Cross contamination in dual inlet isotope ratio mass spectrometers

Since the early days of geochemical isotope ratio mass spectrometry there has always been the problem of cross contamination, i.e. the contamination of the sample gas with traces of reference gas (and vice versa) in a dual inlet system and the analyzer itself. This was attributable to valve leakages and could be corrected for. In modern leak-free machines this problem still influences the measurements by scale contraction. It limits the interlaboratory comparability to values unacceptable for, e.g. global carbon cycle investigations on atmospheric dCO2, with the need to integrate global datasets at a 0.01‰ precision level. No matter what the reason of apparent cross contamination (physical leakage, low pump efficiency, gas–wall interactions), we present a method to measure the actual cross contamination as well as algorithms to correct for it. By using isotopically enriched sample gas the effect on the reference gas is directly shown. Especially for biomedical applications, employing highly enriched materials, the nonlinear correction has to be applied. A long-time record of two machines shows the influence of vacuum breakage and the inlet system setup. We present a survey of cross contamination in some of the frequently used machines. The relatively poor raw results of two recent interlaboratory comparison experiments are explained in terms of cross contamination. Whereas the linear two-point normalization procedure considered by itself may be appropriate for some routine applications at natural isotope abundances, it always relies on relative d values of two reference materials. There is an urgent need to establish absolute values for all scale-determining reference materials. We recommend the use of cross contamination measurements to at least determine the isotopic composition of all normalization materials as long as absolute measurements are not available. In principle, the cross contamination measurements should be used as a diagnostic tool on a regular basis. (Int J Mass Spectrom 198 (2000) 45–61) © 2000 Elsevier Science B.V.