This measurement has to performed before and after ingestion of a test meal and an accuracy within a few parts per mille needs to be reached for a new instrument to be considered as an alternative to the other, more complex and expensive techniques currently used as a diagnostic test.
The considered instrument is based on a 29m Herriott type multi pass cell, direct absorption setup with a VCSEL source and piezoelectric fringe dithering. Effects of temperature and pressure of the sample on line cross sections and broadening are discussed.
The use of a beam splitter and a second photodetector for a live background signal acquisition and spectra normalization was considered and tested against a simpler approach based on only one channel and using more complex fitting techniques to extract the background waveform from the absorption signal itself. For each case, an evaluation of the obtainable performance with different measurement and/or fitting algorithms is presented.
The work has been done with the ultimate goal of measuring small (0.1 liters) samples collected from exhaled breath (usually saturated in water vapor and containing 4 to 10 vol. % carbon dioxide) without any preprocessing other than lowering the measurement cell pressure; for this reason particular attention to isotope ratio measurement stability against variable sample concentration has been put. At the desired accuracy levels, even the differences in collisional broadening coefficients due to gas mixture composition are non negligible; temperature effects can be neglected only by considering proper line pairs for the ratio evaluation.
The current setup shows a resolution of about 0.2% on the δ13CO2 isotopic ratio measurement around the ambient level with a day-to-day reproducibility better than 1.5% on the single channel approach.