nozzle is greatest (≈ 1000 K) and most prolonged for the largest inlet holes (diam. > 150 µm). The upshot is that with a small hole (diam. < 100 µm), a sample is cooled by both the sampling nozzle and the acceleration to sonic velocity in the throat of the inlet. However, with a large orifice (diam. > 150 µm), cooling happens in the acceleration to a Mach number of unity and the following supersonic expansion. Analysis shows that, if a positive ion reacts exothermally in a reversible reaction with a time constant briefer than ≈ 0.5 µs, that reaction will be equilibrated early in the flame. In addition, if the orifice is small, the equilibrium will be just fast enough to shift position to that for a temperature reduced in both the thermal boundary layer around the inlet, and in accelerating to the speed of sound. Consequently, the sample begins the expansion with new species. When using a big orifice, the reaction's time constant (in the flame) must be less than ≈ 0.05 µs (depending on the flame) to generate new ions in the supersonic expansion. It follows that, if there is not an exothermic chemical reaction with a time constant less than ≈ 0.5 µs, sampling is most probably genuine. A similar criterion usually means that no fake neutral species are observed in a low-pressure flame. Negative ions are complicated by their reactions often involving free electrons. Being mobile, they often leave a sample to attach to the metallic sampling nozzle. This loss of free electrons changes the ionic composition, because rapid, steady-state relationships for individual negative ions are thereby shifted, with the change in composition depending on the orifice's size. A theme of this review is that these problems can be identified and resolved by repeating observations using sampling orifices with a range of different diameters. The resulting measurements are then extrapolated to a hole size of either zero or infinity, when there is no effect of a perturbation in either the expansion or the boundary layer; this yields a measurement, e.g. of an equilibrium constant, for the known conditions in the neck of the orifice. In addition, applying a voltage between the burner and the sampling nozzle considerably improves the accuracy, with which ionic concentrations are measured. Consequently, many rapid reactions have had their equilibrium constants measured, yielding the proton affinities of e.g. H2O, CO, CO2 and NH3, the hydration energies of many ions, the stability of ions like MnOH+ (resulting from seeding a flame with Mn) and also the standard changes of enthalpy and entropy for e.g. OH– + CO2 + M = HCO3– + M and HCO3– + OH = CO3– + H2O. In addition, rate constants have been deduced for reactions like the mono-hydration of H3O+, for LiOH + H3O+ → Li+ + H2O and Li+ + CO + M → Li+.CO + M, as well as those for the forward and reverse steps in: e– + O2 + M ⇄ O2– + M, Cl– + H ⇄ HCl + e–, O2– + H ←→ HO2 + e– and O2– + OH ←→OH– + O2. The design of mass spectrometers was discussed, as well as the sampling of neutral species.